How to put it plainly? Analysis
Kerwer et al.
2021
Study 1
Read data files
Note: Anomynized datasets were created by removing possibly identifying information and data that are not relevant to confirmatory analyses.
options(width=120)
df_resilience <- read.csv2(file = "20210901_study_1_resilience_df_long.csv",fileEncoding="UTF-8", stringsAsFactors = T)
df_cama <- read.csv2(file = "20210901_study_1_cama_df_long.csv",fileEncoding="UTF-8", stringsAsFactors = T)
df <- read.csv2(file = "20210901_study_1_df_long.csv",fileEncoding="UTF-8", stringsAsFactors = T)
data <- read.csv2(file = "20210702_study_1_df_wide.csv",fileEncoding="UTF-8", stringsAsFactors = T)
data$s_iv_technical_terms <- factor(data$s_iv_technical_terms, levels = c("Expl None","Replace", "Expl Glossar"))
df_resilience$s_iv_technical_terms <- factor(df_resilience$s_iv_technical_terms, levels = c("Expl None","Replace", "Expl Glossar"))
df$s_iv_technical_terms <- factor(df$s_iv_technical_terms, levels = c("Expl None","Replace", "Expl Glossar"))
df_cama$pls <- factor(df_cama$pls, levels = c("Dunst","Christodoulou"))
df_resilience$pls <- factor(df_resilience$pls, levels = c("Rasmussen","Groth"))
df$pls <- factor(df$pls, levels = c("Dunst","Christodoulou", "Rasmussen","Groth"))Descriptive Within Level
CAMA
require(psych)## Lade nötiges Paket: psychnames(df_cama)## [1] "id" "s_study_arm" "s_iv_quality_of_evidence" "s_iv_operationalization"
## [5] "s_iv_technical_terms" "s_int_psy" "s_knowledge_evidence" "pls"
## [9] "accessibility" "understanding" "empowerment" "pref_quality_of_evidence"
## [13] "knowledge_evidence" "knowledge"df1 <- df_cama[,c("accessibility", "understanding", "empowerment", "pref_quality_of_evidence", "knowledge_evidence","knowledge", "s_iv_quality_of_evidence", "s_iv_operationalization")]
describe(df1)## vars n mean sd median trimmed mad min max range skew kurtosis se
## accessibility 1 2262 6.08 1.89 7 6.33 1.48 1 8 7 -0.86 -0.09 0.04
## understanding 2 2258 6.08 1.87 7 6.32 1.48 1 8 7 -0.85 -0.07 0.04
## empowerment 3 2260 4.86 2.01 5 4.91 1.48 1 8 7 -0.17 -0.80 0.04
## pref_quality_of_evidence 4 2235 4.69 1.87 5 4.68 1.48 1 8 7 0.03 -0.58 0.04
## knowledge_evidence 5 2230 0.39 0.49 0 0.37 0.00 0 1 1 0.43 -1.81 0.01
## knowledge 6 2230 0.65 0.48 1 0.68 0.00 0 1 1 -0.62 -1.62 0.01
## s_iv_quality_of_evidence* 7 2274 1.50 0.50 1 1.50 0.00 1 2 1 0.02 -2.00 0.01
## s_iv_operationalization* 8 2274 1.50 0.50 1 1.49 0.00 1 2 1 0.02 -2.00 0.01df5 <- data.frame(psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu Yes" & df1$s_iv_operationalization == "Op Yes",], na.rm = T)$mean,
psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu Yes" & df1$s_iv_operationalization == "Op Yes",], na.rm = T)$sd,
psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu No" & df1$s_iv_operationalization == "Op Yes",], na.rm = T)$mean,
psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu No" & df1$s_iv_operationalization == "Op Yes",], na.rm = T)$sd,
psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu Yes" & df1$s_iv_operationalization == "Op No",], na.rm = T)$mean,
psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu Yes" & df1$s_iv_operationalization == "Op No",], na.rm = T)$sd,
psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu No" & df1$s_iv_operationalization == "Op No",], na.rm = T)$mean,
psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu No" & df1$s_iv_operationalization == "Op No",], na.rm = T)$sd)
rownames(df5) <- colnames(df1)
colnames(df5) <- c("Mean_Qu_Yes_Op_Yes", "SD_Qu_Yes_Op_Yes","Mean_Qu_No_Op_Yes", "SD_Qu_No_Op_Yes", "Mean_Qu_Yes_Op_No", "SD_Qu_Yes_Op_No", "Mean_Qu_No_Op_No", "SD_Qu_No_Op_No")
df5 <- df5[1:(nrow(df5)-2),]
df5## Mean_Qu_Yes_Op_Yes SD_Qu_Yes_Op_Yes Mean_Qu_No_Op_Yes SD_Qu_No_Op_Yes Mean_Qu_Yes_Op_No
## accessibility 5.9198543 1.9211052 5.8947368 2.0672712 6.1448517
## understanding 5.9253188 1.8805274 5.8929825 2.0246599 6.1083916
## empowerment 4.7868852 1.9724682 4.6830123 2.0722160 4.8706294
## pref_quality_of_evidence 4.8704380 1.8419255 4.5666667 1.9274936 4.8064516
## knowledge_evidence 0.4493554 0.4978872 0.3132743 0.4642359 0.4575972
## knowledge 0.6593002 0.4743816 0.6566372 0.4752521 0.6219081
## SD_Qu_Yes_Op_No Mean_Qu_No_Op_No SD_Qu_No_Op_No
## accessibility 1.7446107 6.3578947 1.7987447
## understanding 1.7015907 6.3827160 1.8059338
## empowerment 1.9550225 5.0950704 2.0188120
## pref_quality_of_evidence 1.8992093 4.5366726 1.7995411
## knowledge_evidence 0.4986394 0.3561151 0.4792810
## knowledge 0.4853397 0.6510791 0.4770581psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu Yes" & df1$s_iv_operationalization == "Op Yes",], na.rm = T)## vars n mean sd median trimmed mad min max range skew kurtosis se
## accessibility 1 549 5.92 1.92 6 6.12 2.97 1 8 7 -0.64 -0.58 0.08
## understanding 2 549 5.93 1.88 6 6.12 2.97 1 8 7 -0.65 -0.55 0.08
## empowerment 3 549 4.79 1.97 5 4.81 1.48 1 8 7 -0.11 -0.82 0.08
## pref_quality_of_evidence 4 548 4.87 1.84 5 4.88 1.48 1 8 7 0.00 -0.64 0.08
## knowledge_evidence 5 543 0.45 0.50 0 0.44 0.00 0 1 1 0.20 -1.96 0.02
## knowledge 6 543 0.66 0.47 1 0.70 0.00 0 1 1 -0.67 -1.55 0.02
## s_iv_quality_of_evidence* 7 552 2.00 0.00 2 2.00 0.00 2 2 0 NaN NaN 0.00
## s_iv_operationalization* 8 552 2.00 0.00 2 2.00 0.00 2 2 0 NaN NaN 0.00psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu Yes" & df1$s_iv_operationalization == "Op No",], na.rm = T)## vars n mean sd median trimmed mad min max range skew kurtosis se
## accessibility 1 573 6.14 1.74 6 6.35 1.48 1 8 7 -0.80 -0.16 0.07
## understanding 2 572 6.11 1.70 6 6.28 1.48 1 8 7 -0.75 -0.08 0.07
## empowerment 3 572 4.87 1.96 5 4.91 1.48 1 8 7 -0.17 -0.77 0.08
## pref_quality_of_evidence 4 558 4.81 1.90 5 4.83 1.48 1 8 7 -0.13 -0.56 0.08
## knowledge_evidence 5 566 0.46 0.50 0 0.45 0.00 0 1 1 0.17 -1.97 0.02
## knowledge 6 566 0.62 0.49 1 0.65 0.00 0 1 1 -0.50 -1.75 0.02
## s_iv_quality_of_evidence* 7 576 2.00 0.00 2 2.00 0.00 2 2 0 NaN NaN 0.00
## s_iv_operationalization* 8 576 1.00 0.00 1 1.00 0.00 1 1 0 NaN NaN 0.00psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu No" & df1$s_iv_operationalization == "Op Yes",], na.rm = T)## vars n mean sd median trimmed mad min max range skew kurtosis se
## accessibility 1 570 5.89 2.07 6 6.18 2.97 1 8 7 -0.88 -0.18 0.09
## understanding 2 570 5.89 2.02 6 6.15 2.97 1 8 7 -0.83 -0.23 0.08
## empowerment 3 571 4.68 2.07 5 4.72 1.48 1 8 7 -0.12 -0.87 0.09
## pref_quality_of_evidence 4 570 4.57 1.93 4 4.53 1.48 1 8 7 0.14 -0.63 0.08
## knowledge_evidence 5 565 0.31 0.46 0 0.27 0.00 0 1 1 0.80 -1.36 0.02
## knowledge 6 565 0.66 0.48 1 0.70 0.00 0 1 1 -0.66 -1.57 0.02
## s_iv_quality_of_evidence* 7 575 1.00 0.00 1 1.00 0.00 1 1 0 NaN NaN 0.00
## s_iv_operationalization* 8 575 2.00 0.00 2 2.00 0.00 2 2 0 NaN NaN 0.00psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu No" & df1$s_iv_operationalization == "Op No",], na.rm = T)## vars n mean sd median trimmed mad min max range skew kurtosis se
## accessibility 1 570 6.36 1.80 7 6.62 1.48 1 8 7 -1.05 0.39 0.08
## understanding 2 567 6.38 1.81 7 6.66 1.48 1 8 7 -1.10 0.54 0.08
## empowerment 3 568 5.10 2.02 5 5.20 1.48 1 8 7 -0.28 -0.73 0.08
## pref_quality_of_evidence 4 559 4.54 1.80 4 4.50 1.48 1 8 7 0.12 -0.45 0.08
## knowledge_evidence 5 556 0.36 0.48 0 0.32 0.00 0 1 1 0.60 -1.64 0.02
## knowledge 6 556 0.65 0.48 1 0.69 0.00 0 1 1 -0.63 -1.60 0.02
## s_iv_quality_of_evidence* 7 571 1.00 0.00 1 1.00 0.00 1 1 0 NaN NaN 0.00
## s_iv_operationalization* 8 571 1.00 0.00 1 1.00 0.00 1 1 0 NaN NaN 0.00names(df_cama)## [1] "id" "s_study_arm" "s_iv_quality_of_evidence" "s_iv_operationalization"
## [5] "s_iv_technical_terms" "s_int_psy" "s_knowledge_evidence" "pls"
## [9] "accessibility" "understanding" "empowerment" "pref_quality_of_evidence"
## [13] "knowledge_evidence" "knowledge"df1 <- df_cama[,c("accessibility", "understanding", "empowerment", "pref_quality_of_evidence","s_iv_operationalization")]
describe(df1)## vars n mean sd median trimmed mad min max range skew kurtosis se
## accessibility 1 2262 6.08 1.89 7 6.33 1.48 1 8 7 -0.86 -0.09 0.04
## understanding 2 2258 6.08 1.87 7 6.32 1.48 1 8 7 -0.85 -0.07 0.04
## empowerment 3 2260 4.86 2.01 5 4.91 1.48 1 8 7 -0.17 -0.80 0.04
## pref_quality_of_evidence 4 2235 4.69 1.87 5 4.68 1.48 1 8 7 0.03 -0.58 0.04
## s_iv_operationalization* 5 2274 1.50 0.50 1 1.49 0.00 1 2 1 0.02 -2.00 0.01df5 <- data.frame(psych::describe(df1[df1$s_iv_operationalization == "Op Yes",], na.rm = T)$mean,
psych::describe(df1[df1$s_iv_operationalization == "Op Yes",], na.rm = T)$sd,
psych::describe(df1[df1$s_iv_operationalization == "Op No",], na.rm = T)$mean,
psych::describe(df1[df1$s_iv_operationalization == "Op No",], na.rm = T)$sd)
psych::describe(df1[df1$s_iv_operationalization == "Op Yes",], na.rm = T)## vars n mean sd median trimmed mad min max range skew kurtosis se
## accessibility 1 1119 5.91 2.00 6 6.15 2.97 1 8 7 -0.78 -0.32 0.06
## understanding 2 1119 5.91 1.95 6 6.14 2.97 1 8 7 -0.76 -0.34 0.06
## empowerment 3 1120 4.73 2.02 5 4.76 1.48 1 8 7 -0.12 -0.84 0.06
## pref_quality_of_evidence 4 1118 4.72 1.89 5 4.70 1.48 1 8 7 0.07 -0.64 0.06
## s_iv_operationalization* 5 1127 2.00 0.00 2 2.00 0.00 2 2 0 NaN NaN 0.00psych::describe(df1[df1$s_iv_operationalization == "Op No",], na.rm = T)## vars n mean sd median trimmed mad min max range skew kurtosis se
## accessibility 1 1143 6.25 1.77 7 6.49 1.48 1 8 7 -0.92 0.10 0.05
## understanding 2 1139 6.24 1.76 7 6.47 1.48 1 8 7 -0.92 0.20 0.05
## empowerment 3 1140 4.98 1.99 5 5.05 1.48 1 8 7 -0.22 -0.75 0.06
## pref_quality_of_evidence 4 1117 4.67 1.85 5 4.66 1.48 1 8 7 0.00 -0.53 0.06
## s_iv_operationalization* 5 1147 1.00 0.00 1 1.00 0.00 1 1 0 NaN NaN 0.00rownames(df5) <- colnames(df1)
colnames(df5) <- c("Mean_Op_Yes", "SD_Op_Yes", "Mean_Op_No", "SD_Op_No")
df5 <- df5[1:(nrow(df5)-1),]
df5## Mean_Op_Yes SD_Op_Yes Mean_Op_No SD_Op_No
## accessibility 5.907060 1.996046 6.251094 1.774241
## understanding 5.908847 1.954469 6.244952 1.758899
## empowerment 4.733929 2.023700 4.982456 1.989359
## pref_quality_of_evidence 4.715564 1.891305 4.671441 1.854090#correlations
cor(df_cama$understanding, df_cama$accessibility, use = "pairwise.complete.obs")## [1] 0.8221927cor(df_cama$understanding, df_cama$empowerment, use = "pairwise.complete.obs")## [1] 0.6402011cor(df_cama$accessibility, df_cama$empowerment, use = "pairwise.complete.obs")## [1] 0.5586959df1 <- df[,c("accessibility", "understanding", "empowerment", "pref_quality_of_evidence","s_iv_quality_of_evidence")]
describe(df1)## vars n mean sd median trimmed mad min max range skew kurtosis se
## accessibility 1 4326 5.57 2.07 6 5.77 2.97 1 8 7 -0.55 -0.72 0.03
## understanding 2 4323 5.60 2.02 6 5.80 2.97 1 8 7 -0.58 -0.60 0.03
## empowerment 3 4324 4.49 2.07 5 4.50 2.97 1 8 7 -0.04 -0.90 0.03
## pref_quality_of_evidence 4 4276 4.74 1.80 5 4.74 1.48 1 8 7 0.02 -0.48 0.03
## s_iv_quality_of_evidence* 5 4350 1.49 0.50 1 1.48 0.00 1 2 1 0.06 -2.00 0.01df5 <- data.frame(psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu Yes",], na.rm = T)$mean,
psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu Yes",], na.rm = T)$sd,
psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu No",], na.rm = T)$mean,
psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu No",], na.rm = T)$sd)
psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu Yes",], na.rm = T)## vars n mean sd median trimmed mad min max range skew kurtosis se
## accessibility 1 2100 5.53 2.04 6 5.70 2.97 1 8 7 -0.50 -0.73 0.04
## understanding 2 2101 5.54 1.97 6 5.70 2.97 1 8 7 -0.51 -0.63 0.04
## empowerment 3 2099 4.47 2.04 4 4.48 2.97 1 8 7 -0.03 -0.87 0.04
## pref_quality_of_evidence 4 2078 4.85 1.82 5 4.86 1.48 1 8 7 -0.06 -0.56 0.04
## s_iv_quality_of_evidence* 5 2112 2.00 0.00 2 2.00 0.00 2 2 0 NaN NaN 0.00psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu No",], na.rm = T)## vars n mean sd median trimmed mad min max range skew kurtosis se
## accessibility 1 2226 5.61 2.11 6 5.83 2.97 1 8 7 -0.60 -0.70 0.04
## understanding 2 2222 5.67 2.06 6 5.89 2.97 1 8 7 -0.65 -0.56 0.04
## empowerment 3 2225 4.51 2.11 5 4.51 2.97 1 8 7 -0.04 -0.93 0.04
## pref_quality_of_evidence 4 2198 4.64 1.77 5 4.62 1.48 1 8 7 0.09 -0.38 0.04
## s_iv_quality_of_evidence* 5 2238 1.00 0.00 1 1.00 0.00 1 1 0 NaN NaN 0.00rownames(df5) <- colnames(df1)
colnames(df5) <- c("Mean_Qu_Yes", "SD_Qu_Yes", "Mean_Qu_No", "SD_Qu_No")
df5 <- df5[1:(nrow(df5)-1),]
df5## Mean_Qu_Yes SD_Qu_Yes Mean_Qu_No SD_Qu_No
## accessibility 5.527143 2.035588 5.613657 2.106570
## understanding 5.536411 1.971278 5.666067 2.057943
## empowerment 4.472130 2.036502 4.508764 2.106226
## pref_quality_of_evidence 4.850337 1.817697 4.637398 1.769103Resilience
names(df_resilience)## [1] "id" "s_study_arm" "s_iv_quality_of_evidence" "s_iv_operationalization"
## [5] "s_iv_technical_terms" "s_int_psy" "s_knowledge_evidence" "pls"
## [9] "accessibility" "understanding" "empowerment" "pref_quality_of_evidence"
## [13] "knowledge_evidence" "knowledge"df1 <- df_resilience[,c("accessibility", "understanding", "empowerment", "pref_quality_of_evidence", "knowledge_evidence","knowledge", "s_iv_quality_of_evidence", "s_iv_technical_terms")]
require(psych)
describe(df1)## vars n mean sd median trimmed mad min max range skew kurtosis se
## accessibility 1 2064 5.01 2.12 5 5.10 2.97 1 8 7 -0.25 -0.99 0.05
## understanding 2 2065 5.08 2.05 5 5.18 2.97 1 8 7 -0.33 -0.84 0.05
## empowerment 3 2064 4.09 2.07 4 4.03 2.97 1 8 7 0.13 -0.91 0.05
## pref_quality_of_evidence 4 2041 4.79 1.71 5 4.79 1.48 1 8 7 0.01 -0.38 0.04
## knowledge_evidence 5 1990 0.42 0.49 0 0.40 0.00 0 1 1 0.33 -1.89 0.01
## knowledge 6 1990 0.65 0.48 1 0.69 0.00 0 1 1 -0.63 -1.61 0.01
## s_iv_quality_of_evidence* 7 2076 1.47 0.50 1 1.47 0.00 1 2 1 0.10 -1.99 0.01
## s_iv_technical_terms* 8 2076 1.97 0.80 2 1.97 1.48 1 3 2 0.05 -1.46 0.02df5 <- data.frame(psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu Yes" & df1$s_iv_technical_terms == "Expl Glossar",], na.rm = T)$mean,
psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu Yes" & df1$s_iv_technical_terms == "Expl Glossar",], na.rm = T)$sd,
psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu Yes" & df1$s_iv_technical_terms == "Replace",], na.rm = T)$mean,
psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu Yes" & df1$s_iv_technical_terms == "Replace",], na.rm = T)$sd,
psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu Yes" & df1$s_iv_technical_terms == "Expl None",], na.rm = T)$mean,
psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu Yes" & df1$s_iv_technical_terms == "Expl None",], na.rm = T)$sd,
psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu No" & df1$s_iv_technical_terms == "Expl Glossar",], na.rm = T)$mean,
psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu No" & df1$s_iv_technical_terms == "Expl Glossar",], na.rm = T)$sd,
psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu No" & df1$s_iv_technical_terms == "Replace",], na.rm = T)$mean,
psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu No" & df1$s_iv_technical_terms == "Replace",], na.rm = T)$sd,
psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu No" & df1$s_iv_technical_terms == "Expl None",], na.rm = T)$mean,
psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu No" & df1$s_iv_technical_terms == "Expl None",], na.rm = T)$sd)
rownames(df5) <- colnames(df1)
colnames(df5) <- c("Mean_Qu_Yes_Tech_Glo", "SD_Qu_Yes_Tech_Glo","Mean_Qu_Yes_Tech_Repl", "SD_Qu_Yes_Tech_Repl", "Mean_Qu_Yes_Tech_None", "SD_Qu_Yes_Tech_None", "Mean_Qu_No_Tech_Glo", "SD_Qu_No_Tech_Glo","Mean_Qu_No_Tech_Repl", "SD_Qu_No_Tech_Repl", "Mean_Qu_No_Tech_None", "SD_Qu_No_Tech_None")
df5 <- df5[1:(nrow(df5)-2),]
df5## Mean_Qu_Yes_Tech_Glo SD_Qu_Yes_Tech_Glo Mean_Qu_Yes_Tech_Repl SD_Qu_Yes_Tech_Repl
## accessibility 4.9036545 2.0576734 5.2168675 2.1062313
## understanding 5.0897010 1.9668741 5.2462462 1.9801787
## empowerment 4.2033333 2.0237690 4.2349398 2.0680819
## pref_quality_of_evidence 5.0872483 1.6714392 4.8851964 1.7885473
## knowledge_evidence 0.4265734 0.4954461 0.4923077 0.5007117
## knowledge 0.6503497 0.4776954 0.6584615 0.4749569
## Mean_Qu_Yes_Tech_None SD_Qu_Yes_Tech_None Mean_Qu_No_Tech_Glo SD_Qu_No_Tech_Glo
## accessibility 4.7188406 2.1004758 5.3274854 2.0358385
## understanding 4.6416185 2.0652829 5.5877193 1.9198519
## empowerment 3.7745665 2.0089771 4.3918129 2.0171530
## pref_quality_of_evidence 4.6501458 1.7770610 5.0029940 1.6004851
## knowledge_evidence 0.4700599 0.4998516 0.4000000 0.4906534
## knowledge 0.6197605 0.4861739 0.5938462 0.4918712
## Mean_Qu_No_Tech_Repl SD_Qu_No_Tech_Repl Mean_Qu_No_Tech_None SD_Qu_No_Tech_None
## accessibility 5.2208122 2.1064480 4.6657143 2.1939540
## understanding 5.3274112 2.0407826 4.5902579 2.1188853
## empowerment 4.3832487 2.0878477 3.5285714 2.0363560
## pref_quality_of_evidence 4.7046632 1.6279472 4.4899713 1.7113674
## knowledge_evidence 0.3884514 0.4880390 0.3421829 0.4751418
## knowledge 0.7139108 0.4525259 0.6519174 0.4770665psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu Yes" & df1$s_iv_technical_terms == "Expl Glossar",], na.rm = T)## vars n mean sd median trimmed mad min max range skew kurtosis se
## accessibility 1 301 4.90 2.06 5 4.97 2.97 1 8 7 -0.21 -0.94 0.12
## understanding 2 301 5.09 1.97 5 5.17 1.48 1 8 7 -0.29 -0.79 0.11
## empowerment 3 300 4.20 2.02 4 4.19 2.97 1 8 7 0.01 -0.91 0.12
## pref_quality_of_evidence 4 298 5.09 1.67 5 5.07 1.48 1 8 7 0.06 -0.69 0.10
## knowledge_evidence 5 286 0.43 0.50 0 0.41 0.00 0 1 1 0.30 -1.92 0.03
## knowledge 6 286 0.65 0.48 1 0.69 0.00 0 1 1 -0.63 -1.61 0.03
## s_iv_quality_of_evidence* 7 301 2.00 0.00 2 2.00 0.00 2 2 0 NaN NaN 0.00
## s_iv_technical_terms* 8 301 3.00 0.00 3 3.00 0.00 3 3 0 NaN NaN 0.00psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu Yes" & df1$s_iv_technical_terms == "Replace",], na.rm = T)## vars n mean sd median trimmed mad min max range skew kurtosis se
## accessibility 1 332 5.22 2.11 5.5 5.36 2.22 1 8 7 -0.43 -0.83 0.12
## understanding 2 333 5.25 1.98 6.0 5.37 1.48 1 8 7 -0.42 -0.73 0.11
## empowerment 3 332 4.23 2.07 4.0 4.22 2.97 1 8 7 0.04 -0.99 0.11
## pref_quality_of_evidence 4 331 4.89 1.79 5.0 4.91 1.48 1 8 7 -0.11 -0.55 0.10
## knowledge_evidence 5 325 0.49 0.50 0.0 0.49 0.00 0 1 1 0.03 -2.01 0.03
## knowledge 6 325 0.66 0.47 1.0 0.70 0.00 0 1 1 -0.67 -1.56 0.03
## s_iv_quality_of_evidence* 7 335 2.00 0.00 2.0 2.00 0.00 2 2 0 NaN NaN 0.00
## s_iv_technical_terms* 8 335 2.00 0.00 2.0 2.00 0.00 2 2 0 NaN NaN 0.00psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu Yes" & df1$s_iv_technical_terms == "Expl None",], na.rm = T)## vars n mean sd median trimmed mad min max range skew kurtosis se
## accessibility 1 345 4.72 2.10 5 4.76 2.97 1 8 7 -0.06 -0.98 0.11
## understanding 2 346 4.64 2.07 5 4.67 2.97 1 8 7 -0.10 -0.89 0.11
## empowerment 3 346 3.77 2.01 4 3.66 1.48 1 8 7 0.28 -0.74 0.11
## pref_quality_of_evidence 4 343 4.65 1.78 5 4.66 1.48 1 8 7 -0.01 -0.53 0.10
## knowledge_evidence 5 334 0.47 0.50 0 0.46 0.00 0 1 1 0.12 -1.99 0.03
## knowledge 6 334 0.62 0.49 1 0.65 0.00 0 1 1 -0.49 -1.76 0.03
## s_iv_quality_of_evidence* 7 348 2.00 0.00 2 2.00 0.00 2 2 0 NaN NaN 0.00
## s_iv_technical_terms* 8 348 1.00 0.00 1 1.00 0.00 1 1 0 NaN NaN 0.00psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu No" & df1$s_iv_technical_terms == "Expl Glossar",], na.rm = T)## vars n mean sd median trimmed mad min max range skew kurtosis se
## accessibility 1 342 5.33 2.04 5 5.47 2.97 1 8 7 -0.38 -0.78 0.11
## understanding 2 342 5.59 1.92 6 5.76 1.48 1 8 7 -0.58 -0.40 0.10
## empowerment 3 342 4.39 2.02 4 4.41 2.97 1 8 7 -0.05 -0.88 0.11
## pref_quality_of_evidence 4 334 5.00 1.60 5 4.99 1.48 1 8 7 0.02 -0.19 0.09
## knowledge_evidence 5 325 0.40 0.49 0 0.38 0.00 0 1 1 0.41 -1.84 0.03
## knowledge 6 325 0.59 0.49 1 0.62 0.00 0 1 1 -0.38 -1.86 0.03
## s_iv_quality_of_evidence* 7 343 1.00 0.00 1 1.00 0.00 1 1 0 NaN NaN 0.00
## s_iv_technical_terms* 8 343 3.00 0.00 3 3.00 0.00 3 3 0 NaN NaN 0.00psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu No" & df1$s_iv_technical_terms == "Replace",], na.rm = T)## vars n mean sd median trimmed mad min max range skew kurtosis se
## accessibility 1 394 5.22 2.11 6 5.34 2.97 1 8 7 -0.34 -1.02 0.11
## understanding 2 394 5.33 2.04 6 5.47 2.97 1 8 7 -0.45 -0.76 0.10
## empowerment 3 394 4.38 2.09 4 4.36 2.97 1 8 7 0.02 -0.91 0.11
## pref_quality_of_evidence 4 386 4.70 1.63 5 4.67 1.48 1 8 7 0.13 -0.27 0.08
## knowledge_evidence 5 381 0.39 0.49 0 0.36 0.00 0 1 1 0.46 -1.80 0.03
## knowledge 6 381 0.71 0.45 1 0.77 0.00 0 1 1 -0.94 -1.11 0.02
## s_iv_quality_of_evidence* 7 395 1.00 0.00 1 1.00 0.00 1 1 0 NaN NaN 0.00
## s_iv_technical_terms* 8 395 2.00 0.00 2 2.00 0.00 2 2 0 NaN NaN 0.00psych::describe(df1[df1$s_iv_quality_of_evidence == "Qu No" & df1$s_iv_technical_terms == "Expl None",], na.rm = T)## vars n mean sd median trimmed mad min max range skew kurtosis se
## accessibility 1 350 4.67 2.19 5 4.70 2.97 1 8 7 -0.08 -1.19 0.12
## understanding 2 349 4.59 2.12 5 4.62 2.97 1 8 7 -0.09 -1.10 0.11
## empowerment 3 350 3.53 2.04 3 3.35 2.97 1 8 7 0.47 -0.70 0.11
## pref_quality_of_evidence 4 349 4.49 1.71 4 4.48 1.48 1 8 7 0.07 -0.18 0.09
## knowledge_evidence 5 339 0.34 0.48 0 0.30 0.00 0 1 1 0.66 -1.57 0.03
## knowledge 6 339 0.65 0.48 1 0.69 0.00 0 1 1 -0.64 -1.60 0.03
## s_iv_quality_of_evidence* 7 354 1.00 0.00 1 1.00 0.00 1 1 0 NaN NaN 0.00
## s_iv_technical_terms* 8 354 1.00 0.00 1 1.00 0.00 1 1 0 NaN NaN 0.00#Correlations
cor(df_resilience$understanding, df_resilience$accessibility, use = "pairwise.complete.obs")## [1] 0.8323697cor(df_resilience$understanding, df_resilience$empowerment, use = "pairwise.complete.obs")## [1] 0.7123068cor(df_resilience$accessibility, df_resilience$empowerment, use = "pairwise.complete.obs")## [1] 0.6439495df1 <- df_resilience[,c("accessibility", "understanding", "empowerment", "pref_quality_of_evidence", "s_iv_technical_terms")]
require(psych)
describe(df1)## vars n mean sd median trimmed mad min max range skew kurtosis se
## accessibility 1 2064 5.01 2.12 5 5.10 2.97 1 8 7 -0.25 -0.99 0.05
## understanding 2 2065 5.08 2.05 5 5.18 2.97 1 8 7 -0.33 -0.84 0.05
## empowerment 3 2064 4.09 2.07 4 4.03 2.97 1 8 7 0.13 -0.91 0.05
## pref_quality_of_evidence 4 2041 4.79 1.71 5 4.79 1.48 1 8 7 0.01 -0.38 0.04
## s_iv_technical_terms* 5 2076 1.97 0.80 2 1.97 1.48 1 3 2 0.05 -1.46 0.02df5 <- data.frame(psych::describe(df1[df1$s_iv_technical_terms == "Expl Glossar",], na.rm = T)$mean,
psych::describe(df1[df1$s_iv_technical_terms == "Expl Glossar",], na.rm = T)$sd,
psych::describe(df1[df1$s_iv_technical_terms == "Replace",], na.rm = T)$mean,
psych::describe(df1[df1$s_iv_technical_terms == "Replace",], na.rm = T)$sd,
psych::describe(df1[df1$s_iv_technical_terms == "Expl None",], na.rm = T)$mean,
psych::describe(df1[df1$s_iv_technical_terms == "Expl None",], na.rm = T)$sd)
psych::describe(df1[df1$s_iv_technical_terms == "Expl Glossar",], na.rm = T)## vars n mean sd median trimmed mad min max range skew kurtosis se
## accessibility 1 643 5.13 2.06 5 5.24 2.97 1 8 7 -0.30 -0.86 0.08
## understanding 2 643 5.35 1.96 6 5.50 1.48 1 8 7 -0.44 -0.63 0.08
## empowerment 3 642 4.30 2.02 4 4.30 2.97 1 8 7 -0.02 -0.89 0.08
## pref_quality_of_evidence 4 632 5.04 1.63 5 5.02 1.48 1 8 7 0.04 -0.43 0.06
## s_iv_technical_terms* 5 644 3.00 0.00 3 3.00 0.00 3 3 0 NaN NaN 0.00psych::describe(df1[df1$s_iv_technical_terms == "Replace",], na.rm = T)## vars n mean sd median trimmed mad min max range skew kurtosis se
## accessibility 1 726 5.22 2.10 6 5.35 2.97 1 8 7 -0.38 -0.93 0.08
## understanding 2 727 5.29 2.01 6 5.42 2.97 1 8 7 -0.44 -0.75 0.07
## empowerment 3 726 4.32 2.08 4 4.29 2.97 1 8 7 0.03 -0.94 0.08
## pref_quality_of_evidence 4 717 4.79 1.71 5 4.78 1.48 1 8 7 0.02 -0.42 0.06
## s_iv_technical_terms* 5 730 2.00 0.00 2 2.00 0.00 2 2 0 NaN NaN 0.00psych::describe(df1[df1$s_iv_technical_terms == "Expl None",], na.rm = T)## vars n mean sd median trimmed mad min max range skew kurtosis se
## accessibility 1 695 4.69 2.15 5 4.73 2.97 1 8 7 -0.07 -1.09 0.08
## understanding 2 695 4.62 2.09 5 4.64 2.97 1 8 7 -0.10 -0.99 0.08
## empowerment 3 696 3.65 2.03 4 3.51 2.97 1 8 7 0.37 -0.73 0.08
## pref_quality_of_evidence 4 692 4.57 1.74 4 4.57 1.48 1 8 7 0.04 -0.36 0.07
## s_iv_technical_terms* 5 702 1.00 0.00 1 1.00 0.00 1 1 0 NaN NaN 0.00rownames(df5) <- colnames(df1)
colnames(df5) <- c("Mean_Tech_Glo", "SD_Tech_Glo","Mean_Tech_Repl", "SD_Tech_Repl", "Mean_Tech_None", "SD_Tech_None")
df5 <- df5[1:(nrow(df5)-1),]
df5## Mean_Tech_Glo SD_Tech_Glo Mean_Tech_Repl SD_Tech_Repl Mean_Tech_None SD_Tech_None
## accessibility 5.129082 2.055418 5.219008 2.104897 4.692086 2.146677
## understanding 5.354588 1.956360 5.290234 2.012277 4.615827 2.091022
## empowerment 4.303738 2.020863 4.315427 2.078716 3.650862 2.025080
## pref_quality_of_evidence 5.042722 1.633566 4.788006 1.705142 4.569364 1.744817Confirmatory Analyses
Load Packages
require(lme4)## Lade nötiges Paket: lme4## Lade nötiges Paket: Matrixrequire(lmerTest)## Lade nötiges Paket: lmerTest##
## Attache Paket: 'lmerTest'## Das folgende Objekt ist maskiert 'package:lme4':
##
## lmer## Das folgende Objekt ist maskiert 'package:stats':
##
## steprequire(sjstats)## Lade nötiges Paket: sjstats## Registered S3 methods overwritten by 'parameters':
## method from
## as.double.parameters_kurtosis datawizard
## as.double.parameters_skewness datawizard
## as.double.parameters_smoothness datawizard
## as.numeric.parameters_kurtosis datawizard
## as.numeric.parameters_skewness datawizard
## as.numeric.parameters_smoothness datawizard
## print.parameters_distribution datawizard
## print.parameters_kurtosis datawizard
## print.parameters_skewness datawizard
## summary.parameters_kurtosis datawizard
## summary.parameters_skewness datawizard##
## Attache Paket: 'sjstats'## Das folgende Objekt ist maskiert 'package:psych':
##
## phirequire(MuMIn)## Lade nötiges Paket: MuMInlibrary(multcomp)## Lade nötiges Paket: mvtnorm## Lade nötiges Paket: survival## Lade nötiges Paket: TH.data## Lade nötiges Paket: MASS##
## Attache Paket: 'TH.data'## Das folgende Objekt ist maskiert 'package:MASS':
##
## geyserlibrary(r2glmm)
library(performance)##
## Attache Paket: 'performance'## Die folgenden Objekte sind maskiert von 'package:sjstats':
##
## icc, r2require(effectsize)## Lade nötiges Paket: effectsize##
## Attache Paket: 'effectsize'## Die folgenden Objekte sind maskiert von 'package:sjstats':
##
## cohens_f, phi## Das folgende Objekt ist maskiert 'package:psych':
##
## phiH1 and H2 Technical Terms
A Accessibility
m1 <- lmer(accessibility ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id), data = df_resilience)
summary(m1)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: accessibility ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id)
## Data: df_resilience
##
## REML criterion at convergence: 8442.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.02541 -0.48177 0.03719 0.52881 2.87533
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.675 1.636
## Residual 1.673 1.293
## Number of obs: 2064, groups: id, 1101
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.01046 0.11752 1228.90462 42.637 < 2e-16 ***
## s_iv_technical_termsReplace 0.54451 0.13803 1091.60442 3.945 8.49e-05 ***
## s_iv_technical_termsExpl Glossar 0.43651 0.14213 1093.93513 3.071 0.00218 **
## s_iv_quality_of_evidenceQu Yes -0.09228 0.11457 1093.37180 -0.806 0.42070
## plsGroth -0.55018 0.05815 1011.55853 -9.462 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v__R s___EG s____Y
## s_v_tchnc_R -0.615
## s_v_tchn_EG -0.594 0.497
## s_v_qlt__QY -0.484 0.032 0.023
## plsGroth -0.247 -0.002 -0.002 0.003test = glht(m1,linfct=mcp(s_iv_technical_terms="Tukey"))
summary(test, test = adjusted("none"))##
## Simultaneous Tests for General Linear Hypotheses
##
## Multiple Comparisons of Means: Tukey Contrasts
##
##
## Fit: lmer(formula = accessibility ~ s_iv_technical_terms + s_iv_quality_of_evidence +
## pls + (1 | id), data = df_resilience)
##
## Linear Hypotheses:
## Estimate Std. Error z value Pr(>|z|)
## Replace - Expl None == 0 0.5445 0.1380 3.945 7.98e-05 ***
## Expl Glossar - Expl None == 0 0.4365 0.1421 3.071 0.00213 **
## Expl Glossar - Replace == 0 -0.1080 0.1405 -0.769 0.44218
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- none method)summary(test, test = adjusted("BH"))##
## Simultaneous Tests for General Linear Hypotheses
##
## Multiple Comparisons of Means: Tukey Contrasts
##
##
## Fit: lmer(formula = accessibility ~ s_iv_technical_terms + s_iv_quality_of_evidence +
## pls + (1 | id), data = df_resilience)
##
## Linear Hypotheses:
## Estimate Std. Error z value Pr(>|z|)
## Replace - Expl None == 0 0.5445 0.1380 3.945 0.000239 ***
## Expl Glossar - Expl None == 0 0.4365 0.1421 3.071 0.003197 **
## Expl Glossar - Replace == 0 -0.1080 0.1405 -0.769 0.442179
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- BH method)r.squaredGLMM(m1)## Warning: 'r.squaredGLMM' now calculates a revised statistic. See the help page.## R2m R2c
## [1,] 0.02999658 0.626808r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.030 0.047 0.018
## 5 plsGroth 0.017 0.030 0.008
## 2 s_iv_technical_termsReplace 0.012 0.022 0.004
## 3 s_iv_technical_termsExpl Glossar 0.007 0.016 0.002
## 4 s_iv_quality_of_evidenceQu Yes 0.000 0.004 0.000anova_stats(m1)## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## term | sumsq | meansq | NumDF | DenDF | statistic | p.value | df | etasq | partial.etasq | omegasq | partial.omegasq | epsilonsq | cohens.f | power
## --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
## s_iv_technical_terms | 28.776 | 14.388 | 2 | 1093.597 | 8.600 | < .001 | | 0.013 | 0.014 | | | | 0.118 |
## s_iv_quality_of_evidence | 1.085 | 1.085 | 1 | 1093.372 | 0.649 | 0.421 | | 0.000 | 0.001 | | | | 0.023 |
## pls | 149.764 | 149.764 | 1 | 1011.559 | 89.521 | < .001 | | 0.066 | 0.067 | | | | 0.268 |
## Residuals | 2080.261 | 1.008 | | | | | 2064 | | | | | | |anova_stats(m1)$term## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "s_iv_technical_terms" "s_iv_quality_of_evidence" "pls" "Residuals"interpret_omega_squared(anova_stats(m1)$partial.etasq, rules = "field2013")## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "small" "very small" "medium" NA
## (Rules: field2013)# Interest in psychological research
m1 <- lmer(accessibility ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id), data = df_resilience[!is.na(df_resilience$s_int_psy),])
summary(m1)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: accessibility ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id)
## Data: df_resilience[!is.na(df_resilience$s_int_psy), ]
##
## REML criterion at convergence: 8442.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.02541 -0.48177 0.03719 0.52881 2.87533
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.675 1.636
## Residual 1.673 1.293
## Number of obs: 2064, groups: id, 1101
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.01046 0.11752 1228.90462 42.637 < 2e-16 ***
## s_iv_technical_termsReplace 0.54451 0.13803 1091.60442 3.945 8.49e-05 ***
## s_iv_technical_termsExpl Glossar 0.43651 0.14213 1093.93513 3.071 0.00218 **
## s_iv_quality_of_evidenceQu Yes -0.09228 0.11457 1093.37180 -0.806 0.42070
## plsGroth -0.55018 0.05815 1011.55853 -9.462 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v__R s___EG s____Y
## s_v_tchnc_R -0.615
## s_v_tchn_EG -0.594 0.497
## s_v_qlt__QY -0.484 0.032 0.023
## plsGroth -0.247 -0.002 -0.002 0.003r.squaredGLMM(m1)## R2m R2c
## [1,] 0.02999658 0.626808r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.030 0.047 0.018
## 5 plsGroth 0.017 0.030 0.008
## 2 s_iv_technical_termsReplace 0.012 0.022 0.004
## 3 s_iv_technical_termsExpl Glossar 0.007 0.016 0.002
## 4 s_iv_quality_of_evidenceQu Yes 0.000 0.004 0.000m6 <- lmer(accessibility ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id) + s_int_psy, data = df_resilience[!is.na(df_resilience$s_int_psy),])
summary(m6)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: accessibility ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id) + s_int_psy
## Data: df_resilience[!is.na(df_resilience$s_int_psy), ]
##
## REML criterion at convergence: 8310.8
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.2278 -0.4836 0.0575 0.5139 2.7345
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.250 1.500
## Residual 1.677 1.295
## Number of obs: 2064, groups: id, 1101
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.01966 0.11080 1240.12256 45.303 < 2e-16 ***
## s_iv_technical_termsReplace 0.54343 0.12961 1085.34403 4.193 2.98e-05 ***
## s_iv_technical_termsExpl Glossar 0.45064 0.13348 1087.80582 3.376 0.000761 ***
## s_iv_quality_of_evidenceQu Yes -0.10679 0.10759 1087.13307 -0.993 0.321159
## plsGroth -0.54654 0.05814 1012.22283 -9.401 < 2e-16 ***
## s_int_psy 0.64344 0.05346 1099.02550 12.035 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v__R s___EG s____Y plsGrt
## s_v_tchnc_R -0.613
## s_v_tchn_EG -0.591 0.497
## s_v_qlt__QY -0.483 0.032 0.023
## plsGroth -0.262 -0.002 -0.002 0.003
## s_int_psy 0.006 0.000 0.009 -0.010 0.007r.squaredGLMM(m6)## R2m R2c
## [1,] 0.1230622 0.6255119r2beta(m6, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.123 0.151 0.100
## 6 s_int_psy 0.095 0.120 0.073
## 5 plsGroth 0.019 0.032 0.009
## 2 s_iv_technical_termsReplace 0.013 0.024 0.005
## 3 s_iv_technical_termsExpl Glossar 0.008 0.018 0.002
## 4 s_iv_quality_of_evidenceQu Yes 0.001 0.005 0.000m7 <- lmer(accessibility ~ s_iv_technical_terms * s_int_psy + s_iv_quality_of_evidence + pls + (1 | id), data = df_resilience[!is.na(df_resilience$s_int_psy),])
summary(m7)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: accessibility ~ s_iv_technical_terms * s_int_psy + s_iv_quality_of_evidence + pls + (1 | id)
## Data: df_resilience[!is.na(df_resilience$s_int_psy), ]
##
## REML criterion at convergence: 8314.2
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.2039 -0.4859 0.0552 0.5125 2.7195
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.253 1.501
## Residual 1.677 1.295
## Number of obs: 2064, groups: id, 1101
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.02033 0.11086 1237.82813 45.287 < 2e-16 ***
## s_iv_technical_termsReplace 0.54288 0.12967 1083.60834 4.187 3.06e-05 ***
## s_iv_technical_termsExpl Glossar 0.44795 0.13356 1085.63521 3.354 0.000824 ***
## s_int_psy 0.71876 0.09135 1087.96312 7.868 8.63e-15 ***
## s_iv_quality_of_evidenceQu Yes -0.10832 0.10767 1085.51069 -1.006 0.314656
## plsGroth -0.54548 0.05814 1012.26696 -9.383 < 2e-16 ***
## s_iv_technical_termsReplace:s_int_psy -0.08362 0.12748 1091.89295 -0.656 0.511998
## s_iv_technical_termsExpl Glossar:s_int_psy -0.15271 0.13432 1100.36284 -1.137 0.255820
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v__R s_v__EG s_nt_p s____Y plsGrt s___R:
## s_v_tchnc_R -0.613
## s_v_tchn_EG -0.591 0.497
## s_int_psy 0.012 -0.005 -0.005
## s_v_qlt__QY -0.483 0.032 0.023 -0.023
## plsGroth -0.262 -0.002 -0.002 0.015 0.003
## s_v_tc_R:__ -0.013 0.007 0.003 -0.717 0.025 -0.009
## s_v_t_EG:__ -0.004 0.003 0.018 -0.680 0.010 -0.016 0.487r.squaredGLMM(m7)## R2m R2c
## [1,] 0.1238579 0.626189r2beta(m7, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.124 0.153 0.101
## 4 s_int_psy 0.043 0.062 0.028
## 6 plsGroth 0.019 0.032 0.009
## 2 s_iv_technical_termsReplace 0.013 0.024 0.005
## 3 s_iv_technical_termsExpl Glossar 0.008 0.018 0.002
## 8 s_iv_technical_termsExpl Glossar:s_int_psy 0.001 0.005 0.000
## 5 s_iv_quality_of_evidenceQu Yes 0.001 0.005 0.000
## 7 s_iv_technical_termsReplace:s_int_psy 0.000 0.004 0.000anova(m1,m6, m7)## refitting model(s) with ML (instead of REML)## Data: df_resilience[!is.na(df_resilience$s_int_psy), ]
## Models:
## m1: accessibility ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id)
## m6: accessibility ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id) + s_int_psy
## m7: accessibility ~ s_iv_technical_terms * s_int_psy + s_iv_quality_of_evidence + pls + (1 | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## m1 7 8441.9 8481.3 -4214.0 8427.9
## m6 8 8307.6 8352.6 -4145.8 8291.6 136.3477 1 <2e-16 ***
## m7 10 8310.3 8366.6 -4145.1 8290.3 1.3127 2 0.5187
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1#ICC Subject
m0 <- lmer(accessibility ~ (1 | id), data = df_resilience)
summary(m0)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: accessibility ~ (1 | id)
## Data: df_resilience
##
## REML criterion at convergence: 8535.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.69336 -0.55614 0.08356 0.56335 2.49701
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.658 1.630
## Residual 1.819 1.349
## Number of obs: 2064, groups: id, 1101
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.019e+00 5.766e-02 1.095e+03 87.05 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1performance::icc(m0)## # Intraclass Correlation Coefficient
##
## Adjusted ICC: 0.594
## Conditional ICC: 0.594B Understanding
m1 <- lmer(understanding ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id), data = df_resilience)
summary(m1)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: understanding ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id)
## Data: df_resilience
##
## REML criterion at convergence: 8277.1
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.2210 -0.4793 0.0598 0.5198 3.0099
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.476 1.574
## Residual 1.536 1.239
## Number of obs: 2065, groups: id, 1101
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.9493 0.1130 1230.8055 43.806 < 2e-16 ***
## s_iv_technical_termsReplace 0.6827 0.1326 1091.9235 5.147 3.14e-07 ***
## s_iv_technical_termsExpl Glossar 0.7511 0.1366 1094.7145 5.498 4.77e-08 ***
## s_iv_quality_of_evidenceQu Yes -0.1359 0.1101 1093.7269 -1.235 0.217
## plsGroth -0.5442 0.0557 1012.3304 -9.771 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v__R s___EG s____Y
## s_v_tchnc_R -0.616
## s_v_tchn_EG -0.594 0.497
## s_v_qlt__QY -0.485 0.032 0.024
## plsGroth -0.248 -0.001 -0.001 0.003r.squaredGLMM(m1)## R2m R2c
## [1,] 0.0462724 0.6349152r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.046 0.067 0.032
## 3 s_iv_technical_termsExpl Glossar 0.022 0.036 0.011
## 2 s_iv_technical_termsReplace 0.020 0.033 0.010
## 5 plsGroth 0.018 0.031 0.009
## 4 s_iv_quality_of_evidenceQu Yes 0.001 0.006 0.000test = glht(m1,linfct=mcp(s_iv_technical_terms="Tukey"))
summary(test, test = adjusted("none"))##
## Simultaneous Tests for General Linear Hypotheses
##
## Multiple Comparisons of Means: Tukey Contrasts
##
##
## Fit: lmer(formula = understanding ~ s_iv_technical_terms + s_iv_quality_of_evidence +
## pls + (1 | id), data = df_resilience)
##
## Linear Hypotheses:
## Estimate Std. Error z value Pr(>|z|)
## Replace - Expl None == 0 0.6827 0.1326 5.147 2.65e-07 ***
## Expl Glossar - Expl None == 0 0.7511 0.1366 5.498 3.83e-08 ***
## Expl Glossar - Replace == 0 0.0684 0.1351 0.506 0.613
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- none method)summary(test, test = adjusted("BH"))##
## Simultaneous Tests for General Linear Hypotheses
##
## Multiple Comparisons of Means: Tukey Contrasts
##
##
## Fit: lmer(formula = understanding ~ s_iv_technical_terms + s_iv_quality_of_evidence +
## pls + (1 | id), data = df_resilience)
##
## Linear Hypotheses:
## Estimate Std. Error z value Pr(>|z|)
## Replace - Expl None == 0 0.6827 0.1326 5.147 3.98e-07 ***
## Expl Glossar - Expl None == 0 0.7511 0.1366 5.498 1.15e-07 ***
## Expl Glossar - Replace == 0 0.0684 0.1351 0.506 0.613
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- BH method)anova_stats(m1)## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## term | sumsq | meansq | NumDF | DenDF | statistic | p.value | df | etasq | partial.etasq | omegasq | partial.omegasq | epsilonsq | cohens.f | power
## --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
## s_iv_technical_terms | 58.314 | 29.157 | 2 | 1094.057 | 18.984 | < .001 | | 0.028 | 0.030 | | | | 0.175 |
## s_iv_quality_of_evidence | 2.342 | 2.342 | 1 | 1093.727 | 1.525 | 0.217 | | 0.001 | 0.001 | | | | 0.035 |
## pls | 146.629 | 146.629 | 1 | 1012.330 | 95.468 | < .001 | | 0.069 | 0.071 | | | | 0.277 |
## Residuals | 1908.572 | 0.924 | | | | | 2065 | | | | | | |anova_stats(m1)$term## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "s_iv_technical_terms" "s_iv_quality_of_evidence" "pls" "Residuals"interpret_omega_squared(anova_stats(m1)$partial.etasq, rules = "field2013")## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "small" "very small" "medium" NA
## (Rules: field2013)# Interest in psychological research
m1 <- lmer(understanding ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id), data = df_resilience[!is.na(df_resilience$s_int_psy),])
summary(m1)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: understanding ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id)
## Data: df_resilience[!is.na(df_resilience$s_int_psy), ]
##
## REML criterion at convergence: 8277.1
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.2210 -0.4793 0.0598 0.5198 3.0099
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.476 1.574
## Residual 1.536 1.239
## Number of obs: 2065, groups: id, 1101
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.9493 0.1130 1230.8055 43.806 < 2e-16 ***
## s_iv_technical_termsReplace 0.6827 0.1326 1091.9235 5.147 3.14e-07 ***
## s_iv_technical_termsExpl Glossar 0.7511 0.1366 1094.7145 5.498 4.77e-08 ***
## s_iv_quality_of_evidenceQu Yes -0.1359 0.1101 1093.7269 -1.235 0.217
## plsGroth -0.5442 0.0557 1012.3304 -9.771 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v__R s___EG s____Y
## s_v_tchnc_R -0.616
## s_v_tchn_EG -0.594 0.497
## s_v_qlt__QY -0.485 0.032 0.024
## plsGroth -0.248 -0.001 -0.001 0.003r.squaredGLMM(m1)## R2m R2c
## [1,] 0.0462724 0.6349152r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.046 0.067 0.032
## 3 s_iv_technical_termsExpl Glossar 0.022 0.036 0.011
## 2 s_iv_technical_termsReplace 0.020 0.033 0.010
## 5 plsGroth 0.018 0.031 0.009
## 4 s_iv_quality_of_evidenceQu Yes 0.001 0.006 0.000m6 <- lmer(understanding ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id) + s_int_psy, data = df_resilience[!is.na(df_resilience$s_int_psy),])
summary(m6)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: understanding ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id) + s_int_psy
## Data: df_resilience[!is.na(df_resilience$s_int_psy), ]
##
## REML criterion at convergence: 8090.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.2386 -0.4818 0.0590 0.5118 3.1263
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.944 1.394
## Residual 1.538 1.240
## Number of obs: 2065, groups: id, 1101
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.95949 0.10411 1250.05216 47.635 < 2e-16 ***
## s_iv_technical_termsReplace 0.68157 0.12153 1086.45097 5.608 2.59e-08 ***
## s_iv_technical_termsExpl Glossar 0.76595 0.12517 1089.51371 6.119 1.31e-09 ***
## s_iv_quality_of_evidenceQu Yes -0.15405 0.10088 1088.35118 -1.527 0.127
## plsGroth -0.53865 0.05563 1016.09681 -9.683 < 2e-16 ***
## s_int_psy 0.72298 0.05012 1099.54000 14.425 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v__R s___EG s____Y plsGrt
## s_v_tchnc_R -0.612
## s_v_tchn_EG -0.590 0.497
## s_v_qlt__QY -0.482 0.032 0.024
## plsGroth -0.268 -0.001 -0.001 0.003
## s_int_psy 0.006 0.000 0.009 -0.010 0.007r.squaredGLMM(m6)## R2m R2c
## [1,] 0.1715256 0.6340409r2beta(m6, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.172 0.202 0.145
## 6 s_int_psy 0.131 0.158 0.105
## 3 s_iv_technical_termsExpl Glossar 0.027 0.042 0.015
## 2 s_iv_technical_termsReplace 0.022 0.037 0.012
## 5 plsGroth 0.020 0.034 0.010
## 4 s_iv_quality_of_evidenceQu Yes 0.002 0.007 0.000m7 <- lmer(understanding ~ s_iv_technical_terms * s_int_psy + s_iv_quality_of_evidence + pls + (1 | id), data = df_resilience[!is.na(df_resilience$s_int_psy),])
summary(m7)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: understanding ~ s_iv_technical_terms * s_int_psy + s_iv_quality_of_evidence + pls + (1 | id)
## Data: df_resilience[!is.na(df_resilience$s_int_psy), ]
##
## REML criterion at convergence: 8094.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.2463 -0.4891 0.0612 0.5049 3.0876
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.947 1.395
## Residual 1.538 1.240
## Number of obs: 2065, groups: id, 1101
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.959e+00 1.042e-01 1.248e+03 47.609 < 2e-16 ***
## s_iv_technical_termsReplace 6.817e-01 1.216e-01 1.085e+03 5.607 2.61e-08 ***
## s_iv_technical_termsExpl Glossar 7.632e-01 1.252e-01 1.087e+03 6.094 1.53e-09 ***
## s_int_psy 7.590e-01 8.561e-02 1.088e+03 8.866 < 2e-16 ***
## s_iv_quality_of_evidenceQu Yes -1.537e-01 1.010e-01 1.087e+03 -1.522 0.128
## plsGroth -5.378e-01 5.563e-02 1.016e+03 -9.668 < 2e-16 ***
## s_iv_technical_termsReplace:s_int_psy 5.865e-03 1.195e-01 1.092e+03 0.049 0.961
## s_iv_technical_termsExpl Glossar:s_int_psy -1.294e-01 1.259e-01 1.102e+03 -1.028 0.304
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v__R s_v__EG s_nt_p s____Y plsGrt s___R:
## s_v_tchnc_R -0.612
## s_v_tchn_EG -0.590 0.497
## s_int_psy 0.011 -0.004 -0.004
## s_v_qlt__QY -0.482 0.033 0.024 -0.023
## plsGroth -0.268 -0.001 -0.001 0.015 0.003
## s_v_tc_R:__ -0.013 0.006 0.003 -0.717 0.025 -0.008
## s_v_t_EG:__ -0.004 0.003 0.018 -0.680 0.011 -0.015 0.487r.squaredGLMM(m7)## R2m R2c
## [1,] 0.1722773 0.6347116r2beta(m7, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.172 0.203 0.147
## 4 s_int_psy 0.054 0.074 0.037
## 3 s_iv_technical_termsExpl Glossar 0.026 0.041 0.014
## 2 s_iv_technical_termsReplace 0.022 0.037 0.012
## 6 plsGroth 0.020 0.034 0.010
## 5 s_iv_quality_of_evidenceQu Yes 0.002 0.007 0.000
## 8 s_iv_technical_termsExpl Glossar:s_int_psy 0.001 0.005 0.000
## 7 s_iv_technical_termsReplace:s_int_psy 0.000 0.002 0.000anova(m1,m6, m7)## refitting model(s) with ML (instead of REML)## Data: df_resilience[!is.na(df_resilience$s_int_psy), ]
## Models:
## m1: understanding ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id)
## m6: understanding ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id) + s_int_psy
## m7: understanding ~ s_iv_technical_terms * s_int_psy + s_iv_quality_of_evidence + pls + (1 | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## m1 7 8276.0 8315.4 -4131.0 8262.0
## m6 8 8086.9 8132.0 -4035.5 8070.9 191.046 1 <2e-16 ***
## m7 10 8089.5 8145.8 -4034.7 8069.5 1.459 2 0.4822
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1#ICC Subject
m0 <- lmer(understanding ~ (1 | id), data = df_resilience)
summary(m0)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: understanding ~ (1 | id)
## Data: df_resilience
##
## REML criterion at convergence: 8396.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.81113 -0.49898 0.07828 0.56346 2.58388
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.512 1.585
## Residual 1.683 1.297
## Number of obs: 2065, groups: id, 1101
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.087e+00 5.589e-02 1.096e+03 91.02 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1performance::icc(m0)## # Intraclass Correlation Coefficient
##
## Adjusted ICC: 0.599
## Conditional ICC: 0.599C Knowledge
table(df_resilience$knowledge, df_resilience$s_iv_technical_terms)/sum(table(df_resilience$knowledge, df_resilience$s_iv_technical_terms)) * 100##
## Expl None Replace Expl Glossar
## 0 12.31156 11.05528 11.65829
## 1 21.50754 24.42211 19.04523table(df_resilience$knowledge[df_resilience$s_iv_technical_terms == "Replace"])/sum(table(df_resilience$knowledge[df_resilience$s_iv_technical_terms == "Replace"])) * 100##
## 0 1
## 31.16147 68.83853table(df_resilience$knowledge[df_resilience$s_iv_technical_terms == "Expl None"])/sum(table(df_resilience$knowledge[df_resilience$s_iv_technical_terms == "Expl None"])) * 100##
## 0 1
## 36.40416 63.59584table(df_resilience$knowledge[df_resilience$s_iv_technical_terms == "Expl Glossar"])/sum(table(df_resilience$knowledge[df_resilience$s_iv_technical_terms == "Expl Glossar"])) * 100##
## 0 1
## 37.97054 62.02946m1 <- glmer(knowledge ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id), family = binomial("logit"), data = df_resilience)
summary(m1)## Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
## Family: binomial ( logit )
## Formula: knowledge ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id)
## Data: df_resilience
##
## AIC BIC logLik deviance df.resid
## 2439.4 2473.0 -1213.7 2427.4 1984
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.8769 -0.8946 0.3655 0.5982 1.1178
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.617 1.272
## Number of obs: 1990, groups: id, 1026
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.41519 0.16374 8.643 <2e-16 ***
## s_iv_technical_termsReplace 0.32753 0.16791 1.951 0.0511 .
## s_iv_technical_termsExpl Glossar -0.10669 0.17057 -0.625 0.5317
## s_iv_quality_of_evidenceQu Yes -0.08229 0.13841 -0.595 0.5521
## plsGroth -1.17890 0.12431 -9.483 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v__R s___EG s____Y
## s_v_tchnc_R -0.473
## s_v_tchn_EG -0.526 0.483
## s_v_qlt__QY -0.437 0.024 0.027
## plsGroth -0.511 -0.055 0.020 0.016test = glht(m1,linfct=mcp(s_iv_technical_terms="Tukey"))
summary(test, test = adjusted("none"))##
## Simultaneous Tests for General Linear Hypotheses
##
## Multiple Comparisons of Means: Tukey Contrasts
##
##
## Fit: glmer(formula = knowledge ~ s_iv_technical_terms + s_iv_quality_of_evidence +
## pls + (1 | id), data = df_resilience, family = binomial("logit"))
##
## Linear Hypotheses:
## Estimate Std. Error z value Pr(>|z|)
## Replace - Expl None == 0 0.3275 0.1679 1.951 0.0511 .
## Expl Glossar - Expl None == 0 -0.1067 0.1706 -0.625 0.5317
## Expl Glossar - Replace == 0 -0.4342 0.1721 -2.523 0.0116 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- none method)summary(test, test = adjusted("BH"))##
## Simultaneous Tests for General Linear Hypotheses
##
## Multiple Comparisons of Means: Tukey Contrasts
##
##
## Fit: glmer(formula = knowledge ~ s_iv_technical_terms + s_iv_quality_of_evidence +
## pls + (1 | id), data = df_resilience, family = binomial("logit"))
##
## Linear Hypotheses:
## Estimate Std. Error z value Pr(>|z|)
## Replace - Expl None == 0 0.3275 0.1679 1.951 0.0767 .
## Expl Glossar - Expl None == 0 -0.1067 0.1706 -0.625 0.5317
## Expl Glossar - Replace == 0 -0.4342 0.1721 -2.523 0.0349 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- BH method)r.squaredGLMM(m1)## Warning: The null model is correct only if all variables used by the original model remain unchanged.## R2m R2c
## theoretical 0.07259137 0.3782418
## delta 0.05996717 0.3124626r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.049 0.071 0.034
## 5 plsGroth 0.045 0.065 0.029
## 2 s_iv_technical_termsReplace 0.002 0.008 0.000
## 3 s_iv_technical_termsExpl Glossar 0.000 0.004 0.000
## 4 s_iv_quality_of_evidenceQu Yes 0.000 0.003 0.000anova_stats(m1)## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: npar## term | npar | sumsq | meansq | statistic | df | etasq | partial.etasq | omegasq | partial.omegasq | epsilonsq | cohens.f | power
## ----------------------------------------------------------------------------------------------------------------------------------------------------------
## s_iv_technical_terms | 2 | 6.301 | 3.150 | 3.150 | | 0.004 | 0.004 | | | | 0.062 |
## s_iv_quality_of_evidence | 1 | 0.349 | 0.349 | 0.349 | | 0.000 | 0.000 | | | | 0.015 |
## pls | 1 | 122.559 | 122.559 | 122.559 | | 0.070 | 0.070 | | | | 0.275 |
## Residuals | | 1625.649 | 0.124 | | 1990 | | | | | | |exp(fixef(m1))## (Intercept) s_iv_technical_termsReplace s_iv_technical_termsExpl Glossar
## 4.1172612 1.3875345 0.8988072
## s_iv_quality_of_evidenceQu Yes plsGroth
## 0.9210017 0.3076168exp(summary(test, test = adjusted("none"))$test$coefficients)## Replace - Expl None Expl Glossar - Expl None Expl Glossar - Replace
## 1.3875345 0.8988072 0.6477729interpret_oddsratio(exp(summary(test, test = adjusted("none"))$test$coefficients), rules = "cohen1988")## Replace - Expl None Expl Glossar - Expl None Expl Glossar - Replace
## "very small" "very small" "small"
## (Rules: cohen1988)# Interest in psychological research
m6 <- glmer(knowledge ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id) + s_int_psy, family = binomial("logit"), data = df_resilience)
summary(m6)## Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
## Family: binomial ( logit )
## Formula: knowledge ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id) + s_int_psy
## Data: df_resilience
##
## AIC BIC logLik deviance df.resid
## 2382.8 2422.0 -1184.4 2368.8 1983
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.3459 -0.7626 0.3522 0.5679 1.4836
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.471 1.213
## Number of obs: 1990, groups: id, 1026
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.42311 0.16264 8.750 < 2e-16 ***
## s_iv_technical_termsReplace 0.31830 0.16581 1.920 0.0549 .
## s_iv_technical_termsExpl Glossar -0.09473 0.16854 -0.562 0.5741
## s_iv_quality_of_evidenceQu Yes -0.08209 0.13675 -0.600 0.5483
## plsGroth -1.19426 0.12499 -9.555 < 2e-16 ***
## s_int_psy 0.52404 0.07329 7.150 8.66e-13 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v__R s___EG s____Y plsGrt
## s_v_tchnc_R -0.473
## s_v_tchn_EG -0.520 0.485
## s_v_qlt__QY -0.434 0.023 0.024
## plsGroth -0.518 -0.053 0.016 0.017
## s_int_psy 0.178 0.037 -0.004 -0.013 -0.199m7 <- glmer(knowledge ~ s_iv_technical_terms * s_int_psy + s_iv_quality_of_evidence + pls + (1 | id), family = binomial("logit"), data = df_resilience, glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 100000)))
summary(m7)## Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
## Family: binomial ( logit )
## Formula: knowledge ~ s_iv_technical_terms * s_int_psy + s_iv_quality_of_evidence + pls + (1 | id)
## Data: df_resilience
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 2383.9 2434.2 -1182.9 2365.9 1981
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.2619 -0.7512 0.3448 0.5637 1.6052
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.458 1.208
## Number of obs: 1990, groups: id, 1026
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.42411 0.16247 8.765 < 2e-16 ***
## s_iv_technical_termsReplace 0.31000 0.16539 1.874 0.0609 .
## s_iv_technical_termsExpl Glossar -0.08676 0.16895 -0.514 0.6076
## s_int_psy 0.55152 0.11856 4.652 3.29e-06 ***
## s_iv_quality_of_evidenceQu Yes -0.08315 0.13656 -0.609 0.5426
## plsGroth -1.19525 0.12494 -9.567 < 2e-16 ***
## s_iv_technical_termsReplace:s_int_psy -0.17410 0.16297 -1.068 0.2854
## s_iv_technical_termsExpl Glossar:s_int_psy 0.11651 0.17255 0.675 0.4995
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v__R s_v__EG s_nt_p s____Y plsGrt s___R:
## s_v_tchnc_R -0.474
## s_v_tchn_EG -0.518 0.485
## s_int_psy 0.137 -0.010 -0.037
## s_v_qlt__QY -0.434 0.024 0.025 -0.025
## plsGroth -0.518 -0.053 0.014 -0.118 0.016
## s_v_tc_R:__ -0.049 0.036 0.024 -0.702 0.022 0.019
## s_v_t_EG:__ -0.009 0.025 0.057 -0.647 0.017 -0.031 0.475anova(m1,m6,m7)## Data: df_resilience
## Models:
## m1: knowledge ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id)
## m6: knowledge ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id) + s_int_psy
## m7: knowledge ~ s_iv_technical_terms * s_int_psy + s_iv_quality_of_evidence + pls + (1 | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## m1 6 2439.4 2473.0 -1213.7 2427.4
## m6 7 2382.8 2422.0 -1184.4 2368.8 58.5999 1 1.932e-14 ***
## m7 9 2383.9 2434.2 -1182.9 2365.9 2.9631 2 0.2273
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1#ICC Subject
m0 <- glmer(knowledge ~ (1 | id), family = binomial("logit"), data = df_resilience)
summary(m0)## Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
## Family: binomial ( logit )
## Formula: knowledge ~ (1 | id)
## Data: df_resilience
##
## AIC BIC logLik deviance df.resid
## 2546.6 2557.8 -1271.3 2542.6 1988
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.2928 -0.9289 0.5462 0.5462 0.7735
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 0.9791 0.9895
## Number of obs: 1990, groups: id, 1026
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.75961 0.06828 11.12 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1performance::icc(m0)## # Intraclass Correlation Coefficient
##
## Adjusted ICC: 0.229
## Conditional ICC: 0.229D Empowerment
m1 <- lmer(empowerment ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id), data = df_resilience)
summary(m1)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: empowerment ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id)
## Data: df_resilience
##
## REML criterion at convergence: 8196.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.2070 -0.4795 -0.0377 0.4650 3.3815
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.824 1.681
## Residual 1.315 1.147
## Number of obs: 2064, groups: id, 1101
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 3.85345 0.11608 1201.13824 33.197 < 2e-16 ***
## s_iv_technical_termsReplace 0.67952 0.13715 1089.44717 4.955 8.39e-07 ***
## s_iv_technical_termsExpl Glossar 0.66392 0.14122 1091.93598 4.701 2.91e-06 ***
## s_iv_quality_of_evidenceQu Yes -0.01451 0.11384 1091.40493 -0.127 0.899
## plsGroth -0.43105 0.05168 1000.27711 -8.340 2.44e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v__R s___EG s____Y
## s_v_tchnc_R -0.619
## s_v_tchn_EG -0.598 0.497
## s_v_qlt__QY -0.488 0.032 0.024
## plsGroth -0.223 -0.001 -0.001 0.004test = glht(m1,linfct=mcp(s_iv_technical_terms="Tukey"))
summary(test, test = adjusted("none"))##
## Simultaneous Tests for General Linear Hypotheses
##
## Multiple Comparisons of Means: Tukey Contrasts
##
##
## Fit: lmer(formula = empowerment ~ s_iv_technical_terms + s_iv_quality_of_evidence +
## pls + (1 | id), data = df_resilience)
##
## Linear Hypotheses:
## Estimate Std. Error z value Pr(>|z|)
## Replace - Expl None == 0 0.6795 0.1371 4.955 7.24e-07 ***
## Expl Glossar - Expl None == 0 0.6639 0.1412 4.701 2.58e-06 ***
## Expl Glossar - Replace == 0 -0.0156 0.1396 -0.112 0.911
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- none method)summary(test, test = adjusted("BH"))##
## Simultaneous Tests for General Linear Hypotheses
##
## Multiple Comparisons of Means: Tukey Contrasts
##
##
## Fit: lmer(formula = empowerment ~ s_iv_technical_terms + s_iv_quality_of_evidence +
## pls + (1 | id), data = df_resilience)
##
## Linear Hypotheses:
## Estimate Std. Error z value Pr(>|z|)
## Replace - Expl None == 0 0.6795 0.1371 4.955 2.17e-06 ***
## Expl Glossar - Expl None == 0 0.6639 0.1412 4.701 3.88e-06 ***
## Expl Glossar - Replace == 0 -0.0156 0.1396 -0.112 0.911
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- BH method)r.squaredGLMM(m1)## R2m R2c
## [1,] 0.03441642 0.6931759r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.034 0.053 0.022
## 2 s_iv_technical_termsReplace 0.019 0.032 0.009
## 3 s_iv_technical_termsExpl Glossar 0.017 0.030 0.008
## 5 plsGroth 0.011 0.022 0.004
## 4 s_iv_quality_of_evidenceQu Yes 0.000 0.002 0.000anova_stats(m1)## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## term | sumsq | meansq | NumDF | DenDF | statistic | p.value | df | etasq | partial.etasq | omegasq | partial.omegasq | epsilonsq | cohens.f | power
## -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
## s_iv_technical_terms | 41.052 | 20.526 | 2 | 1091.600 | 15.603 | < .001 | | 0.024 | 0.026 | | | | 0.162 |
## s_iv_quality_of_evidence | 0.021 | 0.021 | 1 | 1091.405 | 0.016 | 0.899 | | 0.000 | 0.000 | | | | 0.004 |
## pls | 91.502 | 91.502 | 1 | 1000.277 | 69.558 | < .001 | | 0.054 | 0.055 | | | | 0.242 |
## Residuals | 1561.427 | 0.757 | | | | | 2064 | | | | | | |anova_stats(m1)$term## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "s_iv_technical_terms" "s_iv_quality_of_evidence" "pls" "Residuals"interpret_omega_squared(anova_stats(m1)$partial.etasq, rules = "field2013")## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "small" "very small" "small" NA
## (Rules: field2013)# Interest in psychological research
m1 <- lmer(empowerment ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id), data = df_resilience[!is.na(df_resilience$s_int_psy),])
summary(m1)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: empowerment ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id)
## Data: df_resilience[!is.na(df_resilience$s_int_psy), ]
##
## REML criterion at convergence: 8196.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.2070 -0.4795 -0.0377 0.4650 3.3815
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.824 1.681
## Residual 1.315 1.147
## Number of obs: 2064, groups: id, 1101
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 3.85345 0.11608 1201.13824 33.197 < 2e-16 ***
## s_iv_technical_termsReplace 0.67952 0.13715 1089.44717 4.955 8.39e-07 ***
## s_iv_technical_termsExpl Glossar 0.66392 0.14122 1091.93598 4.701 2.91e-06 ***
## s_iv_quality_of_evidenceQu Yes -0.01451 0.11384 1091.40493 -0.127 0.899
## plsGroth -0.43105 0.05168 1000.27711 -8.340 2.44e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v__R s___EG s____Y
## s_v_tchnc_R -0.619
## s_v_tchn_EG -0.598 0.497
## s_v_qlt__QY -0.488 0.032 0.024
## plsGroth -0.223 -0.001 -0.001 0.004r.squaredGLMM(m1)## R2m R2c
## [1,] 0.03441642 0.6931759r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.034 0.053 0.022
## 2 s_iv_technical_termsReplace 0.019 0.032 0.009
## 3 s_iv_technical_termsExpl Glossar 0.017 0.030 0.008
## 5 plsGroth 0.011 0.022 0.004
## 4 s_iv_quality_of_evidenceQu Yes 0.000 0.002 0.000m6 <- lmer(empowerment ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id) + s_int_psy, data = df_resilience[!is.na(df_resilience$s_int_psy),])
summary(m6)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: empowerment ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id) + s_int_psy
## Data: df_resilience[!is.na(df_resilience$s_int_psy), ]
##
## REML criterion at convergence: 7965
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.4346 -0.4920 -0.0359 0.4758 3.2519
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.141 1.463
## Residual 1.315 1.147
## Number of obs: 2064, groups: id, 1101
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 3.86694 0.10487 1223.68011 36.872 < 2e-16 ***
## s_iv_technical_termsReplace 0.67889 0.12319 1087.15313 5.511 4.45e-08 ***
## s_iv_technical_termsExpl Glossar 0.68246 0.12687 1089.99230 5.379 9.16e-08 ***
## s_iv_quality_of_evidenceQu Yes -0.03348 0.10227 1089.31094 -0.327 0.743
## plsGroth -0.42649 0.05157 1007.90377 -8.271 4.18e-16 ***
## s_int_psy 0.82291 0.05078 1098.91902 16.205 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v__R s___EG s____Y plsGrt
## s_v_tchnc_R -0.616
## s_v_tchn_EG -0.594 0.497
## s_v_qlt__QY -0.485 0.032 0.024
## plsGroth -0.247 -0.001 -0.001 0.004
## s_int_psy 0.006 0.000 0.010 -0.011 0.007r.squaredGLMM(m6)## R2m R2c
## [1,] 0.193389 0.6930716r2beta(m6, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.193 0.224 0.166
## 6 s_int_psy 0.164 0.193 0.137
## 2 s_iv_technical_termsReplace 0.022 0.037 0.012
## 3 s_iv_technical_termsExpl Glossar 0.021 0.035 0.011
## 5 plsGroth 0.013 0.024 0.005
## 4 s_iv_quality_of_evidenceQu Yes 0.000 0.003 0.000m7 <- lmer(empowerment ~ s_iv_technical_terms * s_int_psy + s_iv_quality_of_evidence + pls + (1 | id), data = df_resilience[!is.na(df_resilience$s_int_psy),])
summary(m7)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: empowerment ~ s_iv_technical_terms * s_int_psy + s_iv_quality_of_evidence + pls + (1 | id)
## Data: df_resilience[!is.na(df_resilience$s_int_psy), ]
##
## REML criterion at convergence: 7969.1
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.4521 -0.4924 -0.0341 0.4801 3.2571
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.144 1.464
## Residual 1.315 1.147
## Number of obs: 2064, groups: id, 1101
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 3.86579 0.10494 1221.07672 36.839 < 2e-16 ***
## s_iv_technical_termsReplace 0.67957 0.12326 1085.04471 5.513 4.40e-08 ***
## s_iv_technical_termsExpl Glossar 0.68135 0.12697 1087.47655 5.366 9.82e-08 ***
## s_int_psy 0.79702 0.08679 1087.63076 9.184 < 2e-16 ***
## s_iv_quality_of_evidenceQu Yes -0.03155 0.10236 1087.32755 -0.308 0.758
## plsGroth -0.42641 0.05157 1007.62663 -8.268 4.26e-16 ***
## s_iv_technical_termsReplace:s_int_psy 0.08621 0.12110 1091.30634 0.712 0.477
## s_iv_technical_termsExpl Glossar:s_int_psy -0.01809 0.12760 1100.02009 -0.142 0.887
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v__R s_v__EG s_nt_p s____Y plsGrt s___R:
## s_v_tchnc_R -0.616
## s_v_tchn_EG -0.594 0.497
## s_int_psy 0.013 -0.005 -0.004
## s_v_qlt__QY -0.485 0.033 0.024 -0.023
## plsGroth -0.246 -0.001 -0.001 0.014 0.004
## s_v_tc_R:__ -0.014 0.007 0.003 -0.717 0.025 -0.007
## s_v_t_EG:__ -0.005 0.003 0.019 -0.680 0.011 -0.014 0.487r.squaredGLMM(m7)## R2m R2c
## [1,] 0.1937032 0.6934813r2beta(m7, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.194 0.225 0.167
## 4 s_int_psy 0.060 0.081 0.042
## 2 s_iv_technical_termsReplace 0.022 0.037 0.012
## 3 s_iv_technical_termsExpl Glossar 0.021 0.035 0.011
## 6 plsGroth 0.013 0.024 0.005
## 7 s_iv_technical_termsReplace:s_int_psy 0.000 0.004 0.000
## 5 s_iv_quality_of_evidenceQu Yes 0.000 0.003 0.000
## 8 s_iv_technical_termsExpl Glossar:s_int_psy 0.000 0.003 0.000anova(m1,m6, m7)## refitting model(s) with ML (instead of REML)## Data: df_resilience[!is.na(df_resilience$s_int_psy), ]
## Models:
## m1: empowerment ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id)
## m6: empowerment ~ s_iv_technical_terms + s_iv_quality_of_evidence + pls + (1 | id) + s_int_psy
## m7: empowerment ~ s_iv_technical_terms * s_int_psy + s_iv_quality_of_evidence + pls + (1 | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## m1 7 8195.5 8234.9 -4090.7 8181.5
## m6 8 7961.0 8006.0 -3972.5 7945.0 236.492 1 <2e-16 ***
## m7 10 7964.1 8020.5 -3972.1 7944.1 0.825 2 0.662
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1#ICC Subject
m0 <- lmer(empowerment ~ (1 | id), data = df_resilience)
summary(m0)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: empowerment ~ (1 | id)
## Data: df_resilience
##
## REML criterion at convergence: 8283.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.88191 -0.51027 -0.01293 0.49165 3.02182
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.873 1.695
## Residual 1.406 1.186
## Number of obs: 2064, groups: id, 1101
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.078e+00 5.759e-02 1.093e+03 70.81 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1performance::icc(m0)## # Intraclass Correlation Coefficient
##
## Adjusted ICC: 0.671
## Conditional ICC: 0.671H3 Quality of Evidence
A Understanding of the quality of evidence - decision task
m1 <- lmer(pref_quality_of_evidence ~ s_iv_quality_of_evidence + pls + (1 | id), data = df)
summary(m1)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: pref_quality_of_evidence ~ s_iv_quality_of_evidence + pls + (1 | id)
## Data: df
##
## REML criterion at convergence: 16496.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.99612 -0.43383 -0.01972 0.47015 2.89585
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.676 1.295
## Residual 1.534 1.238
## Number of obs: 4276, groups: id, 2264
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.68528 0.06266 3136.44400 74.778 < 2e-16 ***
## s_iv_quality_of_evidenceQu Yes 0.21911 0.06663 2235.73258 3.288 0.00102 **
## plsChristodoulou -0.22001 0.05310 2091.21912 -4.143 3.56e-05 ***
## plsRasmussen 0.07706 0.07708 3528.17926 1.000 0.31748
## plsGroth -0.09622 0.07692 3517.02349 -1.251 0.21108
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s____Y plsChr plsRsm
## s_v_qlt__QY -0.527
## plsChristdl -0.425 0.001
## plsRasmussn -0.595 0.015 0.345
## plsGroth -0.598 0.018 0.345 0.737r.squaredGLMM(m1)## R2m R2c
## [1,] 0.007509718 0.5258153r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.008 0.014 0.004
## 2 s_iv_quality_of_evidenceQu Yes 0.004 0.008 0.001
## 3 plsChristodoulou 0.002 0.006 0.000
## 5 plsGroth 0.000 0.002 0.000
## 4 plsRasmussen 0.000 0.002 0.000anova_stats(m1)## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## term | sumsq | meansq | NumDF | DenDF | statistic | p.value | df | etasq | partial.etasq | omegasq | partial.omegasq | epsilonsq | cohens.f | power
## -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
## s_iv_quality_of_evidence | 16.585 | 16.585 | 1 | 2235.733 | 10.813 | 0.001 | | 0.004 | 0.004 | | | | 0.063 |
## pls | 44.546 | 14.849 | 3 | 2782.515 | 9.681 | < .001 | | 0.010 | 0.010 | | | | 0.103 |
## Residuals | 4234.647 | 0.990 | | | | | 4276 | | | | | | |anova_stats(m1)$term## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "s_iv_quality_of_evidence" "pls" "Residuals"interpret_omega_squared(anova_stats(m1)$partial.etasq, rules = "field2013")## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "very small" "small" NA
## (Rules: field2013)#ICC Subject
m0 <- lmer(pref_quality_of_evidence ~ (1 | id), data = df)
summary(m0)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: pref_quality_of_evidence ~ (1 | id)
## Data: df
##
## REML criterion at convergence: 16521
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.86807 -0.45962 0.06856 0.48959 2.75165
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.681 1.296
## Residual 1.552 1.246
## Number of obs: 4276, groups: id, 2264
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.730e+00 3.339e-02 2.237e+03 141.7 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1performance::icc(m0)## # Intraclass Correlation Coefficient
##
## Adjusted ICC: 0.520
## Conditional ICC: 0.520B Knowledge of the quality of evidence - pls-specific
table(df$knowledge_evidence, df$s_iv_quality_of_evidence)/sum(table(df$knowledge_evidence, df$s_iv_quality_of_evidence)) * 100##
## Qu No Qu Yes
## 0 33.10427 26.35071
## 1 18.22275 22.32227table(df$knowledge_evidence[df$s_iv_quality_of_evidence == "Qu No"])/sum(table(df$knowledge_evidence[df$s_iv_quality_of_evidence == "Qu No"])) * 100##
## 0 1
## 64.49677 35.50323table(df$knowledge_evidence[df$s_iv_quality_of_evidence == "Qu Yes"])/sum(table(df$knowledge_evidence[df$s_iv_quality_of_evidence == "Qu Yes"])) * 100##
## 0 1
## 54.13827 45.86173m1 <- glmer(knowledge_evidence ~ s_iv_quality_of_evidence + pls + (1 | id), family = binomial("logit"), data = df, glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 100000)))
summary(m1)## Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
## Family: binomial ( logit )
## Formula: knowledge_evidence ~ s_iv_quality_of_evidence + pls + (1 | id)
## Data: df
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 5560.6 5598.7 -2774.3 5548.6 4214
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -0.9657 -0.6518 -0.5368 0.8519 1.3561
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.164 1.079
## Number of obs: 4220, groups: id, 2173
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.75417 0.09067 -8.318 < 2e-16 ***
## s_iv_quality_of_evidenceQu Yes 0.54752 0.08679 6.309 2.81e-10 ***
## plsChristodoulou -0.15647 0.09860 -1.587 0.113
## plsRasmussen 0.08643 0.11120 0.777 0.437
## plsGroth 0.05270 0.11131 0.473 0.636
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s____Y plsChr plsRsm
## s_v_qlt__QY -0.516
## plsChristdl -0.524 -0.012
## plsRasmussn -0.594 0.023 0.437
## plsGroth -0.591 0.020 0.437 0.566anova_stats(m1)## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: npar## term | npar | sumsq | meansq | statistic | df | etasq | partial.etasq | omegasq | partial.omegasq | epsilonsq | cohens.f | power
## ---------------------------------------------------------------------------------------------------------------------------------------------------------
## s_iv_quality_of_evidence | 1 | 44.824 | 44.824 | 44.824 | | 0.011 | 0.011 | | | | 0.105 |
## pls | 3 | 6.295 | 2.098 | 2.098 | | 0.002 | 0.002 | | | | 0.039 |
## Residuals | | 4046.421 | 0.150 | | 4220 | | | | | | |#r.squaredGLMM(m1)
r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.012 0.020 0.007
## 2 s_iv_quality_of_evidenceQu Yes 0.011 0.018 0.006
## 3 plsChristodoulou 0.001 0.003 0.000
## 4 plsRasmussen 0.000 0.002 0.000
## 5 plsGroth 0.000 0.001 0.000exp(fixef(m1))## (Intercept) s_iv_quality_of_evidenceQu Yes plsChristodoulou
## 0.4703989 1.7289524 0.8551547
## plsRasmussen plsGroth
## 1.0902734 1.0541081interpret_oddsratio(exp(fixef(m1)), rules = "cohen1988")## (Intercept) s_iv_quality_of_evidenceQu Yes plsChristodoulou
## "small" "small" "very small"
## plsRasmussen plsGroth
## "very small" "very small"
## (Rules: cohen1988)#ICC Subject
m0 <- glmer(knowledge_evidence ~ (1 | id), family = binomial("logit"), data = df)
summary(m0)## Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
## Family: binomial ( logit )
## Formula: knowledge_evidence ~ (1 | id)
## Data: df
##
## AIC BIC logLik deviance df.resid
## 5598.8 5611.5 -2797.4 5594.8 4218
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -0.8567 -0.5787 -0.5787 0.8012 1.1672
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.197 1.094
## Number of obs: 4220, groups: id, 2173
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.4930 0.0445 -11.08 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1performance::icc(m0)## # Intraclass Correlation Coefficient
##
## Adjusted ICC: 0.267
## Conditional ICC: 0.267B Knowledge on the quality of evidence - general
data$s_int_psy <- scale(data$s_int_psy)
table(data$s_knowledge_evidence, data$s_iv_quality_of_evidence)##
## Qu No Qu Yes
## 0 389 313
## 1 659 683table(data$s_knowledge_evidence, data$s_iv_quality_of_evidence)/sum(table(data$s_knowledge_evidence, data$s_iv_quality_of_evidence)) * 100##
## Qu No Qu Yes
## 0 19.03131 15.31311
## 1 32.24070 33.41487table(data$s_knowledge_evidence[data$s_iv_quality_of_evidence == "Qu No"])/sum(table(data$s_knowledge_evidence[data$s_iv_quality_of_evidence == "Qu No"])) * 100##
## 0 1
## 37.11832 62.88168table(data$s_knowledge_evidence[data$s_iv_quality_of_evidence == "Qu Yes"])/sum(table(data$s_knowledge_evidence[data$s_iv_quality_of_evidence == "Qu Yes"])) * 100##
## 0 1
## 31.4257 68.5743table(data$s_knowledge_evidence)##
## 0 1
## 702 1342sum(table(data$s_knowledge_evidence))## [1] 2044m1 <- glm(s_knowledge_evidence ~ s_iv_quality_of_evidence, data = data, family = "binomial")
summary(m1)##
## Call:
## glm(formula = s_knowledge_evidence ~ s_iv_quality_of_evidence,
## family = "binomial", data = data)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.5215 -1.4079 0.8686 0.9632 0.9632
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.52714 0.06394 8.245 <2e-16 ***
## s_iv_quality_of_evidenceQu Yes 0.25315 0.09353 2.707 0.0068 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 2629.8 on 2043 degrees of freedom
## Residual deviance: 2622.4 on 2042 degrees of freedom
## (244 Beobachtungen als fehlend gelöscht)
## AIC: 2626.4
##
## Number of Fisher Scoring iterations: 4exp(coef(m1))## (Intercept) s_iv_quality_of_evidenceQu Yes
## 1.694087 1.288073interpret_oddsratio(exp(coef(m1)), rules = "cohen1988")## (Intercept) s_iv_quality_of_evidenceQu Yes
## "small" "very small"
## (Rules: cohen1988)m4 <- glm(s_knowledge_evidence ~ s_iv_quality_of_evidence + s_int_psy, data = data, family = "binomial")
summary(m4)##
## Call:
## glm(formula = s_knowledge_evidence ~ s_iv_quality_of_evidence +
## s_int_psy, family = "binomial", data = data)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.7179 -1.3454 0.8146 0.9167 1.1936
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.52060 0.06458 8.061 7.55e-16 ***
## s_iv_quality_of_evidenceQu Yes 0.26312 0.09450 2.784 0.00536 **
## s_int_psy 0.30270 0.04803 6.302 2.93e-10 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 2629.8 on 2043 degrees of freedom
## Residual deviance: 2582.3 on 2041 degrees of freedom
## (244 Beobachtungen als fehlend gelöscht)
## AIC: 2588.3
##
## Number of Fisher Scoring iterations: 4m5 <- glm(s_knowledge_evidence ~ s_iv_quality_of_evidence * s_int_psy, data = data, family = "binomial")
summary(m5)##
## Call:
## glm(formula = s_knowledge_evidence ~ s_iv_quality_of_evidence *
## s_int_psy, family = "binomial", data = data)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.7223 -1.3466 0.8133 0.9176 1.1887
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.52055 0.06455 8.064 7.4e-16 ***
## s_iv_quality_of_evidenceQu Yes 0.26359 0.09457 2.787 0.00532 **
## s_int_psy 0.29638 0.06711 4.416 1.0e-05 ***
## s_iv_quality_of_evidenceQu Yes:s_int_psy 0.01294 0.09608 0.135 0.89283
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 2629.8 on 2043 degrees of freedom
## Residual deviance: 2582.2 on 2040 degrees of freedom
## (244 Beobachtungen als fehlend gelöscht)
## AIC: 2590.2
##
## Number of Fisher Scoring iterations: 4anova(m1,m4,m5)## Analysis of Deviance Table
##
## Model 1: s_knowledge_evidence ~ s_iv_quality_of_evidence
## Model 2: s_knowledge_evidence ~ s_iv_quality_of_evidence + s_int_psy
## Model 3: s_knowledge_evidence ~ s_iv_quality_of_evidence * s_int_psy
## Resid. Df Resid. Dev Df Deviance
## 1 2042 2622.4
## 2 2041 2582.3 1 40.169
## 3 2040 2582.2 1 0.018Accessibility
m1 <- lmer(accessibility ~ s_iv_quality_of_evidence + pls + (1 | id), data = df)
summary(m1)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: accessibility ~ s_iv_quality_of_evidence + pls + (1 | id)
## Data: df
##
## REML criterion at convergence: 17051.2
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.3950 -0.4229 0.1020 0.3894 3.2173
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.678 1.637
## Residual 1.310 1.145
## Number of obs: 4326, groups: id, 2284
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 6.17517 0.07007 2860.15780 88.128 <2e-16 ***
## s_iv_quality_of_evidenceQu Yes -0.10028 0.07713 2264.75789 -1.300 0.1937
## plsChristodoulou -0.11271 0.04889 2090.82160 -2.305 0.0212 *
## plsRasmussen -0.83337 0.08493 3158.64246 -9.813 <2e-16 ***
## plsGroth -1.38174 0.08493 3158.81290 -16.269 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s____Y plsChr plsRsm
## s_v_qlt__QY -0.544
## plsChristdl -0.348 -0.001
## plsRasmussn -0.591 0.019 0.287
## plsGroth -0.591 0.020 0.287 0.816r.squaredGLMM(m1)## R2m R2c
## [1,] 0.07311593 0.6955514r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.073 0.089 0.060
## 5 plsGroth 0.056 0.070 0.044
## 4 plsRasmussen 0.021 0.031 0.014
## 2 s_iv_quality_of_evidenceQu Yes 0.001 0.003 0.000
## 3 plsChristodoulou 0.000 0.003 0.000anova_stats(m1)## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## term | sumsq | meansq | NumDF | DenDF | statistic | p.value | df | etasq | partial.etasq | omegasq | partial.omegasq | epsilonsq | cohens.f | power
## --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
## s_iv_quality_of_evidence | 2.214 | 2.214 | 1 | 2264.758 | 1.690 | 0.194 | | 0.001 | 0.001 | | | | 0.026 |
## pls | 398.364 | 132.788 | 3 | 2528.406 | 101.365 | < .001 | | 0.108 | 0.108 | | | | 0.347 |
## Residuals | 3301.401 | 0.763 | | | | | 4326 | | | | | | |Understanding
m1 <- lmer(understanding ~ s_iv_quality_of_evidence + pls + (1 | id), data = df)
summary(m1)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: understanding ~ s_iv_quality_of_evidence + pls + (1 | id)
## Data: df
##
## REML criterion at convergence: 16962.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.3970 -0.4492 0.0883 0.4375 3.1136
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.431 1.559
## Residual 1.359 1.166
## Number of obs: 4323, groups: id, 2285
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 6.23796 0.06820 2922.53712 91.459 < 2e-16 ***
## s_iv_quality_of_evidenceQu Yes -0.13625 0.07457 2265.71594 -1.827 0.0678 .
## plsChristodoulou -0.20543 0.04985 2097.09535 -4.121 3.92e-05 ***
## plsRasmussen -0.81360 0.08289 3242.66775 -9.816 < 2e-16 ***
## plsGroth -1.35754 0.08287 3241.77724 -16.381 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s____Y plsChr plsRsm
## s_v_qlt__QY -0.541
## plsChristdl -0.364 -0.001
## plsRasmussn -0.592 0.018 0.300
## plsGroth -0.593 0.020 0.300 0.800r.squaredGLMM(m1)## R2m R2c
## [1,] 0.06998235 0.6665696r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.070 0.086 0.057
## 5 plsGroth 0.057 0.071 0.045
## 4 plsRasmussen 0.021 0.031 0.014
## 3 plsChristodoulou 0.001 0.005 0.000
## 2 s_iv_quality_of_evidenceQu Yes 0.001 0.004 0.000anova_stats(m1)## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## term | sumsq | meansq | NumDF | DenDF | statistic | p.value | df | etasq | partial.etasq | omegasq | partial.omegasq | epsilonsq | cohens.f | power
## --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
## s_iv_quality_of_evidence | 4.537 | 4.537 | 1 | 2265.716 | 3.339 | 0.068 | | 0.001 | 0.001 | | | | 0.036 |
## pls | 405.456 | 135.152 | 3 | 2581.664 | 99.461 | < .001 | | 0.104 | 0.104 | | | | 0.341 |
## Residuals | 3490.875 | 0.808 | | | | | 4323 | | | | | | |Knowledge
m1 <- glmer(knowledge ~ s_iv_quality_of_evidence + pls + (1 | id), family = binomial("logit"), data = df)
summary(m1)## Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
## Family: binomial ( logit )
## Formula: knowledge ~ s_iv_quality_of_evidence + pls + (1 | id)
## Data: df
##
## AIC BIC logLik deviance df.resid
## 5232.9 5271.0 -2610.5 5220.9 4214
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.7385 -0.9814 0.4019 0.6739 1.0190
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.102 1.05
## Number of obs: 4220, groups: id, 2173
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.34612 0.10027 13.425 <2e-16 ***
## s_iv_quality_of_evidenceQu Yes -0.07517 0.08665 -0.868 0.386
## plsChristodoulou -1.05180 0.10615 -9.909 <2e-16 ***
## plsRasmussen 0.04058 0.11944 0.340 0.734
## plsGroth -1.05234 0.11784 -8.930 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s____Y plsChr plsRsm
## s_v_qlt__QY -0.443
## plsChristdl -0.647 0.009
## plsRasmussn -0.559 0.016 0.445
## plsGroth -0.670 0.024 0.529 0.549Empowerment
m1 <- lmer(empowerment ~ s_iv_quality_of_evidence + pls + (1 | id), data = df)
summary(m1)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: empowerment ~ s_iv_quality_of_evidence + pls + (1 | id)
## Data: df
##
## REML criterion at convergence: 17087.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.2348 -0.4920 -0.0097 0.4726 3.3710
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.868 1.693
## Residual 1.268 1.126
## Number of obs: 4324, groups: id, 2285
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.99787 0.07145 2818.10409 69.952 < 2e-16 ***
## s_iv_quality_of_evidenceQu Yes -0.04076 0.07900 2266.46337 -0.516 0.606
## plsChristodoulou -0.27488 0.04818 2086.11696 -5.706 1.32e-08 ***
## plsRasmussen -0.68588 0.08639 3098.19956 -7.939 2.82e-15 ***
## plsGroth -1.11652 0.08642 3100.48310 -12.920 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s____Y plsChr plsRsm
## s_v_qlt__QY -0.547
## plsChristdl -0.335 -0.001
## plsRasmussn -0.590 0.019 0.278
## plsGroth -0.590 0.020 0.278 0.827r.squaredGLMM(m1)## R2m R2c
## [1,] 0.04118597 0.7060639r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.041 0.054 0.031
## 5 plsGroth 0.036 0.048 0.026
## 4 plsRasmussen 0.014 0.022 0.008
## 3 plsChristodoulou 0.002 0.006 0.000
## 2 s_iv_quality_of_evidenceQu Yes 0.000 0.002 0.000anova_stats(m1)## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## term | sumsq | meansq | NumDF | DenDF | statistic | p.value | df | etasq | partial.etasq | omegasq | partial.omegasq | epsilonsq | cohens.f | power
## -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
## s_iv_quality_of_evidence | 0.337 | 0.337 | 1 | 2266.463 | 0.266 | 0.606 | | 0.000 | 0.000 | | | | 0.010 |
## pls | 250.962 | 83.654 | 3 | 2489.532 | 65.981 | < .001 | | 0.074 | 0.074 | | | | 0.282 |
## Residuals | 3145.492 | 0.727 | | | | | 4324 | | | | | | |H4 Operationalization
A Accessibility
m1 <- lmer(accessibility ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + (1 | id), data = df_cama)
summary(m1)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: accessibility ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + (1 | id)
## Data: df_cama
##
## REML criterion at convergence: 8512.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.8731 -0.3673 0.1713 0.3470 3.3567
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.6104 1.6157
## Residual 0.9826 0.9912
## Number of obs: 2262, groups: id, 1183
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 6.34136 0.09145 1298.54431 69.343 < 2e-16 ***
## s_iv_operationalizationOp Yes -0.33786 0.10304 1169.17411 -3.279 0.00107 **
## s_iv_quality_of_evidenceQu Yes -0.09955 0.10304 1169.12932 -0.966 0.33420
## plsChristodoulou -0.11306 0.04240 1095.61897 -2.667 0.00777 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_OY s____Y
## s_v_prtnlOY -0.564
## s_v_qlt__QY -0.564 0.012
## plsChristdl -0.230 -0.001 -0.002r.squaredGLMM(m1)## R2m R2c
## [1,] 0.009399288 0.7290993r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.009 0.020 0.004
## 2 s_iv_operationalizationOp Yes 0.008 0.017 0.002
## 4 plsChristodoulou 0.001 0.005 0.000
## 3 s_iv_quality_of_evidenceQu Yes 0.001 0.005 0.000anova_stats(m1)## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## term | sumsq | meansq | NumDF | DenDF | statistic | p.value | df | etasq | partial.etasq | omegasq | partial.omegasq | epsilonsq | cohens.f | power
## -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
## s_iv_operationalization | 10.564 | 10.564 | 1 | 1169.174 | 10.751 | 0.001 | | 0.008 | 0.008 | | | | 0.092 |
## s_iv_quality_of_evidence | 0.917 | 0.917 | 1 | 1169.129 | 0.933 | 0.334 | | 0.001 | 0.001 | | | | 0.027 |
## pls | 6.987 | 6.987 | 1 | 1095.619 | 7.111 | 0.008 | | 0.005 | 0.006 | | | | 0.075 |
## Residuals | 1254.603 | 0.555 | | | | | 2262 | | | | | | |anova_stats(m1)$term## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "s_iv_operationalization" "s_iv_quality_of_evidence" "pls" "Residuals"interpret_omega_squared(anova_stats(m1)$partial.etasq, rules = "field2013")## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "very small" "very small" "very small" NA
## (Rules: field2013)# Interest in psychological research
m1 <- lmer(accessibility ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + (1 | id), data = df_cama[!is.na(df_cama$s_int_psy),])
summary(m1)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: accessibility ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + (1 | id)
## Data: df_cama[!is.na(df_cama$s_int_psy), ]
##
## REML criterion at convergence: 8512.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.8731 -0.3673 0.1713 0.3470 3.3567
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.6104 1.6157
## Residual 0.9826 0.9912
## Number of obs: 2262, groups: id, 1183
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 6.34136 0.09145 1298.54431 69.343 < 2e-16 ***
## s_iv_operationalizationOp Yes -0.33786 0.10304 1169.17411 -3.279 0.00107 **
## s_iv_quality_of_evidenceQu Yes -0.09955 0.10304 1169.12932 -0.966 0.33420
## plsChristodoulou -0.11306 0.04240 1095.61897 -2.667 0.00777 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_OY s____Y
## s_v_prtnlOY -0.564
## s_v_qlt__QY -0.564 0.012
## plsChristdl -0.230 -0.001 -0.002r.squaredGLMM(m1)## R2m R2c
## [1,] 0.009399288 0.7290993r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.009 0.020 0.004
## 2 s_iv_operationalizationOp Yes 0.008 0.017 0.002
## 4 plsChristodoulou 0.001 0.005 0.000
## 3 s_iv_quality_of_evidenceQu Yes 0.001 0.005 0.000m6 <- lmer(accessibility ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + s_int_psy + (1 | id), data = df_cama[!is.na(df_cama$s_int_psy),])
summary(m6)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: accessibility ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + s_int_psy + (1 | id)
## Data: df_cama[!is.na(df_cama$s_int_psy), ]
##
## REML criterion at convergence: 8412.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.0270 -0.3715 0.1240 0.3718 3.3889
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.3498 1.5329
## Residual 0.9819 0.9909
## Number of obs: 2262, groups: id, 1183
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 6.33927 0.08777 1309.94782 72.230 < 2e-16 ***
## s_iv_operationalizationOp Yes -0.30770 0.09870 1169.59311 -3.118 0.00187 **
## s_iv_quality_of_evidenceQu Yes -0.11243 0.09867 1169.17651 -1.139 0.25474
## plsChristodoulou -0.11207 0.04236 1098.99177 -2.645 0.00827 **
## s_int_psy 0.51019 0.04892 1181.27630 10.428 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_OY s____Y plsChr
## s_v_prtnlOY -0.563
## s_v_qlt__QY -0.563 0.012
## plsChristdl -0.239 -0.001 -0.002
## s_int_psy -0.003 0.029 -0.013 0.002r.squaredGLMM(m6)## R2m R2c
## [1,] 0.08163412 0.7293484r2beta(m6, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.082 0.105 0.063
## 5 s_int_psy 0.073 0.094 0.054
## 2 s_iv_operationalizationOp Yes 0.007 0.016 0.002
## 3 s_iv_quality_of_evidenceQu Yes 0.001 0.005 0.000
## 4 plsChristodoulou 0.001 0.005 0.000m7 <- lmer(accessibility ~ s_iv_operationalization * s_int_psy + s_iv_quality_of_evidence + pls + (1 | id), data = df_cama[!is.na(df_cama$s_int_psy),])
summary(m7)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: accessibility ~ s_iv_operationalization * s_int_psy + s_iv_quality_of_evidence + pls + (1 | id)
## Data: df_cama[!is.na(df_cama$s_int_psy), ]
##
## REML criterion at convergence: 8407.8
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.9976 -0.3768 0.1395 0.3709 3.4315
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.3319 1.5270
## Residual 0.9825 0.9912
## Number of obs: 2262, groups: id, 1183
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 6.34885 0.08758 1309.23258 72.490 < 2e-16 ***
## s_iv_operationalizationOp Yes -0.30410 0.09841 1167.09311 -3.090 0.00205 **
## s_int_psy 0.37956 0.06804 1178.21774 5.579 3.01e-08 ***
## s_iv_quality_of_evidenceQu Yes -0.12725 0.09851 1167.33435 -1.292 0.19672
## plsChristodoulou -0.11199 0.04237 1097.95573 -2.643 0.00834 **
## s_iv_operationalizationOp Yes:s_int_psy 0.26909 0.09773 1179.69595 2.753 0.00599 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_OY s_nt_p s____Y plsChr
## s_v_prtnlOY -0.562
## s_int_psy -0.030 0.012
## s_v_qlt__QY -0.563 0.011 0.029
## plsChristdl -0.240 -0.001 0.001 -0.002
## s_v_prOY:__ 0.040 0.013 -0.697 -0.055 0.000r.squaredGLMM(m7)## R2m R2c
## [1,] 0.0863755 0.7291743r2beta(m7, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.086 0.111 0.067
## 3 s_int_psy 0.022 0.035 0.012
## 2 s_iv_operationalizationOp Yes 0.007 0.015 0.002
## 6 s_iv_operationalizationOp Yes:s_int_psy 0.005 0.013 0.001
## 4 s_iv_quality_of_evidenceQu Yes 0.001 0.006 0.000
## 5 plsChristodoulou 0.001 0.005 0.000anova(m1,m6,m7)## refitting model(s) with ML (instead of REML)## Data: df_cama[!is.na(df_cama$s_int_psy), ]
## Models:
## m1: accessibility ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + (1 | id)
## m6: accessibility ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + s_int_psy + (1 | id)
## m7: accessibility ~ s_iv_operationalization * s_int_psy + s_iv_quality_of_evidence + pls + (1 | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## m1 6 8510.6 8544.9 -4249.3 8498.6
## m6 7 8408.1 8448.2 -4197.1 8394.1 104.4646 1 < 2.2e-16 ***
## m7 8 8402.5 8448.3 -4193.3 8386.5 7.5869 1 0.005879 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1#ICC Subject
m0 <- lmer(accessibility ~ (1 | id), data = df_cama)
summary(m0)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: accessibility ~ (1 | id)
## Data: df_cama
##
## REML criterion at convergence: 8521.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.7701 -0.3288 0.1481 0.3071 3.2715
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.6327 1.6226
## Residual 0.9882 0.9941
## Number of obs: 2262, groups: id, 1183
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 6.068e+00 5.172e-02 1.172e+03 117.3 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1performance::icc(m0)## # Intraclass Correlation Coefficient
##
## Adjusted ICC: 0.727
## Conditional ICC: 0.727B Understanding
m1 <- lmer(understanding ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + (1 | id), data = df_cama)
summary(m1)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: understanding ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + (1 | id)
## Data: df_cama
##
## REML criterion at convergence: 8624.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.5711 -0.4254 0.1251 0.4269 3.0112
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.268 1.506
## Residual 1.198 1.095
## Number of obs: 2258, groups: id, 1184
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 6.38915 0.08882 1338.07729 71.930 < 2e-16 ***
## s_iv_operationalizationOp Yes -0.32029 0.09922 1168.57595 -3.228 0.00128 **
## s_iv_quality_of_evidenceQu Yes -0.12150 0.09923 1168.53379 -1.224 0.22103
## plsChristodoulou -0.20532 0.04683 1098.58679 -4.385 1.27e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_OY s____Y
## s_v_prtnlOY -0.560
## s_v_qlt__QY -0.559 0.012
## plsChristdl -0.261 -0.002 -0.002r.squaredGLMM(m1)## R2m R2c
## [1,] 0.01134843 0.6582069r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.011 0.023 0.005
## 2 s_iv_operationalizationOp Yes 0.007 0.016 0.002
## 4 plsChristodoulou 0.003 0.009 0.000
## 3 s_iv_quality_of_evidenceQu Yes 0.001 0.005 0.000anova_stats(m1)## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## term | sumsq | meansq | NumDF | DenDF | statistic | p.value | df | etasq | partial.etasq | omegasq | partial.omegasq | epsilonsq | cohens.f | power
## -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
## s_iv_operationalization | 12.485 | 12.485 | 1 | 1168.576 | 10.419 | 0.001 | | 0.008 | 0.008 | | | | 0.088 |
## s_iv_quality_of_evidence | 1.797 | 1.797 | 1 | 1168.534 | 1.499 | 0.221 | | 0.001 | 0.001 | | | | 0.034 |
## pls | 23.038 | 23.038 | 1 | 1098.587 | 19.226 | < .001 | | 0.014 | 0.014 | | | | 0.120 |
## Residuals | 1599.460 | 0.708 | | | | | 2258 | | | | | | |anova_stats(m1)$term## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "s_iv_operationalization" "s_iv_quality_of_evidence" "pls" "Residuals"interpret_omega_squared(anova_stats(m1)$partial.etasq, rules = "field2013")## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "very small" "very small" "small" NA
## (Rules: field2013)# Interest in psychological research
m1 <- lmer(understanding ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + (1 | id), data = df_cama[!is.na(df_cama$s_int_psy),])
summary(m1)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: understanding ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + (1 | id)
## Data: df_cama[!is.na(df_cama$s_int_psy), ]
##
## REML criterion at convergence: 8624.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.5711 -0.4254 0.1251 0.4269 3.0112
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.268 1.506
## Residual 1.198 1.095
## Number of obs: 2258, groups: id, 1184
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 6.38915 0.08882 1338.07729 71.930 < 2e-16 ***
## s_iv_operationalizationOp Yes -0.32029 0.09922 1168.57595 -3.228 0.00128 **
## s_iv_quality_of_evidenceQu Yes -0.12150 0.09923 1168.53379 -1.224 0.22103
## plsChristodoulou -0.20532 0.04683 1098.58679 -4.385 1.27e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_OY s____Y
## s_v_prtnlOY -0.560
## s_v_qlt__QY -0.559 0.012
## plsChristdl -0.261 -0.002 -0.002r.squaredGLMM(m1)## R2m R2c
## [1,] 0.01134843 0.6582069r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.011 0.023 0.005
## 2 s_iv_operationalizationOp Yes 0.007 0.016 0.002
## 4 plsChristodoulou 0.003 0.009 0.000
## 3 s_iv_quality_of_evidenceQu Yes 0.001 0.005 0.000m6 <- lmer(understanding ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + s_int_psy + (1 | id), data = df_cama[!is.na(df_cama$s_int_psy),])
summary(m6)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: understanding ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + s_int_psy + (1 | id)
## Data: df_cama[!is.na(df_cama$s_int_psy), ]
##
## REML criterion at convergence: 8498.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.6247 -0.3847 0.1232 0.4376 3.0311
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.970 1.404
## Residual 1.196 1.094
## Number of obs: 2258, groups: id, 1184
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 6.38496 0.08445 1357.74238 75.602 < 2e-16 ***
## s_iv_operationalizationOp Yes -0.28769 0.09403 1170.32286 -3.060 0.00227 **
## s_iv_quality_of_evidenceQu Yes -0.13331 0.09399 1169.82406 -1.418 0.15635
## plsChristodoulou -0.20405 0.04675 1104.56877 -4.365 1.39e-05 ***
## s_int_psy 0.54753 0.04667 1185.84457 11.731 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_OY s____Y plsChr
## s_v_prtnlOY -0.557
## s_v_qlt__QY -0.557 0.011
## plsChristdl -0.274 -0.002 -0.002
## s_int_psy -0.005 0.030 -0.011 0.001r.squaredGLMM(m6)## R2m R2c
## [1,] 0.09719461 0.6588952r2beta(m6, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.097 0.122 0.077
## 5 s_int_psy 0.086 0.109 0.066
## 2 s_iv_operationalizationOp Yes 0.006 0.015 0.002
## 4 plsChristodoulou 0.003 0.010 0.000
## 3 s_iv_quality_of_evidenceQu Yes 0.001 0.006 0.000m7 <- lmer(understanding ~ s_iv_operationalization * s_int_psy + s_iv_quality_of_evidence + pls + (1 | id), data = df_cama[!is.na(df_cama$s_int_psy),])
summary(m7)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: understanding ~ s_iv_operationalization * s_int_psy + s_iv_quality_of_evidence + pls + (1 | id)
## Data: df_cama[!is.na(df_cama$s_int_psy), ]
##
## REML criterion at convergence: 8497.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.5979 -0.3848 0.1316 0.4296 3.0368
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.962 1.401
## Residual 1.197 1.094
## Number of obs: 2258, groups: id, 1184
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 6.39166 0.08441 1356.91835 75.717 < 2e-16 ***
## s_iv_operationalizationOp Yes -0.28559 0.09390 1167.67040 -3.041 0.00241 **
## s_int_psy 0.46010 0.06504 1184.38944 7.074 2.58e-12 ***
## s_iv_quality_of_evidenceQu Yes -0.14326 0.09400 1168.01998 -1.524 0.12774
## plsChristodoulou -0.20412 0.04676 1103.36752 -4.365 1.39e-05 ***
## s_iv_operationalizationOp Yes:s_int_psy 0.17991 0.09338 1184.24046 1.927 0.05427 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_OY s_nt_p s____Y plsChr
## s_v_prtnlOY -0.556
## s_int_psy -0.032 0.013
## s_v_qlt__QY -0.558 0.011 0.031
## plsChristdl -0.274 -0.002 0.002 -0.002
## s_v_prOY:__ 0.041 0.012 -0.698 -0.055 -0.001r.squaredGLMM(m7)## R2m R2c
## [1,] 0.09936772 0.6587064r2beta(m7, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.099 0.125 0.079
## 3 s_int_psy 0.033 0.049 0.020
## 2 s_iv_operationalizationOp Yes 0.006 0.015 0.002
## 5 plsChristodoulou 0.003 0.010 0.000
## 6 s_iv_operationalizationOp Yes:s_int_psy 0.003 0.008 0.000
## 4 s_iv_quality_of_evidenceQu Yes 0.002 0.007 0.000anova(m1,m6,m7)## refitting model(s) with ML (instead of REML)## Data: df_cama[!is.na(df_cama$s_int_psy), ]
## Models:
## m1: understanding ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + (1 | id)
## m6: understanding ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + s_int_psy + (1 | id)
## m7: understanding ~ s_iv_operationalization * s_int_psy + s_iv_quality_of_evidence + pls + (1 | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## m1 6 8622.5 8656.8 -4305.2 8610.5
## m6 7 8493.7 8533.7 -4239.8 8479.7 130.8210 1 < 2e-16 ***
## m7 8 8492.0 8537.7 -4238.0 8476.0 3.7215 1 0.05372 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1#ICC Subject
m0 <- lmer(understanding ~ (1 | id), data = df_cama)
summary(m0)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: understanding ~ (1 | id)
## Data: df_cama
##
## REML criterion at convergence: 8645.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.4699 -0.3949 0.1780 0.3690 2.8710
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.281 1.510
## Residual 1.219 1.104
## Number of obs: 2258, groups: id, 1184
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 6.068e+00 4.981e-02 1.172e+03 121.8 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1performance::icc(m0)## # Intraclass Correlation Coefficient
##
## Adjusted ICC: 0.652
## Conditional ICC: 0.652C Knowledge
table(df_cama$knowledge, df_cama$s_iv_operationalization)/sum(table(df_cama$knowledge, df_cama$s_iv_operationalization)) * 100##
## Op No Op Yes
## 0 18.29596 16.99552
## 1 32.01794 32.69058table(df_cama$knowledge[df_cama$s_iv_operationalization == "Op Yes"])##
## 0 1
## 379 729table(df_cama$knowledge[df_cama$s_iv_operationalization == "Op Yes"])/sum(table(df_cama$knowledge[df_cama$s_iv_operationalization == "Op Yes"])) * 100##
## 0 1
## 34.20578 65.79422table(df_cama$knowledge[df_cama$s_iv_operationalization == "Op No"])##
## 0 1
## 408 714table(df_cama$knowledge[df_cama$s_iv_operationalization == "Op No"])/sum(table(df_cama$knowledge[df_cama$s_iv_operationalization == "Op No"])) * 100##
## 0 1
## 36.36364 63.63636table(df_cama$knowledge, df_cama$s_iv_operationalization)/sum(table(df_cama$knowledge, df_cama$s_iv_operationalization)) * 100##
## Op No Op Yes
## 0 18.29596 16.99552
## 1 32.01794 32.69058m1 <- glmer(knowledge ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + (1 | id), family = binomial("logit"), data = df_cama)
summary(m1)## Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
## Family: binomial ( logit )
## Formula: knowledge ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + (1 | id)
## Data: df_cama
##
## AIC BIC logLik deviance df.resid
## 2790.2 2818.8 -1390.1 2780.2 2225
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.7455 -0.9916 0.4476 0.7026 1.0085
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 0.7623 0.8731
## Number of obs: 2230, groups: id, 1147
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.21282 0.11881 10.208 <2e-16 ***
## s_iv_operationalizationOp Yes 0.11695 0.11115 1.052 0.293
## s_iv_quality_of_evidenceQu Yes -0.06672 0.11105 -0.601 0.548
## plsChristodoulou -0.99358 0.10552 -9.416 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_OY s____Y
## s_v_prtnlOY -0.446
## s_v_qlt__QY -0.482 0.015
## plsChristdl -0.558 -0.025 0.009r.squaredGLMM(m1)## Warning: The null model is correct only if all variables used by the original model remain unchanged.## R2m R2c
## theoretical 0.05840006 0.2355287
## delta 0.04658602 0.1878824r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.042 0.060 0.028
## 4 plsChristodoulou 0.041 0.058 0.027
## 2 s_iv_operationalizationOp Yes 0.001 0.004 0.000
## 3 s_iv_quality_of_evidenceQu Yes 0.000 0.003 0.000anova_stats(m1)## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: npar## term | npar | sumsq | meansq | statistic | df | etasq | partial.etasq | omegasq | partial.omegasq | epsilonsq | cohens.f | power
## ----------------------------------------------------------------------------------------------------------------------------------------------------------
## s_iv_operationalization | 1 | 0.935 | 0.935 | 0.935 | | 0.000 | 0.000 | | | | 0.021 |
## s_iv_quality_of_evidence | 1 | 0.331 | 0.331 | 0.331 | | 0.000 | 0.000 | | | | 0.012 |
## pls | 1 | 109.786 | 109.786 | 109.786 | | 0.047 | 0.047 | | | | 0.222 |
## Residuals | | 2220.857 | 0.161 | | 2230 | | | | | | |exp(fixef(m1))## (Intercept) s_iv_operationalizationOp Yes s_iv_quality_of_evidenceQu Yes
## 3.3629391 1.1240591 0.9354533
## plsChristodoulou
## 0.3702491interpret_oddsratio(exp(fixef(m1)), rules = "cohen1988")## (Intercept) s_iv_operationalizationOp Yes s_iv_quality_of_evidenceQu Yes
## "medium" "very small" "very small"
## plsChristodoulou
## "medium"
## (Rules: cohen1988)m1 <- glmer(knowledge ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + (1 | id), family = binomial("logit"), data = df_cama[!is.na(df_cama$s_int_psy),])
summary(m1)## Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
## Family: binomial ( logit )
## Formula: knowledge ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + (1 | id)
## Data: df_cama[!is.na(df_cama$s_int_psy), ]
##
## AIC BIC logLik deviance df.resid
## 2790.2 2818.8 -1390.1 2780.2 2225
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.7455 -0.9916 0.4476 0.7026 1.0085
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 0.7623 0.8731
## Number of obs: 2230, groups: id, 1147
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.21282 0.11881 10.208 <2e-16 ***
## s_iv_operationalizationOp Yes 0.11695 0.11115 1.052 0.293
## s_iv_quality_of_evidenceQu Yes -0.06672 0.11105 -0.601 0.548
## plsChristodoulou -0.99358 0.10552 -9.416 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_OY s____Y
## s_v_prtnlOY -0.446
## s_v_qlt__QY -0.482 0.015
## plsChristdl -0.558 -0.025 0.009# Interest in psychological research
m6 <- glmer(knowledge ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + s_int_psy + (1 | id), family = binomial("logit"), data = df_cama[!is.na(df_cama$s_int_psy),], glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 100000)))
summary(m6)## Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
## Family: binomial ( logit )
## Formula: knowledge ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + s_int_psy + (1 | id)
## Data: df_cama[!is.na(df_cama$s_int_psy), ]
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 2772.0 2806.3 -1380.0 2760.0 2224
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.0150 -0.9276 0.4529 0.6722 1.2087
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 0.7265 0.8524
## Number of obs: 2230, groups: id, 1147
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.20244 0.11834 10.161 < 2e-16 ***
## s_iv_operationalizationOp Yes 0.13505 0.11081 1.219 0.223
## s_iv_quality_of_evidenceQu Yes -0.06786 0.11061 -0.614 0.540
## plsChristodoulou -0.99565 0.10554 -9.434 < 2e-16 ***
## s_int_psy 0.24902 0.05617 4.433 9.28e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_OY s____Y plsChr
## s_v_prtnlOY -0.444
## s_v_qlt__QY -0.484 0.016
## plsChristdl -0.558 -0.027 0.010
## s_int_psy 0.065 0.046 -0.006 -0.085m7 <- glmer(knowledge ~ s_iv_operationalization * s_int_psy + s_iv_quality_of_evidence + pls + (1 | id), family = binomial("logit"), data = df_cama[!is.na(df_cama$s_int_psy),], glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 100000)))
summary(m7)## Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
## Family: binomial ( logit )
## Formula: knowledge ~ s_iv_operationalization * s_int_psy + s_iv_quality_of_evidence + pls + (1 | id)
## Data: df_cama[!is.na(df_cama$s_int_psy), ]
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 2773.4 2813.4 -1379.7 2759.4 2223
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.0699 -0.9333 0.4521 0.6663 1.1865
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 0.727 0.8527
## Number of obs: 2230, groups: id, 1147
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.20622 0.11851 10.178 < 2e-16 ***
## s_iv_operationalizationOp Yes 0.13687 0.11089 1.234 0.21708
## s_int_psy 0.20750 0.07742 2.680 0.00736 **
## s_iv_quality_of_evidenceQu Yes -0.07313 0.11084 -0.660 0.50942
## plsChristodoulou -0.99597 0.10557 -9.434 < 2e-16 ***
## s_iv_operationalizationOp Yes:s_int_psy 0.08611 0.11126 0.774 0.43898
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_OY s_nt_p s____Y plsChr
## s_v_prtnlOY -0.442
## s_int_psy 0.007 0.021
## s_v_qlt__QY -0.486 0.014 0.038
## plsChristdl -0.559 -0.027 -0.049 0.011
## s_v_prOY:__ 0.056 0.023 -0.688 -0.062 -0.018anova(m1,m6,m7)## Data: df_cama[!is.na(df_cama$s_int_psy), ]
## Models:
## m1: knowledge ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + (1 | id)
## m6: knowledge ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + s_int_psy + (1 | id)
## m7: knowledge ~ s_iv_operationalization * s_int_psy + s_iv_quality_of_evidence + pls + (1 | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## m1 5 2790.2 2818.8 -1390.1 2780.2
## m6 6 2772.0 2806.3 -1380.0 2760.0 20.1909 1 7.009e-06 ***
## m7 7 2773.4 2813.4 -1379.7 2759.4 0.5999 1 0.4386
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1#ICC Subject
m0 <- glmer(knowledge ~ (1 | id), family = binomial("logit"), data = df_cama, glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 100000)))
summary(m0)## Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
## Family: binomial ( logit )
## Formula: knowledge ~ (1 | id)
## Data: df_cama
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 2886.3 2897.7 -1441.1 2882.3 2228
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.3169 -1.0937 0.6287 0.6287 0.7594
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 0.4524 0.6726
## Number of obs: 2230, groups: id, 1147
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.67196 0.05486 12.25 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1performance::icc(m0)## # Intraclass Correlation Coefficient
##
## Adjusted ICC: 0.121
## Conditional ICC: 0.121D Empowerment
m1 <- lmer(empowerment ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + (1 | id), data = df_cama)
summary(m1)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: empowerment ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + (1 | id)
## Data: df_cama
##
## REML criterion at convergence: 8863.1
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.11460 -0.45527 0.00239 0.45932 2.96964
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.814 1.677
## Residual 1.225 1.107
## Number of obs: 2260, groups: id, 1184
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.11035 0.09650 1316.33003 52.955 < 2e-16 ***
## s_iv_operationalizationOp Yes -0.21953 0.10835 1171.12888 -2.026 0.043 *
## s_iv_quality_of_evidenceQu Yes -0.04851 0.10835 1171.09519 -0.448 0.654
## plsChristodoulou -0.27469 0.04736 1097.86035 -5.800 8.65e-09 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_OY s____Y
## s_v_prtnlOY -0.563
## s_v_qlt__QY -0.562 0.012
## plsChristdl -0.242 -0.003 -0.002r.squaredGLMM(m1)## R2m R2c
## [1,] 0.007757616 0.6990427r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.008 0.018 0.003
## 4 plsChristodoulou 0.005 0.012 0.001
## 2 s_iv_operationalizationOp Yes 0.003 0.009 0.000
## 3 s_iv_quality_of_evidenceQu Yes 0.000 0.003 0.000anova_stats(m1)## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## term | sumsq | meansq | NumDF | DenDF | statistic | p.value | df | etasq | partial.etasq | omegasq | partial.omegasq | epsilonsq | cohens.f | power
## -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
## s_iv_operationalization | 5.029 | 5.029 | 1 | 1171.129 | 4.105 | 0.043 | | 0.003 | 0.003 | | | | 0.056 |
## s_iv_quality_of_evidence | 0.246 | 0.246 | 1 | 1171.095 | 0.200 | 0.654 | | 0.000 | 0.000 | | | | 0.012 |
## pls | 41.212 | 41.212 | 1 | 1097.860 | 33.643 | < .001 | | 0.025 | 0.025 | | | | 0.161 |
## Residuals | 1591.937 | 0.704 | | | | | 2260 | | | | | | |anova_stats(m1)$term## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "s_iv_operationalization" "s_iv_quality_of_evidence" "pls" "Residuals"interpret_omega_squared(anova_stats(m1)$partial.etasq, rules = "field2013")## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "very small" "very small" "small" NA
## (Rules: field2013)m1 <- lmer(empowerment ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + (1 | id), data = df_cama[!is.na(df_cama$s_int_psy),])
summary(m1)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: empowerment ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + (1 | id)
## Data: df_cama[!is.na(df_cama$s_int_psy), ]
##
## REML criterion at convergence: 8863.1
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.11460 -0.45527 0.00239 0.45932 2.96964
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.814 1.677
## Residual 1.225 1.107
## Number of obs: 2260, groups: id, 1184
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.11035 0.09650 1316.33003 52.955 < 2e-16 ***
## s_iv_operationalizationOp Yes -0.21953 0.10835 1171.12888 -2.026 0.043 *
## s_iv_quality_of_evidenceQu Yes -0.04851 0.10835 1171.09519 -0.448 0.654
## plsChristodoulou -0.27469 0.04736 1097.86035 -5.800 8.65e-09 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_OY s____Y
## s_v_prtnlOY -0.563
## s_v_qlt__QY -0.562 0.012
## plsChristdl -0.242 -0.003 -0.002r.squaredGLMM(m1)## R2m R2c
## [1,] 0.007757616 0.6990427r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.008 0.018 0.003
## 4 plsChristodoulou 0.005 0.012 0.001
## 2 s_iv_operationalizationOp Yes 0.003 0.009 0.000
## 3 s_iv_quality_of_evidenceQu Yes 0.000 0.003 0.000# Interest in psychological research
m6 <- lmer(empowerment ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + s_int_psy + (1 | id), data = df_cama[!is.na(df_cama$s_int_psy),])
summary(m6)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: empowerment ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + s_int_psy + (1 | id)
## Data: df_cama[!is.na(df_cama$s_int_psy), ]
##
## REML criterion at convergence: 8679
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.3109 -0.4479 0.0078 0.4857 2.8654
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.312 1.521
## Residual 1.222 1.106
## Number of obs: 2260, groups: id, 1184
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.10726 0.08966 1342.29549 56.961 < 2e-16 ***
## s_iv_operationalizationOp Yes -0.17849 0.10023 1174.12249 -1.781 0.0752 .
## s_iv_quality_of_evidenceQu Yes -0.06604 0.10019 1173.70480 -0.659 0.5099
## plsChristodoulou -0.27293 0.04725 1106.19081 -5.776 9.95e-09 ***
## s_int_psy 0.70926 0.04973 1188.17611 14.263 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_OY s____Y plsChr
## s_v_prtnlOY -0.560
## s_v_qlt__QY -0.559 0.011
## plsChristdl -0.260 -0.003 -0.002
## s_int_psy -0.004 0.030 -0.012 0.001r.squaredGLMM(m6)## R2m R2c
## [1,] 0.1316673 0.6997381r2beta(m6, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.132 0.159 0.108
## 5 s_int_psy 0.125 0.150 0.101
## 4 plsChristodoulou 0.005 0.013 0.001
## 2 s_iv_operationalizationOp Yes 0.002 0.008 0.000
## 3 s_iv_quality_of_evidenceQu Yes 0.000 0.003 0.000m7 <- lmer(empowerment ~ s_iv_operationalization * s_int_psy + s_iv_quality_of_evidence + pls + (1 | id), data = df_cama[!is.na(df_cama$s_int_psy),])
summary(m7)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: empowerment ~ s_iv_operationalization * s_int_psy + s_iv_quality_of_evidence + pls + (1 | id)
## Data: df_cama[!is.na(df_cama$s_int_psy), ]
##
## REML criterion at convergence: 8678.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.2867 -0.4482 0.0158 0.4827 2.8389
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.305 1.518
## Residual 1.223 1.106
## Number of obs: 2260, groups: id, 1184
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.11375 0.08964 1341.44472 57.048 < 2e-16 ***
## s_iv_operationalizationOp Yes -0.17635 0.10012 1171.76880 -1.761 0.0784 .
## s_int_psy 0.62283 0.06931 1186.10354 8.986 < 2e-16 ***
## s_iv_quality_of_evidenceQu Yes -0.07587 0.10022 1172.16054 -0.757 0.4492
## plsChristodoulou -0.27286 0.04727 1105.25549 -5.773 1.01e-08 ***
## s_iv_operationalizationOp Yes:s_int_psy 0.17792 0.09952 1186.83321 1.788 0.0741 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_OY s_nt_p s____Y plsChr
## s_v_prtnlOY -0.559
## s_int_psy -0.031 0.013
## s_v_qlt__QY -0.560 0.011 0.030
## plsChristdl -0.260 -0.003 0.001 -0.002
## s_v_prOY:__ 0.040 0.012 -0.697 -0.055 0.001r.squaredGLMM(m7)## R2m R2c
## [1,] 0.1334905 0.6996498r2beta(m7, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.133 0.161 0.110
## 3 s_int_psy 0.053 0.073 0.037
## 5 plsChristodoulou 0.005 0.013 0.001
## 6 s_iv_operationalizationOp Yes:s_int_psy 0.002 0.008 0.000
## 2 s_iv_operationalizationOp Yes 0.002 0.008 0.000
## 4 s_iv_quality_of_evidenceQu Yes 0.000 0.004 0.000anova(m1,m6,m7)## refitting model(s) with ML (instead of REML)## Data: df_cama[!is.na(df_cama$s_int_psy), ]
## Models:
## m1: empowerment ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + (1 | id)
## m6: empowerment ~ s_iv_operationalization + s_iv_quality_of_evidence + pls + s_int_psy + (1 | id)
## m7: empowerment ~ s_iv_operationalization * s_int_psy + s_iv_quality_of_evidence + pls + (1 | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## m1 6 8861.6 8895.9 -4424.8 8849.6
## m6 7 8674.9 8714.9 -4330.4 8660.9 188.7249 1 <2e-16 ***
## m7 8 8673.7 8719.4 -4328.8 8657.7 3.2053 1 0.0734 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1#ICC Subject
m0 <- lmer(empowerment ~ (1 | id), data = df_cama)
summary(m0)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: empowerment ~ (1 | id)
## Data: df_cama
##
## REML criterion at convergence: 8891.1
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.1705 -0.4653 0.0261 0.5175 3.0588
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.799 1.673
## Residual 1.263 1.124
## Number of obs: 2260, groups: id, 1184
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.841e+00 5.421e-02 1.174e+03 89.3 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1performance::icc(m0)## # Intraclass Correlation Coefficient
##
## Adjusted ICC: 0.689
## Conditional ICC: 0.689Plots
library(readr)
library(tidyr)##
## Attache Paket: 'tidyr'## Die folgenden Objekte sind maskiert von 'package:Matrix':
##
## expand, pack, unpacklibrary(ggplot2)##
## Attache Paket: 'ggplot2'## Die folgenden Objekte sind maskiert von 'package:psych':
##
## %+%, alphalibrary(Hmisc)## Lade nötiges Paket: lattice## Lade nötiges Paket: Formula##
## Attache Paket: 'Hmisc'## Das folgende Objekt ist maskiert 'package:psych':
##
## describe## Die folgenden Objekte sind maskiert von 'package:base':
##
## format.pval, unitslibrary(plyr)##
## Attache Paket: 'plyr'## Die folgenden Objekte sind maskiert von 'package:Hmisc':
##
## is.discrete, summarizelibrary(RColorBrewer)
library(reshape2)##
## Attache Paket: 'reshape2'## Das folgende Objekt ist maskiert 'package:tidyr':
##
## smiths#R Code for Raincloud Plots
source("https://gist.githubusercontent.com/benmarwick/2a1bb0133ff568cbe28d/raw/fb53bd97121f7f9ce947837ef1a4c65a73bffb3f/geom_flat_violin.R")##
## Attache Paket: 'dplyr'## Die folgenden Objekte sind maskiert von 'package:plyr':
##
## arrange, count, desc, failwith, id, mutate, rename, summarise, summarize## Die folgenden Objekte sind maskiert von 'package:Hmisc':
##
## src, summarize## Das folgende Objekt ist maskiert 'package:MASS':
##
## select## Die folgenden Objekte sind maskiert von 'package:stats':
##
## filter, lag## Die folgenden Objekte sind maskiert von 'package:base':
##
## intersect, setdiff, setequal, unionraincloud_theme = theme(
text = element_text(size = 10),
axis.title.x = element_blank(),
axis.title.y = element_text(size = 16),
axis.text = element_text(size = 12),
axis.text.x = element_text(angle = 90, vjust = 0.5),
legend.title=element_text(size=16),
legend.text=element_text(size=16),
legend.position = "right",
plot.title = element_text(lineheight=.8, face="bold", size = 16),
panel.border = element_blank(),
panel.grid.minor = element_blank(),
panel.grid.major = element_blank(),
axis.line.x = element_line(colour = 'black', size=0.5, linetype='solid'),
axis.line.y = element_line(colour = 'black', size=0.5, linetype='solid'))
m000 <- lmer(accessibility ~ pls + (1 | id) + s_iv_quality_of_evidence, data = df_cama)
summary(m000)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: accessibility ~ pls + (1 | id) + s_iv_quality_of_evidence
## Data: df_cama
##
## REML criterion at convergence: 8520.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.8443 -0.3771 0.1624 0.3557 3.3325
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.6370 1.6239
## Residual 0.9824 0.9911
## Number of obs: 2262, groups: id, 1183
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 6.17208 0.07578 1366.07505 81.448 < 2e-16 ***
## plsChristodoulou -0.11325 0.04240 1095.81293 -2.671 0.00767 **
## s_iv_quality_of_evidenceQu Yes -0.09543 0.10347 1170.56191 -0.922 0.35659
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) plsChr
## plsChristdl -0.278
## s_v_qlt__QY -0.675 -0.002nrow(df_cama)## [1] 2274ranef_accessibility <- ranef(m000)$id[,1]
ranef_id <- as.numeric(rownames(ranef(m000)$id))
rain <- data[data$id %in% ranef_id,c("id","s_iv_operationalization", "s_iv_quality_of_evidence")]
rain <- rain[
with(rain, order(id)),
]
rain$accessibility <- ranef_accessibility
rain$PLS <- rain$s_iv_operationalization
levels(rain$PLS)## [1] "Op No" "Op Yes"levels(rain$PLS) <- c("Subsection Excluded", "Subsection Included")
p1 <- ggplot(data = rain, aes(y = accessibility, x = PLS, fill = PLS)) +
geom_flat_violin(position = position_nudge(x = .2, y = 0), alpha = .8) +
geom_point(aes(y = accessibility, color = PLS), position = position_jitter(width = .15), size = .5, alpha = 0.8) +
geom_boxplot(width = .1, guides = FALSE, outlier.shape = NA, alpha = 0.5) +
#expand_limits(y = 5.25) +
guides(fill = FALSE) +
guides(color = FALSE) +
scale_color_brewer(palette = "Dark2") +
scale_fill_brewer(palette = "Dark2") +
# coord_flip() +
theme_bw() +
raincloud_theme + ylab("Accessibility") + scale_x_discrete(guide = guide_axis(angle = 60)) + theme(axis.text.x=element_text(size=10))## Warning: Ignoring unknown parameters: guidesp1
# Understanding
m000 <- lmer(understanding ~ pls +(1 | id) + s_iv_quality_of_evidence, data = df_cama)
summary(m000)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: understanding ~ pls + (1 | id) + s_iv_quality_of_evidence
## Data: df_cama
##
## REML criterion at convergence: 8632.1
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.5998 -0.4578 0.0937 0.4486 2.9845
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.293 1.514
## Residual 1.198 1.094
## Number of obs: 2258, groups: id, 1184
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 6.22864 0.07388 1424.19254 84.304 < 2e-16 ***
## plsChristodoulou -0.20566 0.04682 1099.28526 -4.393 1.23e-05 ***
## s_iv_quality_of_evidenceQu Yes -0.11771 0.09964 1170.51435 -1.181 0.238
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) plsChr
## plsChristdl -0.315
## s_v_qlt__QY -0.668 -0.002nrow(df_cama)## [1] 2274ranef_understanding <- ranef(m000)$id[,1]
ranef_id <- as.numeric(rownames(ranef(m000)$id))
rain <- data[data$id %in% ranef_id,c("id","s_iv_operationalization", "s_iv_quality_of_evidence")]
rain <- rain[
with(rain, order(id)),
]
rain$understanding <- ranef_understanding
rain$PLS <- rain$s_iv_operationalization
levels(rain$PLS)## [1] "Op No" "Op Yes"levels(rain$PLS) <- c("Subsection Excluded", "Subsection Included")
p2 <- ggplot(data = rain, aes(y = understanding, x = PLS, fill = PLS)) +
geom_flat_violin(position = position_nudge(x = .2, y = 0), alpha = .8) +
geom_point(aes(y = understanding, color = PLS), position = position_jitter(width = .15), size = .5, alpha = 0.8) +
geom_boxplot(width = .1, guides = FALSE, outlier.shape = NA, alpha = 0.5) +
#expand_limits(x = 5.25) +
guides(fill = FALSE) +
guides(color = FALSE) +
scale_color_brewer(palette = "Dark2") +
scale_fill_brewer(palette = "Dark2") +
# coord_flip() +
theme_bw() +
raincloud_theme + ylab("Understanding") + scale_x_discrete(guide = guide_axis(angle = 60)) + theme(axis.text.x=element_text(size=10)) #+ labs(caption="Operationalization") + ## Warning: Ignoring unknown parameters: guides # theme(plot.caption = element_text(hjust=0.5, size=rel(1.2)))
p2
# Empowerment
m000 <- lmer(empowerment ~ pls + (1 | id) + s_iv_quality_of_evidence, data = df_cama)
summary(m000)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: empowerment ~ pls + (1 | id) + s_iv_quality_of_evidence
## Data: df_cama
##
## REML criterion at convergence: 8864.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.0971 -0.4675 -0.0073 0.4683 2.9876
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.824 1.681
## Residual 1.225 1.107
## Number of obs: 2260, groups: id, 1184
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.00034 0.07990 1392.43451 62.586 < 2e-16 ***
## plsChristodoulou -0.27495 0.04735 1098.52844 -5.807 8.33e-09 ***
## s_iv_quality_of_evidenceQu Yes -0.04590 0.10851 1172.85147 -0.423 0.672
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) plsChr
## plsChristdl -0.294
## s_v_qlt__QY -0.672 -0.002nrow(df_cama)## [1] 2274ranef_empowerment <- ranef(m000)$id[,1]
ranef_id <- as.numeric(rownames(ranef(m000)$id))
rain <- data[data$id %in% ranef_id,c("id","s_iv_operationalization", "s_iv_quality_of_evidence")]
rain <- rain[
with(rain, order(id)),
]
rain$empowerment <- ranef_empowerment
rain$PLS <- rain$s_iv_operationalization
levels(rain$PLS)## [1] "Op No" "Op Yes"levels(rain$PLS) <- c("Subsection Excluded", "Subsection Included")
p3 <- ggplot(data = rain, aes(y = empowerment, x = PLS, fill = PLS)) +
geom_flat_violin(position = position_nudge(x = .2, y = 0), alpha = .8) +
geom_point(aes(y = empowerment, color = PLS), position = position_jitter(width = .15), size = .5, alpha = 0.8) +
geom_boxplot(width = .1, guides = FALSE, outlier.shape = NA, alpha = 0.5) +
#expand_limits(x = 5.25) +
guides(fill = FALSE) +
guides(color = FALSE) +
scale_color_brewer(palette = "Dark2") +
scale_fill_brewer(palette = "Dark2") +
# coord_flip() +
theme_bw() +
raincloud_theme + ylab("Empowerment") + scale_x_discrete(guide = guide_axis(angle = 60)) + theme(axis.text.x=element_text(size=10)) ## Warning: Ignoring unknown parameters: guidesp3
# Accessibility
m000 <- lmer(accessibility ~ pls + (1 | id) + s_iv_quality_of_evidence, data = df_resilience)
summary(m000)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: accessibility ~ pls + (1 | id) + s_iv_quality_of_evidence
## Data: df_resilience
##
## REML criterion at convergence: 8455.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.02142 -0.50125 0.03743 0.54772 2.83362
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.723 1.650
## Residual 1.674 1.294
## Number of obs: 2064, groups: id, 1101
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.34550 0.08454 1377.13171 63.233 <2e-16 ***
## plsGroth -0.54952 0.05817 1010.33584 -9.447 <2e-16 ***
## s_iv_quality_of_evidenceQu Yes -0.10790 0.11527 1094.69919 -0.936 0.349
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) plsGrt
## plsGroth -0.346
## s_v_qlt__QY -0.647 0.003nrow(df_resilience)## [1] 2076ranef_accessibility <- ranef(m000)$id[,1]
ranef_id <- as.numeric(rownames(ranef(m000)$id))
rain <- data[data$id %in% ranef_id,c("id","s_iv_technical_terms", "s_iv_quality_of_evidence")]
rain <- rain[
with(rain, order(id)),
]
rain$accessibility <- ranef_accessibility
rain$PLS <- rain$s_iv_technical_terms
levels(rain$PLS)## [1] "Expl None" "Replace" "Expl Glossar"levels(rain$PLS) <- c("No Explanation", "Replacement", "Glossary")
p4 <- ggplot(data = rain, aes(y = accessibility, x = PLS, fill = PLS)) +
geom_flat_violin(position = position_nudge(x = .2, y = 0), alpha = .8) +
geom_point(aes(y = accessibility, color = PLS), position = position_jitter(width = .15), size = .5, alpha = 0.8) +
geom_boxplot(width = .1, guides = FALSE, outlier.shape = NA, alpha = 0.5) +
#expand_limits(x = 5.25) +
guides(fill = FALSE) +
guides(color = FALSE) +
scale_color_brewer(palette = "Dark2") +
scale_fill_brewer(palette = "Dark2") +
# coord_flip() +
theme_bw() +
raincloud_theme + ylab("Accessibility") + scale_x_discrete(guide = guide_axis(angle = 60)) + theme(axis.text.x=element_text(size=10))## Warning: Ignoring unknown parameters: guidesp4
#understanding
m000 <- lmer(understanding ~ pls + (1 | id) + s_iv_quality_of_evidence, data = df_resilience)
summary(m000)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: understanding ~ pls + (1 | id) + s_iv_quality_of_evidence
## Data: df_resilience
##
## REML criterion at convergence: 8309.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.1648 -0.4624 0.0636 0.5270 2.9198
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.582 1.607
## Residual 1.537 1.240
## Number of obs: 2065, groups: id, 1101
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.43412 0.08190 1372.46175 66.350 <2e-16 ***
## plsGroth -0.54391 0.05574 1009.93724 -9.758 <2e-16 ***
## s_iv_quality_of_evidenceQu Yes -0.15709 0.11179 1094.61595 -1.405 0.16
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) plsGrt
## plsGroth -0.343
## s_v_qlt__QY -0.648 0.003nrow(df_resilience)## [1] 2076ranef_understanding <- ranef(m000)$id[,1]
ranef_id <- as.numeric(rownames(ranef(m000)$id))
rain <- data[data$id %in% ranef_id,c("id","s_iv_technical_terms", "s_iv_quality_of_evidence")]
rain <- rain[
with(rain, order(id)),
]
rain$understanding <- ranef_understanding
rain$PLS <- rain$s_iv_technical_terms
levels(rain$PLS)## [1] "Expl None" "Replace" "Expl Glossar"levels(rain$PLS) <- c("No Explanation", "Replacement", "Glossary")
p5 <- ggplot(data = rain, aes(y = understanding, x = PLS, fill = PLS)) +
geom_flat_violin(position = position_nudge(x = .2, y = 0), alpha = .8) +
geom_point(aes(y = understanding, color = PLS), position = position_jitter(width = .15), size = .5, alpha = 0.8) +
geom_boxplot(width = .1, guides = FALSE, outlier.shape = NA, alpha = 0.5) +
#expand_limits(x = 5.25) +
guides(fill = FALSE) +
guides(color = FALSE) +
scale_color_brewer(palette = "Dark2") +
scale_fill_brewer(palette = "Dark2") +
# coord_flip() +
theme_bw() +
raincloud_theme + ylab("Understanding") + scale_x_discrete(guide = guide_axis(angle = 60)) + theme(axis.text.x=element_text(size=10)) + theme(axis.text.x=element_text(size=10)) #+ labs(caption="Technical Terms") + theme(plot.caption = element_text(hjust=0.5, size=rel(1.2)))## Warning: Ignoring unknown parameters: guidesp5
# empowerment
m000 <- lmer(empowerment ~ pls + (1 | id) + s_iv_quality_of_evidence, data = df_resilience)
summary(m000)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: empowerment ~ pls + (1 | id) + s_iv_quality_of_evidence
## Data: df_resilience
##
## REML criterion at convergence: 8222.8
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.1736 -0.5052 -0.0169 0.4751 3.3095
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 2.917 1.708
## Residual 1.316 1.147
## Number of obs: 2064, groups: id, 1101
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.30944 0.08346 1319.79525 51.635 < 2e-16 ***
## plsGroth -0.43069 0.05171 998.79986 -8.330 2.66e-16 ***
## s_iv_quality_of_evidenceQu Yes -0.03520 0.11526 1092.83988 -0.305 0.76
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) plsGrt
## plsGroth -0.312
## s_v_qlt__QY -0.655 0.004nrow(df_resilience)## [1] 2076ranef_empowerment <- ranef(m000)$id[,1]
ranef_id <- as.numeric(rownames(ranef(m000)$id))
rain <- data[data$id %in% ranef_id,c("id","s_iv_technical_terms", "s_iv_quality_of_evidence")]
rain <- rain[
with(rain, order(id)),
]
rain$empowerment <- ranef_empowerment
rain$PLS <- rain$s_iv_technical_terms
levels(rain$PLS)## [1] "Expl None" "Replace" "Expl Glossar"levels(rain$PLS) <- c("No Explanation", "Replacement", "Glossary")
p6 <- ggplot(data = rain, aes(y = empowerment, x = PLS, fill = PLS)) +
geom_flat_violin(position = position_nudge(x = .2, y = 0), alpha = .8) +
geom_point(aes(y = empowerment, color = PLS), position = position_jitter(width = .15), size = .5, alpha = 0.8) +
geom_boxplot(width = .1, guides = FALSE, outlier.shape = NA, alpha = 0.5) +
#expand_limits(x = 5.25) +
guides(fill = FALSE) +
guides(color = FALSE) +
scale_color_brewer(palette = "Dark2") +
scale_fill_brewer(palette = "Dark2") +
# coord_flip() +
theme_bw() +
raincloud_theme + ylab("Empowerment") + scale_x_discrete(guide = guide_axis(angle = 60)) + theme(axis.text.x=element_text(size=10))## Warning: Ignoring unknown parameters: guidesp6
require(ggpubr)## Lade nötiges Paket: ggpubr## Registered S3 methods overwritten by 'car':
## method from
## influence.merMod lme4
## cooks.distance.influence.merMod lme4
## dfbeta.influence.merMod lme4
## dfbetas.influence.merMod lme4##
## Attache Paket: 'ggpubr'## Das folgende Objekt ist maskiert 'package:plyr':
##
## mutatelibrary(cowplot)##
## Attache Paket: 'cowplot'## Das folgende Objekt ist maskiert 'package:ggpubr':
##
## get_legendlibrary(ggimage)##
## Attache Paket: 'ggimage'## Das folgende Objekt ist maskiert 'package:cowplot':
##
## theme_nothing## Das folgende Objekt ist maskiert 'package:ggpubr':
##
## theme_transparentp9 <- plot_grid(p1, p2, p3, p4, p5, p6, ncol=3, labels=LETTERS[1:6])
p7 <- plot_grid(p1, p2, p3, ncol=3, labels=LETTERS[4:6])
p7 <- p7 + labs(caption="OPERATIONALIZATION") +
theme(plot.caption = element_text(hjust=0.5, size=rel(1.2)))
p8 <- plot_grid(p4, p5, p6, ncol=3, labels=LETTERS[1:3]) + labs(caption="TECHNICAL TERMS") + theme(plot.caption = element_text(hjust=0.5, size=rel(1.2)))
p10 <- plot_grid(p8, p7, ncol=1)
#ggsave("Figure_Study_1.tiff", p10, width = 16, height = 12, compression = "lzw")
#ggsave("Figure_Study_1.png", p10, width = 16, height = 12)
#pref_quality_of_evidence
m000 <- lmer(pref_quality_of_evidence ~ (1|pls) + (1 | id), data = df)
summary(m000)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: pref_quality_of_evidence ~ (1 | pls) + (1 | id)
## Data: df
##
## REML criterion at convergence: 16501.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.95716 -0.40792 -0.00403 0.46097 2.85804
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.68698 1.2988
## pls (Intercept) 0.01472 0.1213
## Residual 1.53392 1.2385
## Number of obs: 4276, groups: id, 2264; pls, 4
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.73135 0.06924 3.99626 68.33 2.78e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1ranef_accessibility <- ranef(m000)$id[,1]
ranef_id <- as.numeric(rownames(ranef(m000)$id))
rain <- data[data$id %in% ranef_id,c("id","s_iv_operationalization", "s_iv_quality_of_evidence")]
rain <- rain[
with(rain, order(id)),
]
rain$accessibility <- ranef_accessibility
rain$PLS <- rain$s_iv_quality_of_evidence
levels(rain$PLS)## [1] "Qu No" "Qu Yes"levels(rain$PLS) <- c("No Quality of Evidence Statement", "Quality of Evidence Statement")
p1 <- ggplot(data = rain, aes(y = accessibility, x = PLS, fill = PLS)) +
geom_flat_violin(position = position_nudge(x = .2, y = 0), alpha = .8) +
geom_point(aes(y = accessibility, color = PLS), position = position_jitter(width = .15), size = .5, alpha = 0.8) +
geom_boxplot(width = .1, guides = FALSE, outlier.shape = NA, alpha = 0.5) +
#expand_limits(y = 5.25) +
guides(fill = FALSE) +
guides(color = FALSE) +
scale_color_brewer(palette = "Dark2") +
scale_fill_brewer(palette = "Dark2") +
# coord_flip() +
theme_bw() +
raincloud_theme + ylab("Preference of Meta-Analytical Evidence") + xlab("Preference of Meta-Analytical Evidence") + scale_x_discrete(guide = guide_axis(title = "Quality of Evidence", angle = 0)) + theme(axis.text.x=element_text(size=10)) + theme(axis.title = element_text(size=10)) ## Warning: Ignoring unknown parameters: guidesp1
#ggsave("Figure_Study_1_Quality_of_Evidence.tiff", p1, width = 8, height = 6, compression = "lzw")
#ggsave("Figure_Study_1_Quality_of_Evidence.png", p1, width = 8, height = 6)Study 2
Read data files
Note: Anomynized datasets were created by removing possibly identifying information and data that are not relevant to confirmatory analyses.
options(width=120)
df <- read.csv2(file = "20210901_study_2_df_long.csv",fileEncoding="UTF-8", stringsAsFactors = T)
df$s_iv_statistics <- factor(df$s_iv_statistics, levels = c("eff + qual + gloss","eff + qual","eff + gloss", "qual"))
df$s_iv_quality_of_evidence <- factor(df$s_iv_quality_of_evidence, levels = c("normal", "extended"))
df$s_iv_complex <- factor(df$s_iv_complex, levels = c("uncomplex","complex"))
df$s_iv_structured <- factor(df$s_iv_structured, levels = c("unstructured","structured"))
df$pls <- factor(df$pls, levels = c("Schwalm","Bucher", "Bergmann","Yule"))
df_complex <- df[df$s_iv_complex == "complex",]
df_uncomplex <- df[df$s_iv_complex == "uncomplex",]
data <- read.csv2(file = "20210719_study_2_df_wide.csv",fileEncoding="UTF-8", stringsAsFactors = T)
data$s_iv_complex <- factor(data$s_iv_complex, levels = c("uncomplex","complex"))
data$s_iv_quality_of_evidence <- factor(data$s_iv_quality_of_evidence, levels = c("normal", "extended"))
data$s_iv_statistics <- factor(data$s_iv_statistics, levels = c("eff + qual + gloss","eff + qual","eff + gloss", "qual"))
data$s_iv_structured <- factor(data$s_iv_structured, levels = c("unstructured","structured"))Descriptive Within Level
Complex
names(df_complex)## [1] "id" "s_iv_complex" "s_iv_quality_of_evidence" "s_iv_structured"
## [5] "s_iv_statistics" "s_knowledge_evidence" "pls" "accessibility"
## [9] "understanding" "empowerment" "knowledge_evidence" "knowledge"df1 <- df_complex[,c("accessibility", "understanding", "empowerment","knowledge_evidence","knowledge")]
describe(df1)## df1
##
## 5 Variables 2116 Observations
## ------------------------------------------------------------------------------------------------------------------------
## accessibility
## n missing distinct Info Mean Gmd
## 2100 16 8 0.968 5.768 2.016
##
## lowest : 1 2 3 4 5, highest: 4 5 6 7 8
##
## Value 1 2 3 4 5 6 7 8
## Frequency 57 57 134 271 288 472 384 437
## Proportion 0.027 0.027 0.064 0.129 0.137 0.225 0.183 0.208
## ------------------------------------------------------------------------------------------------------------------------
## understanding
## n missing distinct Info Mean Gmd
## 2098 18 8 0.961 6.034 1.89
##
## lowest : 1 2 3 4 5, highest: 4 5 6 7 8
##
## Value 1 2 3 4 5 6 7 8
## Frequency 45 37 100 203 310 472 402 529
## Proportion 0.021 0.018 0.048 0.097 0.148 0.225 0.192 0.252
## ------------------------------------------------------------------------------------------------------------------------
## empowerment
## n missing distinct Info Mean Gmd
## 2106 10 8 0.967 5.072 1.977
##
## lowest : 1 2 3 4 5, highest: 4 5 6 7 8
##
## Value 1 2 3 4 5 6 7 8
## Frequency 104 93 191 309 469 526 230 184
## Proportion 0.049 0.044 0.091 0.147 0.223 0.250 0.109 0.087
## ------------------------------------------------------------------------------------------------------------------------
## knowledge_evidence
## n missing distinct Info Sum Mean Gmd
## 2089 27 2 0.749 1008 0.4825 0.4996
##
## ------------------------------------------------------------------------------------------------------------------------
## knowledge
## n missing distinct Info Sum Mean Gmd
## 2089 27 2 0.659 1408 0.674 0.4397
##
## ------------------------------------------------------------------------------------------------------------------------describeBy(df_complex$accessibility, group = df_complex$s_iv_statistics)##
## Descriptive statistics by group
## group: eff + qual + gloss
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 500 5.52 1.9 6 5.66 1.48 1 8 7 -0.49 -0.57 0.09
## ------------------------------------------------------------------------------------------
## group: eff + qual
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 494 6.01 1.71 6 6.18 1.48 1 8 7 -0.64 -0.28 0.08
## ------------------------------------------------------------------------------------------
## group: eff + gloss
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 573 5.43 1.81 6 5.56 1.48 1 8 7 -0.59 -0.25 0.08
## ------------------------------------------------------------------------------------------
## group: qual
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 533 6.13 1.74 6 6.34 1.48 1 8 7 -0.87 0.25 0.08describeBy(df_complex$understanding, group = df_complex$s_iv_statistics)##
## Descriptive statistics by group
## group: eff + qual + gloss
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 498 5.76 1.84 6 5.96 1.48 1 8 7 -0.76 0 0.08
## ------------------------------------------------------------------------------------------
## group: eff + qual
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 496 6.32 1.57 7 6.51 1.48 1 8 7 -0.84 0.24 0.07
## ------------------------------------------------------------------------------------------
## group: eff + gloss
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 573 5.82 1.76 6 5.99 1.48 1 8 7 -0.71 0 0.07
## ------------------------------------------------------------------------------------------
## group: qual
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 531 6.25 1.66 6 6.43 1.48 1 8 7 -0.8 0.19 0.07describeBy(df_complex$empowerment, group = df_complex$s_iv_statistics)##
## Descriptive statistics by group
## group: eff + qual + gloss
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 500 4.85 1.95 5 4.95 1.48 1 8 7 -0.4 -0.62 0.09
## ------------------------------------------------------------------------------------------
## group: eff + qual
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 499 5.25 1.61 5 5.28 1.48 1 8 7 -0.28 -0.17 0.07
## ------------------------------------------------------------------------------------------
## group: eff + gloss
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 574 4.93 1.85 5 5.01 1.48 1 8 7 -0.35 -0.43 0.08
## ------------------------------------------------------------------------------------------
## group: qual
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 533 5.26 1.66 5 5.34 1.48 1 8 7 -0.51 0.03 0.07a <- aggregate(df_complex[, c("accessibility","understanding","empowerment")], list(df_complex$s_iv_statistics), mean, na.rm = TRUE)
b <- aggregate(df_complex[, c("accessibility","understanding","empowerment")], list(df_complex$s_iv_statistics), sd, na.rm = TRUE)
df5 <- cbind(t(a[,-c(1)]),t(b[,-c(1)]))
df5 <- as.data.frame(df5)
names(df5) <- c(paste(a[,1], "mean"),paste(b[,1], "sd"))
df5 <- df5[, order(names(df5))]
describeBy(df_complex$accessibility, group = df_complex$s_iv_quality_of_evidence)##
## Descriptive statistics by group
## group: normal
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 1028 5.73 1.86 6 5.92 1.48 1 8 7 -0.68 -0.17 0.06
## ------------------------------------------------------------------------------------------
## group: extended
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 1072 5.8 1.77 6 5.96 1.48 1 8 7 -0.59 -0.35 0.05describeBy(df_complex$understanding, group = df_complex$s_iv_quality_of_evidence)##
## Descriptive statistics by group
## group: normal
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 1030 5.99 1.78 6 6.2 1.48 1 8 7 -0.83 0.17 0.06
## ------------------------------------------------------------------------------------------
## group: extended
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 1068 6.08 1.68 6 6.25 1.48 1 8 7 -0.75 0.12 0.05describeBy(df_complex$empowerment, group = df_complex$s_iv_quality_of_evidence)##
## Descriptive statistics by group
## group: normal
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 1033 5 1.82 5 5.1 1.48 1 8 7 -0.43 -0.33 0.06
## ------------------------------------------------------------------------------------------
## group: extended
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 1073 5.14 1.74 5 5.22 1.48 1 8 7 -0.42 -0.2 0.05a <- aggregate(df_complex[, c("accessibility","understanding","empowerment")], list(df_complex$s_iv_quality_of_evidence), mean, na.rm = TRUE)
b <- aggregate(df_complex[, c("accessibility","understanding","empowerment")], list(df_complex$s_iv_quality_of_evidence), sd, na.rm = TRUE)
df5 <- cbind(t(a[,-c(1)]),t(b[,-c(1)]))
df5 <- as.data.frame(df5)
names(df5) <- c(paste(a[,1], "mean"),paste(b[,1], "sd"))
df5 <- df5[, order(names(df5))]
describeBy(df_complex$accessibility, group = df_complex$s_iv_structured)##
## Descriptive statistics by group
## group: unstructured
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 1026 5.66 1.83 6 5.82 1.48 1 8 7 -0.64 -0.21 0.06
## ------------------------------------------------------------------------------------------
## group: structured
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 1074 5.87 1.8 6 6.04 1.48 1 8 7 -0.65 -0.28 0.05describeBy(df_complex$understanding, group = df_complex$s_iv_structured)##
## Descriptive statistics by group
## group: unstructured
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 1026 5.95 1.73 6 6.14 1.48 1 8 7 -0.79 0.17 0.05
## ------------------------------------------------------------------------------------------
## group: structured
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 1072 6.11 1.72 6 6.31 1.48 1 8 7 -0.81 0.18 0.05describeBy(df_complex$empowerment, group = df_complex$s_iv_structured)##
## Descriptive statistics by group
## group: unstructured
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 1031 5 1.77 5 5.08 1.48 1 8 7 -0.4 -0.29 0.06
## ------------------------------------------------------------------------------------------
## group: structured
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 1075 5.14 1.78 5 5.24 1.48 1 8 7 -0.46 -0.22 0.05a <- aggregate(df_complex[, c("accessibility","understanding","empowerment")], list(df_complex$s_iv_structured), mean, na.rm = TRUE)
b <- aggregate(df_complex[, c("accessibility","understanding","empowerment")], list(df_complex$s_iv_structured), sd, na.rm = TRUE)
df5 <- cbind(t(a[,-c(1)]),t(b[,-c(1)]))
df5 <- as.data.frame(df5)
names(df5) <- c(paste(a[,1], "mean"),paste(b[,1], "sd"))
df5 <- df5[, order(names(df5))]
#correlations
cor(df_complex$understanding, df_complex$accessibility, use = "pairwise.complete.obs")## [1] 0.6930322cor(df_complex$understanding, df_complex$empowerment, use = "pairwise.complete.obs")## [1] 0.6248939cor(df_complex$accessibility, df_complex$empowerment, use = "pairwise.complete.obs")## [1] 0.5557023Uncomplex
names(df_uncomplex)## [1] "id" "s_iv_complex" "s_iv_quality_of_evidence" "s_iv_structured"
## [5] "s_iv_statistics" "s_knowledge_evidence" "pls" "accessibility"
## [9] "understanding" "empowerment" "knowledge_evidence" "knowledge"df1 <- df_uncomplex[,c("accessibility", "understanding", "empowerment", "knowledge_evidence","knowledge")]
describe(df1)## df1
##
## 5 Variables 2190 Observations
## ------------------------------------------------------------------------------------------------------------------------
## accessibility
## n missing distinct Info Mean Gmd
## 2164 26 8 0.965 5.927 1.983
##
## lowest : 1 2 3 4 5, highest: 4 5 6 7 8
##
## Value 1 2 3 4 5 6 7 8
## Frequency 48 51 126 260 287 443 428 521
## Proportion 0.022 0.024 0.058 0.120 0.133 0.205 0.198 0.241
## ------------------------------------------------------------------------------------------------------------------------
## understanding
## n missing distinct Info Mean Gmd
## 2173 17 8 0.957 6.126 1.909
##
## lowest : 1 2 3 4 5, highest: 4 5 6 7 8
##
## Value 1 2 3 4 5 6 7 8
## Frequency 32 48 98 232 299 414 417 633
## Proportion 0.015 0.022 0.045 0.107 0.138 0.191 0.192 0.291
## ------------------------------------------------------------------------------------------------------------------------
## empowerment
## n missing distinct Info Mean Gmd
## 2175 15 8 0.97 5.103 2.008
##
## lowest : 1 2 3 4 5, highest: 4 5 6 7 8
##
## Value 1 2 3 4 5 6 7 8
## Frequency 105 100 189 328 477 506 265 205
## Proportion 0.048 0.046 0.087 0.151 0.219 0.233 0.122 0.094
## ------------------------------------------------------------------------------------------------------------------------
## knowledge_evidence
## n missing distinct Info Sum Mean Gmd
## 2161 29 2 0.749 1037 0.4799 0.4994
##
## ------------------------------------------------------------------------------------------------------------------------
## knowledge
## n missing distinct Info Sum Mean Gmd
## 2161 29 2 0.577 1600 0.7404 0.3846
##
## ------------------------------------------------------------------------------------------------------------------------describeBy(df_uncomplex$accessibility, group = df_uncomplex$s_iv_statistics)##
## Descriptive statistics by group
## group: eff + qual + gloss
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 542 5.69 1.84 6 5.85 1.48 1 8 7 -0.59 -0.3 0.08
## ------------------------------------------------------------------------------------------
## group: eff + qual
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 601 6.12 1.76 6 6.33 1.48 1 8 7 -0.79 -0.13 0.07
## ------------------------------------------------------------------------------------------
## group: eff + gloss
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 491 5.54 1.86 6 5.67 1.48 1 8 7 -0.52 -0.52 0.08
## ------------------------------------------------------------------------------------------
## group: qual
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 530 6.3 1.61 7 6.48 1.48 1 8 7 -0.91 0.45 0.07describeBy(df_uncomplex$understanding, group = df_uncomplex$s_iv_statistics)##
## Descriptive statistics by group
## group: eff + qual + gloss
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 544 5.93 1.84 6 6.14 1.48 1 8 7 -0.74 -0.13 0.08
## ------------------------------------------------------------------------------------------
## group: eff + qual
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 601 6.24 1.74 7 6.45 1.48 1 8 7 -0.83 -0.03 0.07
## ------------------------------------------------------------------------------------------
## group: eff + gloss
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 491 5.93 1.72 6 6.09 1.48 1 8 7 -0.62 -0.2 0.08
## ------------------------------------------------------------------------------------------
## group: qual
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 537 6.37 1.63 7 6.58 1.48 1 8 7 -0.86 -0.03 0.07describeBy(df_uncomplex$empowerment, group = df_uncomplex$s_iv_statistics)##
## Descriptive statistics by group
## group: eff + qual + gloss
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 547 5.03 1.89 5 5.12 1.48 1 8 7 -0.43 -0.45 0.08
## ------------------------------------------------------------------------------------------
## group: eff + qual
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 602 5.23 1.84 5 5.34 1.48 1 8 7 -0.5 -0.29 0.08
## ------------------------------------------------------------------------------------------
## group: eff + gloss
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 493 4.85 1.75 5 4.93 1.48 1 8 7 -0.32 -0.44 0.08
## ------------------------------------------------------------------------------------------
## group: qual
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 533 5.27 1.68 5 5.33 1.48 1 8 7 -0.35 -0.13 0.07a <- aggregate(df_uncomplex[, c("accessibility","understanding","empowerment")], list(df_uncomplex$s_iv_statistics), mean, na.rm = TRUE)
b <- aggregate(df_uncomplex[, c("accessibility","understanding","empowerment")], list(df_uncomplex$s_iv_statistics), sd, na.rm = TRUE)
df5 <- cbind(t(a[,-c(1)]),t(b[,-c(1)]))
df5 <- as.data.frame(df5)
names(df5) <- c(paste(a[,1], "mean"),paste(b[,1], "sd"))
df5 <- df5[, order(names(df5))]
describeBy(df_uncomplex$accessibility, group = df_uncomplex$s_iv_quality_of_evidence)##
## Descriptive statistics by group
## group: normal
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 1065 5.93 1.83 6 6.13 1.48 1 8 7 -0.73 -0.12 0.06
## ------------------------------------------------------------------------------------------
## group: extended
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 1099 5.92 1.76 6 6.1 1.48 1 8 7 -0.69 -0.25 0.05describeBy(df_uncomplex$understanding, group = df_uncomplex$s_iv_quality_of_evidence)##
## Descriptive statistics by group
## group: normal
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 1072 6.04 1.78 6 6.25 1.48 1 8 7 -0.77 -0.06 0.05
## ------------------------------------------------------------------------------------------
## group: extended
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 1101 6.21 1.7 7 6.4 1.48 1 8 7 -0.77 -0.13 0.05describeBy(df_uncomplex$empowerment, group = df_uncomplex$s_iv_quality_of_evidence)##
## Descriptive statistics by group
## group: normal
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 1074 5.03 1.82 5 5.13 1.48 1 8 7 -0.44 -0.36 0.06
## ------------------------------------------------------------------------------------------
## group: extended
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 1101 5.18 1.78 5 5.25 1.48 1 8 7 -0.39 -0.29 0.05a <- aggregate(df_uncomplex[, c("accessibility","understanding","empowerment")], list(df_uncomplex$s_iv_quality_of_evidence), mean, na.rm = TRUE)
b <- aggregate(df_uncomplex[, c("accessibility","understanding","empowerment")], list(df_uncomplex$s_iv_quality_of_evidence), sd, na.rm = TRUE)
df5 <- cbind(t(a[,-c(1)]),t(b[,-c(1)]))
df5 <- as.data.frame(df5)
names(df5) <- c(paste(a[,1], "mean"),paste(b[,1], "sd"))
df5 <- df5[, order(names(df5))]
describeBy(df_uncomplex$accessibility, group = df_uncomplex$s_iv_structured)##
## Descriptive statistics by group
## group: unstructured
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 1128 5.99 1.71 6 6.16 1.48 1 8 7 -0.69 -0.11 0.05
## ------------------------------------------------------------------------------------------
## group: structured
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 1036 5.86 1.88 6 6.05 1.48 1 8 7 -0.7 -0.3 0.06describeBy(df_uncomplex$understanding, group = df_uncomplex$s_iv_structured)##
## Descriptive statistics by group
## group: unstructured
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 1136 6.21 1.68 6 6.39 1.48 1 8 7 -0.76 -0.09 0.05
## ------------------------------------------------------------------------------------------
## group: structured
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 1037 6.04 1.8 6 6.25 1.48 1 8 7 -0.77 -0.12 0.06describeBy(df_uncomplex$empowerment, group = df_uncomplex$s_iv_structured)##
## Descriptive statistics by group
## group: unstructured
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 1135 5.19 1.78 5 5.27 1.48 1 8 7 -0.42 -0.3 0.05
## ------------------------------------------------------------------------------------------
## group: structured
## vars n mean sd median trimmed mad min max range skew kurtosis se
## X1 1 1040 5 1.82 5 5.09 1.48 1 8 7 -0.41 -0.35 0.06a <- aggregate(df_uncomplex[, c("accessibility","understanding","empowerment")], list(df_uncomplex$s_iv_structured), mean, na.rm = TRUE)
b <- aggregate(df_uncomplex[, c("accessibility","understanding","empowerment")], list(df_uncomplex$s_iv_structured), sd, na.rm = TRUE)
df5 <- cbind(t(a[,-c(1)]),t(b[,-c(1)]))
df5 <- as.data.frame(df5)
names(df5) <- c(paste(a[,1], "mean"),paste(b[,1], "sd"))
df5 <- df5[, order(names(df5))]
#correlations
cor(df_uncomplex$understanding, df_uncomplex$accessibility, use = "pairwise.complete.obs")## [1] 0.6781916cor(df_uncomplex$understanding, df_uncomplex$empowerment, use = "pairwise.complete.obs")## [1] 0.6064299cor(df_uncomplex$accessibility, df_uncomplex$empowerment, use = "pairwise.complete.obs")## [1] 0.5200545Confirmatory analysis
Complex
A Accessibility
m1 <- lmer(accessibility ~ s_iv_statistics + s_iv_quality_of_evidence + s_iv_structured + pls + (1 | id), data = df_complex)
summary(m1)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: accessibility ~ s_iv_statistics + s_iv_quality_of_evidence + s_iv_structured + pls + (1 | id)
## Data: df_complex
##
## REML criterion at convergence: 8013.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.3394 -0.4614 0.0925 0.5061 2.6429
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.844 1.358
## Residual 1.363 1.168
## Number of obs: 2100, groups: id, 1085
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.38422 0.12591 1177.37674 42.762 < 2e-16 ***
## s_iv_statisticseff + qual 0.49952 0.14108 1083.99090 3.541 0.000416 ***
## s_iv_statisticseff + gloss -0.06578 0.13646 1081.98645 -0.482 0.629841
## s_iv_statisticsqual 0.62976 0.13859 1081.27190 4.544 6.14e-06 ***
## s_iv_quality_of_evidenceextended 0.06255 0.09721 1080.84430 0.644 0.520032
## s_iv_structuredstructured 0.19660 0.09739 1080.63125 2.019 0.043772 *
## plsYule -0.01513 0.05145 1045.25710 -0.294 0.768846
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_+q s_v_+g s_v_stt s_v___ s_v_str
## s_v_sttst+q -0.561
## s_v_sttst+g -0.605 0.514
## s_v_sttstcs -0.575 0.505 0.523
## s_v_qlty_f_ -0.393 -0.013 0.011 0.003
## s_v_strctrd -0.426 0.024 0.064 0.020 -0.002
## plsYule -0.202 -0.002 -0.006 -0.004 0.001 0.004r.squaredGLMM(m1)## R2m R2c
## [1,] 0.03184354 0.588453r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.032 0.050 0.021
## 4 s_iv_statisticsqual 0.015 0.027 0.006
## 2 s_iv_statisticseff + qual 0.009 0.019 0.003
## 6 s_iv_structuredstructured 0.003 0.009 0.000
## 5 s_iv_quality_of_evidenceextended 0.000 0.004 0.000
## 3 s_iv_statisticseff + gloss 0.000 0.003 0.000
## 7 plsYule 0.000 0.002 0.000anova_stats(m1)## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## term | sumsq | meansq | NumDF | DenDF | statistic | p.value | df | etasq | partial.etasq | omegasq | partial.omegasq | epsilonsq | cohens.f | power
## -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
## s_iv_statistics | 53.790 | 17.930 | 3 | 1080.828 | 13.152 | < .001 | | 0.029 | 0.029 | | | | 0.173 |
## s_iv_quality_of_evidence | 0.565 | 0.565 | 1 | 1080.844 | 0.414 | 0.520 | | 0.000 | 0.000 | | | | 0.018 |
## s_iv_structured | 5.555 | 5.555 | 1 | 1080.631 | 4.075 | 0.044 | | 0.003 | 0.003 | | | | 0.056 |
## pls | 0.118 | 0.118 | 1 | 1045.257 | 0.086 | 0.769 | | 0.000 | 0.000 | | | | 0.008 |
## Residuals | 1794.135 | 0.854 | | | | | 2100 | | | | | | |anova_stats(m1)$term## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "s_iv_statistics" "s_iv_quality_of_evidence" "s_iv_structured" "pls"
## [5] "Residuals"interpret_omega_squared(anova_stats(m1)$partial.etasq, rules = "field2013")## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "small" "very small" "very small" "very small" NA
## (Rules: field2013)test = glht(m1,linfct=mcp(s_iv_statistics="Tukey"))
summary(test, test = adjusted("none"))##
## Simultaneous Tests for General Linear Hypotheses
##
## Multiple Comparisons of Means: Tukey Contrasts
##
##
## Fit: lmer(formula = accessibility ~ s_iv_statistics + s_iv_quality_of_evidence +
## s_iv_structured + pls + (1 | id), data = df_complex)
##
## Linear Hypotheses:
## Estimate Std. Error z value Pr(>|z|)
## eff + qual - eff + qual + gloss == 0 0.49952 0.14108 3.541 0.000399 ***
## eff + gloss - eff + qual + gloss == 0 -0.06578 0.13646 -0.482 0.629743
## qual - eff + qual + gloss == 0 0.62976 0.13859 4.544 5.52e-06 ***
## eff + gloss - eff + qual == 0 -0.56530 0.13683 -4.132 3.60e-05 ***
## qual - eff + qual == 0 0.13025 0.13909 0.936 0.349051
## qual - eff + gloss == 0 0.69555 0.13427 5.180 2.22e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- none method)summary(test, test = adjusted("BH"))##
## Simultaneous Tests for General Linear Hypotheses
##
## Multiple Comparisons of Means: Tukey Contrasts
##
##
## Fit: lmer(formula = accessibility ~ s_iv_statistics + s_iv_quality_of_evidence +
## s_iv_structured + pls + (1 | id), data = df_complex)
##
## Linear Hypotheses:
## Estimate Std. Error z value Pr(>|z|)
## eff + qual - eff + qual + gloss == 0 0.49952 0.14108 3.541 0.000599 ***
## eff + gloss - eff + qual + gloss == 0 -0.06578 0.13646 -0.482 0.629743
## qual - eff + qual + gloss == 0 0.62976 0.13859 4.544 1.66e-05 ***
## eff + gloss - eff + qual == 0 -0.56530 0.13683 -4.132 7.21e-05 ***
## qual - eff + qual == 0 0.13025 0.13909 0.936 0.418862
## qual - eff + gloss == 0 0.69555 0.13427 5.180 1.33e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- BH method)#ICC Subject
m0 <- lmer(accessibility ~ (1 | id), data = df_complex)
summary(m0)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: accessibility ~ (1 | id)
## Data: df_complex
##
## REML criterion at convergence: 8040.8
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.2807 -0.4881 0.0517 0.4978 2.7149
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.936 1.391
## Residual 1.363 1.168
## Number of obs: 2100, groups: id, 1085
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.768e+00 4.944e-02 1.085e+03 116.7 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1performance::icc(m0)## # Intraclass Correlation Coefficient
##
## Adjusted ICC: 0.587
## Conditional ICC: 0.587B Understanding
m1 <- lmer(understanding ~ s_iv_statistics + s_iv_quality_of_evidence + s_iv_structured + pls + (1 | id), data = df_complex)
summary(m1)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: understanding ~ s_iv_statistics + s_iv_quality_of_evidence + s_iv_structured + pls + (1 | id)
## Data: df_complex
##
## REML criterion at convergence: 7761.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.6324 -0.4330 0.1114 0.4866 2.9880
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.760 1.327
## Residual 1.162 1.078
## Number of obs: 2098, groups: id, 1086
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.617e+00 1.210e-01 1.168e+03 46.439 < 2e-16 ***
## s_iv_statisticseff + qual 5.721e-01 1.357e-01 1.084e+03 4.216 2.69e-05 ***
## s_iv_statisticseff + gloss 7.758e-02 1.314e-01 1.082e+03 0.591 0.554975
## s_iv_statisticsqual 5.071e-01 1.335e-01 1.082e+03 3.799 0.000154 ***
## s_iv_quality_of_evidenceextended 9.070e-02 9.353e-02 1.080e+03 0.970 0.332420
## s_iv_structuredstructured 1.680e-01 9.372e-02 1.080e+03 1.792 0.073334 .
## plsYule -6.832e-03 4.757e-02 1.041e+03 -0.144 0.885821
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_+q s_v_+g s_v_stt s_v___ s_v_str
## s_v_sttst+q -0.563
## s_v_sttst+g -0.607 0.515
## s_v_sttstcs -0.576 0.506 0.524
## s_v_qlty_f_ -0.393 -0.013 0.010 0.002
## s_v_strctrd -0.428 0.024 0.064 0.021 0.000
## plsYule -0.192 -0.003 -0.006 -0.007 -0.001 0.003anova_stats(m1)## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## term | sumsq | meansq | NumDF | DenDF | statistic | p.value | df | etasq | partial.etasq | omegasq | partial.omegasq | epsilonsq | cohens.f | power
## -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
## s_iv_statistics | 33.501 | 11.167 | 3 | 1080.542 | 9.614 | < .001 | | 0.022 | 0.022 | | | | 0.150 |
## s_iv_quality_of_evidence | 1.092 | 1.092 | 1 | 1080.418 | 0.940 | 0.332 | | 0.001 | 0.001 | | | | 0.027 |
## s_iv_structured | 3.732 | 3.732 | 1 | 1080.304 | 3.213 | 0.073 | | 0.002 | 0.002 | | | | 0.050 |
## pls | 0.024 | 0.024 | 1 | 1040.544 | 0.021 | 0.886 | | 0.000 | 0.000 | | | | 0.004 |
## Residuals | 1498.331 | 0.714 | | | | | 2098 | | | | | | |anova_stats(m1)$term## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "s_iv_statistics" "s_iv_quality_of_evidence" "s_iv_structured" "pls"
## [5] "Residuals"interpret_omega_squared(anova_stats(m1)$partial.etasq, rules = "field2013")## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "small" "very small" "very small" "very small" NA
## (Rules: field2013)test = glht(m1,linfct=mcp(s_iv_statistics="Tukey"))
summary(test, test = adjusted("none"))##
## Simultaneous Tests for General Linear Hypotheses
##
## Multiple Comparisons of Means: Tukey Contrasts
##
##
## Fit: lmer(formula = understanding ~ s_iv_statistics + s_iv_quality_of_evidence +
## s_iv_structured + pls + (1 | id), data = df_complex)
##
## Linear Hypotheses:
## Estimate Std. Error z value Pr(>|z|)
## eff + qual - eff + qual + gloss == 0 0.57207 0.13568 4.216 2.48e-05 ***
## eff + gloss - eff + qual + gloss == 0 0.07758 0.13137 0.591 0.554852
## qual - eff + qual + gloss == 0 0.50706 0.13348 3.799 0.000145 ***
## eff + gloss - eff + qual == 0 -0.49449 0.13153 -3.759 0.000170 ***
## qual - eff + qual == 0 -0.06500 0.13376 -0.486 0.627002
## qual - eff + gloss == 0 0.42949 0.12927 3.322 0.000892 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- none method)summary(test, test = adjusted("BH"))##
## Simultaneous Tests for General Linear Hypotheses
##
## Multiple Comparisons of Means: Tukey Contrasts
##
##
## Fit: lmer(formula = understanding ~ s_iv_statistics + s_iv_quality_of_evidence +
## s_iv_structured + pls + (1 | id), data = df_complex)
##
## Linear Hypotheses:
## Estimate Std. Error z value Pr(>|z|)
## eff + qual - eff + qual + gloss == 0 0.57207 0.13568 4.216 0.000149 ***
## eff + gloss - eff + qual + gloss == 0 0.07758 0.13137 0.591 0.627002
## qual - eff + qual + gloss == 0 0.50706 0.13348 3.799 0.000341 ***
## eff + gloss - eff + qual == 0 -0.49449 0.13153 -3.759 0.000341 ***
## qual - eff + qual == 0 -0.06500 0.13376 -0.486 0.627002
## qual - eff + gloss == 0 0.42949 0.12927 3.322 0.001338 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- BH method)r.squaredGLMM(m1)## R2m R2c
## [1,] 0.02427837 0.6121035r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.024 0.041 0.015
## 2 s_iv_statisticseff + qual 0.013 0.024 0.005
## 4 s_iv_statisticsqual 0.011 0.021 0.004
## 6 s_iv_structuredstructured 0.002 0.008 0.000
## 5 s_iv_quality_of_evidenceextended 0.001 0.005 0.000
## 3 s_iv_statisticseff + gloss 0.000 0.003 0.000
## 7 plsYule 0.000 0.002 0.000#ICC Subject
m0 <- lmer(understanding ~ (1 | id), data = df_complex)
summary(m0)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: understanding ~ (1 | id)
## Data: df_complex
##
## REML criterion at convergence: 7777.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.5909 -0.4558 0.1204 0.4422 2.9038
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.820 1.349
## Residual 1.162 1.078
## Number of obs: 2098, groups: id, 1086
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 6.030e+00 4.733e-02 1.084e+03 127.4 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1performance::icc(m0)## # Intraclass Correlation Coefficient
##
## Adjusted ICC: 0.610
## Conditional ICC: 0.610C Knowledge
table(df_complex$knowledge, df_complex$s_iv_statistics)/sum(table(df_complex$knowledge, df_complex$s_iv_statistics)) * 100##
## eff + qual + gloss eff + qual eff + gloss qual
## 0 9.047391 7.467688 8.329344 7.754907
## 1 14.648157 16.275730 19.004308 17.472475table(df_complex[,c("knowledge", "s_iv_statistics")])## s_iv_statistics
## knowledge eff + qual + gloss eff + qual eff + gloss qual
## 0 189 156 174 162
## 1 306 340 397 365# richtige Antworten je Bedingung
table(df_complex[,c("knowledge", "s_iv_statistics")])[2,]/colSums(table(df_complex[,c("knowledge", "s_iv_statistics")]))## eff + qual + gloss eff + qual eff + gloss qual
## 0.6181818 0.6854839 0.6952715 0.6925996#test
397/571## [1] 0.6952715table(df_complex[,c("knowledge", "s_iv_structured")])## s_iv_structured
## knowledge unstructured structured
## 0 329 352
## 1 691 717table(df_complex[,c("knowledge", "s_iv_structured")])[2,]/colSums(table(df_complex[,c("knowledge", "s_iv_structured")]))## unstructured structured
## 0.6774510 0.6707203table(df_complex[,c("knowledge", "s_iv_quality_of_evidence")])## s_iv_quality_of_evidence
## knowledge normal extended
## 0 307 374
## 1 721 687table(df_complex[,c("knowledge", "s_iv_quality_of_evidence")])[2,]/colSums(table(df_complex[,c("knowledge", "s_iv_quality_of_evidence")]))## normal extended
## 0.7013619 0.6475024m1 <- glmer(knowledge ~ s_iv_statistics + s_iv_quality_of_evidence + s_iv_structured + pls + (1 | id), family = binomial("logit"), data = df_complex, glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1000000)))
summary(m1)## Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
## Family: binomial ( logit )
## Formula: knowledge ~ s_iv_statistics + s_iv_quality_of_evidence + s_iv_structured + pls + (1 | id)
## Data: df_complex
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+06))
##
## AIC BIC logLik deviance df.resid
## 2575.1 2620.2 -1279.5 2559.1 2081
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.5494 -0.8584 0.4447 0.5145 0.9349
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.514 1.23
## Number of obs: 2089, groups: id, 1064
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.68192 0.17761 3.839 0.000123 ***
## s_iv_statisticseff + qual 0.40094 0.19204 2.088 0.036813 *
## s_iv_statisticseff + gloss 0.44108 0.18616 2.369 0.017818 *
## s_iv_statisticsqual 0.44826 0.19018 2.357 0.018423 *
## s_iv_quality_of_evidenceextended -0.32329 0.13386 -2.415 0.015728 *
## s_iv_structuredstructured -0.01943 0.13328 -0.146 0.884089
## plsYule 0.27716 0.10827 2.560 0.010468 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_+q s_v_+g s_v_stt s_v___ s_v_str
## s_v_sttst+q -0.507
## s_v_sttst+g -0.548 0.507
## s_v_sttstcs -0.513 0.497 0.514
## s_v_qlty_f_ -0.400 -0.033 -0.006 -0.018
## s_v_strctrd -0.421 0.026 0.064 0.023 0.004
## plsYule -0.274 0.013 0.011 0.012 -0.016 0.004test = glht(m1,linfct=mcp(s_iv_statistics ="Tukey"))
summary(test, test = adjusted("none"))##
## Simultaneous Tests for General Linear Hypotheses
##
## Multiple Comparisons of Means: Tukey Contrasts
##
##
## Fit: glmer(formula = knowledge ~ s_iv_statistics + s_iv_quality_of_evidence +
## s_iv_structured + pls + (1 | id), data = df_complex, family = binomial("logit"),
## control = glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+06)))
##
## Linear Hypotheses:
## Estimate Std. Error z value Pr(>|z|)
## eff + qual - eff + qual + gloss == 0 0.400945 0.192039 2.088 0.0368 *
## eff + gloss - eff + qual + gloss == 0 0.441084 0.186160 2.369 0.0178 *
## qual - eff + qual + gloss == 0 0.448258 0.190181 2.357 0.0184 *
## eff + gloss - eff + qual == 0 0.040139 0.187831 0.214 0.8308
## qual - eff + qual == 0 0.047313 0.191644 0.247 0.8050
## qual - eff + gloss == 0 0.007174 0.185516 0.039 0.9692
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- none method)summary(test, test = adjusted("BH"))##
## Simultaneous Tests for General Linear Hypotheses
##
## Multiple Comparisons of Means: Tukey Contrasts
##
##
## Fit: glmer(formula = knowledge ~ s_iv_statistics + s_iv_quality_of_evidence +
## s_iv_structured + pls + (1 | id), data = df_complex, family = binomial("logit"),
## control = glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+06)))
##
## Linear Hypotheses:
## Estimate Std. Error z value Pr(>|z|)
## eff + qual - eff + qual + gloss == 0 0.400945 0.192039 2.088 0.0736 .
## eff + gloss - eff + qual + gloss == 0 0.441084 0.186160 2.369 0.0553 .
## qual - eff + qual + gloss == 0 0.448258 0.190181 2.357 0.0553 .
## eff + gloss - eff + qual == 0 0.040139 0.187831 0.214 0.9692
## qual - eff + qual == 0 0.047313 0.191644 0.247 0.9692
## qual - eff + gloss == 0 0.007174 0.185516 0.039 0.9692
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- BH method)exp(fixef(m1))## (Intercept) s_iv_statisticseff + qual s_iv_statisticseff + gloss
## 1.9776761 1.4932351 1.5543919
## s_iv_statisticsqual s_iv_quality_of_evidenceextended s_iv_structuredstructured
## 1.5655826 0.7237674 0.9807567
## plsYule
## 1.3193828interpret_oddsratio(exp(fixef(m1)), rules = "cohen1988")## (Intercept) s_iv_statisticseff + qual s_iv_statisticseff + gloss
## "small" "small" "small"
## s_iv_statisticsqual s_iv_quality_of_evidenceextended s_iv_structuredstructured
## "small" "very small" "very small"
## plsYule
## "very small"
## (Rules: cohen1988)exp(summary(test, test = adjusted("none"))$test$coefficients)## eff + qual - eff + qual + gloss eff + gloss - eff + qual + gloss qual - eff + qual + gloss
## 1.493235 1.554392 1.565583
## eff + gloss - eff + qual qual - eff + qual qual - eff + gloss
## 1.040956 1.048450 1.007199interpret_oddsratio(exp(summary(test, test = adjusted("none"))$test$coefficients), rules = "cohen1988")## eff + qual - eff + qual + gloss eff + gloss - eff + qual + gloss qual - eff + qual + gloss
## "small" "small" "small"
## eff + gloss - eff + qual qual - eff + qual qual - eff + gloss
## "very small" "very small" "very small"
## (Rules: cohen1988)anova_stats(m1)## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: npar## term | npar | sumsq | meansq | statistic | df | etasq | partial.etasq | omegasq | partial.omegasq | epsilonsq | cohens.f | power
## ---------------------------------------------------------------------------------------------------------------------------------------------------------
## s_iv_statistics | 3 | 8.642 | 2.881 | 2.881 | | 0.005 | 0.005 | | | | 0.070 |
## s_iv_quality_of_evidence | 1 | 6.359 | 6.359 | 6.359 | | 0.004 | 0.004 | | | | 0.060 |
## s_iv_structured | 1 | 0.028 | 0.028 | 0.028 | | 0.000 | 0.000 | | | | 0.004 |
## pls | 1 | 7.479 | 7.479 | 7.479 | | 0.004 | 0.004 | | | | 0.065 |
## Residuals | | 1754.922 | 0.129 | | 2089 | | | | | | |r.squaredGLMM(m1)## Warning: The null model is correct only if all variables used by the original model remain unchanged.## R2m R2c
## theoretical 0.01635049 0.3262973
## delta 0.01295702 0.2585758r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.010 0.024 0.005
## 4 s_iv_statisticsqual 0.003 0.010 0.000
## 5 s_iv_quality_of_evidenceextended 0.003 0.010 0.000
## 3 s_iv_statisticseff + gloss 0.003 0.010 0.000
## 7 plsYule 0.003 0.009 0.000
## 2 s_iv_statisticseff + qual 0.002 0.008 0.000
## 6 s_iv_structuredstructured 0.000 0.002 0.000#ICC Subject
m0 <- glmer(knowledge ~ (1 | id), family = binomial("logit"), data = df_complex, glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1000000)))
summary(m0)## Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
## Family: binomial ( logit )
## Formula: knowledge ~ (1 | id)
## Data: df_complex
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+06))
##
## AIC BIC logLik deviance df.resid
## 2583.5 2594.8 -1289.8 2579.5 2087
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.3208 -0.8562 0.4698 0.4698 0.7571
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.508 1.228
## Number of obs: 2089, groups: id, 1064
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.96556 0.07888 12.24 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1performance::icc(m0)## # Intraclass Correlation Coefficient
##
## Adjusted ICC: 0.314
## Conditional ICC: 0.314D Empowerment
m1 <- lmer(empowerment ~ s_iv_statistics + s_iv_quality_of_evidence + s_iv_structured + pls + (1 | id), data = df_complex)
summary(m1)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: empowerment ~ s_iv_statistics + s_iv_quality_of_evidence + s_iv_structured + pls + (1 | id)
## Data: df_complex
##
## REML criterion at convergence: 7912
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.2877 -0.4763 0.0351 0.5213 3.0965
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.919 1.385
## Residual 1.210 1.100
## Number of obs: 2106, groups: id, 1086
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.70379 0.12542 1168.04629 37.505 < 2e-16 ***
## s_iv_statisticseff + qual 0.40388 0.14071 1084.53016 2.870 0.00418 **
## s_iv_statisticseff + gloss 0.09751 0.13630 1084.15704 0.715 0.47452
## s_iv_statisticsqual 0.40526 0.13846 1084.04787 2.927 0.00349 **
## s_iv_quality_of_evidenceextended 0.14705 0.09703 1082.24031 1.516 0.12992
## s_iv_structuredstructured 0.14902 0.09722 1082.02863 1.533 0.12561
## plsYule -0.02372 0.04842 1046.52480 -0.490 0.62427
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_+q s_v_+g s_v_stt s_v___ s_v_str
## s_v_sttst+q -0.563
## s_v_sttst+g -0.607 0.515
## s_v_sttstcs -0.576 0.506 0.523
## s_v_qlty_f_ -0.394 -0.012 0.011 0.002
## s_v_strctrd -0.427 0.024 0.064 0.021 -0.001
## plsYule -0.190 -0.002 -0.005 -0.005 0.000 0.003test = glht(m1,linfct=mcp(s_iv_statistics="Tukey"))
summary(test, test = adjusted("none"))##
## Simultaneous Tests for General Linear Hypotheses
##
## Multiple Comparisons of Means: Tukey Contrasts
##
##
## Fit: lmer(formula = empowerment ~ s_iv_statistics + s_iv_quality_of_evidence +
## s_iv_structured + pls + (1 | id), data = df_complex)
##
## Linear Hypotheses:
## Estimate Std. Error z value Pr(>|z|)
## eff + qual - eff + qual + gloss == 0 0.403878 0.140709 2.870 0.00410 **
## eff + gloss - eff + qual + gloss == 0 0.097507 0.136298 0.715 0.47436
## qual - eff + qual + gloss == 0 0.405259 0.138460 2.927 0.00342 **
## eff + gloss - eff + qual == 0 -0.306371 0.136443 -2.245 0.02474 *
## qual - eff + qual == 0 0.001381 0.138727 0.010 0.99206
## qual - eff + gloss == 0 0.307752 0.134127 2.294 0.02176 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- none method)summary(test, test = adjusted("BH"))##
## Simultaneous Tests for General Linear Hypotheses
##
## Multiple Comparisons of Means: Tukey Contrasts
##
##
## Fit: lmer(formula = empowerment ~ s_iv_statistics + s_iv_quality_of_evidence +
## s_iv_structured + pls + (1 | id), data = df_complex)
##
## Linear Hypotheses:
## Estimate Std. Error z value Pr(>|z|)
## eff + qual - eff + qual + gloss == 0 0.403878 0.140709 2.870 0.0123 *
## eff + gloss - eff + qual + gloss == 0 0.097507 0.136298 0.715 0.5692
## qual - eff + qual + gloss == 0 0.405259 0.138460 2.927 0.0123 *
## eff + gloss - eff + qual == 0 -0.306371 0.136443 -2.245 0.0371 *
## qual - eff + qual == 0 0.001381 0.138727 0.010 0.9921
## qual - eff + gloss == 0 0.307752 0.134127 2.294 0.0371 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- BH method)r.squaredGLMM(m1)## R2m R2c
## [1,] 0.0137975 0.6184957r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.014 0.028 0.008
## 4 s_iv_statisticsqual 0.006 0.015 0.001
## 2 s_iv_statisticseff + qual 0.006 0.015 0.001
## 6 s_iv_structuredstructured 0.002 0.007 0.000
## 5 s_iv_quality_of_evidenceextended 0.002 0.007 0.000
## 3 s_iv_statisticseff + gloss 0.000 0.004 0.000
## 7 plsYule 0.000 0.003 0.000anova_stats(m1)## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## term | sumsq | meansq | NumDF | DenDF | statistic | p.value | df | etasq | partial.etasq | omegasq | partial.omegasq | epsilonsq | cohens.f | power
## -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
## s_iv_statistics | 16.616 | 5.539 | 3 | 1082.248 | 4.576 | 0.003 | | 0.011 | 0.011 | | | | 0.103 |
## s_iv_quality_of_evidence | 2.780 | 2.780 | 1 | 1082.240 | 2.297 | 0.130 | | 0.002 | 0.002 | | | | 0.042 |
## s_iv_structured | 2.844 | 2.844 | 1 | 1082.029 | 2.350 | 0.126 | | 0.002 | 0.002 | | | | 0.043 |
## pls | 0.291 | 0.291 | 1 | 1046.525 | 0.240 | 0.624 | | 0.000 | 0.000 | | | | 0.014 |
## Residuals | 1558.557 | 0.740 | | | | | 2106 | | | | | | |anova_stats(m1)$term## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "s_iv_statistics" "s_iv_quality_of_evidence" "s_iv_structured" "pls"
## [5] "Residuals"interpret_omega_squared(anova_stats(m1)$partial.etasq, rules = "field2013")## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "small" "very small" "very small" "very small" NA
## (Rules: field2013)#ICC Subject
m0 <- lmer(empowerment ~ (1 | id), data = df_complex)
summary(m0)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: empowerment ~ (1 | id)
## Data: df_complex
##
## REML criterion at convergence: 7913.8
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.3042 -0.4450 -0.0145 0.5477 3.0599
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.95 1.396
## Residual 1.21 1.100
## Number of obs: 2106, groups: id, 1086
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.067e+00 4.879e-02 1.087e+03 103.9 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1performance::icc(m0)## # Intraclass Correlation Coefficient
##
## Adjusted ICC: 0.617
## Conditional ICC: 0.617Uncomplex
A Accessibility
m1 <- lmer(accessibility ~ s_iv_statistics + s_iv_quality_of_evidence + s_iv_structured + pls + (1 | id), data = df_uncomplex)
summary(m1)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: accessibility ~ s_iv_statistics + s_iv_quality_of_evidence + s_iv_structured + pls + (1 | id)
## Data: df_uncomplex
##
## REML criterion at convergence: 8264
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.4018 -0.4762 0.1075 0.5181 2.5986
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.738 1.318
## Residual 1.416 1.190
## Number of obs: 2164, groups: id, 1123
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.65779 0.11834 1216.59014 47.811 < 2e-16 ***
## s_iv_statisticseff + qual 0.44310 0.12995 1098.40767 3.410 0.000674 ***
## s_iv_statisticseff + gloss -0.13991 0.13653 1100.52375 -1.025 0.305719
## s_iv_statisticsqual 0.60960 0.13367 1106.50501 4.561 5.67e-06 ***
## s_iv_quality_of_evidenceextended -0.01968 0.09434 1101.55215 -0.209 0.834783
## s_iv_structuredstructured -0.09793 0.09450 1101.46638 -1.036 0.300309
## plsBucher 0.13926 0.05170 1060.57281 2.694 0.007179 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_+q s_v_+g s_v_stt s_v___ s_v_str
## s_v_sttst+q -0.581
## s_v_sttst+g -0.528 0.499
## s_v_sttstcs -0.567 0.511 0.485
## s_v_qlty_f_ -0.392 -0.030 -0.046 -0.030
## s_v_strctrd -0.425 0.041 -0.006 0.045 0.049
## plsBucher -0.222 0.008 0.006 0.005 -0.004 -0.005test = glht(m1,linfct=mcp(s_iv_statistics="Tukey"))
summary(test, test = adjusted("none"))##
## Simultaneous Tests for General Linear Hypotheses
##
## Multiple Comparisons of Means: Tukey Contrasts
##
##
## Fit: lmer(formula = accessibility ~ s_iv_statistics + s_iv_quality_of_evidence +
## s_iv_structured + pls + (1 | id), data = df_uncomplex)
##
## Linear Hypotheses:
## Estimate Std. Error z value Pr(>|z|)
## eff + qual - eff + qual + gloss == 0 0.4431 0.1299 3.410 0.00065 ***
## eff + gloss - eff + qual + gloss == 0 -0.1399 0.1365 -1.025 0.30549
## qual - eff + qual + gloss == 0 0.6096 0.1337 4.561 5.10e-06 ***
## eff + gloss - eff + qual == 0 -0.5830 0.1335 -4.366 1.26e-05 ***
## qual - eff + qual == 0 0.1665 0.1304 1.277 0.20160
## qual - eff + gloss == 0 0.7495 0.1372 5.465 4.64e-08 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- none method)summary(test, test = adjusted("BH"))##
## Simultaneous Tests for General Linear Hypotheses
##
## Multiple Comparisons of Means: Tukey Contrasts
##
##
## Fit: lmer(formula = accessibility ~ s_iv_statistics + s_iv_quality_of_evidence +
## s_iv_structured + pls + (1 | id), data = df_uncomplex)
##
## Linear Hypotheses:
## Estimate Std. Error z value Pr(>|z|)
## eff + qual - eff + qual + gloss == 0 0.4431 0.1299 3.410 0.000975 ***
## eff + gloss - eff + qual + gloss == 0 -0.1399 0.1365 -1.025 0.305494
## qual - eff + qual + gloss == 0 0.6096 0.1337 4.561 1.53e-05 ***
## eff + gloss - eff + qual == 0 -0.5830 0.1335 -4.366 2.53e-05 ***
## qual - eff + qual == 0 0.1665 0.1304 1.277 0.241921
## qual - eff + gloss == 0 0.7495 0.1372 5.465 2.78e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- BH method)r.squaredGLMM(m1)## R2m R2c
## [1,] 0.03104364 0.5650063r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.031 0.049 0.020
## 4 s_iv_statisticsqual 0.014 0.026 0.006
## 2 s_iv_statisticseff + qual 0.008 0.017 0.002
## 7 plsBucher 0.002 0.007 0.000
## 6 s_iv_structuredstructured 0.001 0.005 0.000
## 3 s_iv_statisticseff + gloss 0.001 0.005 0.000
## 5 s_iv_quality_of_evidenceextended 0.000 0.002 0.000anova_stats(m1)## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## term | sumsq | meansq | NumDF | DenDF | statistic | p.value | df | etasq | partial.etasq | omegasq | partial.omegasq | epsilonsq | cohens.f | power
## -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
## s_iv_statistics | 58.705 | 19.568 | 3 | 1101.132 | 13.823 | < .001 | | 0.029 | 0.029 | | | | 0.174 |
## s_iv_quality_of_evidence | 0.062 | 0.062 | 1 | 1101.552 | 0.044 | 0.835 | | 0.000 | 0.000 | | | | 0.006 |
## s_iv_structured | 1.520 | 1.520 | 1 | 1101.466 | 1.074 | 0.300 | | 0.001 | 0.001 | | | | 0.028 |
## pls | 10.271 | 10.271 | 1 | 1060.573 | 7.256 | 0.007 | | 0.005 | 0.005 | | | | 0.073 |
## Residuals | 1948.347 | 0.900 | | | | | 2164 | | | | | | |anova_stats(m1)$term## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "s_iv_statistics" "s_iv_quality_of_evidence" "s_iv_structured" "pls"
## [5] "Residuals"interpret_omega_squared(anova_stats(m1)$partial.etasq, rules = "field2013")## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "small" "very small" "very small" "very small" NA
## (Rules: field2013)#ICC Subject
m0 <- lmer(accessibility ~ (1 | id), data = df_uncomplex)
summary(m0)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: accessibility ~ (1 | id)
## Data: df_uncomplex
##
## REML criterion at convergence: 8296.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.2653 -0.4512 0.0208 0.4928 2.5989
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.817 1.348
## Residual 1.425 1.194
## Number of obs: 2164, groups: id, 1123
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.912e+00 4.784e-02 1.107e+03 123.6 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1performance::icc(m0)## # Intraclass Correlation Coefficient
##
## Adjusted ICC: 0.560
## Conditional ICC: 0.560B Understanding
m1 <- lmer(understanding ~ s_iv_statistics + s_iv_quality_of_evidence + s_iv_structured + pls + (1 | id), data = df_uncomplex)
summary(m1)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: understanding ~ s_iv_statistics + s_iv_quality_of_evidence + s_iv_structured + pls + (1 | id)
## Data: df_uncomplex
##
## REML criterion at convergence: 8047
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.7105 -0.4113 0.1365 0.4482 3.0289
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.884 1.372
## Residual 1.121 1.059
## Number of obs: 2173, groups: id, 1121
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.861e+00 1.175e-01 1.198e+03 49.897 < 2e-16 ***
## s_iv_statisticseff + qual 3.004e-01 1.298e-01 1.106e+03 2.314 0.02087 *
## s_iv_statisticseff + gloss -5.239e-03 1.364e-01 1.108e+03 -0.038 0.96937
## s_iv_statisticsqual 4.107e-01 1.332e-01 1.111e+03 3.084 0.00209 **
## s_iv_quality_of_evidenceextended 1.545e-01 9.414e-02 1.108e+03 1.642 0.10095
## s_iv_structuredstructured -1.564e-01 9.432e-02 1.108e+03 -1.659 0.09747 .
## plsBucher 1.361e-01 4.589e-02 1.069e+03 2.967 0.00308 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_+q s_v_+g s_v_stt s_v___ s_v_str
## s_v_sttst+q -0.584
## s_v_sttst+g -0.531 0.498
## s_v_sttstcs -0.572 0.512 0.486
## s_v_qlty_f_ -0.395 -0.028 -0.045 -0.027
## s_v_strctrd -0.425 0.039 -0.006 0.045 0.047
## plsBucher -0.199 0.007 0.004 0.004 -0.001 -0.004test = glht(m1,linfct=mcp(s_iv_statistics="Tukey"))
summary(test, test = adjusted("none"))##
## Simultaneous Tests for General Linear Hypotheses
##
## Multiple Comparisons of Means: Tukey Contrasts
##
##
## Fit: lmer(formula = understanding ~ s_iv_statistics + s_iv_quality_of_evidence +
## s_iv_structured + pls + (1 | id), data = df_uncomplex)
##
## Linear Hypotheses:
## Estimate Std. Error z value Pr(>|z|)
## eff + qual - eff + qual + gloss == 0 0.300357 0.129821 2.314 0.02069 *
## eff + gloss - eff + qual + gloss == 0 -0.005239 0.136419 -0.038 0.96937
## qual - eff + qual + gloss == 0 0.410742 0.133190 3.084 0.00204 **
## eff + gloss - eff + qual == 0 -0.305596 0.133479 -2.289 0.02205 *
## qual - eff + qual == 0 0.110385 0.129978 0.849 0.39574
## qual - eff + gloss == 0 0.415981 0.136767 3.042 0.00235 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- none method)summary(test, test = adjusted("BH"))##
## Simultaneous Tests for General Linear Hypotheses
##
## Multiple Comparisons of Means: Tukey Contrasts
##
##
## Fit: lmer(formula = understanding ~ s_iv_statistics + s_iv_quality_of_evidence +
## s_iv_structured + pls + (1 | id), data = df_uncomplex)
##
## Linear Hypotheses:
## Estimate Std. Error z value Pr(>|z|)
## eff + qual - eff + qual + gloss == 0 0.300357 0.129821 2.314 0.03308 *
## eff + gloss - eff + qual + gloss == 0 -0.005239 0.136419 -0.038 0.96937
## qual - eff + qual + gloss == 0 0.410742 0.133190 3.084 0.00706 **
## eff + gloss - eff + qual == 0 -0.305596 0.133479 -2.289 0.03308 *
## qual - eff + qual == 0 0.110385 0.129978 0.849 0.47489
## qual - eff + gloss == 0 0.415981 0.136767 3.042 0.00706 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- BH method)r.squaredGLMM(m1)## R2m R2c
## [1,] 0.01710391 0.6332309r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.017 0.032 0.010
## 4 s_iv_statisticsqual 0.007 0.016 0.002
## 2 s_iv_statisticseff + qual 0.004 0.011 0.000
## 6 s_iv_structuredstructured 0.002 0.008 0.000
## 5 s_iv_quality_of_evidenceextended 0.002 0.007 0.000
## 7 plsBucher 0.002 0.007 0.000
## 3 s_iv_statisticseff + gloss 0.000 0.002 0.000anova_stats(m1)## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## term | sumsq | meansq | NumDF | DenDF | statistic | p.value | df | etasq | partial.etasq | omegasq | partial.omegasq | epsilonsq | cohens.f | power
## -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
## s_iv_statistics | 16.738 | 5.579 | 3 | 1107.715 | 4.976 | 0.002 | | 0.011 | 0.011 | | | | 0.106 |
## s_iv_quality_of_evidence | 3.022 | 3.022 | 1 | 1108.047 | 2.695 | 0.101 | | 0.002 | 0.002 | | | | 0.045 |
## s_iv_structured | 3.085 | 3.085 | 1 | 1107.992 | 2.751 | 0.097 | | 0.002 | 0.002 | | | | 0.046 |
## pls | 9.869 | 9.869 | 1 | 1069.129 | 8.802 | 0.003 | | 0.007 | 0.007 | | | | 0.082 |
## Residuals | 1476.245 | 0.679 | | | | | 2173 | | | | | | |anova_stats(m1)$term## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "s_iv_statistics" "s_iv_quality_of_evidence" "s_iv_structured" "pls"
## [5] "Residuals"interpret_omega_squared(anova_stats(m1)$partial.etasq, rules = "field2013")## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "small" "very small" "very small" "very small" NA
## (Rules: field2013)#ICC Subject
m0 <- lmer(understanding ~ (1 | id), data = df_uncomplex)
summary(m0)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: understanding ~ (1 | id)
## Data: df_uncomplex
##
## REML criterion at convergence: 8059.8
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.6390 -0.3881 0.1894 0.4038 2.9459
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.914 1.384
## Residual 1.130 1.063
## Number of obs: 2173, groups: id, 1121
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 6.1165 0.0473 1112.8386 129.3 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1performance::icc(m0)## # Intraclass Correlation Coefficient
##
## Adjusted ICC: 0.629
## Conditional ICC: 0.629C Knowledge
table(df_uncomplex$knowledge, df_uncomplex$s_iv_statistics)/sum(table(df_uncomplex$knowledge, df_uncomplex$s_iv_statistics)) * 100##
## eff + qual + gloss eff + qual eff + gloss qual
## 0 6.941231 7.172605 6.108283 5.738084
## 1 18.186025 20.592318 16.658954 18.602499table(df_uncomplex[,c("knowledge", "s_iv_statistics")])## s_iv_statistics
## knowledge eff + qual + gloss eff + qual eff + gloss qual
## 0 150 155 132 124
## 1 393 445 360 402table(df_uncomplex[,c("knowledge", "s_iv_statistics")])[2,]/colSums(table(df_uncomplex[,c("knowledge", "s_iv_statistics")]))## eff + qual + gloss eff + qual eff + gloss qual
## 0.7237569 0.7416667 0.7317073 0.7642586table(df_uncomplex[,c("knowledge", "s_iv_structured")])## s_iv_structured
## knowledge unstructured structured
## 0 285 276
## 1 846 754table(df_uncomplex[,c("knowledge", "s_iv_structured")])[2,]/colSums(table(df_uncomplex[,c("knowledge", "s_iv_structured")]))## unstructured structured
## 0.7480106 0.7320388table(df_uncomplex[,c("knowledge", "s_iv_quality_of_evidence")])## s_iv_quality_of_evidence
## knowledge normal extended
## 0 298 263
## 1 765 835table(df_uncomplex[,c("knowledge", "s_iv_quality_of_evidence")])[2,]/colSums(table(df_uncomplex[,c("knowledge", "s_iv_quality_of_evidence")]))## normal extended
## 0.7196613 0.7604736m1 <- glmer(knowledge ~ s_iv_statistics + s_iv_quality_of_evidence + s_iv_structured + pls + (1 | id), family = binomial("logit"), data = df_uncomplex, glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 100000)))
summary(m1)## Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
## Family: binomial ( logit )
## Formula: knowledge ~ s_iv_statistics + s_iv_quality_of_evidence + s_iv_structured + pls + (1 | id)
## Data: df_uncomplex
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 2352.1 2397.5 -1168.0 2336.1 2153
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.6854 -0.5729 0.2741 0.3434 0.8926
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 3.877 1.969
## Number of obs: 2161, groups: id, 1100
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.28587 0.24300 5.292 1.21e-07 ***
## s_iv_statisticseff + qual 0.17533 0.24651 0.711 0.4769
## s_iv_statisticseff + gloss 0.08457 0.25839 0.327 0.7434
## s_iv_statisticsqual 0.30010 0.25462 1.179 0.2386
## s_iv_quality_of_evidenceextended 0.32267 0.18029 1.790 0.0735 .
## s_iv_structuredstructured -0.12875 0.18019 -0.715 0.4749
## plsBucher 0.54143 0.12659 4.277 1.89e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_+q s_v_+g s_v_stt s_v___ s_v_str
## s_v_sttst+q -0.511
## s_v_sttst+g -0.466 0.492
## s_v_sttstcs -0.496 0.503 0.476
## s_v_qlty_f_ -0.328 -0.010 -0.038 -0.020
## s_v_strctrd -0.417 0.045 -0.005 0.047 0.037
## plsBucher -0.165 0.022 0.012 0.021 0.025 -0.017test = glht(m1,linfct=mcp(s_iv_statistics ="Tukey"))
summary(test, test = adjusted("none"))##
## Simultaneous Tests for General Linear Hypotheses
##
## Multiple Comparisons of Means: Tukey Contrasts
##
##
## Fit: glmer(formula = knowledge ~ s_iv_statistics + s_iv_quality_of_evidence +
## s_iv_structured + pls + (1 | id), data = df_uncomplex, family = binomial("logit"),
## control = glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05)))
##
## Linear Hypotheses:
## Estimate Std. Error z value Pr(>|z|)
## eff + qual - eff + qual + gloss == 0 0.17533 0.24651 0.711 0.477
## eff + gloss - eff + qual + gloss == 0 0.08457 0.25839 0.327 0.743
## qual - eff + qual + gloss == 0 0.30010 0.25462 1.179 0.239
## eff + gloss - eff + qual == 0 -0.09076 0.25468 -0.356 0.722
## qual - eff + qual == 0 0.12477 0.25003 0.499 0.618
## qual - eff + gloss == 0 0.21553 0.26250 0.821 0.412
## (Adjusted p values reported -- none method)summary(test, test = adjusted("BH"))##
## Simultaneous Tests for General Linear Hypotheses
##
## Multiple Comparisons of Means: Tukey Contrasts
##
##
## Fit: glmer(formula = knowledge ~ s_iv_statistics + s_iv_quality_of_evidence +
## s_iv_structured + pls + (1 | id), data = df_uncomplex, family = binomial("logit"),
## control = glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05)))
##
## Linear Hypotheses:
## Estimate Std. Error z value Pr(>|z|)
## eff + qual - eff + qual + gloss == 0 0.17533 0.24651 0.711 0.743
## eff + gloss - eff + qual + gloss == 0 0.08457 0.25839 0.327 0.743
## qual - eff + qual + gloss == 0 0.30010 0.25462 1.179 0.743
## eff + gloss - eff + qual == 0 -0.09076 0.25468 -0.356 0.743
## qual - eff + qual == 0 0.12477 0.25003 0.499 0.743
## qual - eff + gloss == 0 0.21553 0.26250 0.821 0.743
## (Adjusted p values reported -- BH method)exp(fixef(m1))## (Intercept) s_iv_statisticseff + qual s_iv_statisticseff + gloss
## 3.6178299 1.1916428 1.0882509
## s_iv_statisticsqual s_iv_quality_of_evidenceextended s_iv_structuredstructured
## 1.3499963 1.3808112 0.8791897
## plsBucher
## 1.7184543interpret_oddsratio(exp(fixef(m1)), rules = "cohen1988")## (Intercept) s_iv_statisticseff + qual s_iv_statisticseff + gloss
## "medium" "very small" "very small"
## s_iv_statisticsqual s_iv_quality_of_evidenceextended s_iv_structuredstructured
## "very small" "very small" "very small"
## plsBucher
## "small"
## (Rules: cohen1988)exp(summary(test, test = adjusted("none"))$test$coefficients)## eff + qual - eff + qual + gloss eff + gloss - eff + qual + gloss qual - eff + qual + gloss
## 1.1916428 1.0882509 1.3499963
## eff + gloss - eff + qual qual - eff + qual qual - eff + gloss
## 0.9132358 1.1328867 1.2405193interpret_oddsratio(exp(summary(test, test = adjusted("none"))$test$coefficients), rules = "cohen1988")## eff + qual - eff + qual + gloss eff + gloss - eff + qual + gloss qual - eff + qual + gloss
## "very small" "very small" "very small"
## eff + gloss - eff + qual qual - eff + qual qual - eff + gloss
## "very small" "very small" "very small"
## (Rules: cohen1988)anova_stats(m1)## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: npar## term | npar | sumsq | meansq | statistic | df | etasq | partial.etasq | omegasq | partial.omegasq | epsilonsq | cohens.f | power
## ---------------------------------------------------------------------------------------------------------------------------------------------------------
## s_iv_statistics | 3 | 1.645 | 0.548 | 0.548 | | 0.001 | 0.001 | | | | 0.037 |
## s_iv_quality_of_evidence | 1 | 3.361 | 3.361 | 3.361 | | 0.003 | 0.003 | | | | 0.053 |
## s_iv_structured | 1 | 0.506 | 0.506 | 0.506 | | 0.000 | 0.000 | | | | 0.021 |
## pls | 1 | 20.034 | 20.034 | 20.034 | | 0.016 | 0.016 | | | | 0.129 |
## Residuals | | 1202.001 | 0.081 | | 2161 | | | | | | |r.squaredGLMM(m1)## Warning: The null model is correct only if all variables used by the original model remain unchanged.## R2m R2c
## theoretical 0.01619252 0.5483932
## delta 0.01270776 0.4303746r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.011 0.024 0.006
## 7 plsBucher 0.008 0.018 0.002
## 5 s_iv_quality_of_evidenceextended 0.002 0.007 0.000
## 4 s_iv_statisticsqual 0.001 0.005 0.000
## 6 s_iv_structuredstructured 0.000 0.004 0.000
## 2 s_iv_statisticseff + qual 0.000 0.004 0.000
## 3 s_iv_statisticseff + gloss 0.000 0.003 0.000#ICC Subject
m0 <- glmer(knowledge ~ (1 | id), family = binomial("logit"), data = df_uncomplex, glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1000000)))
summary(m0)## Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
## Family: binomial ( logit )
## Formula: knowledge ~ (1 | id)
## Data: df_uncomplex
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+06))
##
## AIC BIC logLik deviance df.resid
## 2364.6 2375.9 -1180.3 2360.6 2159
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.3744 -0.7118 0.3064 0.3064 0.7276
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 3.608 1.899
## Number of obs: 2161, groups: id, 1100
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.7463 0.1396 12.51 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1performance::icc(m0)## # Intraclass Correlation Coefficient
##
## Adjusted ICC: 0.523
## Conditional ICC: 0.523D Empowerment
m1 <- lmer(empowerment ~ s_iv_statistics + s_iv_quality_of_evidence + s_iv_structured + pls + (1 | id), data = df_uncomplex)
summary(m1)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: empowerment ~ s_iv_statistics + s_iv_quality_of_evidence + s_iv_structured + pls + (1 | id)
## Data: df_uncomplex
##
## REML criterion at convergence: 8241.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.3598 -0.4361 0.0440 0.5185 3.0483
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.952 1.397
## Residual 1.262 1.123
## Number of obs: 2175, groups: id, 1123
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.98718 0.12081 1209.92447 41.281 <2e-16 ***
## s_iv_statisticseff + qual 0.18926 0.13328 1111.47218 1.420 0.1559
## s_iv_statisticseff + gloss -0.18242 0.14008 1111.44206 -1.302 0.1931
## s_iv_statisticsqual 0.22074 0.13692 1118.59298 1.612 0.1072
## s_iv_quality_of_evidenceextended 0.13201 0.09675 1114.78535 1.364 0.1727
## s_iv_structuredstructured -0.17448 0.09693 1114.50092 -1.800 0.0721 .
## plsBucher 0.12257 0.04865 1075.91347 2.519 0.0119 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_+q s_v_+g s_v_stt s_v___ s_v_str
## s_v_sttst+q -0.582
## s_v_sttst+g -0.530 0.498
## s_v_sttstcs -0.570 0.511 0.484
## s_v_qlty_f_ -0.394 -0.030 -0.045 -0.027
## s_v_strctrd -0.427 0.040 -0.005 0.046 0.049
## plsBucher -0.205 0.007 0.005 0.004 -0.004 -0.003test = glht(m1,linfct=mcp(s_iv_statistics="Tukey"))
summary(test, test = adjusted("none"))##
## Simultaneous Tests for General Linear Hypotheses
##
## Multiple Comparisons of Means: Tukey Contrasts
##
##
## Fit: lmer(formula = empowerment ~ s_iv_statistics + s_iv_quality_of_evidence +
## s_iv_structured + pls + (1 | id), data = df_uncomplex)
##
## Linear Hypotheses:
## Estimate Std. Error z value Pr(>|z|)
## eff + qual - eff + qual + gloss == 0 0.18926 0.13328 1.420 0.15559
## eff + gloss - eff + qual + gloss == 0 -0.18242 0.14008 -1.302 0.19284
## qual - eff + qual + gloss == 0 0.22074 0.13692 1.612 0.10693
## eff + gloss - eff + qual == 0 -0.37168 0.13711 -2.711 0.00671 **
## qual - eff + qual == 0 0.03148 0.13368 0.236 0.81382
## qual - eff + gloss == 0 0.40316 0.14068 2.866 0.00416 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- none method)summary(test, test = adjusted("BH"))##
## Simultaneous Tests for General Linear Hypotheses
##
## Multiple Comparisons of Means: Tukey Contrasts
##
##
## Fit: lmer(formula = empowerment ~ s_iv_statistics + s_iv_quality_of_evidence +
## s_iv_structured + pls + (1 | id), data = df_uncomplex)
##
## Linear Hypotheses:
## Estimate Std. Error z value Pr(>|z|)
## eff + qual - eff + qual + gloss == 0 0.18926 0.13328 1.420 0.2314
## eff + gloss - eff + qual + gloss == 0 -0.18242 0.14008 -1.302 0.2314
## qual - eff + qual + gloss == 0 0.22074 0.13692 1.612 0.2139
## eff + gloss - eff + qual == 0 -0.37168 0.13711 -2.711 0.0201 *
## qual - eff + qual == 0 0.03148 0.13368 0.236 0.8138
## qual - eff + gloss == 0 0.40316 0.14068 2.866 0.0201 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## (Adjusted p values reported -- BH method)r.squaredGLMM(m1)## R2m R2c
## [1,] 0.01319522 0.6126359r2beta(m1, method = "nsj")## Effect Rsq upper.CL lower.CL
## 1 Model 0.013 0.027 0.007
## 6 s_iv_structuredstructured 0.002 0.008 0.000
## 4 s_iv_statisticsqual 0.002 0.007 0.000
## 2 s_iv_statisticseff + qual 0.001 0.006 0.000
## 5 s_iv_quality_of_evidenceextended 0.001 0.006 0.000
## 3 s_iv_statisticseff + gloss 0.001 0.006 0.000
## 7 plsBucher 0.001 0.006 0.000anova_stats(m1)## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## term | sumsq | meansq | NumDF | DenDF | statistic | p.value | df | etasq | partial.etasq | omegasq | partial.omegasq | epsilonsq | cohens.f | power
## -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
## s_iv_statistics | 13.566 | 4.522 | 3 | 1114.179 | 3.585 | 0.013 | | 0.008 | 0.008 | | | | 0.090 |
## s_iv_quality_of_evidence | 2.349 | 2.349 | 1 | 1114.785 | 1.862 | 0.173 | | 0.001 | 0.001 | | | | 0.037 |
## s_iv_structured | 4.088 | 4.088 | 1 | 1114.501 | 3.240 | 0.072 | | 0.002 | 0.002 | | | | 0.049 |
## pls | 8.006 | 8.006 | 1 | 1075.913 | 6.346 | 0.012 | | 0.005 | 0.005 | | | | 0.069 |
## Residuals | 1683.200 | 0.774 | | | | | 2175 | | | | | | |anova_stats(m1)$term## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "s_iv_statistics" "s_iv_quality_of_evidence" "s_iv_structured" "pls"
## [5] "Residuals"interpret_omega_squared(anova_stats(m1)$partial.etasq, rules = "field2013")## Warning in tidy.anova(model): The following column names in ANOVA output were not recognized or transformed: NumDF,
## DenDF## [1] "very small" "very small" "very small" "very small" NA
## (Rules: field2013)#ICC Subject
m0 <- lmer(empowerment ~ (1 | id), data = df_uncomplex)
summary(m0)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: empowerment ~ (1 | id)
## Data: df_uncomplex
##
## REML criterion at convergence: 8247.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.2372 -0.4525 -0.0211 0.5308 2.9793
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.976 1.406
## Residual 1.268 1.126
## Number of obs: 2175, groups: id, 1123
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.098e+00 4.851e-02 1.120e+03 105.1 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1performance::icc(m0)## # Intraclass Correlation Coefficient
##
## Adjusted ICC: 0.609
## Conditional ICC: 0.609Plots
library(readr)
library(tidyr)
library(ggplot2)
library(Hmisc)
library(plyr)
library(RColorBrewer)
library(reshape2)
library(gridExtra)##
## Attache Paket: 'gridExtra'## Das folgende Objekt ist maskiert 'package:dplyr':
##
## combinesource("https://gist.githubusercontent.com/benmarwick/2a1bb0133ff568cbe28d/raw/fb53bd97121f7f9ce947837ef1a4c65a73bffb3f/geom_flat_violin.R")
raincloud_theme = theme(
text = element_text(size = 10),
axis.title.x = element_blank(),
axis.title.y = element_text(size = 16),
axis.text = element_text(size = 12),
axis.text.x = element_text(angle = 90, vjust = 0.5),
legend.title=element_text(size=16),
legend.text=element_text(size=16),
legend.position = "right",
plot.title = element_text(lineheight=.8, face="bold", size = 16),
panel.border = element_blank(),
panel.grid.minor = element_blank(),
panel.grid.major = element_blank(),
axis.line.x = element_line(colour = 'black', size=0.5, linetype='solid'),
axis.line.y = element_line(colour = 'black', size=0.5, linetype='solid'))
#Accessibility
m000 <- lmer(accessibility ~ (1|pls) + (1 | id) + s_iv_complex * s_iv_structured + + s_iv_complex * s_iv_quality_of_evidence, data = df)
summary(m000)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: accessibility ~ (1 | pls) + (1 | id) + s_iv_complex * s_iv_structured +
## +s_iv_complex * s_iv_quality_of_evidence
## Data: df
##
## REML criterion at convergence: 16339.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.3484 -0.4628 0.0744 0.5090 2.7063
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.87586 1.36962
## pls (Intercept) 0.00357 0.05975
## Residual 1.38973 1.17887
## Number of obs: 4264, groups: id, 2208; pls, 4
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.98551 0.09419 26.18111 63.546 <2e-16 ***
## s_iv_complexcomplex -0.36046 0.13436 27.08647 -2.683 0.0123 *
## s_iv_structuredstructured -0.13653 0.09665 2192.62507 -1.413 0.1579
## s_iv_quality_of_evidenceextended -0.01794 0.09657 2192.69061 -0.186 0.8526
## s_iv_complexcomplex:s_iv_structuredstructured 0.34589 0.13768 2188.80023 2.512 0.0121 *
## s_iv_complexcomplex:s_iv_quality_of_evidenceextended 0.08886 0.13762 2188.86071 0.646 0.5185
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_cm s_v_st s_v___ s_v_:__
## s_v_cmplxcm -0.701
## s_v_strctrd -0.517 0.363
## s_v_qlty_f_ -0.544 0.381 0.050
## s_v_cmpl:__ 0.363 -0.519 -0.702 -0.035
## s_v_cm:____ 0.382 -0.532 -0.035 -0.702 0.023nrow(df)## [1] 4306ranef_accessibility <- ranef(m000)$id[,1]
ranef_id <- as.numeric(rownames(ranef(m000)$id))
rain <- data[data$id %in% ranef_id,c("id","s_iv_statistics")]
rain <- rain[
with(rain, order(id)),
]
rain$accessibility <- ranef_accessibility
rain$PLS <- rain$s_iv_statistics
levels(rain$PLS)## [1] "eff + qual + gloss" "eff + qual" "eff + gloss" "qual"levels(rain$PLS) <- c("Statement + Effect + Glossary","Statement + Effect", "Effect + Glossary", "Statement")
p1 <- ggplot(data = rain, aes(y = accessibility, x = PLS, fill = PLS)) +
geom_flat_violin(position = position_nudge(x = .2, y = 0), alpha = .8) +
geom_point(aes(y = accessibility, color = PLS), position = position_jitter(width = .15), size = .5, alpha = 0.8) +
geom_boxplot(width = .1, guides = FALSE, outlier.shape = NA, alpha = 0.5) +
#expand_limits(y = 5.25) +
guides(fill = FALSE) +
guides(color = FALSE) +
scale_color_brewer(palette = "Dark2") +
scale_fill_brewer(palette = "Dark2") +
# coord_flip() +
theme_bw() +
raincloud_theme + ylab("Accessibility") + scale_x_discrete(guide = guide_axis(angle = 60)) + theme(axis.text.x=element_text(size=10))## Warning: Ignoring unknown parameters: guidesp1
# Understanding
m000 <- lmer(understanding ~ (1|pls) + (1 | id) + s_iv_complex * s_iv_structured + + s_iv_complex * s_iv_quality_of_evidence, data = df)
summary(m000)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: understanding ~ (1 | pls) + (1 | id) + s_iv_complex * s_iv_structured +
## +s_iv_complex * s_iv_quality_of_evidence
## Data: df
##
## REML criterion at convergence: 15834.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.6415 -0.4132 0.1429 0.4606 3.0078
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.86327 1.36502
## pls (Intercept) 0.00353 0.05941
## Residual 1.14137 1.06835
## Number of obs: 4271, groups: id, 2207; pls, 4
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 6.12249 0.09181 26.62704 66.689 < 2e-16 ***
## s_iv_complexcomplex -0.23046 0.13108 27.65751 -1.758 0.08980 .
## s_iv_structuredstructured -0.17884 0.09394 2191.95368 -1.904 0.05707 .
## s_iv_quality_of_evidenceextended 0.15697 0.09385 2191.80563 1.673 0.09456 .
## s_iv_complexcomplex:s_iv_structuredstructured 0.35109 0.13383 2193.82076 2.623 0.00877 **
## s_iv_complexcomplex:s_iv_quality_of_evidenceextended -0.05835 0.13377 2193.77782 -0.436 0.66274
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_cm s_v_st s_v___ s_v_:__
## s_v_cmplxcm -0.700
## s_v_strctrd -0.513 0.359
## s_v_qlty_f_ -0.540 0.378 0.047
## s_v_cmpl:__ 0.360 -0.517 -0.702 -0.033
## s_v_cm:____ 0.379 -0.529 -0.033 -0.702 0.023nrow(df)## [1] 4306ranef_understanding <- ranef(m000)$id[,1]
ranef_id <- as.numeric(rownames(ranef(m000)$id))
rain <- data[data$id %in% ranef_id,c("id","s_iv_statistics")]
rain <- rain[
with(rain, order(id)),
]
rain$understanding <- ranef_understanding
rain$PLS <- rain$s_iv_statistics
levels(rain$PLS)## [1] "eff + qual + gloss" "eff + qual" "eff + gloss" "qual"levels(rain$PLS) <- c("Statement + Effect + Glossary","Statement + Effect", "Effect + Glossary", "Statement")
p2 <- ggplot(data = rain, aes(y = understanding, x = PLS, fill = PLS)) +
geom_flat_violin(position = position_nudge(x = .2, y = 0), alpha = .8) +
geom_point(aes(y = understanding, color = PLS), position = position_jitter(width = .15), size = .5, alpha = 0.8) +
geom_boxplot(width = .1, guides = FALSE, outlier.shape = NA, alpha = 0.5) +
#expand_limits(y = 5.25) +
guides(fill = FALSE) +
guides(color = FALSE) +
scale_color_brewer(palette = "Dark2") +
scale_fill_brewer(palette = "Dark2") +
# coord_flip() +
theme_bw() +
raincloud_theme + ylab("Understanding") + scale_x_discrete(guide = guide_axis(angle = 60)) + theme(axis.text.x=element_text(size=10))## Warning: Ignoring unknown parameters: guidesp2
# Empowerment
m000 <- lmer(empowerment ~ (1|pls) + (1 | id) + s_iv_complex * s_iv_structured + + s_iv_complex * s_iv_quality_of_evidence, data = df)
summary(m000)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: empowerment ~ (1 | pls) + (1 | id) + s_iv_complex * s_iv_structured +
## +s_iv_complex * s_iv_quality_of_evidence
## Data: df
##
## REML criterion at convergence: 16160.1
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.3493 -0.4418 0.0260 0.5338 3.0935
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.957692 1.39918
## pls (Intercept) 0.002736 0.05231
## Residual 1.236177 1.11183
## Number of obs: 4281, groups: id, 2209; pls, 4
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.12476 0.09179 36.93301 55.831 <2e-16 ***
## s_iv_complexcomplex -0.21126 0.13107 38.35870 -1.612 0.1152
## s_iv_structuredstructured -0.19354 0.09659 2204.71796 -2.004 0.0452 *
## s_iv_quality_of_evidenceextended 0.12903 0.09650 2204.79545 1.337 0.1813
## s_iv_complexcomplex:s_iv_structuredstructured 0.34373 0.13760 2203.29099 2.498 0.0126 *
## s_iv_complexcomplex:s_iv_quality_of_evidenceextended 0.02264 0.13753 2203.41526 0.165 0.8693
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_cm s_v_st s_v___ s_v_:__
## s_v_cmplxcm -0.700
## s_v_strctrd -0.530 0.371
## s_v_qlty_f_ -0.556 0.390 0.050
## s_v_cmpl:__ 0.372 -0.532 -0.702 -0.035
## s_v_cm:____ 0.390 -0.544 -0.035 -0.702 0.024nrow(df)## [1] 4306ranef_empowerment <- ranef(m000)$id[,1]
ranef_id <- as.numeric(rownames(ranef(m000)$id))
rain <- data[data$id %in% ranef_id,c("id","s_iv_statistics")]
rain <- rain[
with(rain, order(id)),
]
rain$empowerment <- ranef_empowerment
rain$PLS <- rain$s_iv_statistics
levels(rain$PLS)## [1] "eff + qual + gloss" "eff + qual" "eff + gloss" "qual"levels(rain$PLS) <- c("Statement + Effect + Glossary","Statement + Effect", "Effect + Glossary", "Statement")
p3 <- ggplot(data = rain, aes(y = empowerment, x = PLS, fill = PLS)) +
geom_flat_violin(position = position_nudge(x = .2, y = 0), alpha = .8) +
geom_point(aes(y = empowerment, color = PLS), position = position_jitter(width = .15), size = .5, alpha = 0.8) +
geom_boxplot(width = .1, guides = FALSE, outlier.shape = NA, alpha = 0.5) +
#expand_limits(y = 5.25) +
guides(fill = FALSE) +
guides(color = FALSE) +
scale_color_brewer(palette = "Dark2") +
scale_fill_brewer(palette = "Dark2") +
# coord_flip() +
theme_bw() +
raincloud_theme + ylab("Empowerment") + scale_x_discrete(guide = guide_axis(angle = 60)) + theme(axis.text.x=element_text(size=10))## Warning: Ignoring unknown parameters: guidesp3
## Structured x Complex
# Accessibility
m000 <- lmer(accessibility ~ (1|pls) + (1 | id) + s_iv_complex * s_iv_statistics + + s_iv_complex * s_iv_quality_of_evidence, data = df)
summary(m000)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: accessibility ~ (1 | pls) + (1 | id) + s_iv_complex * s_iv_statistics +
## +s_iv_complex * s_iv_quality_of_evidence
## Data: df
##
## REML criterion at convergence: 16274.1
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.3934 -0.4721 0.1145 0.5072 2.6314
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.793898 1.33936
## pls (Intercept) 0.003617 0.06014
## Residual 1.389648 1.17883
## Number of obs: 4264, groups: id, 2208; pls, 4
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.675e+00 1.130e-01 5.264e+01 50.226 < 2e-16 ***
## s_iv_complexcomplex -1.902e-01 1.635e-01 5.760e+01 -1.163 0.249523
## s_iv_statisticseff + qual 4.481e-01 1.310e-01 2.183e+03 3.422 0.000633 ***
## s_iv_statisticseff + gloss -1.410e-01 1.377e-01 2.187e+03 -1.024 0.305867
## s_iv_statisticsqual 6.154e-01 1.347e-01 2.198e+03 4.570 5.15e-06 ***
## s_iv_quality_of_evidenceextended -1.457e-02 9.503e-02 2.188e+03 -0.153 0.878190
## s_iv_complexcomplex:s_iv_statisticseff + qual 4.441e-02 1.917e-01 2.186e+03 0.232 0.816847
## s_iv_complexcomplex:s_iv_statisticseff + gloss 5.756e-02 1.930e-01 2.185e+03 0.298 0.765519
## s_iv_complexcomplex:s_iv_statisticsqual 8.341e-03 1.925e-01 2.190e+03 0.043 0.965446
## s_iv_complexcomplex:s_iv_quality_of_evidenceextended 7.734e-02 1.354e-01 2.184e+03 0.571 0.568037
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_cm s_v_+q s_v_+g s_v_st s_v___ s__:+q s__:+g s_v_:__
## s_v_cmplxcm -0.691
## s_v_sttst+q -0.594 0.411
## s_v_sttst+g -0.559 0.386 0.499
## s_v_sttstcs -0.578 0.399 0.510 0.486
## s_v_qlty_f_ -0.394 0.272 -0.032 -0.046 -0.032
## s_v_cm:__+q 0.406 -0.588 -0.683 -0.341 -0.348 0.022
## s_v_cm:__+g 0.399 -0.587 -0.356 -0.714 -0.347 0.033 0.506
## s_v_cmpl:__ 0.404 -0.589 -0.357 -0.340 -0.700 0.022 0.507 0.504
## s_v_cm:____ 0.276 -0.406 0.023 0.032 0.022 -0.702 -0.022 -0.017 -0.014nrow(df)## [1] 4306ranef_accessibility <- ranef(m000)$id[,1]
ranef_id <- as.numeric(rownames(ranef(m000)$id))
rain <- data[data$id %in% ranef_id,c("id","s_iv_structured","s_iv_complex")]
rain <- rain[
with(rain, order(id)),
]
rain$accessibility <- ranef_accessibility
rain$PLS <- interaction(rain$s_iv_structured, rain$s_iv_complex)
levels(rain$PLS)## [1] "unstructured.uncomplex" "structured.uncomplex" "unstructured.complex" "structured.complex"levels(rain$PLS) <- c("Unstructured + Uncomplex", "Structured + Uncomplex","Unstructured + Complex", "Structured + Complex")
p4 <- ggplot(data = rain, aes(y = accessibility, x = PLS, fill = PLS)) +
geom_flat_violin(position = position_nudge(x = .2, y = 0), alpha = .8) +
geom_point(aes(y = accessibility, color = PLS), position = position_jitter(width = .15), size = .5, alpha = 0.8) +
geom_boxplot(width = .1, guides = FALSE, outlier.shape = NA, alpha = 0.5) +
#expand_limits(y = 5.25) +
guides(fill = FALSE) +
guides(color = FALSE) +
scale_color_brewer(palette = "Dark2") +
scale_fill_brewer(palette = "Dark2") +
# coord_flip() +
theme_bw() +
raincloud_theme + ylab("Accessibility") + scale_x_discrete(guide = guide_axis(angle = 60)) + theme(axis.text.x=element_text(size=10))## Warning: Ignoring unknown parameters: guidesp4
# Understanding
m000 <- lmer(understanding ~ (1|pls) + (1 | id) + s_iv_complex * s_iv_statistics + + s_iv_complex * s_iv_quality_of_evidence, data = df)
summary(m000)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: understanding ~ (1 | pls) + (1 | id) + s_iv_complex * s_iv_statistics +
## +s_iv_complex * s_iv_quality_of_evidence
## Data: df
##
## REML criterion at convergence: 15806
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.6935 -0.4120 0.1394 0.4601 3.0045
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.825887 1.35125
## pls (Intercept) 0.003562 0.05968
## Residual 1.141552 1.06843
## Number of obs: 4271, groups: id, 2207; pls, 4
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.847e+00 1.110e-01 5.554e+01 52.661 < 2e-16 ***
## s_iv_complexcomplex -1.408e-01 1.606e-01 6.067e+01 -0.876 0.38424
## s_iv_statisticseff + qual 3.080e-01 1.285e-01 2.183e+03 2.397 0.01659 *
## s_iv_statisticseff + gloss -6.997e-03 1.351e-01 2.187e+03 -0.052 0.95870
## s_iv_statisticsqual 4.206e-01 1.318e-01 2.192e+03 3.191 0.00144 **
## s_iv_quality_of_evidenceextended 1.619e-01 9.313e-02 2.187e+03 1.738 0.08230 .
## s_iv_complexcomplex:s_iv_statisticseff + qual 2.583e-01 1.880e-01 2.191e+03 1.374 0.16961
## s_iv_complexcomplex:s_iv_statisticseff + gloss 6.946e-02 1.893e-01 2.190e+03 0.367 0.71373
## s_iv_complexcomplex:s_iv_statisticsqual 8.172e-02 1.887e-01 2.194e+03 0.433 0.66495
## s_iv_complexcomplex:s_iv_quality_of_evidenceextended -7.106e-02 1.328e-01 2.189e+03 -0.535 0.59255
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_cm s_v_+q s_v_+g s_v_st s_v___ s__:+q s__:+g s_v_:__
## s_v_cmplxcm -0.691
## s_v_sttst+q -0.594 0.410
## s_v_sttst+g -0.558 0.386 0.499
## s_v_sttstcs -0.579 0.400 0.511 0.486
## s_v_qlty_f_ -0.394 0.272 -0.030 -0.045 -0.029
## s_v_cm:__+q 0.406 -0.588 -0.683 -0.341 -0.349 0.021
## s_v_cm:__+g 0.398 -0.586 -0.356 -0.714 -0.347 0.032 0.507
## s_v_cmpl:__ 0.404 -0.589 -0.357 -0.340 -0.699 0.020 0.508 0.505
## s_v_cm:____ 0.276 -0.404 0.021 0.031 0.020 -0.701 -0.021 -0.017 -0.013nrow(df)## [1] 4306ranef_understanding <- ranef(m000)$id[,1]
ranef_id <- as.numeric(rownames(ranef(m000)$id))
rain <- data[data$id %in% ranef_id,c("id","s_iv_structured","s_iv_complex")]
rain <- rain[
with(rain, order(id)),
]
rain$understanding <- ranef_understanding
rain$PLS <- interaction(rain$s_iv_structured, rain$s_iv_complex)
levels(rain$PLS)## [1] "unstructured.uncomplex" "structured.uncomplex" "unstructured.complex" "structured.complex"levels(rain$PLS) <- c("Unstructured + Uncomplex", "Structured + Uncomplex","Unstructured + Complex", "Structured + Complex")
p5 <- ggplot(data = rain, aes(y = understanding, x = PLS, fill = PLS)) +
geom_flat_violin(position = position_nudge(x = .2, y = 0), alpha = .8) +
geom_point(aes(y = understanding, color = PLS), position = position_jitter(width = .15), size = .5, alpha = 0.8) +
geom_boxplot(width = .1, guides = FALSE, outlier.shape = NA, alpha = 0.5) +
#expand_limits(y = 5.25) +
guides(fill = FALSE) +
guides(color = FALSE) +
scale_color_brewer(palette = "Dark2") +
scale_fill_brewer(palette = "Dark2") +
# coord_flip() +
theme_bw() +
raincloud_theme + ylab("Understanding") + scale_x_discrete(guide = guide_axis(angle = 60)) + theme(axis.text.x=element_text(size=10))## Warning: Ignoring unknown parameters: guidesp5
# Empowerment
m000 <- lmer(empowerment ~ (1|pls) + (1 | id) + s_iv_complex * s_iv_statistics + + s_iv_complex * s_iv_quality_of_evidence, data = df)
summary(m000)## Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest']
## Formula: empowerment ~ (1 | pls) + (1 | id) + s_iv_complex * s_iv_statistics +
## +s_iv_complex * s_iv_quality_of_evidence
## Data: df
##
## REML criterion at convergence: 16149.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.3590 -0.4437 0.0318 0.5278 3.0771
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 1.939343 1.39260
## pls (Intercept) 0.002741 0.05235
## Residual 1.236556 1.11201
## Number of obs: 4281, groups: id, 2209; pls, 4
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.956e+00 1.121e-01 8.058e+01 44.198 <2e-16 ***
## s_iv_complexcomplex -1.819e-01 1.625e-01 8.856e+01 -1.119 0.266
## s_iv_statisticseff + qual 1.983e-01 1.325e-01 2.194e+03 1.496 0.135
## s_iv_statisticseff + gloss -1.842e-01 1.394e-01 2.193e+03 -1.321 0.187
## s_iv_statisticsqual 2.315e-01 1.361e-01 2.207e+03 1.701 0.089 .
## s_iv_quality_of_evidenceextended 1.408e-01 9.615e-02 2.199e+03 1.464 0.143
## s_iv_complexcomplex:s_iv_statisticseff + qual 2.005e-01 1.939e-01 2.198e+03 1.034 0.301
## s_iv_complexcomplex:s_iv_statisticseff + gloss 2.682e-01 1.954e-01 2.197e+03 1.373 0.170
## s_iv_complexcomplex:s_iv_statisticsqual 1.692e-01 1.948e-01 2.204e+03 0.869 0.385
## s_iv_complexcomplex:s_iv_quality_of_evidenceextended 6.287e-03 1.371e-01 2.198e+03 0.046 0.963
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s_v_cm s_v_+q s_v_+g s_v_st s_v___ s__:+q s__:+g s_v_:__
## s_v_cmplxcm -0.690
## s_v_sttst+q -0.605 0.417
## s_v_sttst+g -0.569 0.393 0.498
## s_v_sttstcs -0.590 0.407 0.510 0.485
## s_v_qlty_f_ -0.401 0.277 -0.032 -0.045 -0.029
## s_v_cm:__+q 0.413 -0.599 -0.683 -0.341 -0.348 0.022
## s_v_cm:__+g 0.406 -0.597 -0.356 -0.713 -0.346 0.032 0.506
## s_v_cmpl:__ 0.412 -0.600 -0.356 -0.339 -0.699 0.020 0.508 0.504
## s_v_cm:____ 0.282 -0.412 0.023 0.032 0.021 -0.702 -0.022 -0.017 -0.014nrow(df)## [1] 4306ranef_empowerment <- ranef(m000)$id[,1]
ranef_id <- as.numeric(rownames(ranef(m000)$id))
rain <- data[data$id %in% ranef_id,c("id","s_iv_structured","s_iv_complex")]
rain <- rain[
with(rain, order(id)),
]
rain$empowerment <- ranef_empowerment
rain$PLS <- interaction(rain$s_iv_structured, rain$s_iv_complex)
levels(rain$PLS)## [1] "unstructured.uncomplex" "structured.uncomplex" "unstructured.complex" "structured.complex"levels(rain$PLS) <- c("Unstructured + Uncomplex", "Structured + Uncomplex","Unstructured + Complex", "Structured + Complex")
p6 <- ggplot(data = rain, aes(y = empowerment, x = PLS, fill = PLS)) +
geom_flat_violin(position = position_nudge(x = .2, y = 0), alpha = .8) +
geom_point(aes(y = empowerment, color = PLS), position = position_jitter(width = .15), size = .5, alpha = 0.8) +
geom_boxplot(width = .1, guides = FALSE, outlier.shape = NA, alpha = 0.5) +
#expand_limits(y = 5.25) +
guides(fill = FALSE) +
guides(color = FALSE) +
scale_color_brewer(palette = "Dark2") +
scale_fill_brewer(palette = "Dark2") +
# coord_flip() +
theme_bw() +
raincloud_theme + ylab("Empowerment") + scale_x_discrete(guide = guide_axis(angle = 60)) + theme(axis.text.x=element_text(size=10))## Warning: Ignoring unknown parameters: guidesp6
require(ggpubr)
library(cowplot)
library(ggimage)
p9 <- plot_grid(p1, p2, p3, p4, p5, p6, ncol=3, labels=LETTERS[1:6])
p9
p7 <- plot_grid(p1, p2, p3, ncol=3, labels=LETTERS[1:3])
p7 <- p7 + labs(caption="STATISTICAL TERMS") +
theme(plot.caption = element_text(hjust=0.5, size=rel(1.2)))
p8 <- plot_grid(p4, p5, p6, ncol=3, labels=LETTERS[4:6]) + labs(caption="STRUCTURE X COMPLEXITY") + theme(plot.caption = element_text(hjust=0.5, size=rel(1.2)))
p10 <- plot_grid(p7, p8, ncol=1)
p10
#ggsave("Figure_Study_2.tiff", p10, width = 16, height = 12, compression = "lzw")
#ggsave("Figure_Study_2.png", p10, width = 16, height = 12)