1. organize datasets

2. how well were the faces randomized?

descriptives for face randomization

## [1] "average number of times faces rated"
## [1] 27.84722
## [1] "SD for faces rated"
## [1] 10.60236
## [1] "range"
## [1] 10 53

all face trust histogram

hist(dwide$trust)

3. trust ratings

histogram for trust ratings per face

female face trust ratings

male face trust ratings

plot for trust means x SDs for individual faces

male face box plots

plot for trust means x SDs face gender

barchart for means for face ethnicity

barchart for means for face gender

t.test for gender & trust rating

t.test(dmeans$mean ~ dmeans$gender)
## 
##  Welch Two Sample t-test
## 
## data:  dmeans$mean by dmeans$gender
## t = -17.095, df = 807.15, p-value < 2.2e-16
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -0.8118882 -0.6446438
## sample estimates:
## mean in group F mean in group M 
##        2.856825        3.585091

box plots - variance within face measures

male box plots

male face box plots

female box plots

female face box plots

4. Analyses

df$faceLeftTrusted_1 <- ifelse(df$faceLeft == df$trustedFace, 1, 0)
df$faceLeftCued_1 <- ifelse(df$faceLeft == df$cuedFace, 1, 0)
df$faceLeftCued_.5 <- ifelse(df$faceLeft == df$cuedFace, .5, -.5)

df$faceRightTrusted_1 <- ifelse(df$faceRight == df$trustedFace, 1, 0)
df$faceRightCued_1 <- ifelse(df$faceRight == df$cuedFace, 1, 0)
df$faceRightCued_.5 <- ifelse(df$faceRight == df$cuedFace, .5, -.5)


df$trustAvg <- ((df$trustL + df$trustR)/2)
df$trustAvg.c <- df$trustAvg - mean(df$trustAvg, na.rm = T)

df$trustDiff.LR <- df$trustL - df$trustR 
df$trustDiff.RL <- df$trustR - df$trustL
df$trustDiff.LR.c <-  df$trustDiff.LR - mean(df$trustDiff.LR, na.rm = T)
df$trustDiff.RL.c <- df$trustDiff.RL - mean(df$trustDiff.LR, na.rm = T)

descriptives

describe(df$trustDiff.LR, na.rm = T)
##    vars    n mean   sd median trimmed  mad   min  max range  skew kurtosis   se
## X1    1 4789    0 0.63   0.02       0 0.64 -2.04 2.09  4.13 -0.01    -0.28 0.01
describe(df$trustAvg, na.rm = T)
##    vars    n mean   sd median trimmed  mad  min  max range skew kurtosis   se
## X1    1 4789  3.1 0.42   3.09    3.09 0.41 2.17 4.49  2.32 0.37     0.06 0.01

faceLeftTrusted ~ faceLeftCued (trustAvg + trustDiff | pt) singular

m3 <- glmer(faceLeftTrusted_1 ~ faceLeftCued_.5 +
              (trustAvg.c + trustDiff.LR.c | participant), family = binomial, data = df)
## boundary (singular) fit: see ?isSingular
summary(m3)
## Generalized linear mixed model fit by maximum likelihood (Laplace
##   Approximation) [glmerMod]
##  Family: binomial  ( logit )
## Formula: faceLeftTrusted_1 ~ faceLeftCued_.5 + (trustAvg.c + trustDiff.LR.c |  
##     participant)
##    Data: df
## 
##      AIC      BIC   logLik deviance df.resid 
##   6561.6   6613.4  -3272.8   6545.6     4781 
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -1.7631 -0.9693 -0.4491  0.9575  1.9096 
## 
## Random effects:
##  Groups      Name           Variance Std.Dev. Corr       
##  participant (Intercept)    0.004997 0.07069             
##              trustAvg.c     0.003074 0.05544  -1.00      
##              trustDiff.LR.c 0.666914 0.81665   1.00 -1.00
## Number of obs: 4789, groups:  participant, 221
## 
## Fixed effects:
##                 Estimate Std. Error z value Pr(>|z|)   
## (Intercept)      0.03134    0.03985   0.786  0.43161   
## faceLeftCued_.5  0.17428    0.06014   2.898  0.00376 **
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr)
## facLftCd_.5 0.012 
## optimizer (Nelder_Mead) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingular
tab_model(m3)
  faceLeftTrusted_1
Predictors Odds Ratios CI p
(Intercept) 1.03 0.95 – 1.12 0.432
faceLeftCued_.5 1.19 1.06 – 1.34 0.004
Random Effects
σ2 3.29
τ00 participant 0.00
τ11 participant.trustAvg.c 0.00
τ11 participant.trustDiff.LR.c 0.67
ρ01 -1.00
1.00
N participant 221
Observations 4789
Marginal R2 / Conditional R2 0.002 / NA

faceLeftTrusted ~ faceLeftCued_.5 (trustAvg | pt) singular

m3 <- glmer(faceLeftTrusted_1 ~  faceLeftCued_.5 +
              (trustAvg.c | participant), family = binomial, data = df)
## boundary (singular) fit: see ?isSingular
summary(m3)
## Generalized linear mixed model fit by maximum likelihood (Laplace
##   Approximation) [glmerMod]
##  Family: binomial  ( logit )
## Formula: faceLeftTrusted_1 ~ faceLeftCued_.5 + (trustAvg.c | participant)
##    Data: df
## 
##      AIC      BIC   logLik deviance df.resid 
##   6641.7   6674.0  -3315.8   6631.7     4784 
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -1.0811 -0.9682 -0.9307  0.9733  1.0700 
## 
## Random effects:
##  Groups      Name        Variance Std.Dev. Corr 
##  participant (Intercept) 0.001442 0.03798       
##              trustAvg.c  0.027407 0.16555  -1.00
## Number of obs: 4789, groups:  participant, 221
## 
## Fixed effects:
##                  Estimate Std. Error z value Pr(>|z|)   
## (Intercept)     -0.004745   0.029089  -0.163  0.87044   
## faceLeftCued_.5  0.155124   0.057949   2.677  0.00743 **
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr)
## facLftCd_.5 0.002 
## optimizer (Nelder_Mead) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingular
tab_model(m3)
  faceLeftTrusted_1
Predictors Odds Ratios CI p
(Intercept) 1.00 0.94 – 1.05 0.870
faceLeftCued_.5 1.17 1.04 – 1.31 0.007
Random Effects
σ2 3.29
τ00 participant 0.00
τ11 participant.trustAvg.c 0.03
ρ01 participant -1.00
ICC 0.00
N participant 221
Observations 4789
Marginal R2 / Conditional R2 0.002 / 0.002

faceLeftTrusted ~ faceLeftCued_.5 (trustDiff | pt) singular

m3 <- glmer(faceLeftTrusted_1 ~  faceLeftCued_.5 +
              (trustDiff.LR.c | participant), family = binomial, data = df)
## boundary (singular) fit: see ?isSingular
summary(m3)
## Generalized linear mixed model fit by maximum likelihood (Laplace
##   Approximation) [glmerMod]
##  Family: binomial  ( logit )
## Formula: faceLeftTrusted_1 ~ faceLeftCued_.5 + (trustDiff.LR.c | participant)
##    Data: df
## 
##      AIC      BIC   logLik deviance df.resid 
##   6555.9   6588.2  -3272.9   6545.9     4784 
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -1.7924 -0.9690 -0.4529  0.9586  1.8737 
## 
## Random effects:
##  Groups      Name           Variance Std.Dev. Corr
##  participant (Intercept)    0.004758 0.06898      
##              trustDiff.LR.c 0.669568 0.81827  1.00
## Number of obs: 4789, groups:  participant, 221
## 
## Fixed effects:
##                 Estimate Std. Error z value Pr(>|z|)   
## (Intercept)      0.03052    0.04001   0.763  0.44563   
## faceLeftCued_.5  0.17419    0.06014   2.896  0.00378 **
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr)
## facLftCd_.5 0.012 
## optimizer (Nelder_Mead) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingular
tab_model(m3)
  faceLeftTrusted_1
Predictors Odds Ratios CI p
(Intercept) 1.03 0.95 – 1.12 0.446
faceLeftCued_.5 1.19 1.06 – 1.34 0.004
Random Effects
σ2 3.29
τ00 participant 0.00
τ11 participant.trustDiff.LR.c 0.67
ρ01 participant 1.00
ICC 0.00
N participant 221
Observations 4789
Marginal R2 / Conditional R2 0.002 / 0.004

faceLeftTrusted ~ faceLeftCued_.5 (1 | pt) singular

m3 <- glmer(faceLeftTrusted_1 ~  faceLeftCued_.5 +
              (1 | participant), family = binomial, data = df)
## boundary (singular) fit: see ?isSingular
summary(m3)
## Generalized linear mixed model fit by maximum likelihood (Laplace
##   Approximation) [glmerMod]
##  Family: binomial  ( logit )
## Formula: faceLeftTrusted_1 ~ faceLeftCued_.5 + (1 | participant)
##    Data: df
## 
##      AIC      BIC   logLik deviance df.resid 
##   6637.8   6657.2  -3315.9   6631.8     4786 
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -1.0371 -0.9600 -0.9600  0.9642  1.0417 
## 
## Random effects:
##  Groups      Name        Variance  Std.Dev. 
##  participant (Intercept) 2.337e-14 1.529e-07
## Number of obs: 4789, groups:  participant, 221
## 
## Fixed effects:
##                  Estimate Std. Error z value Pr(>|z|)   
## (Intercept)     -0.004358   0.028922  -0.151  0.88022   
## faceLeftCued_.5  0.154571   0.057842   2.672  0.00753 **
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr)
## facLftCd_.5 0.003 
## optimizer (Nelder_Mead) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingular
tab_model(m3)
  faceLeftTrusted_1
Predictors Odds Ratios CI p
(Intercept) 1.00 0.94 – 1.05 0.880
faceLeftCued_.5 1.17 1.04 – 1.31 0.008
Random Effects
σ2 3.29
τ00 participant 0.00
N participant 221
Observations 4789
Marginal R2 / Conditional R2 0.002 / NA

faceLeftTrusted ~ trustDiff * trustAvg * faceLeftCued_.5 + (1|pt) singular

m3 <- glmer(faceLeftTrusted_1 ~ trustDiff.LR.c * trustAvg.c * faceLeftCued_.5 +
              (1 | participant), family = binomial, data = df)
## boundary (singular) fit: see ?isSingular
summary(m3)
## Generalized linear mixed model fit by maximum likelihood (Laplace
##   Approximation) [glmerMod]
##  Family: binomial  ( logit )
## Formula: faceLeftTrusted_1 ~ trustDiff.LR.c * trustAvg.c * faceLeftCued_.5 +  
##     (1 | participant)
##    Data: df
## 
##      AIC      BIC   logLik deviance df.resid 
##   6405.1   6463.4  -3193.6   6387.1     4780 
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -2.0714 -0.9431 -0.5083  0.9463  1.9183 
## 
## Random effects:
##  Groups      Name        Variance Std.Dev.
##  participant (Intercept) 4e-14    2e-07   
## Number of obs: 4789, groups:  participant, 221
## 
## Fixed effects:
##                                            Estimate Std. Error z value Pr(>|z|)
## (Intercept)                               -0.004909   0.029680  -0.165  0.86864
## trustDiff.LR.c                            -0.744362   0.049612 -15.004  < 2e-16
## trustAvg.c                                -0.081766   0.071263  -1.147  0.25122
## faceLeftCued_.5                            0.156426   0.059360   2.635  0.00841
## trustDiff.LR.c:trustAvg.c                  0.379402   0.118286   3.207  0.00134
## trustDiff.LR.c:faceLeftCued_.5             0.005280   0.099224   0.053  0.95757
## trustAvg.c:faceLeftCued_.5                 0.175964   0.142529   1.235  0.21699
## trustDiff.LR.c:trustAvg.c:faceLeftCued_.5  0.263360   0.236561   1.113  0.26559
##                                              
## (Intercept)                                  
## trustDiff.LR.c                            ***
## trustAvg.c                                   
## faceLeftCued_.5                           ** 
## trustDiff.LR.c:trustAvg.c                 ** 
## trustDiff.LR.c:faceLeftCued_.5               
## trustAvg.c:faceLeftCued_.5                   
## trustDiff.LR.c:trustAvg.c:faceLeftCued_.5    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) trD.LR. trstA. fLC_.5 trD.LR.:A. tD.LR.:L tA.:LC
## trstDff.LR.  0.001                                                 
## trustAvg.c  -0.007  0.004                                          
## facLftCd_.5  0.002 -0.011   0.004                                  
## trsD.LR.:A.  0.002 -0.196  -0.011 -0.013                           
## tD.LR.:LC_. -0.011  0.008  -0.002  0.001 -0.016                    
## trsA.:LC_.5  0.004 -0.002   0.010 -0.007 -0.068      0.004         
## tD.LR.:A.:L -0.013 -0.016  -0.068  0.002 -0.030     -0.196   -0.010
## optimizer (Nelder_Mead) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingular
#tab_model(m3)

faceLeftTrusted ~ trustDiff + trustAvg * faceLeftCued_.5 + (1|pt) singular

m3 <- glmer(faceLeftTrusted_1 ~ trustDiff.LR.c + trustAvg.c * faceLeftCued_.5 +
              (1 | participant), family = binomial, data = df)
## boundary (singular) fit: see ?isSingular
summary(m3)
## Generalized linear mixed model fit by maximum likelihood (Laplace
##   Approximation) [glmerMod]
##  Family: binomial  ( logit )
## Formula: faceLeftTrusted_1 ~ trustDiff.LR.c + trustAvg.c * faceLeftCued_.5 +  
##     (1 | participant)
##    Data: df
## 
##      AIC      BIC   logLik deviance df.resid 
##   6410.9   6449.7  -3199.4   6398.9     4783 
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -2.1595 -0.9444 -0.5137  0.9449  1.9206 
## 
## Random effects:
##  Groups      Name        Variance Std.Dev.
##  participant (Intercept) 4e-14    2e-07   
## Number of obs: 4789, groups:  participant, 221
## 
## Fixed effects:
##                             Estimate Std. Error z value Pr(>|z|)    
## (Intercept)                -0.004348   0.029639  -0.147  0.88337    
## trustDiff.LR.c             -0.715560   0.048729 -14.684  < 2e-16 ***
## trustAvg.c                 -0.073333   0.071373  -1.027  0.30421    
## faceLeftCued_.5             0.159232   0.059285   2.686  0.00723 ** 
## trustAvg.c:faceLeftCued_.5  0.209721   0.142772   1.469  0.14185    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) tD.LR. trstA. fLC_.5
## trstDff.LR.  0.001                     
## trustAvg.c   0.010  0.001              
## facLftCd_.5  0.002 -0.014  0.010       
## trsA.:LC_.5  0.010 -0.018  0.012  0.010
## optimizer (Nelder_Mead) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingular
#tab_model(m3)

faceLeftTrusted ~ trustAvg + trustDiff * faceLeftCued_.5 + (1|pt) singular

m3 <- glmer(faceLeftTrusted_1 ~  trustAvg.c + trustDiff.LR.c * faceLeftCued_.5 +
              (1 | participant), family = binomial, data = df)
## boundary (singular) fit: see ?isSingular
summary(m3)
## Generalized linear mixed model fit by maximum likelihood (Laplace
##   Approximation) [glmerMod]
##  Family: binomial  ( logit )
## Formula: faceLeftTrusted_1 ~ trustAvg.c + trustDiff.LR.c * faceLeftCued_.5 +  
##     (1 | participant)
##    Data: df
## 
##      AIC      BIC   logLik deviance df.resid 
##   6413.0   6451.8  -3200.5   6401.0     4783 
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -2.1099 -0.9443 -0.5120  0.9477  1.9115 
## 
## Random effects:
##  Groups      Name        Variance Std.Dev.
##  participant (Intercept) 0        0       
## Number of obs: 4789, groups:  participant, 221
## 
## Fixed effects:
##                                Estimate Std. Error z value Pr(>|z|)    
## (Intercept)                    -0.00491    0.02963  -0.166   0.8684    
## trustAvg.c                     -0.07487    0.07134  -1.049   0.2940    
## trustDiff.LR.c                 -0.71450    0.04871 -14.669   <2e-16 ***
## faceLeftCued_.5                 0.15847    0.05927   2.674   0.0075 ** 
## trustDiff.LR.c:faceLeftCued_.5  0.02937    0.09742   0.301   0.7631    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) trstA. trD.LR. fLC_.5
## trustAvg.c   0.008                      
## trstDff.LR.  0.002  0.002               
## facLftCd_.5  0.001  0.010 -0.013        
## tD.LR.:LC_. -0.013 -0.010  0.001   0.002
## optimizer (Nelder_Mead) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingular
#tab_model(m3)

faceLeftTrusted ~ trustAvg + trustDiff + faceLeftCued

m3 <- glm(faceLeftTrusted_1 ~ trustDiff.LR.c + trustAvg.c + faceLeftCued_.5, family = binomial, data = df)
summary(m3)
## 
## Call:
## glm(formula = faceLeftTrusted_1 ~ trustDiff.LR.c + trustAvg.c + 
##     faceLeftCued_.5, family = binomial, data = df)
## 
## Deviance Residuals: 
##     Min       1Q   Median       3Q      Max  
## -1.8551  -1.1281  -0.6882   1.1331   1.7428  
## 
## Coefficients:
##                  Estimate Std. Error z value Pr(>|z|)    
## (Intercept)     -0.004795   0.029631  -0.162  0.87144    
## trustDiff.LR.c  -0.714560   0.048710 -14.670  < 2e-16 ***
## trustAvg.c      -0.074663   0.071338  -1.047  0.29528    
## faceLeftCued_.5  0.158439   0.059268   2.673  0.00751 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 6638.9  on 4788  degrees of freedom
## Residual deviance: 6401.1  on 4785  degrees of freedom
## AIC: 6409.1
## 
## Number of Fisher Scoring iterations: 4
tab_model(m3)
  faceLeftTrusted_1
Predictors Odds Ratios CI p
(Intercept) 1.00 0.94 – 1.05 0.871
trustDiff.LR.c 0.49 0.44 – 0.54 <0.001
trustAvg.c 0.93 0.81 – 1.07 0.295
faceLeftCued_.5 1.17 1.04 – 1.32 0.008
Observations 4789
R2 Tjur 0.049 
Pdiff <- exp(-0.714560)/ (1 + exp(-0.714560))
Pdiff
## [1] 0.328592
Pcued <- exp(0.158439)/(1+exp(0.158439))
Pcued
## [1] 0.5395271

faceLeftTrusted ~ trustDiff + trustAvg * faceLeftCued

m3 <- glm(faceLeftTrusted_1 ~ trustDiff.LR.c + trustAvg.c * faceLeftCued_.5, family = binomial, data = df)
summary(m3)
## 
## Call:
## glm(formula = faceLeftTrusted_1 ~ trustDiff.LR.c + trustAvg.c * 
##     faceLeftCued_.5, family = binomial, data = df)
## 
## Deviance Residuals: 
##     Min       1Q   Median       3Q      Max  
## -1.8623  -1.1292  -0.6844   1.1297   1.7579  
## 
## Coefficients:
##                             Estimate Std. Error z value Pr(>|z|)    
## (Intercept)                -0.004348   0.029639  -0.147  0.88337    
## trustDiff.LR.c             -0.715560   0.048729 -14.684  < 2e-16 ***
## trustAvg.c                 -0.073333   0.071373  -1.027  0.30421    
## faceLeftCued_.5             0.159232   0.059285   2.686  0.00723 ** 
## trustAvg.c:faceLeftCued_.5  0.209721   0.142770   1.469  0.14185    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 6638.9  on 4788  degrees of freedom
## Residual deviance: 6398.9  on 4784  degrees of freedom
## AIC: 6408.9
## 
## Number of Fisher Scoring iterations: 4
tab_model(m3)
  faceLeftTrusted_1
Predictors Odds Ratios CI p
(Intercept) 1.00 0.94 – 1.06 0.883
trustDiff.LR.c 0.49 0.44 – 0.54 <0.001
trustAvg.c 0.93 0.81 – 1.07 0.304
faceLeftCued_.5 1.17 1.04 – 1.32 0.007
trustAvg.c *
faceLeftCued_.5 		1.23 0.93 – 1.63 0.142
Observations 4789
R2 Tjur 0.049 
Pdiff <- exp(-0.715560)/ (1 + exp(-0.715560))
Pdiff
## [1] 0.3283714
Pcued <- exp(0.159232)/ (1 + exp(0.159232))
Pcued
## [1] 0.5397241

faceLeftTrusted ~ trustAvg + trustDiff * faceLeftCued

m3 <- glm(faceLeftTrusted_1 ~ trustAvg.c + trustDiff.LR.c * faceLeftCued_.5, family = binomial, data = df)
summary(m3)
## 
## Call:
## glm(formula = faceLeftTrusted_1 ~ trustAvg.c + trustDiff.LR.c * 
##     faceLeftCued_.5, family = binomial, data = df)
## 
## Deviance Residuals: 
##     Min       1Q   Median       3Q      Max  
## -1.8417  -1.1291  -0.6824   1.1322   1.7537  
## 
## Coefficients:
##                                Estimate Std. Error z value Pr(>|z|)    
## (Intercept)                    -0.00491    0.02963  -0.166   0.8684    
## trustAvg.c                     -0.07487    0.07134  -1.049   0.2940    
## trustDiff.LR.c                 -0.71450    0.04871 -14.669   <2e-16 ***
## faceLeftCued_.5                 0.15847    0.05927   2.674   0.0075 ** 
## trustDiff.LR.c:faceLeftCued_.5  0.02937    0.09742   0.301   0.7631    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 6638.9  on 4788  degrees of freedom
## Residual deviance: 6401.0  on 4784  degrees of freedom
## AIC: 6411
## 
## Number of Fisher Scoring iterations: 4
tab_model(m3)
  faceLeftTrusted_1
Predictors Odds Ratios CI p
(Intercept) 1.00 0.94 – 1.05 0.868
trustAvg.c 0.93 0.81 – 1.07 0.294
trustDiff.LR.c 0.49 0.44 – 0.54 <0.001
faceLeftCued_.5 1.17 1.04 – 1.32 0.008
trustDiff.LR.c *
faceLeftCued_.5 		1.03 0.85 – 1.25 0.763
Observations 4789
R2 Tjur 0.049 
Pdiff <- exp(-0.71450)/ (1 + exp(-0.71450 ))
Pdiff
## [1] 0.3286053
Pcued <- exp(0.15847)/ (1 + exp(0.15847))
Pcued
## [1] 0.5395348

faceLeftTrusted ~ faceLeftCued_.5 + trustAvg * trustDiff

m3 <- glm(faceLeftTrusted_1 ~ trustDiff.LR.c * trustAvg.c + faceLeftCued_.5, family = binomial, data = df)
summary(m3)
## 
## Call:
## glm(formula = faceLeftTrusted_1 ~ trustDiff.LR.c * trustAvg.c + 
##     faceLeftCued_.5, family = binomial, data = df)
## 
## Deviance Residuals: 
##     Min       1Q   Median       3Q      Max  
## -1.8364  -1.1245  -0.6941   1.1344   1.7446  
## 
## Coefficients:
##                            Estimate Std. Error z value Pr(>|z|)    
## (Intercept)               -0.004732   0.029666  -0.160 0.873273    
## trustDiff.LR.c            -0.743967   0.049596 -15.001  < 2e-16 ***
## trustAvg.c                -0.077198   0.071022  -1.087 0.277054    
## faceLeftCued_.5            0.157019   0.059339   2.646 0.008141 ** 
## trustDiff.LR.c:trustAvg.c  0.392923   0.117737   3.337 0.000846 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 6638.9  on 4788  degrees of freedom
## Residual deviance: 6390.0  on 4784  degrees of freedom
## AIC: 6400
## 
## Number of Fisher Scoring iterations: 4
tab_model(m3)
  faceLeftTrusted_1
Predictors Odds Ratios CI p
(Intercept) 1.00 0.94 – 1.05 0.873
trustDiff.LR.c 0.48 0.43 – 0.52 <0.001
trustAvg.c 0.93 0.81 – 1.06 0.277
faceLeftCued_.5 1.17 1.04 – 1.31 0.008
trustDiff.LR.c *
trustAvg.c 		1.48 1.18 – 1.87 0.001
Observations 4789
R2 Tjur 0.051 
Pcued <- exp(0.157019)/ (1 + exp(0.157019))
Pcued
## [1] 0.5391743
Pdiff <- exp(-0.743967)/ (1 + exp(-0.743967))
Pdiff
## [1] 0.3221373
Pint <- exp(0.392923)/ (1 + exp(0.392923))
Pint
## [1] 0.5969862

faceLeftTrusted ~ trustAvg * trustDiff * faceLeftCued

m3 <- glm(faceLeftTrusted_1 ~ trustAvg.c * trustDiff.LR.c * faceLeftCued_.5, family = binomial, data = df)
summary(m3)
## 
## Call:
## glm(formula = faceLeftTrusted_1 ~ trustAvg.c * trustDiff.LR.c * 
##     faceLeftCued_.5, family = binomial, data = df)
## 
## Deviance Residuals: 
##    Min      1Q  Median      3Q     Max  
## -1.825  -1.128  -0.678   1.131   1.757  
## 
## Coefficients:
##                                            Estimate Std. Error z value Pr(>|z|)
## (Intercept)                               -0.004909   0.029680  -0.165  0.86864
## trustAvg.c                                -0.081766   0.071263  -1.147  0.25122
## trustDiff.LR.c                            -0.744362   0.049612 -15.004  < 2e-16
## faceLeftCued_.5                            0.156426   0.059360   2.635  0.00841
## trustAvg.c:trustDiff.LR.c                  0.379402   0.118286   3.207  0.00134
## trustAvg.c:faceLeftCued_.5                 0.175964   0.142527   1.235  0.21698
## trustDiff.LR.c:faceLeftCued_.5             0.005280   0.099224   0.053  0.95757
## trustAvg.c:trustDiff.LR.c:faceLeftCued_.5  0.263360   0.236573   1.113  0.26561
##                                              
## (Intercept)                                  
## trustAvg.c                                   
## trustDiff.LR.c                            ***
## faceLeftCued_.5                           ** 
## trustAvg.c:trustDiff.LR.c                 ** 
## trustAvg.c:faceLeftCued_.5                   
## trustDiff.LR.c:faceLeftCued_.5               
## trustAvg.c:trustDiff.LR.c:faceLeftCued_.5    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 6638.9  on 4788  degrees of freedom
## Residual deviance: 6387.1  on 4781  degrees of freedom
## AIC: 6403.1
## 
## Number of Fisher Scoring iterations: 4
tab_model(m3)
  faceLeftTrusted_1
Predictors Odds Ratios CI p
(Intercept) 1.00 0.94 – 1.05 0.869
trustAvg.c 0.92 0.80 – 1.06 0.251
trustDiff.LR.c 0.48 0.43 – 0.52 <0.001
faceLeftCued_.5 1.17 1.04 – 1.31 0.008
trustAvg.c *
trustDiff.LR.c 		1.46 1.16 – 1.84 0.001
trustAvg.c *
faceLeftCued_.5 		1.19 0.90 – 1.58 0.217
trustDiff.LR.c *
faceLeftCued_.5 		1.01 0.83 – 1.22 0.958
(trustAvg.c
trustDiff.LR.c)
faceLeftCued_.5 		1.30 0.82 – 2.07 0.266
Observations 4789
R2 Tjur 0.051 
Pdiff <- exp(-0.744362)/ (1 + exp(-0.744362))
Pdiff
## [1] 0.322051
Pcued <- exp(0.156426)/ (1 + exp(0.156426))
Pcued
## [1] 0.539027
Pint <- exp(0.379402)/ (1 + exp(0.379402))
Pint
## [1] 0.5937289