attention trust replication analyses
Dani Grant
4/1/2021
##
## Attaching package: 'ggplot2'## The following objects are masked from 'package:psych':
##
## %+%, alpha## Loading required package: Matrix##
## Attaching package: 'lmerTest'## The following object is masked from 'package:lme4':
##
## lmer## The following object is masked from 'package:stats':
##
## step##
## Attaching package: 'plyr'## The following object is masked from 'package:ggpubr':
##
## mutate##
## Attaching package: 'tidyr'## The following objects are masked from 'package:Matrix':
##
## expand, pack, unpack## 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## Loading required package: carData##
## Attaching package: 'car'## The following object is masked from 'package:psych':
##
## logit## Number of categories should be increased in order to count frequencies.##
## Reliability analysis
## Call: psych::alpha(x = d2[c("trustworthy", "moral", "ethical", "helpful",
## "honest", "sincere")])
##
## raw_alpha std.alpha G6(smc) average_r S/N ase mean sd median_r
## 0.98 0.98 0.97 0.88 45 0.00024 3 0.45 0.88
##
## lower alpha upper 95% confidence boundaries
## 0.98 0.98 0.98
##
## Reliability if an item is dropped:
## raw_alpha std.alpha G6(smc) average_r S/N alpha se var.r med.r
## trustworthy 0.97 0.97 0.97 0.87 35 0.00032 0.00019 0.87
## moral 0.97 0.97 0.97 0.88 37 0.00030 0.00026 0.88
## ethical 0.97 0.97 0.97 0.88 38 0.00029 0.00029 0.88
## helpful 0.97 0.98 0.97 0.89 39 0.00028 0.00029 0.89
## honest 0.97 0.97 0.97 0.88 37 0.00030 0.00038 0.88
## sincere 0.97 0.97 0.97 0.89 39 0.00029 0.00021 0.88
##
## Item statistics
## n raw.r std.r r.cor r.drop mean sd
## trustworthy 18900 0.96 0.96 0.96 0.95 3.2 0.52
## moral 18900 0.95 0.95 0.94 0.93 3.0 0.43
## ethical 18900 0.94 0.95 0.93 0.92 3.0 0.43
## helpful 18900 0.94 0.94 0.92 0.91 2.9 0.48
## honest 18900 0.95 0.95 0.94 0.93 3.1 0.49
## sincere 18900 0.94 0.94 0.93 0.92 3.0 0.47##
## When a personâ\200\231s attention is drawn to a new personâ\200\231s face, they are more likely to find that person trustworthy compared with someone their attention is not drawn to.
## 4025
## When a personâ\200\231s attention is drawn to an attractive personâ\200\231s face, the person will trust the attractive person more quickly than an unattractive person.
## 1425
## When making evaluations of trustworthiness about others, people are more likely to judge attractive individuals are more trustworthy than unattractive individuals.
## 2475
## When people are asked to evaluate trustworthiness of strangers, people are more likely to trust individuals with child-like facial traits (i.e., round face, large eyes).
## 1875## [1] 0.41071431. cft ~ 1 + (1|pt)
m1 <- glmer(cuedFaceTrusted_1 ~ 1 +
(1 | participant), family = binomial(), data = d)
summary(m1)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: cuedFaceTrusted_1 ~ 1 + (1 | participant)
## Data: d
##
## AIC BIC logLik deviance df.resid
## 13584.3 13598.7 -6790.1 13580.3 9798
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.1064 -1.0044 0.9136 0.9850 1.0966
##
## Random effects:
## Groups Name Variance Std.Dev.
## participant (Intercept) 0.0248 0.1575
## Number of obs: 9800, groups: participant, 392
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.02259 0.02177 1.038 0.299beta <- exp(0.02259)
pTrust <-(beta/(1+beta))
beta## [1] 1.022847pTrust## [1] 0.5056473pNotTrust <- (1-pTrust)
g <- cbind(pTrust, pNotTrust)
barplot(g, ylim = c(.48,.52), col = "yellow")
#graphing
plot_new <- data.frame(
choice = c("trusted","not trusted"),
probFaceTrusted = c( pTrust, pNotTrust))
ggplot(data = plot_new, aes(x = choice, y = probFaceTrusted, fill = choice)) +
geom_bar(stat="identity", position = position_dodge()) +
geom_text(aes(label = round(probFaceTrusted, digits = 4)), vjust=1.6, color="black", position = position_dodge(0.9), size=3.5) +
theme_minimal() +
ylab("probability of cued face...") +
coord_cartesian(ylim = c(0.48, .52)) 
Interpretation: \(\beta_0\): collapsing across conditions, there is no difference between trusting cued and uncued face, p_trust = .506, p_nottrust = .494, p = .299 ***
m2 <- glmer(cuedFaceTrusted_1 ~ trust_.5 +
(1 | participant), family = binomial("logit"), data = d)
summary(m2)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: cuedFaceTrusted_1 ~ trust_.5 + (1 | participant)
## Data: d
##
## AIC BIC logLik deviance df.resid
## 13585.7 13607.2 -6789.8 13579.7 9797
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.1133 -1.0074 0.9079 0.9822 1.0924
##
## Random effects:
## Groups Name Variance Std.Dev.
## participant (Intercept) 0.02452 0.1566
## Number of obs: 9800, groups: participant, 392
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.02181 0.02178 1.002 0.317
## trust_.5 0.03416 0.04356 0.784 0.433
##
## Correlation of Fixed Effects:
## (Intr)
## trust_.5 -0.046m2t <- glmer(cuedFaceTrusted_1 ~ trustYes_0 +
(1 | participant), family = binomial("logit"), data = d)
summary(m2t)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: cuedFaceTrusted_1 ~ trustYes_0 + (1 | participant)
## Data: d
##
## AIC BIC logLik deviance df.resid
## 13585.7 13607.2 -6789.8 13579.7 9797
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.1133 -1.0074 0.9079 0.9822 1.0924
##
## Random effects:
## Groups Name Variance Std.Dev.
## participant (Intercept) 0.02452 0.1566
## Number of obs: 9800, groups: participant, 392
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.03889 0.03009 1.293 0.196
## trustYes_0 -0.03416 0.04356 -0.784 0.433
##
## Correlation of Fixed Effects:
## (Intr)
## trustYes_0 -0.691m2nt <- glmer(cuedFaceTrusted_1 ~ trustNo_0 +
(1 | participant), family = binomial("logit"), data = d)
summary(m2nt)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: cuedFaceTrusted_1 ~ trustNo_0 + (1 | participant)
## Data: d
##
## AIC BIC logLik deviance df.resid
## 13585.7 13607.2 -6789.8 13579.7 9797
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.1133 -1.0074 0.9079 0.9822 1.0924
##
## Random effects:
## Groups Name Variance Std.Dev.
## participant (Intercept) 0.02452 0.1566
## Number of obs: 9800, groups: participant, 392
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.004731 0.031494 0.150 0.881
## trustNo_0 0.034159 0.043557 0.784 0.433
##
## Correlation of Fixed Effects:
## (Intr)
## trustNo_0 -0.723#trust condition
beta <- exp(0.03889)
pTtrust <-(beta/(1+beta))
beta## [1] 1.039656pTtrust## [1] 0.5097213pTnottrust <- (1-pTtrust)
#not trust condition
beta <- exp(0.004731)
pNTtrust <-(beta/(1+beta))
beta## [1] 1.004742pNTtrust## [1] 0.5011827pNTnottrust <- (1-pNTtrust)
#graphing
plot_new2 <- data.frame(
condition = c("trust", "trust", "not trust", "not trust"),
probFaceTrusted = c( pTtrust, pTnottrust, pNTtrust, pNTnottrust),
attention = c("cued", "uncued", "cued", "uncued"))
ggplot(data = plot_new2, aes(x = condition, y = probFaceTrusted, fill = attention)) +
geom_bar(stat="identity", position = position_dodge()) +
geom_text(aes(label = round(probFaceTrusted, digits = 4)), vjust=1.6, color="black", position = position_dodge(0.9), size=3.5) +
theme_minimal() +
ylab("probability of face trusted") +
coord_cartesian(ylim = c(0.48, .52)) 
3. cft ~ trust_condition + stim_gender + side_cued + (1|pt)
m3 <- glmer(cuedFaceTrusted_1 ~ trust_.5 + sideCuedRight_.5 + stim_female_.5 +
(1 | participant), family = binomial, data = d)
summary(m3)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: cuedFaceTrusted_1 ~ trust_.5 + sideCuedRight_.5 + stim_female_.5 +
## (1 | participant)
## Data: d
##
## AIC BIC logLik deviance df.resid
## 13587.3 13623.3 -6788.7 13577.3 9795
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.1304 -1.0041 0.9025 0.9856 1.1088
##
## Random effects:
## Groups Name Variance Std.Dev.
## participant (Intercept) 0.02441 0.1562
## Number of obs: 9800, groups: participant, 392
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.020895 0.022365 0.934 0.350
## trust_.5 0.034151 0.043549 0.784 0.433
## sideCuedRight_.5 0.061711 0.040634 1.519 0.129
## stim_female_.5 0.006156 0.041871 0.147 0.883
##
## Correlation of Fixed Effects:
## (Intr) trs_.5 sCR_.5
## trust_.5 -0.045
## sdCdRght_.5 -0.005 0.000
## stim_fml_.5 -0.228 0.003 0.001tab_model(m3)| cuedFaceTrusted_1 | |||
|---|---|---|---|
| Predictors | Odds Ratios | CI | p |
| (Intercept) | 1.02 | 0.98 – 1.07 | 0.350 |
| trust_.5 | 1.03 | 0.95 – 1.13 | 0.433 |
| sideCuedRight_.5 | 1.06 | 0.98 – 1.15 | 0.129 |
| stim_female_.5 | 1.01 | 0.93 – 1.09 | 0.883 |
| Random Effects | |||
| σ2 | 3.29 | ||
| τ00 participant | 0.02 | ||
| ICC | 0.01 | ||
| N participant | 392 | ||
| Observations | 9800 | ||
| Marginal R2 / Conditional R2 | 0.000 / 0.008 | ||
plot(ggpredict(m3, "trust_.5"))
beta<-exp(0.034151)
beta/(1+beta) #[1] 0.5085369## [1] 0.50853694. cft ~ trust + sim_gender + side cued + (side cued | pt) | & || singular
m4 <- glmer(cuedFaceTrusted_1 ~ trust_.5 + stim_female_.5 + sideCuedRight_.5 +
(sideCuedRight_.5 | participant), family=binomial, data=d)## boundary (singular) fit: see ?isSingularsummary(m4)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: cuedFaceTrusted_1 ~ trust_.5 + stim_female_.5 + sideCuedRight_.5 +
## (sideCuedRight_.5 | participant)
## Data: d
##
## AIC BIC logLik deviance df.resid
## 13590.7 13641.0 -6788.3 13576.7 9793
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.1599 -1.0014 0.8886 0.9865 1.1106
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## participant (Intercept) 0.02553 0.15979
## sideCuedRight_.5 0.00374 0.06115 1.00
## Number of obs: 9800, groups: participant, 392
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.020915 0.022435 0.932 0.351
## trust_.5 0.034121 0.043677 0.781 0.435
## stim_female_.5 0.006377 0.041885 0.152 0.879
## sideCuedRight_.5 0.061903 0.040763 1.519 0.129
##
## Correlation of Fixed Effects:
## (Intr) trs_.5 st__.5
## trust_.5 -0.045
## stim_fml_.5 -0.228 0.004
## sdCdRght_.5 0.025 0.000 0.001
## optimizer (Nelder_Mead) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingularplot(ggpredict(m3, "trust_.5"))
5. cft ~ trust + side cued + stim gender + (1|pt) + (1 | faceLeft) + (1 | face Right) | & || singular
m5 <- glmer(cuedFaceTrusted_1 ~ trust_.5 + sideCuedRight_.5 + stim_female_.5 +
(1 | participant) + (1 | faceRight) + (1 | faceLeft), family = binomial, data = d)## boundary (singular) fit: see ?isSingularsummary(m5) ## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: cuedFaceTrusted_1 ~ trust_.5 + sideCuedRight_.5 + stim_female_.5 +
## (1 | participant) + (1 | faceRight) + (1 | faceLeft)
## Data: d
##
## AIC BIC logLik deviance df.resid
## 13591.3 13641.7 -6788.7 13577.3 9793
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.1301 -1.0041 0.9020 0.9851 1.1089
##
## Random effects:
## Groups Name Variance Std.Dev.
## participant (Intercept) 0.0243943 0.15619
## faceRight (Intercept) 0.0005001 0.02236
## faceLeft (Intercept) 0.0000000 0.00000
## Number of obs: 9800, groups: participant, 392; faceRight, 93; faceLeft, 93
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.020909 0.022504 0.929 0.353
## trust_.5 0.034173 0.043551 0.785 0.433
## sideCuedRight_.5 0.061709 0.040637 1.519 0.129
## stim_female_.5 0.006167 0.042171 0.146 0.884
##
## Correlation of Fixed Effects:
## (Intr) trs_.5 sCR_.5
## trust_.5 -0.045
## sdCdRght_.5 -0.005 0.000
## stim_fml_.5 -0.230 0.003 0.001
## optimizer (Nelder_Mead) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingularcft ~ trust + (1|pt) + (1|face)
m7 <- glmer(cuedFaceTrusted_1 ~ trust_.5 +
(1|pt) +
(1 | face), family = binomial("logit"), data = d2)
summary(m7)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: cuedFaceTrusted_1 ~ trust_.5 + (1 | pt) + (1 | face)
## Data: d2
##
## AIC BIC logLik deviance df.resid
## 26985.2 27016.8 -13488.6 26977.2 19596
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.5280 -0.9811 0.6905 0.9726 1.5812
##
## Random effects:
## Groups Name Variance Std.Dev.
## pt (Intercept) 0.1133255 0.33664
## face (Intercept) 0.0006741 0.02596
## Number of obs: 19600, groups: pt, 392; face, 93
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.02263 0.02253 1.005 0.315
## trust_.5 0.03591 0.04472 0.803 0.422
##
## Correlation of Fixed Effects:
## (Intr)
## trust_.5 -0.044tab_model(m7)| cuedFaceTrusted_1 | |||
|---|---|---|---|
| Predictors | Odds Ratios | CI | p |
| (Intercept) | 1.02 | 0.98 – 1.07 | 0.315 |
| trust_.5 | 1.04 | 0.95 – 1.13 | 0.422 |
| Random Effects | |||
| σ2 | 3.29 | ||
| τ00 pt | 0.11 | ||
| τ00 face | 0.00 | ||
| ICC | 0.03 | ||
| N pt | 392 | ||
| N face | 93 | ||
| Observations | 19600 | ||
| Marginal R2 / Conditional R2 | 0.000 / 0.034 | ||
ggpredict(m7, "trust_.5")cft ~ trust + (1|pt) + (trustworthy | face) | & || isSingular
trustworthy.m <- glmer(cuedFaceTrusted_1 ~ trust_.5 +
(1|pt) +
(trustworthy | face), family = binomial("logit"), data = d2)## boundary (singular) fit: see ?isSingularsummary(trustworthy.m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: cuedFaceTrusted_1 ~ trust_.5 + (1 | pt) + (trustworthy | face)
## Data: d2
##
## AIC BIC logLik deviance df.resid
## 26031.5 26078.6 -13009.7 26019.5 18894
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.5633 -0.9816 0.6936 0.9770 1.6063
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## pt (Intercept) 0.111673 0.33417
## face (Intercept) 0.008707 0.09331
## trustworthy 0.001754 0.04188 -1.00
## Number of obs: 18900, groups: pt, 392; face, 90
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.01963 0.02298 0.854 0.393
## trust_.5 0.04727 0.04488 1.053 0.292
##
## Correlation of Fixed Effects:
## (Intr)
## trust_.5 -0.044
## optimizer (Nelder_Mead) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingularggpredict(trustworthy.m, "trust_.5")cft ~ trust + (1|pt) + ( overallTrust | face) | & || Singular
overall.m <- glmer(cuedFaceTrusted_1 ~ trust_.5 +
(1|pt) +
(overallTrust || face), family = binomial("logit"), data = d2)## boundary (singular) fit: see ?isSingularsummary(overall.m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: cuedFaceTrusted_1 ~ trust_.5 + (1 | pt) + (overallTrust || face)
## Data: d2
##
## AIC BIC logLik deviance df.resid
## 26029.8 26069.0 -13009.9 26019.8 18895
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.5635 -0.9810 0.6945 0.9756 1.5829
##
## Random effects:
## Groups Name Variance Std.Dev.
## pt (Intercept) 1.117e-01 3.342e-01
## face (Intercept) 1.277e-09 3.573e-05
## face.1 overallTrust 2.046e-04 1.430e-02
## Number of obs: 18900, groups: pt, 392; face, 90
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.02020 0.02294 0.881 0.379
## trust_.5 0.04719 0.04489 1.051 0.293
##
## Correlation of Fixed Effects:
## (Intr)
## trust_.5 -0.043
## optimizer (Nelder_Mead) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingulartab_model(overall.m)| cuedFaceTrusted_1 | |||
|---|---|---|---|
| Predictors | Odds Ratios | CI | p |
| (Intercept) | 1.02 | 0.98 – 1.07 | 0.379 |
| trust_.5 | 1.05 | 0.96 – 1.14 | 0.293 |
| Random Effects | |||
| σ2 | 3.29 | ||
| τ00 pt | 0.11 | ||
| τ00 face | 0.00 | ||
| τ11 face.overallTrust | 0.00 | ||
| ρ01 | |||
| ρ01 | |||
| ICC | 0.03 | ||
| N pt | 392 | ||
| N face | 90 | ||
| Observations | 18900 | ||
| Marginal R2 / Conditional R2 | 0.000 / 0.033 | ||
ggpredict(overall.m, "trust_.5")