R Notebook
- 1 Question 1: Do people choose more desirable traits?
- 2 Question 2: Does the tendency to choose the most desirable trait change across trials?
- 3 Question 3: Do people self-derogate for traits that are representative of themselves and the group norm?
- 4 Question 4: On congruent pairs (Pos-Pos, Neg-Neg), are people more likely to choose the group-consistent choice than not?
- 5 Question 6: Does the model (mixture of category representativeness and desirability) predict choice, and is it not just similarity or desirability?
- 6 Question 7: Do broad feedback category predict self-evaluations?
- 7 Question 9: Does final value estimate predict re-evaluations?
- 8 Question 10: Does the effect of category representativeness on re-evaluations depend on centrality?
1 Question 1: Do people choose more desirable traits?
1.1 Study 1
1.1.1 For incongruent pairs, categorical choice: Desirable or not
m <- glmer( chooseDes ~ 1 + ( 1 | subID), data = incongDf1, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ 1 + (1 | subID)
## Data: incongDf1
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 3922.9 3935.1 -1959.4 3918.9 3430
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.7198 -1.1551 0.5270 0.6214 0.9794
##
## Random effects:
## Groups Name Variance Std.Dev.
## subID (Intercept) 0.2622 0.512
## Number of obs: 3432, groups: subID, 80
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.07687 0.07061 15.25 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 11.1.2 For all data, continuous desirability probability
m <- lmer( desAP ~ 1 + ( 1 | subID), data = fullTD1)
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: desAP ~ 1 + (1 | subID)
## Data: fullTD1
##
## REML criterion at convergence: -2587.7
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.4250 -0.5421 -0.1281 0.7090 1.9015
##
## Random effects:
## Groups Name Variance Std.Dev.
## subID (Intercept) 0.0004977 0.02231
## Residual 0.0451700 0.21253
## Number of obs: 10279, groups: subID, 80
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.571405 0.003259 78.944125 175.3 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 11.2 Study 2
1.2.1 For incongruent pairs, categorical choice: Desirable or not
m <- glmer( chooseDes ~ 1 + ( 1 | subID), data = incongDf2, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ 1 + (1 | subID)
## Data: incongDf2
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 5995.9 6009.1 -2996.0 5991.9 5248
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.0112 -1.0240 0.4381 0.6743 1.4060
##
## Random effects:
## Groups Name Variance Std.Dev.
## subID (Intercept) 0.6059 0.7784
## Number of obs: 5250, groups: subID, 105
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.01302 0.08332 12.16 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 11.2.2 For all data, continuous desirability probability
m <- lmer( desAP ~ 1 + ( 1 | subID), data = fullTD2)
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: desAP ~ 1 + (1 | subID)
## Data: fullTD2
##
## REML criterion at convergence: -3600.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.4711 -0.5448 -0.1023 0.6683 2.1255
##
## Random effects:
## Groups Name Variance Std.Dev.
## subID (Intercept) 0.001646 0.04057
## Residual 0.045989 0.21445
## Number of obs: 15750, groups: subID, 105
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.631e-01 4.312e-03 1.040e+02 130.6 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 11.3 Study 3
1.3.1 For incongruent pairs, categorical choice: Desirable or not
m <- glmer( chooseDes ~ 1 + ( 1 | subID), data = incongDf3, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ 1 + (1 | subID)
## Data: incongDf3
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 10296.8 10311.3 -5146.4 10292.8 10231
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.0984 0.2440 0.3948 0.5545 1.5907
##
## Random effects:
## Groups Name Variance Std.Dev.
## subID (Intercept) 0.7997 0.8942
## Number of obs: 10233, groups: subID, 208
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.3742 0.0679 20.24 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 11.3.2 For all data, continuous desirability probability
m <- lmer( desAP ~ 1 + ( 1 | subID), data = fullTD3)
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: desAP ~ 1 + (1 | subID)
## Data: fullTD3
##
## REML criterion at convergence: -8699.1
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.6986 -0.5852 -0.1396 0.8355 2.3049
##
## Random effects:
## Groups Name Variance Std.Dev.
## subID (Intercept) 0.001872 0.04327
## Residual 0.043448 0.20844
## Number of obs: 30578, groups: subID, 208
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.803e-01 3.229e-03 2.057e+02 179.7 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 11.4 Combined
1.4.1 For incongruent pairs, categorical choice: Desirable or not
m <- glmer( chooseDes ~ condition + ( 1 | subID), data = incongDf, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ condition + (1 | subID)
## Data: incongDf
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 20229.9 20269.1 -10109.9 20219.9 18910
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.8821 -0.6767 0.4260 0.6057 1.5408
##
## Random effects:
## Groups Name Variance Std.Dev.
## subID (Intercept) 0.635 0.7969
## Number of obs: 18915, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.34740 0.08632 15.608 < 2e-16 ***
## conditionAsian -0.33226 0.12077 -2.751 0.00594 **
## conditionDemocrat 0.02463 0.12250 0.201 0.84066
## conditionLatino -0.23577 0.13066 -1.804 0.07115 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) cndtnA cndtnD
## conditinAsn -0.712
## condtnDmcrt -0.702 0.502
## conditinLtn -0.659 0.470 0.464ggpredict(m, c("condition")) %>% plot()
m <- glmer( chooseDes ~ conditionEC + ( 1 | subID), data = incongDf, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ conditionEC + (1 | subID)
## Data: incongDf
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 20229.9 20269.1 -10109.9 20219.9 18910
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.8821 -0.6767 0.4260 0.6057 1.5408
##
## Random effects:
## Groups Name Variance Std.Dev.
## subID (Intercept) 0.635 0.7969
## Number of obs: 18915, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.21155 0.04483 27.028 <2e-16 ***
## conditionECAsian -0.19641 0.07463 -2.632 0.0085 **
## conditionECDemocrat 0.16048 0.07600 2.112 0.0347 *
## conditionECLatino -0.09992 0.08254 -1.211 0.2260
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) cndECA cndECD
## condtnECAsn -0.060
## cndtnECDmcr -0.027 -0.299
## condtnECLtn 0.113 -0.359 -0.370ggpredict(m, c("conditionEC")) %>% plot()
1.4.1.1 All ingroup vs. outgroup condition
m <- glmer( chooseDes ~ ingroup+ ( 1 | subID), data = incongDf, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ ingroup + (1 | subID)
## Data: incongDf
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 20234.8 20258.4 -10114.4 20228.8 18912
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.8925 -0.6654 0.4298 0.6070 1.5680
##
## Random effects:
## Groups Name Variance Std.Dev.
## subID (Intercept) 0.6531 0.8082
## Number of obs: 18915, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.16889 0.05241 22.302 <2e-16 ***
## ingroupOut 0.18044 0.10165 1.775 0.0759 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## ingroupOut -0.511ggpredict(m, c("ingroup")) %>% plot()
1.4.1.2 All ingroup vs. outgroup moderated by self-esteem
m <- glmer( chooseDes ~ ingroup * scale(SE) + ( 1 | subID), data = incongDf, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ ingroup * scale(SE) + (1 | subID)
## Data: incongDf
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 19847.1 19886.3 -9918.5 19837.1 18660
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.1945 -0.6616 0.4308 0.6001 1.5571
##
## Random effects:
## Groups Name Variance Std.Dev.
## subID (Intercept) 0.4557 0.6751
## Number of obs: 18665, groups: subID, 388
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.20946 0.04578 26.421 < 2e-16 ***
## ingroupOut 0.09320 0.09086 1.026 0.30500
## scale(SE) 0.48772 0.04776 10.213 < 2e-16 ***
## ingroupOut:scale(SE) -0.22621 0.08712 -2.596 0.00942 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) ingrpO sc(SE)
## ingroupOut -0.498
## scale(SE) 0.135 -0.065
## ingrpO:(SE) -0.072 -0.095 -0.547ggpredict(m, c("SE","ingroup")) %>% plot()## Data were 'prettified'. Consider using `terms="SE [all]"` to get smooth plots.
1.4.2 For all data, continuous desirability probability
m <- lmer( desAP ~ condition + ( 1 | subID), data = fullTD)
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: desAP ~ condition + (1 | subID)
## Data: fullTD
##
## REML criterion at convergence: -14838.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.6608 -0.5662 -0.1271 0.7645 2.2574
##
## Random effects:
## Groups Name Variance Std.Dev.
## subID (Intercept) 0.001546 0.03932
## Residual 0.044467 0.21087
## Number of obs: 56607, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.579635 0.004197 386.381949 138.101 < 2e-16 ***
## conditionAsian -0.016580 0.005930 384.948863 -2.796 0.00543 **
## conditionDemocrat 0.001414 0.005964 386.300697 0.237 0.81267
## conditionLatino -0.008236 0.006425 396.716308 -1.282 0.20061
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) cndtnA cndtnD
## conditinAsn -0.708
## condtnDmcrt -0.704 0.498
## conditinLtn -0.653 0.462 0.460ggpredict(m, c("condition")) %>% plot()
m <- lmer( desAP ~ conditionEC + ( 1 | subID), data = fullTD)
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: desAP ~ conditionEC + (1 | subID)
## Data: fullTD
##
## REML criterion at convergence: -14835.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.6608 -0.5662 -0.1271 0.7645 2.2574
##
## Random effects:
## Groups Name Variance Std.Dev.
## subID (Intercept) 0.001546 0.03932
## Residual 0.044467 0.21087
## Number of obs: 56607, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.573785 0.002191 391.178158 261.928 < 2e-16 ***
## conditionECAsian -0.010730 0.003684 386.201069 -2.912 0.00379 **
## conditionECDemocrat 0.007265 0.003712 387.937631 1.957 0.05102 .
## conditionECLatino -0.002386 0.004078 400.714507 -0.585 0.55887
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) cndECA cndECD
## condtnECAsn -0.051
## cndtnECDmcr -0.038 -0.298
## condtnECLtn 0.125 -0.366 -0.370ggpredict(m, c("conditionEC")) %>% plot()
1.4.2.1 All ingroup vs. outgroup condition
m <- lmer( desAP ~ ingroup+ ( 1 | subID), data = fullTD)
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: desAP ~ ingroup + (1 | subID)
## Data: fullTD
##
## REML criterion at convergence: -14846.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.6615 -0.5663 -0.1268 0.7638 2.2666
##
## Random effects:
## Groups Name Variance Std.Dev.
## subID (Intercept) 0.00158 0.03975
## Residual 0.04447 0.21087
## Number of obs: 56607, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.718e-01 2.564e-03 3.923e+02 223.029 <2e-16 ***
## ingroupOut 7.839e-03 4.951e-03 3.896e+02 1.583 0.114
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## ingroupOut -0.518ggpredict(m, c("ingroup")) %>% plot()
1.4.2.2 All ingroup vs. outgroup moderated by self-esteem
m <- lmer( chooseDes ~ ingroup * scale(SE) + ( 1 | subID), data = fullTD)
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: chooseDes ~ ingroup * scale(SE) + (1 | subID)
## Data: fullTD
##
## REML criterion at convergence: 76777
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.6593 -1.2295 0.6667 0.7925 1.2219
##
## Random effects:
## Groups Name Variance Std.Dev.
## subID (Intercept) 0.003384 0.05817
## Residual 0.229507 0.47907
## Number of obs: 55858, groups: subID, 388
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.624854 0.004179 388.184413 149.530 <2e-16 ***
## ingroupOut 0.005799 0.008334 382.300016 0.696 0.4869
## scale(SE) 0.042560 0.004335 385.739378 9.817 <2e-16 ***
## ingroupOut:scale(SE) -0.018826 0.007907 383.668692 -2.381 0.0177 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) ingrpO sc(SE)
## ingroupOut -0.501
## scale(SE) 0.082 -0.041
## ingrpO:(SE) -0.045 -0.135 -0.548ggpredict(m, c("SE","ingroup")) %>% plot()
2 Question 2: Does the tendency to choose the most desirable trait change across trials?
2.1 Study 1
2.1.1 For incongruent pairs, categorical choice: Desirable or not
m <- glmer( chooseDes ~ scale(trialTotal) + ( scale(trialTotal) | subID), data = incongDf1, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ scale(trialTotal) + (scale(trialTotal) | subID)
## Data: incongDf1
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 3875.1 3905.8 -1932.5 3865.1 3427
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.2154 -1.0617 0.4957 0.6117 1.1217
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.2833 0.5323
## scale(trialTotal) 0.1271 0.3565 -0.26
## Number of obs: 3432, groups: subID, 80
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.12353 0.07370 15.244 < 2e-16 ***
## scale(trialTotal) -0.24572 0.05972 -4.114 3.88e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## scl(trlTtl) -0.2012.1.2 For all data, continuous desirability probability
m <- lmer( desAP ~ scale(trialTotal) + ( scale(trialTotal) | subID), data = fullTD1)
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: desAP ~ scale(trialTotal) + (scale(trialTotal) | subID)
## Data: fullTD1
##
## REML criterion at convergence: -2628.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.5061 -0.5426 -0.1260 0.7087 1.9615
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.0005008 0.02238
## scale(trialTotal) 0.0002158 0.01469 -0.23
## Residual 0.0447829 0.21162
## Number of obs: 10279, groups: subID, 80
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.57135 0.00326 78.84818 175.254 < 2e-16 ***
## scale(trialTotal) -0.01327 0.00266 80.12026 -4.991 3.43e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## scl(trlTtl) -0.1082.2 Study 2
2.2.1 For incongruent pairs, categorical choice: Desirable or not
m <- glmer( chooseDes ~ scale(trialTotal) + ( scale(trialTotal) | subID), data = incongDf2, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ scale(trialTotal) + (scale(trialTotal) | subID)
## Data: incongDf2
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 5994.7 6027.5 -2992.3 5984.7 5245
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.2952 -0.9999 0.4510 0.6766 1.4178
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.611983 0.7823
## scale(trialTotal) 0.008631 0.0929 -0.31
## Number of obs: 5250, groups: subID, 105
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.01759 0.08380 12.142 <2e-16 ***
## scale(trialTotal) -0.09085 0.03540 -2.566 0.0103 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## scl(trlTtl) -0.0982.2.2 For all data, continuous desirability probability
m <- lmer( desAP ~ scale(trialTotal) + ( scale(trialTotal) | subID), data = fullTD2)## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
## Model failed to converge with max|grad| = 0.00580377 (tol = 0.002, component 1)summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: desAP ~ scale(trialTotal) + (scale(trialTotal) | subID)
## Data: fullTD2
##
## REML criterion at convergence: -3595
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.4939 -0.5448 -0.1036 0.6701 2.1397
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 1.647e-03 0.040582
## scale(trialTotal) 7.307e-06 0.002703 -0.18
## Residual 4.597e-02 0.214406
## Number of obs: 15750, groups: subID, 105
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.563055 0.004313 103.917117 130.542 <2e-16 ***
## scale(trialTotal) -0.003762 0.001729 104.035976 -2.176 0.0318 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## scl(trlTtl) -0.025
## optimizer (nloptwrap) convergence code: 0 (OK)
## Model failed to converge with max|grad| = 0.00580377 (tol = 0.002, component 1)2.3 Study 3
2.3.1 For incongruent pairs, categorical choice: Desirable or not
m <- glmer( chooseDes ~ scale(trialTotal) + ( scale(trialTotal) | subID), data = incongDf3, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ scale(trialTotal) + (scale(trialTotal) | subID)
## Data: incongDf3
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 10236.1 10272.2 -5113.0 10226.1 10228
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.4154 0.2333 0.3924 0.5544 1.7973
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.82727 0.9095
## scale(trialTotal) 0.06271 0.2504 -0.04
## Number of obs: 10233, groups: subID, 208
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.40208 0.06921 20.259 < 2e-16 ***
## scale(trialTotal) -0.18307 0.03303 -5.542 2.99e-08 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## scl(trlTtl) -0.0592.3.2 For all data, continuous desirability probability
m <- lmer( desAP ~ scale(trialTotal) + ( scale(trialTotal) | subID), data = fullTD3)## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
## Model failed to converge with max|grad| = 0.00567482 (tol = 0.002, component 1)summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: desAP ~ scale(trialTotal) + (scale(trialTotal) | subID)
## Data: fullTD3
##
## REML criterion at convergence: -8744.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.7207 -0.5852 -0.1402 0.8352 2.3608
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.0018765 0.04332
## scale(trialTotal) 0.0001188 0.01090 0.23
## Residual 0.0432714 0.20802
## Number of obs: 30578, groups: subID, 208
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.580336 0.003231 205.583457 179.600 < 2e-16 ***
## scale(trialTotal) -0.007613 0.001410 208.611782 -5.399 1.81e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## scl(trlTtl) 0.113
## optimizer (nloptwrap) convergence code: 0 (OK)
## Model failed to converge with max|grad| = 0.00567482 (tol = 0.002, component 1)2.4 Combined
2.4.1 For incongruent pairs, categorical choice: Desirable or not
m <- glmer( chooseDes ~ scale(trialTotal)*condition + ( scale(trialTotal) | subID), data = incongDf, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ scale(trialTotal) * condition + (scale(trialTotal) |
## subID)
## Data: incongDf
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 20126.2 20212.6 -10052.1 20104.2 18904
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.3437 -0.6850 0.4217 0.5968 1.7607
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.65435 0.8089
## scale(trialTotal) 0.05479 0.2341 -0.06
## Number of obs: 18915, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.37702 0.08767 15.706 < 2e-16 ***
## scale(trialTotal) -0.14186 0.04270 -3.322 0.000894 ***
## conditionAsian -0.34263 0.12255 -2.796 0.005176 **
## conditionDemocrat 0.03352 0.12434 0.270 0.787462
## conditionLatino -0.29412 0.13270 -2.216 0.026659 *
## scale(trialTotal):conditionAsian 0.05323 0.05775 0.922 0.356713
## scale(trialTotal):conditionDemocrat -0.07554 0.06012 -1.256 0.208951
## scale(trialTotal):conditionLatino -0.11845 0.06782 -1.746 0.080733 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) scl(T) cndtnA cndtnD cndtnL s(T):A s(T):D
## scl(trlTtl) -0.073
## conditinAsn -0.712 0.051
## condtnDmcrt -0.701 0.048 0.501
## conditinLtn -0.658 0.048 0.470 0.463
## scl(trlT):A 0.052 -0.720 -0.067 -0.036 -0.035
## scl(trlT):D 0.049 -0.686 -0.036 -0.078 -0.033 0.506
## scl(trlT):L 0.044 -0.613 -0.032 -0.031 -0.017 0.450 0.431ggpredict(m, c("trialTotal","condition")) %>% plot()## Data were 'prettified'. Consider using `terms="trialTotal [all]"` to get smooth plots.
m <- glmer( chooseDes ~ scale(trialTotal)*conditionEC + ( scale(trialTotal) | subID), data = incongDf, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ scale(trialTotal) * conditionEC + (scale(trialTotal) |
## subID)
## Data: incongDf
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 20126.2 20212.6 -10052.1 20104.2 18904
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.3437 -0.6850 0.4217 0.5968 1.7607
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.65435 0.8089
## scale(trialTotal) 0.05479 0.2341 -0.06
## Number of obs: 18915, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.22621 0.04563 26.870 < 2e-16 ***
## scale(trialTotal) -0.17705 0.02343 -7.555 4.18e-14 ***
## conditionECAsian -0.19182 0.07575 -2.532 0.0113 *
## conditionECDemocrat 0.18433 0.07722 2.387 0.0170 *
## conditionECLatino -0.14331 0.08387 -1.709 0.0875 .
## scale(trialTotal):conditionECAsian 0.08842 0.03592 2.461 0.0138 *
## scale(trialTotal):conditionECDemocrat -0.04035 0.03784 -1.066 0.2863
## scale(trialTotal):conditionECLatino -0.08326 0.04384 -1.899 0.0576 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) scl(T) cndECA cndECD cndECL s(T):ECA s(T):ECD
## scl(trlTtl) -0.052
## condtnECAsn -0.061 -0.005
## cndtnECDmcr -0.025 -0.026 -0.299
## condtnECLtn 0.113 0.040 -0.359 -0.370
## scl(tT):ECA -0.005 -0.138 -0.055 0.033 -0.006
## scl(tT):ECD -0.024 -0.037 0.032 -0.070 0.003 -0.263
## scl(tT):ECL 0.036 0.197 -0.005 0.003 0.003 -0.381 -0.406ggpredict(m, c("trialTotal","conditionEC")) %>% plot()## Data were 'prettified'. Consider using `terms="trialTotal [all]"` to get smooth plots.
2.4.2 For all data, continuous desirability probability
m <- lmer( desAP ~ scale(trialTotal)*condition + ( scale(trialTotal) | subID), data = fullTD)
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: desAP ~ scale(trialTotal) * condition + (scale(trialTotal) |
## subID)
## Data: fullTD
##
## REML criterion at convergence: -14899.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.6819 -0.5648 -0.1288 0.7659 2.3022
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.0015455 0.03931
## scale(trialTotal) 0.0001026 0.01013 0.12
## Residual 0.0443012 0.21048
## Number of obs: 56607, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error df t value
## (Intercept) 0.579926 0.004196 386.340365 138.215
## scale(trialTotal) -0.006369 0.001936 346.852452 -3.290
## conditionAsian -0.016711 0.005928 384.895772 -2.819
## conditionDemocrat 0.001432 0.005962 386.239149 0.240
## conditionLatino -0.011419 0.006438 400.483549 -1.774
## scale(trialTotal):conditionAsian 0.002678 0.002726 341.529107 0.982
## scale(trialTotal):conditionDemocrat -0.002233 0.002751 346.541705 -0.812
## scale(trialTotal):conditionLatino -0.008597 0.003247 510.244990 -2.648
## Pr(>|t|)
## (Intercept) < 2e-16 ***
## scale(trialTotal) 0.00111 **
## conditionAsian 0.00507 **
## conditionDemocrat 0.81037
## conditionLatino 0.07688 .
## scale(trialTotal):conditionAsian 0.32676
## scale(trialTotal):conditionDemocrat 0.41753
## scale(trialTotal):conditionLatino 0.00836 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) scl(T) cndtnA cndtnD cndtnL s(T):A s(T):D
## scl(trlTtl) 0.041
## conditinAsn -0.708 -0.029
## condtnDmcrt -0.704 -0.029 0.498
## conditinLtn -0.652 -0.027 0.461 0.459
## scl(trlT):A -0.029 -0.710 0.042 0.021 0.019
## scl(trlT):D -0.029 -0.704 0.020 0.042 0.019 0.500
## scl(trlT):L -0.025 -0.596 0.017 0.017 0.096 0.423 0.420ggpredict(m, c("trialTotal","condition")) %>% plot()
m <- lmer( desAP ~ scale(trialTotal)*conditionEC + ( scale(trialTotal) | subID), data = fullTD)
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: desAP ~ scale(trialTotal) * conditionEC + (scale(trialTotal) |
## subID)
## Data: fullTD
##
## REML criterion at convergence: -14894.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.6819 -0.5648 -0.1288 0.7659 2.3022
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.0015455 0.03931
## scale(trialTotal) 0.0001026 0.01013 0.12
## Residual 0.0443012 0.21048
## Number of obs: 56607, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error df t value
## (Intercept) 5.733e-01 2.193e-03 3.931e+02 261.429
## scale(trialTotal) -8.407e-03 1.062e-03 4.274e+02 -7.916
## conditionECAsian -1.004e-02 3.685e-03 3.868e+02 -2.724
## conditionECDemocrat 8.106e-03 3.712e-03 3.886e+02 2.184
## conditionECLatino -4.745e-03 4.091e-03 4.060e+02 -1.160
## scale(trialTotal):conditionECAsian 4.716e-03 1.723e-03 3.671e+02 2.736
## scale(trialTotal):conditionECDemocrat -1.947e-04 1.743e-03 3.735e+02 -0.112
## scale(trialTotal):conditionECLatino -6.559e-03 2.127e-03 5.837e+02 -3.083
## Pr(>|t|)
## (Intercept) < 2e-16 ***
## scale(trialTotal) 2.13e-14 ***
## conditionECAsian 0.00675 **
## conditionECDemocrat 0.02957 *
## conditionECLatino 0.24675
## scale(trialTotal):conditionECAsian 0.00652 **
## scale(trialTotal):conditionECDemocrat 0.91112
## scale(trialTotal):conditionECLatino 0.00214 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) scl(T) cndECA cndECD cndECL s(T):ECA s(T):ECD
## scl(trlTtl) 0.072
## condtnECAsn -0.052 -0.021
## cndtnECDmcr -0.040 -0.020 -0.297
## condtnECLtn 0.129 0.057 -0.367 -0.371
## scl(tT):ECA -0.022 -0.113 0.053 -0.001 -0.047
## scl(tT):ECD -0.021 -0.094 -0.001 0.054 -0.047 -0.249
## scl(tT):ECL 0.053 0.253 -0.042 -0.043 0.115 -0.407 -0.411ggpredict(m, c("trialTotal","conditionEC")) %>% plot()
3 Question 3: Do people self-derogate for traits that are representative of themselves and the group norm?
3.1 Study 1
3.1.1 For incongruent pairs, categorical choice: Desirable or not
m <- glmer( chooseDes ~ compMotC + ( compMotC | subID), data = incongDf1, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)## boundary (singular) fit: see ?isSingularsummary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ compMotC + (compMotC | subID)
## Data: incongDf1
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 3925.4 3956.1 -1957.7 3915.4 3427
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.7195 -1.1419 0.5265 0.6226 1.0431
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.238008 0.48786
## compMotCCompete 0.001476 0.03842 1.00
## Number of obs: 3432, groups: subID, 80
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.11073 0.07129 15.58 <2e-16 ***
## compMotCCompete -0.17201 0.11704 -1.47 0.142
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## compMtCCmpt -0.242
## optimizer (bobyqa) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingularggpredict(m, c("compMotC")) %>% plot()
3.1.1.1 Moderated by self-esteem?
m <- glmer( chooseDes ~ compMotC*SE + ( compMotC | subID), data = incongDf1, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)## boundary (singular) fit: see ?isSingularsummary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ compMotC * SE + (compMotC | subID)
## Data: incongDf1
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 3918.9 3961.9 -1952.5 3904.9 3425
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.6941 -1.1324 0.5232 0.6214 1.1176
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.2308042 0.48042
## compMotCCompete 0.0004323 0.02079 -1.00
## Number of obs: 3432, groups: subID, 80
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.5773 0.3881 1.488 0.13688
## compMotCCompete -1.4643 0.5441 -2.691 0.00712 **
## SE 0.1870 0.1344 1.391 0.16415
## compMotCCompete:SE 0.4497 0.1930 2.331 0.01977 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) cmpMCC SE
## compMtCCmpt -0.313
## SE -0.983 0.310
## cmpMtCCm:SE 0.303 -0.978 -0.311
## optimizer (bobyqa) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingularggpredict(m, c("SE","compMotC")) %>% plot()## Data were 'prettified'. Consider using `terms="SE [all]"` to get smooth plots.
3.1.2 For all data, continuous desirability probability
m <- lmer( desAP ~ compMotC + ( compMotC | subID), data = fullTD1)## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
## Model failed to converge with max|grad| = 0.00287085 (tol = 0.002, component 1)summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: desAP ~ compMotC + (compMotC | subID)
## Data: fullTD1
##
## REML criterion at convergence: -2777.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.6085 -0.5194 -0.1095 0.7570 1.9434
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.0009785 0.03128
## compMotCCompete 0.0007011 0.02648 -1.00
## Residual 0.0442508 0.21036
## Number of obs: 10279, groups: subID, 80
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.591676 0.004348 75.871438 136.08 <2e-16 ***
## compMotCCompete -0.058665 0.005269 111.176107 -11.13 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## compMtCCmpt -0.744
## optimizer (nloptwrap) convergence code: 0 (OK)
## Model failed to converge with max|grad| = 0.00287085 (tol = 0.002, component 1)ggpredict(m, c("compMotC")) %>% plot()
3.1.2.1 Moderated by Self-Esteem?
m <- lmer( desAP ~ scale(SE)*compMotC + ( compMotC | subID), data = fullTD1)## boundary (singular) fit: see ?isSingularsummary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: desAP ~ scale(SE) * compMotC + (compMotC | subID)
## Data: fullTD1
##
## REML criterion at convergence: -2777.7
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.6181 -0.5220 -0.1038 0.7461 2.0104
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.0009693 0.03113
## compMotCCompete 0.0008414 0.02901 -1.00
## Residual 0.0441828 0.21020
## Number of obs: 10279, groups: subID, 80
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.591696 0.004333 74.955082 136.540 <2e-16 ***
## scale(SE) 0.005525 0.004332 74.319641 1.275 0.2062
## compMotCCompete -0.058671 0.005429 106.850485 -10.807 <2e-16 ***
## scale(SE):compMotCCompete 0.009522 0.005446 107.626783 1.748 0.0832 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) sc(SE) cmpMCC
## scale(SE) 0.001
## compMtCCmpt -0.763 0.000
## scl(SE):MCC 0.000 -0.760 0.003
## optimizer (nloptwrap) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingularggpredict(m, c("SE","compMotC")) %>% plot()
3.2 Study 2
3.2.1 For incongruent pairs, categorical choice: Desirable or not
m <- glmer( chooseDes ~ compMotC + ( compMotC | subID), data = incongDf2, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ compMotC + (compMotC | subID)
## Data: incongDf2
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 5976.6 6009.5 -2983.3 5966.6 5245
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.1163 -0.9829 0.4496 0.6627 1.4489
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.52215 0.7226
## compMotCCompete 0.02539 0.1593 -0.26
## Number of obs: 5250, groups: subID, 105
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.10640 0.08077 13.699 < 2e-16 ***
## compMotCCompete -0.41559 0.09608 -4.325 1.52e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## compMtCCmpt -0.250ggpredict(m, c("compMotC")) %>% plot()
3.2.1.1 Moderated by self-esteem?
m <- glmer( chooseDes ~ compMotC*SE + ( compMotC | subID), data = incongDf2, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ compMotC * SE + (compMotC | subID)
## Data: incongDf2
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 5946.1 5992.0 -2966.0 5932.1 5243
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.4309 -0.9475 0.4506 0.6597 1.5242
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.37468 0.6121
## compMotCCompete 0.01504 0.1226 -0.92
## Number of obs: 5250, groups: subID, 105
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.7082 0.3504 -2.021 0.0432 *
## compMotCCompete -0.7915 0.3659 -2.163 0.0305 *
## SE 0.7078 0.1345 5.263 1.42e-07 ***
## compMotCCompete:SE 0.1473 0.1509 0.977 0.3288
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) cmpMCC SE
## compMtCCmpt -0.435
## SE -0.979 0.426
## cmpMtCCm:SE 0.398 -0.970 -0.407ggpredict(m, c("SE","compMotC")) %>% plot()## Data were 'prettified'. Consider using `terms="SE [all]"` to get smooth plots.
3.2.2 For all data, continuous desirability probability
m <- lmer( desAP ~ compMotC + ( compMotC | subID), data = fullTD2)
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: desAP ~ compMotC + (compMotC | subID)
## Data: fullTD2
##
## REML criterion at convergence: -4086
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.7091 -0.5176 -0.0657 0.6833 2.1365
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.002421 0.04921
## compMotCCompete 0.001411 0.03757 -0.90
## Residual 0.044485 0.21091
## Number of obs: 15750, groups: subID, 105
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.589718 0.005258 103.604619 112.16 <2e-16 ***
## compMotCCompete -0.074052 0.005067 103.153355 -14.61 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## compMtCCmpt -0.765ggpredict(m, c("compMotC")) %>% plot()
3.2.2.1 Moderated by Self-Esteem?
m <- lmer( desAP ~ scale(SE)*compMotC + ( compMotC | subID), data = fullTD2)
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: desAP ~ scale(SE) * compMotC + (compMotC | subID)
## Data: fullTD2
##
## REML criterion at convergence: -4101.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.69801 -0.52089 -0.06188 0.68647 2.15396
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.001853 0.04304
## compMotCCompete 0.001335 0.03653 -0.93
## Residual 0.044489 0.21092
## Number of obs: 15750, groups: subID, 105
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.589679 0.004715 102.684399 125.074 < 2e-16 ***
## scale(SE) 0.024114 0.004711 102.260629 5.119 1.45e-06 ***
## compMotCCompete -0.073849 0.004994 103.101610 -14.788 < 2e-16 ***
## scale(SE):compMotCCompete -0.008167 0.005012 104.434856 -1.630 0.106
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) sc(SE) cmpMCC
## scale(SE) -0.009
## compMtCCmpt -0.785 0.009
## scl(SE):MCC 0.009 -0.782 0.013ggpredict(m, c("SE","compMotC")) %>% plot()
3.3 Study 3
3.3.1 For incongruent pairs, categorical choice: Desirable or not
m <- glmer( chooseDes ~ compMotC + ( compMotC | subID), data = incongDf3, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)## boundary (singular) fit: see ?isSingularsummary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ compMotC + (compMotC | subID)
## Data: incongDf3
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 10257.8 10294.0 -5123.9 10247.8 10228
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.1226 0.2426 0.4051 0.5552 1.5909
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.729147 0.85390
## compMotCCompete 0.009283 0.09635 -1.00
## Number of obs: 10233, groups: subID, 208
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.46271 0.06685 21.88 < 2e-16 ***
## compMotCCompete -0.48664 0.07522 -6.47 9.81e-11 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## compMtCCmpt -0.278
## optimizer (bobyqa) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingular3.3.1.1 Moderated by Self-Esteem?
m <- glmer( chooseDes ~ compMotC*SE + ( compMotC | subID), data = incongDf3, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)## boundary (singular) fit: see ?isSingularsummary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ compMotC * SE + (compMotC | subID)
## Data: incongDf3
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 9921.3 9971.8 -4953.7 9907.3 9976
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.3551 0.2442 0.4003 0.5604 1.5787
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.499238 0.70657
## compMotCCompete 0.004413 0.06643 -1.00
## Number of obs: 9983, groups: subID, 203
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.60759 0.27623 -2.200 0.0278 *
## compMotCCompete -0.28504 0.29162 -0.977 0.3284
## SE 0.70857 0.09288 7.629 2.37e-14 ***
## compMotCCompete:SE -0.06506 0.10317 -0.631 0.5283
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) cmpMCC SE
## compMtCCmpt -0.351
## SE -0.977 0.344
## cmpMtCCm:SE 0.331 -0.968 -0.340
## optimizer (bobyqa) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingularggpredict(m, c("SE","compMotC")) %>% plot()## Data were 'prettified'. Consider using `terms="SE [all]"` to get smooth plots.
### For all data, continuous desirability probability
m <- lmer( desAP ~ compMotC + ( compMotC | subID), data = fullTD3)
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: desAP ~ compMotC + (compMotC | subID)
## Data: fullTD3
##
## REML criterion at convergence: -9898.1
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.94187 -0.54822 -0.08824 0.80579 2.32637
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.002716 0.05212
## compMotCCompete 0.001554 0.03943 -0.88
## Residual 0.041662 0.20411
## Number of obs: 30578, groups: subID, 208
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.608719 0.003905 198.283431 155.86 <2e-16 ***
## compMotCCompete -0.081765 0.003671 207.094375 -22.27 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## compMtCCmpt -0.760ggpredict(m, c("compMotC")) %>% plot()
3.3.1.2 Moderated by Self-Esteem?
m <- lmer( desAP ~ scale(SE)*compMotC + ( compMotC | subID), data = fullTD3)
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: desAP ~ scale(SE) * compMotC + (compMotC | subID)
## Data: fullTD3
##
## REML criterion at convergence: -9727.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.93352 -0.55219 -0.08943 0.80999 2.32983
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.001892 0.04350
## compMotCCompete 0.001324 0.03639 -0.86
## Residual 0.041622 0.20402
## Number of obs: 29829, groups: subID, 203
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.609609 0.003400 190.848386 179.300 < 2e-16 ***
## scale(SE) 0.026384 0.003403 192.269598 7.753 5.07e-13 ***
## compMotCCompete -0.081944 0.003561 203.621883 -23.013 < 2e-16 ***
## scale(SE):compMotCCompete -0.014807 0.003561 204.271828 -4.158 4.72e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) sc(SE) cmpMCC
## scale(SE) -0.011
## compMtCCmpt -0.737 0.010
## scl(SE):MCC 0.010 -0.739 0.007ggpredict(m, c("SE","compMotC")) %>% plot()
## Combined
3.3.2 For incongruent pairs, categorical choice: Desirable or not
m <- glmer( chooseDes ~ compMotC*condition + ( compMotC | subID), data = incongDf, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ compMotC * condition + (compMotC | subID)
## Data: incongDf
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 20165.8 20252.1 -10071.9 20143.8 18904
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.8757 -0.6826 0.4302 0.5940 1.5625
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.558217 0.74714
## compMotCCompete 0.003766 0.06137 -0.74
## Number of obs: 18915, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.436901 0.084284 17.048 < 2e-16 ***
## compMotCCompete -0.483955 0.094441 -5.124 2.98e-07 ***
## conditionAsian -0.329363 0.117958 -2.792 0.00524 **
## conditionDemocrat 0.029379 0.119863 0.245 0.80638
## conditionLatino -0.298950 0.127218 -2.350 0.01878 *
## compMotCCompete:conditionAsian 0.065793 0.117623 0.559 0.57592
## compMotCCompete:conditionDemocrat 0.001523 0.125824 0.012 0.99034
## compMotCCompete:conditionLatino 0.313386 0.135413 2.314 0.02065 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) cmpMCC cndtnA cndtnD cndtnL cMCC:A cMCC:D
## compMtCCmpt -0.271
## conditinAsn -0.713 0.191
## condtnDmcrt -0.701 0.189 0.501
## conditinLtn -0.661 0.176 0.472 0.464
## cmpMtCCmp:A 0.214 -0.720 -0.288 -0.149 -0.141
## cmpMtCCmp:D 0.200 -0.660 -0.142 -0.291 -0.131 0.523
## cmpMtCCmp:L 0.187 -0.645 -0.132 -0.130 -0.270 0.494 0.456ggpredict(m, c("compMotC","condition")) %>% plot()
m <- glmer( chooseDes ~ compMotC*conditionEC + ( compMotC | subID), data = incongDf, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ compMotC * conditionEC + (compMotC | subID)
## Data: incongDf
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 20165.8 20252.1 -10071.9 20143.8 18904
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.8757 -0.6826 0.4302 0.5940 1.5625
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.558219 0.74714
## compMotCCompete 0.003768 0.06138 -0.74
## Number of obs: 18915, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.28717 0.04371 29.451 < 2e-16 ***
## compMotCCompete -0.38878 0.05290 -7.349 1.99e-13 ***
## conditionECAsian -0.17963 0.07289 -2.465 0.0137 *
## conditionECDemocrat 0.17912 0.07443 2.407 0.0161 *
## conditionECLatino -0.14922 0.08030 -1.858 0.0631 .
## compMotCCompete:conditionECAsian -0.02938 0.07093 -0.414 0.6787
## compMotCCompete:conditionECDemocrat -0.09366 0.07795 -1.201 0.2296
## compMotCCompete:conditionECLatino 0.21821 0.08524 2.560 0.0105 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) cmpMCC cndECA cndECD cndECL cMCC:ECA cMCC:ECD
## compMtCCmpt -0.250
## condtnECAsn -0.059 0.021
## cndtnECDmcr -0.022 0.001 -0.302
## condtnECLtn 0.108 -0.022 -0.357 -0.369
## cmpMtCC:ECA 0.026 -0.142 -0.289 0.093 0.097
## cmpMtCC:ECD -0.005 0.022 0.086 -0.292 0.105 -0.276
## cmpMtCC:ECL -0.019 0.076 0.089 0.103 -0.266 -0.341 -0.396ggpredict(m, c("compMotC","conditionEC")) %>% plot()
3.3.2.1 Moderated by Self-Esteem?
m <- glmer( chooseDes ~ compMotC*condition*scale(SE) + ( compMotC | subID), data = incongDf, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)## boundary (singular) fit: see ?isSingularsummary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ compMotC * condition * scale(SE) + (compMotC | subID)
## Data: incongDf
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 19774.3 19923.2 -9868.2 19736.3 18646
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.5559 -0.6496 0.4348 0.5925 1.6675
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.388608 0.62338
## compMotCCompete 0.007218 0.08496 -1.00
## Number of obs: 18665, groups: subID, 388
##
## Fixed effects:
## Estimate Std. Error z value
## (Intercept) 1.39209 0.07684 18.117
## compMotCCompete -0.52106 0.09576 -5.441
## conditionAsian -0.10603 0.11121 -0.953
## conditionDemocrat -0.01058 0.10790 -0.098
## conditionLatino -0.27251 0.11336 -2.404
## scale(SE) 0.26672 0.07085 3.765
## compMotCCompete:conditionAsian 0.16311 0.13266 1.230
## compMotCCompete:conditionDemocrat 0.07388 0.12928 0.571
## compMotCCompete:conditionLatino 0.32180 0.13553 2.374
## compMotCCompete:scale(SE) -0.09807 0.08815 -1.113
## conditionAsian:scale(SE) 0.16947 0.10955 1.547
## conditionDemocrat:scale(SE) 0.34493 0.10495 3.287
## conditionLatino:scale(SE) -0.14794 0.12110 -1.222
## compMotCCompete:conditionAsian:scale(SE) 0.19480 0.12600 1.546
## compMotCCompete:conditionDemocrat:scale(SE) 0.10156 0.12712 0.799
## compMotCCompete:conditionLatino:scale(SE) 0.37604 0.14727 2.553
## Pr(>|z|)
## (Intercept) < 2e-16 ***
## compMotCCompete 5.29e-08 ***
## conditionAsian 0.340398
## conditionDemocrat 0.921904
## conditionLatino 0.016215 *
## scale(SE) 0.000167 ***
## compMotCCompete:conditionAsian 0.218883
## compMotCCompete:conditionDemocrat 0.567696
## compMotCCompete:conditionLatino 0.017583 *
## compMotCCompete:scale(SE) 0.265869
## conditionAsian:scale(SE) 0.121895
## conditionDemocrat:scale(SE) 0.001014 **
## conditionLatino:scale(SE) 0.221848
## compMotCCompete:conditionAsian:scale(SE) 0.122100
## compMotCCompete:conditionDemocrat:scale(SE) 0.424308
## compMotCCompete:conditionLatino:scale(SE) 0.010667 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1##
## Correlation matrix not shown by default, as p = 16 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need it## optimizer (bobyqa) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingularggpredict(m, c("SE","compMotC","condition")) %>% plot()## Data were 'prettified'. Consider using `terms="SE [all]"` to get smooth plots.
m <- glmer( chooseDes ~ compMotC*conditionEC*scale(SE) + ( compMotC | subID), data = incongDf, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)## boundary (singular) fit: see ?isSingularsummary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ compMotC * conditionEC * scale(SE) + (compMotC |
## subID)
## Data: incongDf
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 19774.3 19923.2 -9868.2 19736.3 18646
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.5559 -0.6496 0.4348 0.5925 1.6675
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.388607 0.62338
## compMotCCompete 0.007218 0.08496 -1.00
## Number of obs: 18665, groups: subID, 388
##
## Fixed effects:
## Estimate Std. Error z value
## (Intercept) 1.294808 0.039792 32.540
## compMotCCompete -0.381365 0.054391 -7.011
## conditionECAsian -0.008747 0.069383 -0.126
## conditionECDemocrat 0.086699 0.066721 1.299
## conditionECLatino -0.175233 0.071084 -2.465
## scale(SE) 0.358331 0.041664 8.601
## compMotCCompete:conditionECAsian 0.023412 0.084095 0.278
## compMotCCompete:conditionECDemocrat -0.065817 0.081072 -0.812
## compMotCCompete:conditionECLatino 0.182096 0.086021 2.117
## compMotCCompete:scale(SE) 0.070028 0.049369 1.418
## conditionECAsian:scale(SE) 0.077863 0.072228 1.078
## conditionECDemocrat:scale(SE) 0.253311 0.068742 3.685
## conditionECLatino:scale(SE) -0.239559 0.080956 -2.959
## compMotCCompete:conditionECAsian:scale(SE) 0.026692 0.080628 0.331
## compMotCCompete:conditionECDemocrat:scale(SE) -0.066530 0.081390 -0.817
## compMotCCompete:conditionECLatino:scale(SE) 0.207937 0.096922 2.145
## Pr(>|z|)
## (Intercept) < 2e-16 ***
## compMotCCompete 2.36e-12 ***
## conditionECAsian 0.899683
## conditionECDemocrat 0.193798
## conditionECLatino 0.013695 *
## scale(SE) < 2e-16 ***
## compMotCCompete:conditionECAsian 0.780709
## compMotCCompete:conditionECDemocrat 0.416890
## compMotCCompete:conditionECLatino 0.034270 *
## compMotCCompete:scale(SE) 0.156052
## conditionECAsian:scale(SE) 0.281029
## conditionECDemocrat:scale(SE) 0.000229 ***
## conditionECLatino:scale(SE) 0.003085 **
## compMotCCompete:conditionECAsian:scale(SE) 0.740603
## compMotCCompete:conditionECDemocrat:scale(SE) 0.413690
## compMotCCompete:conditionECLatino:scale(SE) 0.031921 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1##
## Correlation matrix not shown by default, as p = 16 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need it## optimizer (bobyqa) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingularggpredict(m, c("SE","compMotC","conditionEC")) %>% plot()## Data were 'prettified'. Consider using `terms="SE [all]"` to get smooth plots.
3.3.2.2 Change Across Trials?
m <- glmer( chooseDes ~ compMotC*condition*scale(trialTotal) + ( compMotC + scale(trialTotal) | subID), data = incongDf, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)## boundary (singular) fit: see ?isSingularsummary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ compMotC * condition * scale(trialTotal) + (compMotC +
## scale(trialTotal) | subID)
## Data: incongDf
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 20047.3 20220.0 -10001.7 20003.3 18893
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.2241 -0.6840 0.4214 0.5892 1.8764
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.57717 0.7597
## compMotCCompete 0.02840 0.1685 -0.51
## scale(trialTotal) 0.04697 0.2167 -0.01 -0.85
## Number of obs: 18915, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error z value
## (Intercept) 1.48410 0.08608 17.241
## compMotCCompete -0.53405 0.09811 -5.444
## conditionAsian -0.35439 0.12018 -2.949
## conditionDemocrat 0.03373 0.12240 0.276
## conditionLatino -0.37602 0.12962 -2.901
## scale(trialTotal) -0.19558 0.04719 -4.145
## compMotCCompete:conditionAsian 0.09089 0.12252 0.742
## compMotCCompete:conditionDemocrat -0.04092 0.13145 -0.311
## compMotCCompete:conditionLatino 0.38103 0.14221 2.679
## compMotCCompete:scale(trialTotal) 0.10990 0.08127 1.352
## conditionAsian:scale(trialTotal) 0.12927 0.06457 2.002
## conditionDemocrat:scale(trialTotal) -0.01897 0.06747 -0.281
## conditionLatino:scale(trialTotal) -0.11942 0.07621 -1.567
## compMotCCompete:conditionAsian:scale(trialTotal) -0.21415 0.10841 -1.975
## compMotCCompete:conditionDemocrat:scale(trialTotal) -0.20047 0.11490 -1.745
## compMotCCompete:conditionLatino:scale(trialTotal) 0.05621 0.13219 0.425
## Pr(>|z|)
## (Intercept) < 2e-16 ***
## compMotCCompete 5.22e-08 ***
## conditionAsian 0.00319 **
## conditionDemocrat 0.78290
## conditionLatino 0.00372 **
## scale(trialTotal) 3.40e-05 ***
## compMotCCompete:conditionAsian 0.45818
## compMotCCompete:conditionDemocrat 0.75558
## compMotCCompete:conditionLatino 0.00738 **
## compMotCCompete:scale(trialTotal) 0.17630
## conditionAsian:scale(trialTotal) 0.04528 *
## conditionDemocrat:scale(trialTotal) 0.77858
## conditionLatino:scale(trialTotal) 0.11709
## compMotCCompete:conditionAsian:scale(trialTotal) 0.04821 *
## compMotCCompete:conditionDemocrat:scale(trialTotal) 0.08102 .
## compMotCCompete:conditionLatino:scale(trialTotal) 0.67066
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1##
## Correlation matrix not shown by default, as p = 16 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need it## optimizer (bobyqa) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingularggpredict(m, c("trialTotal","compMotC","condition")) %>% plot()## Data were 'prettified'. Consider using `terms="trialTotal [all]"` to get smooth plots.
m <- glmer( chooseDes ~ compMotC*conditionEC*scale(trialTotal) + ( compMotC + scale(trialTotal) | subID), data = incongDf, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)## boundary (singular) fit: see ?isSingularsummary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ compMotC * conditionEC * scale(trialTotal) + (compMotC +
## scale(trialTotal) | subID)
## Data: incongDf
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 20047.3 20220.0 -10001.7 20003.3 18893
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.2241 -0.6840 0.4214 0.5892 1.8764
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.57717 0.7597
## compMotCCompete 0.02840 0.1685 -0.51
## scale(trialTotal) 0.04697 0.2167 -0.01 -0.85
## Number of obs: 18915, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error
## (Intercept) 1.30993 0.04459
## compMotCCompete -0.42630 0.05546
## conditionECAsian -0.18022 0.07415
## conditionECDemocrat 0.20790 0.07595
## conditionECLatino -0.20185 0.08171
## scale(trialTotal) -0.19786 0.02570
## compMotCCompete:conditionECAsian -0.01686 0.07395
## compMotCCompete:conditionECDemocrat -0.14867 0.08146
## compMotCCompete:conditionECLatino 0.27328 0.08981
## compMotCCompete:scale(trialTotal) 0.02030 0.04481
## conditionECAsian:scale(trialTotal) 0.13156 0.04027
## conditionECDemocrat:scale(trialTotal) -0.01669 0.04258
## conditionECLatino:scale(trialTotal) -0.11715 0.04942
## compMotCCompete:conditionECAsian:scale(trialTotal) -0.12455 0.06724
## compMotCCompete:conditionECDemocrat:scale(trialTotal) -0.11087 0.07244
## compMotCCompete:conditionECLatino:scale(trialTotal) 0.14581 0.08588
## z value Pr(>|z|)
## (Intercept) 29.379 < 2e-16 ***
## compMotCCompete -7.686 1.52e-14 ***
## conditionECAsian -2.430 0.01508 *
## conditionECDemocrat 2.737 0.00619 **
## conditionECLatino -2.470 0.01350 *
## scale(trialTotal) -7.699 1.37e-14 ***
## compMotCCompete:conditionECAsian -0.228 0.81967
## compMotCCompete:conditionECDemocrat -1.825 0.06799 .
## compMotCCompete:conditionECLatino 3.043 0.00234 **
## compMotCCompete:scale(trialTotal) 0.453 0.65064
## conditionECAsian:scale(trialTotal) 3.267 0.00109 **
## conditionECDemocrat:scale(trialTotal) -0.392 0.69510
## conditionECLatino:scale(trialTotal) -2.371 0.01776 *
## compMotCCompete:conditionECAsian:scale(trialTotal) -1.852 0.06395 .
## compMotCCompete:conditionECDemocrat:scale(trialTotal) -1.531 0.12586
## compMotCCompete:conditionECLatino:scale(trialTotal) 1.698 0.08954 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1##
## Correlation matrix not shown by default, as p = 16 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need it## optimizer (bobyqa) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingularggpredict(m, c("trialTotal","compMotC","conditionEC")) %>% plot()## Data were 'prettified'. Consider using `terms="trialTotal [all]"` to get smooth plots.
3.3.2.3 For ingroup vs. outgroup
m <- glmer( chooseDes ~ compMotC*ingroup + ( compMotC | subID), data = incongDf, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ compMotC * ingroup + (compMotC | subID)
## Data: incongDf
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 20172.8 20227.8 -10079.4 20158.8 18908
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.8971 -0.6916 0.4280 0.5987 1.5777
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.583998 0.76420
## compMotCCompete 0.006461 0.08038 -0.60
## Number of obs: 18915, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.24281 0.05142 24.169 < 2e-16 ***
## compMotCCompete -0.37441 0.05784 -6.473 9.59e-11 ***
## ingroupOut 0.19632 0.09983 1.967 0.0492 *
## compMotCCompete:ingroupOut -0.10645 0.10227 -1.041 0.2980
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) cmpMCC ingrpO
## compMtCCmpt -0.257
## ingroupOut -0.511 0.125
## cmpMtCCmp:O 0.139 -0.410 -0.283ggpredict(m, c("compMotC","ingroup")) %>% plot()
3.3.2.4 For ingroup vs. outgroup moderated by self-esteem
m <- glmer( chooseDes ~ compMotC*ingroup*scale(SE) + ( compMotC | subID), data = incongDf, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)## boundary (singular) fit: see ?isSingularsummary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseDes ~ compMotC * ingroup * scale(SE) + (compMotC | subID)
## Data: incongDf
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 19781.4 19867.6 -9879.7 19759.4 18654
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.1068 -0.6473 0.4318 0.5948 1.6922
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.42108 0.6489
## compMotCCompete 0.01173 0.1083 -1.00
## Number of obs: 18665, groups: subID, 388
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.27826 0.04588 27.860 < 2e-16 ***
## compMotCCompete -0.34724 0.05926 -5.860 4.63e-09 ***
## ingroupOut 0.11762 0.09128 1.289 0.1975
## scale(SE) 0.43461 0.04866 8.931 < 2e-16 ***
## compMotCCompete:ingroupOut -0.18131 0.10524 -1.723 0.0849 .
## compMotCCompete:scale(SE) 0.10852 0.05443 1.994 0.0462 *
## ingroupOut:scale(SE) -0.16676 0.08762 -1.903 0.0570 .
## compMotCCompete:ingroupOut:scale(SE) -0.20806 0.10311 -2.018 0.0436 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) cmpMCC ingrpO sc(SE) cmMCC:O cMCC:( iO:(SE
## compMtCCmpt -0.329
## ingroupOut -0.498 0.157
## scale(SE) 0.118 -0.054 -0.055
## cmpMtCCmp:O 0.178 -0.438 -0.363 0.020
## cmpMCC:(SE) -0.046 0.365 0.018 -0.398 -0.173
## ingrpO:(SE) -0.064 0.030 -0.096 -0.554 0.025 0.220
## cMCC:O:(SE) 0.023 -0.156 0.023 0.208 0.041 -0.519 -0.357
## optimizer (bobyqa) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingularggpredict(m, c("SE","compMotC","ingroup")) %>% plot()## Data were 'prettified'. Consider using `terms="SE [all]"` to get smooth plots.
3.3.3 For all data, continuous desirability probability
m <- lmer( desAP ~ compMotC*condition + ( compMotC | subID), data = fullTD)## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
## Model failed to converge with max|grad| = 0.00384047 (tol = 0.002, component 1)summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: desAP ~ compMotC * condition + (compMotC | subID)
## Data: fullTD
##
## REML criterion at convergence: -16695.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.89156 -0.53279 -0.08542 0.76455 2.26276
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.002294 0.04790
## compMotCCompete 0.001359 0.03686 -0.89
## Residual 0.042915 0.20716
## Number of obs: 56607, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error df t value
## (Intercept) 0.607491 0.005129 374.554355 118.448
## compMotCCompete -0.081358 0.005022 385.448522 -16.201
## conditionAsian -0.017761 0.007251 374.161225 -2.450
## conditionDemocrat 0.002738 0.007291 375.048226 0.376
## conditionLatino -0.015957 0.007840 382.531595 -2.035
## compMotCCompete:conditionAsian 0.007307 0.007069 378.820935 1.034
## compMotCCompete:conditionDemocrat -0.001445 0.007130 384.338971 -0.203
## compMotCCompete:conditionLatino 0.023173 0.007792 418.487586 2.974
## Pr(>|t|)
## (Intercept) < 2e-16 ***
## compMotCCompete < 2e-16 ***
## conditionAsian 0.01476 *
## conditionDemocrat 0.70748
## conditionLatino 0.04250 *
## compMotCCompete:conditionAsian 0.30194
## compMotCCompete:conditionDemocrat 0.83954
## compMotCCompete:conditionLatino 0.00311 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) cmpMCC cndtnA cndtnD cndtnL cMCC:A cMCC:D
## compMtCCmpt -0.753
## conditinAsn -0.707 0.533
## condtnDmcrt -0.703 0.530 0.498
## conditinLtn -0.654 0.493 0.463 0.460
## cmpMtCCmp:A 0.535 -0.710 -0.756 -0.377 -0.350
## cmpMtCCmp:D 0.531 -0.704 -0.375 -0.755 -0.347 0.500
## cmpMtCCmp:L 0.486 -0.644 -0.343 -0.342 -0.745 0.458 0.454
## optimizer (nloptwrap) convergence code: 0 (OK)
## Model failed to converge with max|grad| = 0.00384047 (tol = 0.002, component 1)ggpredict(m, c("compMotC","condition")) %>% plot()
m <- lmer( desAP ~ compMotC*conditionEC + ( compMotC | subID), data = fullTD)## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
## Model failed to converge with max|grad| = 0.00384046 (tol = 0.002, component 1)summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: desAP ~ compMotC * conditionEC + (compMotC | subID)
## Data: fullTD
##
## REML criterion at convergence: -16689.8
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.89156 -0.53279 -0.08542 0.76455 2.26276
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.002294 0.04790
## compMotCCompete 0.001359 0.03686 -0.89
## Residual 0.042915 0.20716
## Number of obs: 56607, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error df t value
## (Intercept) 5.997e-01 2.676e-03 3.789e+02 224.110
## compMotCCompete -7.410e-02 2.635e-03 3.993e+02 -28.123
## conditionECAsian -1.002e-02 4.505e-03 3.756e+02 -2.223
## conditionECDemocrat 1.048e-02 4.537e-03 3.767e+02 2.310
## conditionECLatino -8.212e-03 4.974e-03 3.858e+02 -1.651
## compMotCCompete:conditionECAsian 4.785e-05 4.395e-03 3.816e+02 0.011
## compMotCCompete:conditionECDemocrat -8.703e-03 4.444e-03 3.888e+02 -1.958
## compMotCCompete:conditionECLatino 1.591e-02 4.969e-03 4.311e+02 3.203
## Pr(>|t|)
## (Intercept) < 2e-16 ***
## compMotCCompete < 2e-16 ***
## conditionECAsian 0.02679 *
## conditionECDemocrat 0.02141 *
## conditionECLatino 0.09955 .
## compMotCCompete:conditionECAsian 0.99132
## compMotCCompete:conditionECDemocrat 0.05089 .
## compMotCCompete:conditionECLatino 0.00146 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) cmpMCC cndECA cndECD cndECL cMCC:ECA cMCC:ECD
## compMtCCmpt -0.751
## condtnECAsn -0.049 0.038
## cndtnECDmcr -0.037 0.028 -0.299
## condtnECLtn 0.122 -0.094 -0.366 -0.370
## cmpMtCC:ECA 0.038 -0.065 -0.756 0.226 0.278
## cmpMtCC:ECD 0.028 -0.046 0.225 -0.754 0.280 -0.289
## cmpMtCC:ECL -0.092 0.148 0.271 0.274 -0.742 -0.372 -0.377
## optimizer (nloptwrap) convergence code: 0 (OK)
## Model failed to converge with max|grad| = 0.00384046 (tol = 0.002, component 1)ggpredict(m, c("compMotC","conditionEC")) %>% plot()
3.3.3.1 Moderated by Self-Esteem?
m <- lmer( desAP ~ compMotC*condition*scale(SE) + ( compMotC | subID), data = fullTD)
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: desAP ~ compMotC * condition * scale(SE) + (compMotC | subID)
## Data: fullTD
##
## REML criterion at convergence: -16547.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.88227 -0.53690 -0.08302 0.76528 2.27160
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.001670 0.04086
## compMotCCompete 0.001222 0.03496 -0.91
## Residual 0.042910 0.20715
## Number of obs: 55858, groups: subID, 388
##
## Fixed effects:
## Estimate Std. Error df
## (Intercept) 6.052e-01 4.735e-03 3.641e+02
## compMotCCompete -7.859e-02 5.132e-03 3.787e+02
## conditionAsian -4.245e-03 6.855e-03 3.614e+02
## conditionDemocrat -5.852e-04 6.622e-03 3.674e+02
## conditionLatino -1.407e-02 7.066e-03 3.746e+02
## scale(SE) 1.747e-02 4.326e-03 3.611e+02
## compMotCCompete:conditionAsian 8.875e-04 7.433e-03 3.767e+02
## compMotCCompete:conditionDemocrat -1.317e-03 7.149e-03 3.769e+02
## compMotCCompete:conditionLatino 1.957e-02 7.764e-03 4.122e+02
## compMotCCompete:scale(SE) -1.522e-02 4.735e-03 3.898e+02
## conditionAsian:scale(SE) 9.844e-03 6.689e-03 3.623e+02
## conditionDemocrat:scale(SE) 1.645e-02 6.259e-03 3.691e+02
## conditionLatino:scale(SE) -1.094e-02 7.508e-03 3.743e+02
## compMotCCompete:conditionAsian:scale(SE) 5.905e-03 7.261e-03 3.790e+02
## compMotCCompete:conditionDemocrat:scale(SE) 1.784e-03 6.768e-03 3.804e+02
## compMotCCompete:conditionLatino:scale(SE) 2.624e-02 8.317e-03 4.250e+02
## t value Pr(>|t|)
## (Intercept) 127.817 < 2e-16 ***
## compMotCCompete -15.314 < 2e-16 ***
## conditionAsian -0.619 0.53619
## conditionDemocrat -0.088 0.92962
## conditionLatino -1.991 0.04721 *
## scale(SE) 4.038 6.58e-05 ***
## compMotCCompete:conditionAsian 0.119 0.90502
## compMotCCompete:conditionDemocrat -0.184 0.85391
## compMotCCompete:conditionLatino 2.521 0.01208 *
## compMotCCompete:scale(SE) -3.214 0.00142 **
## conditionAsian:scale(SE) 1.472 0.14197
## conditionDemocrat:scale(SE) 2.628 0.00896 **
## conditionLatino:scale(SE) -1.457 0.14602
## compMotCCompete:conditionAsian:scale(SE) 0.813 0.41658
## compMotCCompete:conditionDemocrat:scale(SE) 0.264 0.79221
## compMotCCompete:conditionLatino:scale(SE) 3.155 0.00172 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1##
## Correlation matrix not shown by default, as p = 16 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need itggpredict(m, c("SE","compMotC","condition")) %>% plot()
m <- lmer( desAP ~ compMotC*conditionEC*scale(SE) + ( compMotC | subID), data = fullTD)
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: desAP ~ compMotC * conditionEC * scale(SE) + (compMotC | subID)
## Data: fullTD
##
## REML criterion at convergence: -16536.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.88227 -0.53690 -0.08302 0.76528 2.27160
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.001670 0.04086
## compMotCCompete 0.001222 0.03496 -0.91
## Residual 0.042910 0.20715
## Number of obs: 55858, groups: subID, 388
##
## Fixed effects:
## Estimate Std. Error df
## (Intercept) 6.005e-01 2.449e-03 3.697e+02
## compMotCCompete -7.380e-02 2.669e-03 3.940e+02
## conditionECAsian 4.797e-04 4.276e-03 3.625e+02
## conditionECDemocrat 4.139e-03 4.088e-03 3.705e+02
## conditionECLatino -9.343e-03 4.444e-03 3.793e+02
## scale(SE) 2.131e-02 2.536e-03 3.719e+02
## compMotCCompete:conditionECAsian -3.898e-03 4.645e-03 3.811e+02
## compMotCCompete:conditionECDemocrat -6.103e-03 4.417e-03 3.818e+02
## compMotCCompete:conditionECLatino 1.479e-02 4.909e-03 4.265e+02
## compMotCCompete:scale(SE) -6.734e-03 2.771e-03 4.003e+02
## conditionECAsian:scale(SE) 6.006e-03 4.410e-03 3.659e+02
## conditionECDemocrat:scale(SE) 1.261e-02 4.082e-03 3.747e+02
## conditionECLatino:scale(SE) -1.478e-02 5.026e-03 3.788e+02
## compMotCCompete:conditionECAsian:scale(SE) -2.578e-03 4.778e-03 3.807e+02
## compMotCCompete:conditionECDemocrat:scale(SE) -6.698e-03 4.402e-03 3.826e+02
## compMotCCompete:conditionECLatino:scale(SE) 1.776e-02 5.573e-03 4.322e+02
## t value Pr(>|t|)
## (Intercept) 245.248 < 2e-16 ***
## compMotCCompete -27.653 < 2e-16 ***
## conditionECAsian 0.112 0.91075
## conditionECDemocrat 1.013 0.31191
## conditionECLatino -2.103 0.03616 *
## scale(SE) 8.403 9.35e-16 ***
## compMotCCompete:conditionECAsian -0.839 0.40187
## compMotCCompete:conditionECDemocrat -1.382 0.16785
## compMotCCompete:conditionECLatino 3.012 0.00275 **
## compMotCCompete:scale(SE) -2.430 0.01553 *
## conditionECAsian:scale(SE) 1.362 0.17405
## conditionECDemocrat:scale(SE) 3.089 0.00216 **
## conditionECLatino:scale(SE) -2.940 0.00348 **
## compMotCCompete:conditionECAsian:scale(SE) -0.539 0.58988
## compMotCCompete:conditionECDemocrat:scale(SE) -1.522 0.12887
## compMotCCompete:conditionECLatino:scale(SE) 3.187 0.00154 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1##
## Correlation matrix not shown by default, as p = 16 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need itggpredict(m, c("SE","compMotC","conditionEC")) %>% plot()
3.3.3.2 Change Across Trials?
m <- lmer( desAP ~ compMotC*condition*scale(trialTotal) + ( compMotC + scale(trialTotal) | subID), data = fullTD)
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: desAP ~ compMotC * condition * scale(trialTotal) + (compMotC +
## scale(trialTotal) | subID)
## Data: fullTD
##
## REML criterion at convergence: -16763.1
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.92158 -0.53576 -0.08235 0.76376 2.33287
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.0022919 0.047873
## compMotCCompete 0.0013363 0.036556 -0.89
## scale(trialTotal) 0.0000935 0.009669 0.05 -0.03
## Residual 0.0427328 0.206719
## Number of obs: 56607, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error
## (Intercept) 6.082e-01 5.129e-03
## compMotCCompete -8.223e-02 5.004e-03
## conditionAsian -1.829e-02 7.251e-03
## conditionDemocrat 2.698e-03 7.291e-03
## conditionLatino -2.006e-02 7.857e-03
## scale(trialTotal) -9.003e-03 2.302e-03
## compMotCCompete:conditionAsian 7.792e-03 7.041e-03
## compMotCCompete:conditionDemocrat -1.905e-03 7.103e-03
## compMotCCompete:conditionLatino 2.535e-02 7.812e-03
## compMotCCompete:scale(trialTotal) 1.723e-03 3.395e-03
## conditionAsian:scale(trialTotal) 6.795e-03 3.245e-03
## conditionDemocrat:scale(trialTotal) 5.588e-04 3.281e-03
## conditionLatino:scale(trialTotal) -8.494e-03 3.907e-03
## compMotCCompete:conditionAsian:scale(trialTotal) -9.794e-03 4.763e-03
## compMotCCompete:conditionDemocrat:scale(trialTotal) -8.104e-03 4.809e-03
## compMotCCompete:conditionLatino:scale(trialTotal) 5.012e-03 5.825e-03
## df t value Pr(>|t|)
## (Intercept) 3.750e+02 118.583 < 2e-16
## compMotCCompete 3.856e+02 -16.434 < 2e-16
## conditionAsian 3.744e+02 -2.523 0.01204
## conditionDemocrat 3.754e+02 0.370 0.71157
## conditionLatino 3.862e+02 -2.553 0.01105
## scale(trialTotal) 7.365e+02 -3.911 0.00010
## compMotCCompete:conditionAsian 3.786e+02 1.107 0.26918
## compMotCCompete:conditionDemocrat 3.843e+02 -0.268 0.78867
## compMotCCompete:conditionLatino 4.289e+02 3.245 0.00126
## compMotCCompete:scale(trialTotal) 5.185e+04 0.508 0.61175
## conditionAsian:scale(trialTotal) 7.292e+02 2.094 0.03658
## conditionDemocrat:scale(trialTotal) 7.440e+02 0.170 0.86481
## conditionLatino:scale(trialTotal) 1.139e+03 -2.174 0.02989
## compMotCCompete:conditionAsian:scale(trialTotal) 5.272e+04 -2.056 0.03975
## compMotCCompete:conditionDemocrat:scale(trialTotal) 5.190e+04 -1.685 0.09194
## compMotCCompete:conditionLatino:scale(trialTotal) 5.520e+04 0.860 0.38954
##
## (Intercept) ***
## compMotCCompete ***
## conditionAsian *
## conditionDemocrat
## conditionLatino *
## scale(trialTotal) ***
## compMotCCompete:conditionAsian
## compMotCCompete:conditionDemocrat
## compMotCCompete:conditionLatino **
## compMotCCompete:scale(trialTotal)
## conditionAsian:scale(trialTotal) *
## conditionDemocrat:scale(trialTotal)
## conditionLatino:scale(trialTotal) *
## compMotCCompete:conditionAsian:scale(trialTotal) *
## compMotCCompete:conditionDemocrat:scale(trialTotal) .
## compMotCCompete:conditionLatino:scale(trialTotal)
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1##
## Correlation matrix not shown by default, as p = 16 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need itggpredict(m, c("trialTotal","compMotC","condition")) %>% plot()
m <- lmer( desAP ~ compMotC*conditionEC*scale(trialTotal) + ( compMotC + scale(trialTotal) | subID), data = fullTD)
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: desAP ~ compMotC * conditionEC * scale(trialTotal) + (compMotC +
## scale(trialTotal) | subID)
## Data: fullTD
##
## REML criterion at convergence: -16752
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.92158 -0.53576 -0.08235 0.76376 2.33287
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.0022919 0.047873
## compMotCCompete 0.0013363 0.036556 -0.89
## scale(trialTotal) 0.0000935 0.009669 0.05 -0.03
## Residual 0.0427328 0.206719
## Number of obs: 56607, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error
## (Intercept) 5.993e-01 2.679e-03
## compMotCCompete -7.442e-02 2.634e-03
## conditionECAsian -9.380e-03 4.507e-03
## conditionECDemocrat 1.161e-02 4.539e-03
## conditionECLatino -1.115e-02 4.989e-03
## scale(trialTotal) -9.288e-03 1.274e-03
## compMotCCompete:conditionECAsian -1.816e-05 4.383e-03
## compMotCCompete:conditionECDemocrat -9.715e-03 4.432e-03
## compMotCCompete:conditionECLatino 1.754e-02 4.993e-03
## compMotCCompete:scale(trialTotal) -1.498e-03 1.882e-03
## conditionECAsian:scale(trialTotal) 7.080e-03 2.058e-03
## conditionECDemocrat:scale(trialTotal) 8.438e-04 2.087e-03
## conditionECLatino:scale(trialTotal) -8.209e-03 2.570e-03
## compMotCCompete:conditionECAsian:scale(trialTotal) -6.573e-03 3.020e-03
## compMotCCompete:conditionECDemocrat:scale(trialTotal) -4.882e-03 3.056e-03
## compMotCCompete:conditionECLatino:scale(trialTotal) 8.234e-03 3.840e-03
## df t value
## (Intercept) 3.809e+02 223.684
## compMotCCompete 4.045e+02 -28.254
## conditionECAsian 3.764e+02 -2.081
## conditionECDemocrat 3.776e+02 2.558
## conditionECLatino 3.908e+02 -2.234
## scale(trialTotal) 9.409e+02 -7.292
## compMotCCompete:conditionECAsian 3.831e+02 -0.004
## compMotCCompete:conditionECDemocrat 3.904e+02 -2.192
## compMotCCompete:conditionECLatino 4.458e+02 3.513
## compMotCCompete:scale(trialTotal) 5.438e+04 -0.796
## conditionECAsian:scale(trialTotal) 7.957e+02 3.440
## conditionECDemocrat:scale(trialTotal) 8.147e+02 0.404
## conditionECLatino:scale(trialTotal) 1.325e+03 -3.195
## compMotCCompete:conditionECAsian:scale(trialTotal) 5.387e+04 -2.176
## compMotCCompete:conditionECDemocrat:scale(trialTotal) 5.296e+04 -1.598
## compMotCCompete:conditionECLatino:scale(trialTotal) 5.579e+04 2.144
## Pr(>|t|)
## (Intercept) < 2e-16 ***
## compMotCCompete < 2e-16 ***
## conditionECAsian 0.038080 *
## conditionECDemocrat 0.010906 *
## conditionECLatino 0.026021 *
## scale(trialTotal) 6.48e-13 ***
## compMotCCompete:conditionECAsian 0.996697
## compMotCCompete:conditionECDemocrat 0.028983 *
## compMotCCompete:conditionECLatino 0.000488 ***
## compMotCCompete:scale(trialTotal) 0.426097
## conditionECAsian:scale(trialTotal) 0.000613 ***
## conditionECDemocrat:scale(trialTotal) 0.686109
## conditionECLatino:scale(trialTotal) 0.001434 **
## compMotCCompete:conditionECAsian:scale(trialTotal) 0.029533 *
## compMotCCompete:conditionECDemocrat:scale(trialTotal) 0.110138
## compMotCCompete:conditionECLatino:scale(trialTotal) 0.032016 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1##
## Correlation matrix not shown by default, as p = 16 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need itggpredict(m, c("trialTotal","compMotC","conditionEC")) %>% plot()
3.3.3.3 For ingroup vs. outgroup
m <- lmer( desAP ~ compMotC*ingroup + ( compMotC | subID), data = fullTD)
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: desAP ~ compMotC * ingroup + (compMotC | subID)
## Data: fullTD
##
## REML criterion at convergence: -16709
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.89238 -0.53395 -0.08627 0.76514 2.26903
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.002352 0.04849
## compMotCCompete 0.001402 0.03744 -0.88
## Residual 0.042916 0.20716
## Number of obs: 56607, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.597601 0.003137 381.826646 190.51 <2e-16
## compMotCCompete -0.073209 0.003072 393.620076 -23.83 <2e-16
## ingroupOut 0.009876 0.006058 379.038647 1.63 0.104
## compMotCCompete:ingroupOut -0.008115 0.005922 388.185923 -1.37 0.171
##
## (Intercept) ***
## compMotCCompete ***
## ingroupOut
## compMotCCompete:ingroupOut
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) cmpMCC ingrpO
## compMtCCmpt -0.752
## ingroupOut -0.518 0.389
## cmpMtCCmp:O 0.390 -0.519 -0.753ggpredict(m, c("compMotC","ingroup")) %>% plot()
3.3.3.4 For ingroup vs. outgroup moderated by self-esteem
m <- lmer( desAP ~ compMotC*ingroup*scale(SE) + ( compMotC | subID), data = fullTD)## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
## Model failed to converge with max|grad| = 0.00359685 (tol = 0.002, component 1)summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: desAP ~ compMotC * ingroup * scale(SE) + (compMotC | subID)
## Data: fullTD
##
## REML criterion at convergence: -16592.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.88433 -0.53668 -0.08324 0.76758 2.26546
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.001751 0.04184
## compMotCCompete 0.001315 0.03627 -0.91
## Residual 0.042910 0.20715
## Number of obs: 55858, groups: subID, 388
##
## Fixed effects:
## Estimate Std. Error df t value
## (Intercept) 0.599554 0.002791 373.491412 214.816
## compMotCCompete -0.073311 0.003035 391.441320 -24.153
## ingroupOut 0.005652 0.005574 369.584975 1.014
## scale(SE) 0.025485 0.002905 375.841213 8.771
## compMotCCompete:ingroupOut -0.005296 0.006044 383.104380 -0.876
## compMotCCompete:scale(SE) -0.006007 0.003142 385.159310 -1.912
## ingroupOut:scale(SE) -0.008026 0.005279 368.547175 -1.520
## compMotCCompete:ingroupOut:scale(SE) -0.009223 0.005755 389.328502 -1.603
## Pr(>|t|)
## (Intercept) <2e-16 ***
## compMotCCompete <2e-16 ***
## ingroupOut 0.3112
## scale(SE) <2e-16 ***
## compMotCCompete:ingroupOut 0.3815
## compMotCCompete:scale(SE) 0.0567 .
## ingroupOut:scale(SE) 0.1293
## compMotCCompete:ingroupOut:scale(SE) 0.1098
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) cmpMCC ingrpO sc(SE) cmMCC:O cMCC:( iO:(SE
## compMtCCmpt -0.771
## ingroupOut -0.501 0.386
## scale(SE) 0.071 -0.053 -0.035
## cmpMtCCmp:O 0.387 -0.502 -0.772 0.027
## cmpMCC:(SE) -0.053 0.095 0.027 -0.776 -0.048
## ingrpO:(SE) -0.039 0.029 -0.139 -0.550 0.107 0.427
## cMCC:O:(SE) 0.029 -0.052 0.107 0.424 -0.130 -0.546 -0.768
## optimizer (nloptwrap) convergence code: 0 (OK)
## Model failed to converge with max|grad| = 0.00359685 (tol = 0.002, component 1)ggpredict(m, c("SE","compMotC","ingroup")) %>% plot()
4 Question 4: On congruent pairs (Pos-Pos, Neg-Neg), are people more likely to choose the group-consistent choice than not?
4.1 Study 1
m <- glmer( as.factor(chooseGroup) ~ 1 + pair + ( pair | subID), data = congDf1, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: as.factor(chooseGroup) ~ 1 + pair + (pair | subID)
## Data: congDf1
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 9376.7 9410.9 -4683.4 9366.7 6842
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.5199 -0.9656 0.6807 0.9469 1.6165
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.2159 0.4646
## pairPP 0.3654 0.6045 -0.80
## Number of obs: 6847, groups: subID, 80
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.02074 0.06218 -0.334 0.739
## pairPP 0.12843 0.08383 1.532 0.126
##
## Correlation of Fixed Effects:
## (Intr)
## pairPP -0.7614.2 Study 2
m <- glmer( as.factor(chooseGroup) ~ 1 + pair + ( pair | subID), data = congDf2, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: as.factor(chooseGroup) ~ 1 + pair + (pair | subID)
## Data: congDf2
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 14294.9 14331.2 -7142.4 14284.9 10495
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.9153 -0.9505 0.5874 0.9496 1.4898
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.2597 0.5096
## pairPP 0.3808 0.6171 -0.74
## Number of obs: 10500, groups: subID, 105
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.06272 0.05732 1.094 0.274
## pairPP -0.07821 0.07235 -1.081 0.280
##
## Correlation of Fixed Effects:
## (Intr)
## pairPP -0.7274.3 Study 3
m <- glmer( as.factor(chooseGroup) ~ 1 + pair + ( pair | subID), data = congDf3, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: as.factor(chooseGroup) ~ 1 + pair + (pair | subID)
## Data: congDf3
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 27862.3 27901.9 -13926.1 27852.3 20340
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.5842 -0.9608 0.6312 0.9491 1.5899
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.1784 0.4224
## pairPP 0.3492 0.5909 -0.74
## Number of obs: 20345, groups: subID, 208
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.005711 0.035756 -0.160 0.873
## pairPP 0.013960 0.050038 0.279 0.780
##
## Correlation of Fixed Effects:
## (Intr)
## pairPP -0.7334.4 Combined
m <- glmer( chooseGroup ~ condition + pair + ( pair | subID), data = congDf, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseGroup ~ condition + pair + (pair | subID)
## Data: congDf
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 51526.5 51594.8 -25755.3 51510.5 37684
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.8146 -0.9536 0.6323 0.9480 1.6233
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.2086 0.4567
## pairPP 0.3662 0.6051 -0.75
## Number of obs: 37692, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.03165 0.04117 -0.769 0.442
## conditionAsian 0.04472 0.05052 0.885 0.376
## conditionDemocrat 0.05457 0.05079 1.074 0.283
## conditionLatino 0.07343 0.05519 1.331 0.183
## pairPP 0.01183 0.03714 0.318 0.750
##
## Correlation of Fixed Effects:
## (Intr) cndtnA cndtnD cndtnL
## conditinAsn -0.616
## condtnDmcrt -0.612 0.499
## conditinLtn -0.567 0.456 0.455
## pairPP -0.496 0.002 0.002 0.011ggpredict(m, c("condition")) %>% plot()
m <- glmer( chooseGroup ~ conditionEC + pair + ( pair | subID), data = congDf, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseGroup ~ conditionEC + pair + (pair | subID)
## Data: congDf
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 51526.5 51594.8 -25755.3 51510.5 37684
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.8146 -0.9536 0.6323 0.9480 1.6233
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.2086 0.4567
## pairPP 0.3661 0.6051 -0.75
## Number of obs: 37692, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.011529 0.027569 0.418 0.676
## conditionECAsian 0.001546 0.031466 0.049 0.961
## conditionECDemocrat 0.011393 0.031642 0.360 0.719
## conditionECLatino 0.030240 0.035190 0.859 0.390
## pairPP 0.011826 0.037124 0.319 0.750
##
## Correlation of Fixed Effects:
## (Intr) cndECA cndECD cndECL
## condtnECAsn -0.037
## cndtnECDmcr -0.026 -0.292
## condtnECLtn 0.085 -0.373 -0.374
## pairPP -0.734 -0.003 -0.004 0.011ggpredict(m, c("conditionEC")) %>% plot()
4.4.1 Ingroup conditions compared to outgroup across trials?
m <- glmer( chooseGroup ~ ingroup*scale(trialTotal)*scale(SE) + pair + ( pair | subID), data = congDf, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseGroup ~ ingroup * scale(trialTotal) * scale(SE) + pair +
## (pair | subID)
## Data: congDf
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 50836.5 50938.8 -25406.3 50812.5 37181
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.8278 -0.9569 0.6296 0.9488 1.6344
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.2071 0.4551
## pairPP 0.3698 0.6081 -0.75
## Number of obs: 37193, groups: subID, 388
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.024560 0.029729 0.826 0.40872
## ingroupOut -0.048502 0.043374 -1.118 0.26347
## scale(trialTotal) -0.010882 0.012589 -0.864 0.38738
## scale(SE) -0.024412 0.022580 -1.081 0.27965
## pairPP 0.007266 0.037509 0.194 0.84639
## ingroupOut:scale(trialTotal) 0.005204 0.024423 0.213 0.83127
## ingroupOut:scale(SE) 0.034441 0.041180 0.836 0.40295
## scale(trialTotal):scale(SE) -0.017461 0.012988 -1.344 0.17881
## ingroupOut:scale(trialTotal):scale(SE) 0.072250 0.023203 3.114 0.00185 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) ingrpO scl(T) sc(SE) pairPP inO:(T) iO:(SE s(T):(
## ingroupOut -0.364
## scl(trlTtl) 0.005 -0.007
## scale(SE) 0.058 -0.040 0.005
## pairPP -0.681 -0.005 0.009 0.000
## ingrpOt:(T) -0.002 -0.015 -0.515 -0.002 -0.005
## ingrpO:(SE) -0.030 -0.135 -0.003 -0.547 -0.002 0.008
## scl(T):(SE) 0.002 -0.003 0.091 0.003 0.002 -0.047 -0.001
## iO:(T):(SE) 0.000 0.008 -0.051 -0.002 -0.003 -0.120 -0.019 -0.560ggpredict(m, c("trialTotal","ingroup","SE")) %>% plot()## Data were 'prettified'. Consider using `terms="trialTotal [all]"` to get smooth plots.
m <- glmer( chooseGroup ~ scale(inDiff) + scale(outDiff) + scale(similarity) + pair + ( scale(inDiff) + scale(outDiff) | subID), data = congDf, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: chooseGroup ~ scale(inDiff) + scale(outDiff) + scale(similarity) +
## pair + (scale(inDiff) + scale(outDiff) | subID)
## Data: congDf
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 51497.3 51591.2 -25737.6 51475.3 37681
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.6318 -0.9661 0.5588 0.9548 2.0212
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.09166 0.3028
## scale(inDiff) 0.04744 0.2178 0.17
## scale(outDiff) 0.04652 0.2157 0.11 0.88
## Number of obs: 37692, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.012801 0.021425 0.597 0.5502
## scale(inDiff) 0.028292 0.015531 1.822 0.0685 .
## scale(outDiff) 0.015863 0.015484 1.024 0.3056
## scale(similarity) -0.002297 0.010624 -0.216 0.8288
## pairPP 0.004964 0.021366 0.232 0.8163
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) scl(nD) scl(tD) scl(s)
## scale(nDff) 0.099
## scale(tDff) 0.096 0.361
## scl(smlrty) 0.000 0.038 0.001
## pairPP -0.496 -0.023 -0.077 -0.0015 Question 6: Does the model (mixture of category representativeness and desirability) predict choice, and is it not just similarity or desirability?
5.1 Study 1
m <- glmer( as.factor(choice) ~ scale(rAP) + ( scale(rAP) | subID), data = fullTD1, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)## boundary (singular) fit: see ?isSingularsummary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: as.factor(choice) ~ scale(rAP) + (scale(rAP) | subID)
## Data: fullTD1
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 12783.3 12819.5 -6386.6 12773.3 10274
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.8316 -0.8800 -0.3951 0.9421 2.8201
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 1.166e-04 0.010800
## scale(rAP) 3.056e-06 0.001748 1.00
## Number of obs: 10279, groups: subID, 80
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.09284 0.02124 -4.372 1.23e-05 ***
## scale(rAP) 0.85025 0.02502 33.986 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## scale(rAP) -0.015
## optimizer (bobyqa) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingularr2beta(m,data=fullTD1)## boundary (singular) fit: see ?isSingular
## boundary (singular) fit: see ?isSingularm <- glmer( as.factor(choice) ~ scale(RSV.AP) + scale(rightDesAP) + ( scale(RSV.AP) + scale(rightDesAP) | subID), data = fullTD1, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula:
## as.factor(choice) ~ scale(RSV.AP) + scale(rightDesAP) + (scale(RSV.AP) +
## scale(rightDesAP) | subID)
## Data: fullTD1
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 12953.7 13018.8 -6467.8 12935.7 10270
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.9964 -0.8880 -0.3748 0.9398 3.0511
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.007737 0.08796
## scale(RSV.AP) 0.069691 0.26399 -0.27
## scale(rightDesAP) 0.118818 0.34470 -0.14 -0.44
## Number of obs: 10279, groups: subID, 80
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.10258 0.02354 -4.359 1.31e-05 ***
## scale(RSV.AP) 0.23894 0.04051 5.899 3.66e-09 ***
## scale(rightDesAP) 0.64013 0.04775 13.407 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s(RSV.
## scl(RSV.AP) -0.086
## scl(rghDAP) -0.062 -0.441r2beta(m,data=fullTD1)## boundary (singular) fit: see ?isSingular
## boundary (singular) fit: see ?isSingular
## boundary (singular) fit: see ?isSingular
## boundary (singular) fit: see ?isSingular5.2 Study 2
m <- glmer( as.factor(choice) ~ scale(rAP) + ( scale(rAP) | subID), data = fullTD2, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)## boundary (singular) fit: see ?isSingularsummary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: as.factor(choice) ~ scale(rAP) + (scale(rAP) | subID)
## Data: fullTD2
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 19350.4 19388.7 -9670.2 19340.4 15745
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.3535 -0.8861 -0.3013 0.8983 3.2907
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.0e+00 0.000e+00
## scale(rAP) 2.9e-14 1.703e-07 NaN
## Number of obs: 15750, groups: subID, 105
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.01764 0.01724 -1.023 0.306
## scale(rAP) 0.92542 0.02136 43.334 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## scale(rAP) -0.004
## optimizer (bobyqa) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingularr2beta(m,data=fullTD2)## boundary (singular) fit: see ?isSingularm <- glmer( as.factor(choice) ~ scale(RSV.AP) + scale(rightDesAP) + ( scale(RSV.AP) + scale(rightDesAP) | subID), data = fullTD2, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula:
## as.factor(choice) ~ scale(RSV.AP) + scale(rightDesAP) + (scale(RSV.AP) +
## scale(rightDesAP) | subID)
## Data: fullTD2
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 19797.2 19866.2 -9889.6 19779.2 15741
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.6327 -0.9048 -0.2202 0.9099 4.5146
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.03563 0.1888
## scale(RSV.AP) 0.05641 0.2375 -0.17
## scale(rightDesAP) 0.22282 0.4720 -0.11 -0.30
## Number of obs: 15750, groups: subID, 105
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.02528 0.02530 -0.999 0.318
## scale(RSV.AP) 0.34272 0.03241 10.573 <2e-16 ***
## scale(rightDesAP) 0.54801 0.05138 10.665 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s(RSV.
## scl(RSV.AP) -0.094
## scl(rghDAP) -0.072 -0.319r2beta(m,data=fullTD2)5.3 Study 3
m <- glmer( as.factor(choice) ~ scale(rAP) + ( scale(rAP) | subID), data = fullTD2, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)## boundary (singular) fit: see ?isSingularsummary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: as.factor(choice) ~ scale(rAP) + (scale(rAP) | subID)
## Data: fullTD2
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 19350.4 19388.7 -9670.2 19340.4 15745
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.3535 -0.8861 -0.3013 0.8983 3.2907
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.0e+00 0.000e+00
## scale(rAP) 2.9e-14 1.703e-07 NaN
## Number of obs: 15750, groups: subID, 105
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.01764 0.01724 -1.023 0.306
## scale(rAP) 0.92542 0.02136 43.334 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## scale(rAP) -0.004
## optimizer (bobyqa) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingularr2beta(m,data=fullTD3)## boundary (singular) fit: see ?isSingularm <- glmer( as.factor(choice) ~ scale(RSV.AP) + scale(rightDesAP) + ( scale(RSV.AP) + scale(rightDesAP) | subID), data = fullTD2, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula:
## as.factor(choice) ~ scale(RSV.AP) + scale(rightDesAP) + (scale(RSV.AP) +
## scale(rightDesAP) | subID)
## Data: fullTD2
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 19797.2 19866.2 -9889.6 19779.2 15741
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.6327 -0.9048 -0.2202 0.9099 4.5146
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.03563 0.1888
## scale(RSV.AP) 0.05641 0.2375 -0.17
## scale(rightDesAP) 0.22282 0.4720 -0.11 -0.30
## Number of obs: 15750, groups: subID, 105
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.02528 0.02530 -0.999 0.318
## scale(RSV.AP) 0.34272 0.03241 10.573 <2e-16 ***
## scale(rightDesAP) 0.54801 0.05138 10.665 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) s(RSV.
## scl(RSV.AP) -0.094
## scl(rghDAP) -0.072 -0.319r2beta(m,data=fullTD3)5.4 Combined
m <- glmer( as.factor(choice) ~ scale(rAP)*condition + pair + ( scale(rAP) | subID), data = fullTD, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)## boundary (singular) fit: see ?isSingularsummary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: as.factor(choice) ~ scale(rAP) * condition + pair + (scale(rAP) |
## subID)
## Data: fullTD
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 68337.7 68453.9 -34155.8 68311.7 56594
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.4078 -0.8695 -0.3105 0.8949 3.4046
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.000e+00 0.000e+00
## scale(rAP) 3.224e-16 1.796e-08 NaN
## Number of obs: 56607, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.060495 0.023975 -2.523 0.0116 *
## scale(rAP) 1.009617 0.021813 46.285 <2e-16 ***
## conditionAsian -0.004112 0.024766 -0.166 0.8681
## conditionDemocrat -0.002642 0.025334 -0.104 0.9170
## conditionLatino 0.005279 0.027678 0.191 0.8487
## pairNN 0.028660 0.023639 1.212 0.2253
## pairPP 0.008767 0.023860 0.367 0.7133
## scale(rAP):conditionAsian -0.008642 0.031769 -0.272 0.7856
## scale(rAP):conditionDemocrat -0.011421 0.030134 -0.379 0.7047
## scale(rAP):conditionLatino -0.073696 0.035121 -2.098 0.0359 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) sc(AP) cndtnA cndtnD cndtnL pairNN pairPP s(AP):A s(AP):D
## scale(rAP) -0.020
## conditinAsn -0.545 0.014
## condtnDmcrt -0.516 0.013 0.502
## conditinLtn -0.486 0.012 0.460 0.449
## pairNN -0.602 0.009 0.017 -0.004 0.006
## pairPP -0.601 0.007 0.020 -0.004 0.023 0.600
## scl(rAP):cA 0.011 -0.687 -0.035 -0.009 -0.008 -0.002 -0.002
## scl(rAP):cD 0.009 -0.724 -0.010 -0.015 -0.009 0.001 0.002 0.497
## scl(rAP):cL 0.002 -0.621 -0.008 -0.008 0.025 0.008 0.011 0.426 0.450
## optimizer (bobyqa) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingularggpredict(m, c("rAP", "condition")) %>% plot()## Data were 'prettified'. Consider using `terms="rAP [all]"` to get smooth plots.
r2beta(m,data=fullTD)## boundary (singular) fit: see ?isSingularm <- glmer( as.factor(choice) ~ scale(RSV.AP)*condition + pair + scale(rightDesAP)*condition + ( scale(RSV.AP) + scale(rightDesAP) | subID), data = fullTD, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula:
## as.factor(choice) ~ scale(RSV.AP) * condition + pair + scale(rightDesAP) *
## condition + (scale(RSV.AP) + scale(rightDesAP) | subID)
## Data: fullTD
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 69666.6 69845.5 -34813.3 69626.6 56587
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.8915 -0.8825 -0.2480 0.9045 5.6065
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.03465 0.1861
## scale(RSV.AP) 0.04758 0.2181 -0.10
## scale(rightDesAP) 0.21275 0.4613 -0.04 -0.17
## Number of obs: 56607, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.038266 0.030305 -1.263 0.20670
## scale(RSV.AP) 0.283989 0.032147 8.834 < 2e-16 ***
## conditionAsian 0.012694 0.035730 0.355 0.72237
## conditionDemocrat -0.019077 0.036229 -0.527 0.59850
## conditionLatino -0.062122 0.039259 -1.582 0.11357
## pairNN 0.011235 0.023796 0.472 0.63684
## pairPP -0.009785 0.023988 -0.408 0.68332
## scale(rightDesAP) 0.756290 0.051435 14.704 < 2e-16 ***
## scale(RSV.AP):conditionAsian 0.078536 0.045082 1.742 0.08150 .
## scale(RSV.AP):conditionDemocrat 0.014319 0.045704 0.313 0.75406
## scale(RSV.AP):conditionLatino -0.062299 0.047937 -1.300 0.19374
## conditionAsian:scale(rightDesAP) -0.211342 0.071620 -2.951 0.00317 **
## conditionDemocrat:scale(rightDesAP) 0.034757 0.072659 0.478 0.63239
## conditionLatino:scale(rightDesAP) -0.092894 0.077918 -1.192 0.23318
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1##
## Correlation matrix not shown by default, as p = 14 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need itggpredict(m, c("RSV.AP", "condition")) %>% plot()## Data were 'prettified'. Consider using `terms="RSV.AP [all]"` to get smooth plots.
r2beta(m,data=fullTD)5.4.1 Ingroup vs. Outgroup
m <- glmer( as.factor(choice) ~ scale(rAP)*ingroup + ( scale(rAP) | subID), data = fullTD, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)## boundary (singular) fit: see ?isSingularsummary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: as.factor(choice) ~ scale(rAP) * ingroup + (scale(rAP) | subID)
## Data: fullTD
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 68331.6 68394.2 -34158.8 68317.6 56600
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.4314 -0.8670 -0.3157 0.8966 3.3796
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.000e+00 0.000e+00
## scale(rAP) 1.091e-12 1.044e-06 NaN
## Number of obs: 56607, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.0467239 0.0107525 -4.345 1.39e-05 ***
## scale(rAP) 0.9840977 0.0134533 73.149 < 2e-16 ***
## ingroupOut 0.0004804 0.0207502 0.023 0.982
## scale(rAP):ingroupOut 0.0253273 0.0256271 0.988 0.323
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) sc(AP) ingrpO
## scale(rAP) -0.006
## ingroupOut -0.518 0.003
## scl(rAP):nO 0.003 -0.525 -0.013
## optimizer (bobyqa) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingularggpredict(m, c("rAP", "ingroup")) %>% plot()## Data were 'prettified'. Consider using `terms="rAP [all]"` to get smooth plots.
r2beta(m,data=fullTD)## boundary (singular) fit: see ?isSingularm <- glmer( as.factor(choice) ~ scale(RSV.AP)*ingroup + scale(rightDesAP)*ingroup + ( scale(RSV.AP) + scale(rightDesAP) | subID), data = fullTD, family = binomial, control = glmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)),
nAGQ = 1)
summary(m)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: as.factor(choice) ~ scale(RSV.AP) * ingroup + scale(rightDesAP) *
## ingroup + (scale(RSV.AP) + scale(rightDesAP) | subID)
## Data: fullTD
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## AIC BIC logLik deviance df.resid
## 69674.6 69781.9 -34825.3 69650.6 56595
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.8320 -0.8825 -0.2487 0.9056 5.6445
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.03509 0.1873
## scale(RSV.AP) 0.04959 0.2227 -0.07
## scale(rightDesAP) 0.22269 0.4719 -0.06 -0.19
## Number of obs: 56607, groups: subID, 393
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.05580 0.01546 -3.608 0.000308 ***
## scale(RSV.AP) 0.29928 0.01939 15.434 < 2e-16 ***
## ingroupOut 0.01815 0.02986 0.608 0.543370
## scale(rightDesAP) 0.66515 0.03133 21.234 < 2e-16 ***
## scale(RSV.AP):ingroupOut -0.01607 0.03773 -0.426 0.670129
## ingroupOut:scale(rightDesAP) 0.09313 0.06082 1.531 0.125717
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) sc(RSV.AP) ingrpO s(DAP) s(RSV.AP):
## scl(RSV.AP) -0.035
## ingroupOut -0.516 0.018
## scl(rghDAP) -0.041 -0.250 0.021
## s(RSV.AP):O 0.018 -0.509 -0.040 0.124
## ingrO:(DAP) 0.021 0.124 -0.036 -0.508 -0.257ggpredict(m, c("RSV.AP", "ingroup")) %>% plot()## Data were 'prettified'. Consider using `terms="RSV.AP [all]"` to get smooth plots.
r2beta(m,data=fullTD)## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
## Model failed to converge with max|grad| = 0.013214 (tol = 0.002, component 1)6 Question 7: Do broad feedback category predict self-evaluations?
6.1 Study 1
m <- lmer(selfResp ~ as.factor(propT) +
( as.factor(propT) | subID) + (1 | trait), data = reValDf1, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ as.factor(propT) + (as.factor(propT) | subID) + (1 |
## trait)
## Data: reValDf1
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 40299.2
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.4593 -0.6476 0.0558 0.6764 3.4833
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 1.0521 1.0257
## subID (Intercept) 0.1661 0.4075
## as.factor(propT).L 0.1556 0.3945 -0.24
## as.factor(propT).Q 0.1431 0.3783 -0.03 -0.14
## Residual 1.8374 1.3555
## Number of obs: 11400, groups: trait, 150; subID, 76
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.44781 0.09683 207.40991 45.932 <2e-16 ***
## as.factor(propT).L 0.11764 0.05086 73.31609 2.313 0.0235 *
## as.factor(propT).Q 0.01300 0.04905 76.00932 0.265 0.7917
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) a.(T).L
## as.fct(T).L -0.108
## as.fct(T).Q -0.011 -0.121ggpredict(m, c("propT")) %>% plot()
6.1.1 Moderated by valence
m <- lmer(selfResp ~ as.factor(propT) * valence +
( as.factor(propT) * valence | subID) + (1 | trait), data = reValDf1, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ as.factor(propT) * valence + (as.factor(propT) * valence |
## subID) + (1 | trait)
## Data: reValDf1
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 38916.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.3892 -0.6074 0.0489 0.6377 4.4633
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 0.4010 0.6332
## subID (Intercept) 0.4278 0.6541
## as.factor(propT).L 0.4093 0.6397 -0.10
## as.factor(propT).Q 0.3124 0.5590 -0.15 0.14
## valencePos 0.8296 0.9108 -0.78 0.00 0.05
## as.factor(propT).L:valencePos 0.4164 0.6453 0.07 -0.75 -0.11
## as.factor(propT).Q:valencePos 0.4242 0.6513 0.12 -0.38 -0.79
## Residual 1.5810 1.2574
##
##
##
##
##
##
## -0.05
## 0.02 0.20
##
## Number of obs: 11400, groups: trait, 150; subID, 76
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 3.65513 0.10706 186.90599 34.141 <2e-16
## as.factor(propT).L 0.11308 0.08032 72.96152 1.408 0.1634
## as.factor(propT).Q 0.09289 0.07156 75.82156 1.298 0.1982
## valencePos 1.58955 0.15000 187.68877 10.597 <2e-16
## as.factor(propT).L:valencePos 0.04622 0.08709 76.41034 0.531 0.5971
## as.factor(propT).Q:valencePos -0.15093 0.08691 77.45552 -1.737 0.0864
##
## (Intercept) ***
## as.factor(propT).L
## as.factor(propT).Q
## valencePos ***
## as.factor(propT).L:valencePos
## as.factor(propT).Q:valencePos .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) as.(T).L as.(T).Q vlncPs a.(T).L:
## as.fct(T).L -0.075
## as.fct(T).Q -0.079 0.103
## valencePos -0.743 0.005 0.022
## as.f(T).L:P 0.051 -0.737 -0.071 -0.047
## as.f(T).Q:P 0.061 -0.287 -0.774 0.017 0.132ggpredict(m, c("propT","valence")) %>% plot()
6.2 Study 2
m <- lmer(selfResp ~ as.factor(propT) +
( as.factor(propT) | subID) + (1 | trait), data = reValDf2, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ as.factor(propT) + (as.factor(propT) | subID) + (1 |
## trait)
## Data: reValDf2
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 60319.1
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.5038 -0.6786 0.0387 0.6983 3.0837
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 0.72126 0.8493
## subID (Intercept) 0.12342 0.3513
## as.factor(propT).L 0.07570 0.2751 -0.13
## as.factor(propT).Q 0.08763 0.2960 0.01 -0.04
## Residual 1.93056 1.3894
## Number of obs: 16905, groups: trait, 210; subID, 105
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.30632 0.06878 291.57848 62.613 <2e-16 ***
## as.factor(propT).L 0.05240 0.03285 101.30493 1.595 0.114
## as.factor(propT).Q -0.05365 0.03463 103.71838 -1.549 0.124
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) a.(T).L
## as.fct(T).L -0.056
## as.fct(T).Q 0.001 -0.030ggpredict(m, c("propT")) %>% plot()
6.2.1 Moderated by valence
m <- lmer(selfResp ~ as.factor(propT) * valence +
( as.factor(propT) * valence | subID) + (1 | trait), data = reValDf2, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ as.factor(propT) * valence + (as.factor(propT) * valence |
## subID) + (1 | trait)
## Data: reValDf2
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 58271.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.1762 -0.6273 0.0300 0.6614 4.1043
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 0.3352 0.5790
## subID (Intercept) 0.4218 0.6495
## as.factor(propT).L 0.1622 0.4027 -0.02
## as.factor(propT).Q 0.2329 0.4826 0.00 0.00
## valencePos 0.9186 0.9585 -0.85 -0.11 0.07
## as.factor(propT).L:valencePos 0.2909 0.5393 -0.12 -0.80 0.03
## as.factor(propT).Q:valencePos 0.3326 0.5767 -0.05 -0.03 -0.84
## Residual 1.6662 1.2908
##
##
##
##
##
##
## 0.27
## 0.00 -0.05
##
## Number of obs: 16905, groups: trait, 210; subID, 105
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 3.69226 0.08637 241.45404 42.751 <2e-16
## as.factor(propT).L 0.08247 0.04695 99.08820 1.757 0.0821
## as.factor(propT).Q -0.05108 0.05370 102.33962 -0.951 0.3438
## valencePos 1.21097 0.12493 233.58646 9.693 <2e-16
## as.factor(propT).L:valencePos -0.04644 0.06389 101.93226 -0.727 0.4689
## as.factor(propT).Q:valencePos 0.02061 0.06712 104.26895 0.307 0.7594
##
## (Intercept) ***
## as.factor(propT).L .
## as.factor(propT).Q
## valencePos ***
## as.factor(propT).L:valencePos
## as.factor(propT).Q:valencePos
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) as.(T).L as.(T).Q vlncPs a.(T).L:
## as.fct(T).L -0.014
## as.fct(T).Q -0.002 -0.002
## valencePos -0.785 -0.068 0.046
## as.f(T).L:P -0.069 -0.773 0.022 0.161
## as.f(T).Q:P -0.028 -0.021 -0.803 -0.003 -0.031ggpredict(m, c("propT","valence")) %>% plot()
6.3 Study 3
m <- lmer(selfResp ~ as.factor(propT) +
( as.factor(propT) | subID) + (1 | trait), data = reValDf3, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ as.factor(propT) + (as.factor(propT) | subID) + (1 |
## trait)
## Data: reValDf3
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 112810.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.5391 -0.6881 0.0057 0.6912 3.7087
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 1.2346 1.1111
## subID (Intercept) 0.1230 0.3507
## as.factor(propT).L 0.0888 0.2980 -0.02
## as.factor(propT).Q 0.1075 0.3279 -0.04 -0.14
## Residual 2.1823 1.4773
## Number of obs: 30669, groups: trait, 210; subID, 195
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.103e+00 8.115e-02 2.518e+02 50.559 <2e-16 ***
## as.factor(propT).L -1.283e-02 2.614e-02 1.912e+02 -0.491 0.624
## as.factor(propT).Q -3.896e-04 2.781e-02 1.968e+02 -0.014 0.989
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) a.(T).L
## as.fct(T).L -0.006
## as.fct(T).Q -0.008 -0.102ggpredict(m, c("propT")) %>% plot()
6.3.1 Moderated by valence
m <- lmer(selfResp ~ as.factor(propT) * valence +
( as.factor(propT) * valence | subID) + (1 | trait), data = reValDf3, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ as.factor(propT) * valence + (as.factor(propT) * valence |
## subID) + (1 | trait)
## Data: reValDf3
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 108089.2
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.3782 -0.6150 0.0106 0.6153 4.4447
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 0.3691 0.6075
## subID (Intercept) 0.5163 0.7185
## as.factor(propT).L 0.1720 0.4147 0.07
## as.factor(propT).Q 0.2459 0.4959 -0.01 -0.09
## valencePos 1.3105 1.1448 -0.87 -0.09 0.00
## as.factor(propT).L:valencePos 0.3621 0.6017 -0.08 -0.77 -0.01
## as.factor(propT).Q:valencePos 0.3316 0.5758 -0.02 0.06 -0.80
## Residual 1.8148 1.3472
##
##
##
##
##
##
## 0.06
## 0.01 0.00
##
## Number of obs: 30669, groups: trait, 210; subID, 195
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 3.172230 0.079452 382.413759 39.926 <2e-16
## as.factor(propT).L 0.007586 0.035888 195.504312 0.211 0.833
## as.factor(propT).Q 0.006092 0.040658 197.829716 0.150 0.881
## valencePos 1.857595 0.118492 387.501411 15.677 <2e-16
## as.factor(propT).L:valencePos -0.048061 0.051608 191.385654 -0.931 0.353
## as.factor(propT).Q:valencePos -0.019575 0.049852 194.932658 -0.393 0.695
##
## (Intercept) ***
## as.factor(propT).L
## as.factor(propT).Q
## valencePos ***
## as.factor(propT).L:valencePos
## as.factor(propT).Q:valencePos
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) as.(T).L as.(T).Q vlncPs a.(T).L:
## as.fct(T).L 0.035
## as.fct(T).Q -0.001 -0.069
## valencePos -0.780 -0.048 -0.002
## as.f(T).L:P -0.041 -0.751 -0.003 0.033
## as.f(T).Q:P -0.012 0.046 -0.775 0.011 -0.006ggpredict(m, c("propT","valence")) %>% plot()
6.4 Combined
m <- lmer(selfResp ~ as.factor(propT) * condition +
( as.factor(propT) | subID) + (1 | trait), data = reValDf3, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ as.factor(propT) * condition + (as.factor(propT) |
## subID) + (1 | trait)
## Data: reValDf3
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 112822
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.5415 -0.6874 0.0052 0.6915 3.7106
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 1.23445 1.1111
## subID (Intercept) 0.12355 0.3515
## as.factor(propT).L 0.08942 0.2990 -0.02
## as.factor(propT).Q 0.10808 0.3288 -0.04 -0.14
## Residual 2.18233 1.4773
## Number of obs: 30669, groups: trait, 210; subID, 195
##
## Fixed effects:
## Estimate Std. Error df t value
## (Intercept) 4.11399 0.08554 296.87050 48.093
## as.factor(propT).L -0.01862 0.03734 190.12581 -0.499
## as.factor(propT).Q 0.01474 0.03974 196.19587 0.371
## conditionRepublican -0.02201 0.05322 192.53280 -0.414
## as.factor(propT).L:conditionRepublican 0.01158 0.05246 190.85043 0.221
## as.factor(propT).Q:conditionRepublican -0.02980 0.05576 195.94143 -0.534
## Pr(>|t|)
## (Intercept) <2e-16 ***
## as.factor(propT).L 0.619
## as.factor(propT).Q 0.711
## conditionRepublican 0.680
## as.factor(propT).L:conditionRepublican 0.825
## as.factor(propT).Q:conditionRepublican 0.594
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) as.(T).L as.(T).Q cndtnR a.(T).L:
## as.fct(T).L -0.006
## as.fct(T).Q -0.012 -0.098
## cndtnRpblcn -0.316 0.010 0.020
## as.f(T).L:R 0.004 -0.713 0.070 -0.017
## as.f(T).Q:R 0.009 0.070 -0.713 -0.028 -0.103ggpredict(m, c("propT","condition")) %>% plot()
6.4.1 Moderated by valence
m <- lmer(selfResp ~ as.factor(propT) * valence * condition +
( as.factor(propT) * valence | subID) + (1 | trait), data = reValDf3, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula:
## selfResp ~ as.factor(propT) * valence * condition + (as.factor(propT) *
## valence | subID) + (1 | trait)
## Data: reValDf3
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 108108.1
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.3785 -0.6151 0.0104 0.6151 4.4421
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 0.3690 0.6075
## subID (Intercept) 0.5182 0.7199
## as.factor(propT).L 0.1732 0.4162 0.07
## as.factor(propT).Q 0.2470 0.4970 -0.01 -0.09
## valencePos 1.3165 1.1474 -0.87 -0.09 0.00
## as.factor(propT).L:valencePos 0.3647 0.6039 -0.08 -0.77 -0.01
## as.factor(propT).Q:valencePos 0.3338 0.5777 -0.01 0.06 -0.80
## Residual 1.8148 1.3472
##
##
##
##
##
##
## 0.06
## 0.01 0.00
##
## Number of obs: 30669, groups: trait, 210; subID, 195
##
## Fixed effects:
## Estimate Std. Error
## (Intercept) 3.204e+00 9.590e-02
## as.factor(propT).L 7.918e-03 5.132e-02
## as.factor(propT).Q 2.989e-02 5.803e-02
## valencePos 1.821e+00 1.459e-01
## conditionRepublican -6.185e-02 1.056e-01
## as.factor(propT).L:valencePos -4.801e-02 7.371e-02
## as.factor(propT).Q:valencePos -3.744e-02 7.129e-02
## as.factor(propT).L:conditionRepublican -6.592e-04 7.190e-02
## as.factor(propT).Q:conditionRepublican -4.682e-02 8.141e-02
## valencePos:conditionRepublican 7.191e-02 1.675e-01
## as.factor(propT).L:valencePos:conditionRepublican 2.742e-05 1.036e-01
## as.factor(propT).Q:valencePos:conditionRepublican 3.516e-02 1.000e-01
## df t value Pr(>|t|)
## (Intercept) 3.668e+02 33.405 <2e-16
## as.factor(propT).L 1.957e+02 0.154 0.878
## as.factor(propT).Q 1.962e+02 0.515 0.607
## valencePos 3.467e+02 12.478 <2e-16
## conditionRepublican 1.929e+02 -0.586 0.559
## as.factor(propT).L:valencePos 1.901e+02 -0.651 0.516
## as.factor(propT).Q:valencePos 1.943e+02 -0.525 0.600
## as.factor(propT).L:conditionRepublican 1.943e+02 -0.009 0.993
## as.factor(propT).Q:conditionRepublican 1.955e+02 -0.575 0.566
## valencePos:conditionRepublican 1.930e+02 0.429 0.668
## as.factor(propT).L:valencePos:conditionRepublican 1.913e+02 0.000 1.000
## as.factor(propT).Q:valencePos:conditionRepublican 1.940e+02 0.351 0.726
##
## (Intercept) ***
## as.factor(propT).L
## as.factor(propT).Q
## valencePos ***
## conditionRepublican
## as.factor(propT).L:valencePos
## as.factor(propT).Q:valencePos
## as.factor(propT).L:conditionRepublican
## as.factor(propT).Q:conditionRepublican
## valencePos:conditionRepublican
## as.factor(propT).L:valencePos:conditionRepublican
## as.factor(propT).Q:valencePos:conditionRepublican
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) as.(T).L as.(T).Q vlncPs cndtnR as.(T).L:P as.(T).Q:P
## as.fct(T).L 0.044
## as.fct(T).Q -0.002 -0.068
## valencePos -0.807 -0.058 -0.001
## cndtnRpblcn -0.559 -0.040 0.004 0.504
## as.f(T).L:P -0.050 -0.752 -0.005 0.042 0.046
## as.f(T).Q:P -0.013 0.045 -0.774 0.011 0.010 -0.002
## as.f(T).L:R -0.033 -0.713 0.047 0.042 0.057 0.536 -0.031
## as.f(T).Q:R 0.002 0.048 -0.712 0.000 -0.007 0.004 0.551
## vlncPs:cndR 0.483 0.051 0.000 -0.583 -0.865 -0.037 -0.009
## a.(T).L:P:R 0.037 0.534 0.004 -0.031 -0.064 -0.712 0.002
## a.(T).Q:P:R 0.009 -0.032 0.551 -0.007 -0.013 0.002 -0.713
## a.(T).L:R a.(T).Q:R vlnP:R a.(T).L:P:
## as.fct(T).L
## as.fct(T).Q
## valencePos
## cndtnRpblcn
## as.f(T).L:P
## as.f(T).Q:P
## as.f(T).L:R
## as.f(T).Q:R -0.067
## vlncPs:cndR -0.071 0.001
## a.(T).L:P:R -0.749 -0.005 0.049
## a.(T).Q:P:R 0.044 -0.773 0.012 -0.007ggpredict(m, c("propT","valence","condition")) %>% plot()
6.4.2 Ingroup vs. Outgroup
m <- lmer(selfResp ~ as.factor(propT) * ingroup +
( as.factor(propT) | subID) + (1 | traits), data = reValDf, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ as.factor(propT) * ingroup + (as.factor(propT) | subID) +
## (1 | traits)
## Data: reValDf
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 213553.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.5643 -0.6870 0.0231 0.6971 3.5005
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.1393 0.3732
## as.factor(propT).L 0.1065 0.3263 -0.06
## as.factor(propT).Q 0.1137 0.3372 -0.03 -0.11
## traits (Intercept) 1.0039 1.0020
## Residual 2.0766 1.4411
## Number of obs: 58974, groups: subID, 376; traits, 257
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.29365 0.06684 322.84390 64.235 < 2e-16
## as.factor(propT).L 0.04542 0.02318 366.37634 1.960 0.050786
## as.factor(propT).Q -0.01655 0.02370 378.25948 -0.698 0.485562
## ingroupOut -0.16772 0.04595 376.22010 -3.650 0.000299
## as.factor(propT).L:ingroupOut -0.05258 0.04518 367.00592 -1.164 0.245304
## as.factor(propT).Q:ingroupOut 0.00226 0.04610 375.34451 0.049 0.960932
##
## (Intercept) ***
## as.factor(propT).L .
## as.factor(propT).Q
## ingroupOut ***
## as.factor(propT).L:ingroupOut
## as.factor(propT).Q:ingroupOut
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) as.(T).L as.(T).Q ingrpO a.(T).L:
## as.fct(T).L -0.019
## as.fct(T).Q -0.007 -0.084
## ingroupOut -0.178 0.029 0.011
## as.f(T).L:O 0.010 -0.513 0.043 -0.056
## as.f(T).Q:O 0.004 0.043 -0.514 -0.019 -0.087ggpredict(m, c("propT","ingroup")) %>% plot()
6.4.3 Ingroup vs. Outgroup moderated by valence
m <- lmer(selfResp ~ as.factor(propT) * ingroup * valence +
( as.factor(propT) + valence | subID) + (1 | traits), data = reValDf, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ as.factor(propT) * ingroup * valence + (as.factor(propT) +
## valence | subID) + (1 | traits)
## Data: reValDf
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 206100.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.4096 -0.6323 0.0201 0.6415 4.4522
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 2.92244 1.7095
## as.factor(propT).L 0.09843 0.3137 -0.01
## as.factor(propT).Q 0.11042 0.3323 -0.08 -0.13
## valence 1.14211 1.0687 -0.98 -0.01 0.08
## traits (Intercept) 0.34362 0.5862
## Residual 1.79692 1.3405
## Number of obs: 58974, groups: subID, 376; traits, 257
##
## Fixed effects:
## Estimate Std. Error df t value
## (Intercept) 1.910e+00 1.589e-01 5.490e+02 12.020
## as.factor(propT).L 9.028e-02 4.168e-02 4.240e+03 2.166
## as.factor(propT).Q 6.133e-02 4.186e-02 3.858e+03 1.465
## ingroupOut -6.793e-01 2.045e-01 3.754e+02 -3.322
## valence 1.540e+00 9.858e-02 5.543e+02 15.625
## as.factor(propT).L:ingroupOut -2.337e-02 8.080e-02 4.170e+03 -0.289
## as.factor(propT).Q:ingroupOut -6.273e-02 8.104e-02 3.747e+03 -0.774
## as.factor(propT).L:valence -2.843e-02 2.346e-02 5.814e+04 -1.212
## as.factor(propT).Q:valence -4.694e-02 2.332e-02 5.812e+04 -2.012
## ingroupOut:valence 3.402e-01 1.280e-01 3.764e+02 2.658
## as.factor(propT).L:ingroupOut:valence -2.397e-02 4.570e-02 5.815e+04 -0.525
## as.factor(propT).Q:ingroupOut:valence 3.962e-02 4.505e-02 5.811e+04 0.880
## Pr(>|t|)
## (Intercept) < 2e-16 ***
## as.factor(propT).L 0.030355 *
## as.factor(propT).Q 0.143017
## ingroupOut 0.000981 ***
## valence < 2e-16 ***
## as.factor(propT).L:ingroupOut 0.772433
## as.factor(propT).Q:ingroupOut 0.438977
## as.factor(propT).L:valence 0.225480
## as.factor(propT).Q:valence 0.044174 *
## ingroupOut:valence 0.008207 **
## as.factor(propT).L:ingroupOut:valence 0.599903
## as.factor(propT).Q:ingroupOut:valence 0.379114
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) as.(T).L as.(T).Q ingrpO valenc as.(T).L:O as.(T).Q:O
## as.fct(T).L -0.004
## as.fct(T).Q -0.021 -0.040
## ingroupOut -0.337 0.004 0.017
## valence -0.962 -0.002 0.022 0.331
## as.f(T).L:O 0.002 -0.515 0.020 -0.007 0.002
## as.f(T).Q:O 0.011 0.020 -0.517 -0.033 -0.011 -0.042
## as.fc(T).L: 0.002 -0.846 0.012 -0.004 -0.001 0.436 -0.006
## as.fc(T).Q: -0.002 0.011 -0.834 0.002 0.002 -0.005 0.431
## ingrpOt:vln 0.328 -0.001 -0.016 -0.975 -0.339 -0.001 0.032
## as.(T).L:O: -0.001 0.434 -0.006 0.006 0.000 -0.845 0.016
## as.(T).Q:O: 0.001 -0.005 0.432 -0.004 -0.001 0.016 -0.832
## as.(T).L: as.(T).Q: ingrO: a.(T).L:O:
## as.fct(T).L
## as.fct(T).Q
## ingroupOut
## valence
## as.f(T).L:O
## as.f(T).Q:O
## as.fc(T).L:
## as.fc(T).Q: -0.012
## ingrpOt:vln 0.006 -0.001
## as.(T).L:O: -0.514 0.007 -0.008
## as.(T).Q:O: 0.007 -0.519 0.005 -0.023ggpredict(m, c("propT","valence","ingroup")) %>% plot()
# Question 8: Does average feedback observed during learning for each cue predict re-evaluations?
6.5 Study 1
m <- lmer(selfResp ~ aveCueF +
( aveCueF | subID) + (1 | trait), data = reValDf1, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ aveCueF + (aveCueF | subID) + (1 | trait)
## Data: reValDf1
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 40450.7
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.6338 -0.6549 0.0514 0.6894 3.4683
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 1.066 1.032
## subID (Intercept) 1.225 1.107
## aveCueF 3.427 1.851 -0.93
## Residual 1.882 1.372
## Number of obs: 11400, groups: trait, 150; subID, 76
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.1213 0.1624 128.9855 25.375 < 2e-16 ***
## aveCueF 0.6439 0.2379 73.3219 2.707 0.00845 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## aveCueF -0.802ggpredict(m, c("aveCueF")) %>% plot()
6.5.1 Moderated by valence
m <- lmer(selfResp ~ aveCueF * valence +
( aveCueF * valence | subID) + (1 | trait), data = reValDf1, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ aveCueF * valence + (aveCueF * valence | subID) +
## (1 | trait)
## Data: reValDf1
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 39235.7
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.2481 -0.6200 0.0473 0.6502 4.3110
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 0.4201 0.6481
## subID (Intercept) 2.7546 1.6597
## aveCueF 8.1972 2.8631 -0.92
## valencePos 3.4198 1.8493 -0.76 0.68
## aveCueF:valencePos 9.6577 3.1077 0.70 -0.77 -0.88
## Residual 1.6585 1.2878
## Number of obs: 11400, groups: trait, 150; subID, 76
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 3.30188 0.21938 94.92295 15.051 < 2e-16 ***
## aveCueF 0.64872 0.36157 73.93387 1.794 0.0769 .
## valencePos 1.63427 0.26151 105.07535 6.249 8.96e-09 ***
## aveCueF:valencePos -0.01686 0.41390 74.07535 -0.041 0.9676
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) aveCuF vlncPs
## aveCueF -0.871
## valencePos -0.742 0.622
## avCF:vlncPs 0.651 -0.757 -0.821ggpredict(m, c("aveCueF","valence")) %>% plot()
6.6 Study 2
m <- lmer(selfResp ~ aveCueF +
( aveCueF | subID) + (1 | trait), data = reValDf2, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ aveCueF + (aveCueF | subID) + (1 | trait)
## Data: reValDf2
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 60432.8
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.4920 -0.6829 0.0395 0.7090 3.2081
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 0.7166 0.8465
## subID (Intercept) 0.5114 0.7151
## aveCueF 1.3020 1.1410 -0.87
## Residual 1.9599 1.3999
## Number of obs: 16905, groups: trait, 210; subID, 105
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.2028 0.1001 208.8339 41.993 <2e-16 ***
## aveCueF 0.2079 0.1371 103.0947 1.516 0.132
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## aveCueF -0.729ggpredict(m, c("aveCueF")) %>% plot()
6.6.1 Moderated by valence
m <- lmer(selfResp ~ aveCueF * valence +
( aveCueF * valence | subID) + (1 | trait), data = reValDf2, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ aveCueF * valence + (aveCueF * valence | subID) +
## (1 | trait)
## Data: reValDf2
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 58528.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.0611 -0.6412 0.0290 0.6691 4.0499
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 0.3509 0.5924
## subID (Intercept) 1.1714 1.0823
## aveCueF 2.6540 1.6291 -0.81
## valencePos 1.7603 1.3268 -0.80 0.64
## aveCueF:valencePos 5.1092 2.2604 0.53 -0.76 -0.72
## Residual 1.7177 1.3106
## Number of obs: 16905, groups: trait, 210; subID, 105
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 3.5417 0.1329 151.3867 26.656 < 2e-16 ***
## aveCueF 0.3214 0.1922 100.2925 1.672 0.0977 .
## valencePos 1.2997 0.1726 164.3521 7.529 3.21e-12 ***
## aveCueF:valencePos -0.1806 0.2685 101.6589 -0.673 0.5027
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) aveCuF vlncPs
## aveCueF -0.761
## valencePos -0.760 0.572
## avCF:vlncPs 0.512 -0.742 -0.696ggpredict(m, c("aveCueF","valence")) %>% plot()
6.7 Study 3
m <- lmer(selfResp ~ aveCueF +
( aveCueF | subID) + (1 | trait), data = reValDf3, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ aveCueF + (aveCueF | subID) + (1 | trait)
## Data: reValDf3
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 113039.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.5343 -0.6939 0.0041 0.6967 3.6884
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 1.2350 1.1113
## subID (Intercept) 0.5472 0.7398
## aveCueF 1.6143 1.2706 -0.88
## Residual 2.2169 1.4889
## Number of obs: 30669, groups: trait, 210; subID, 195
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.12713 0.09888 383.92236 41.738 <2e-16 ***
## aveCueF -0.04448 0.11106 190.75013 -0.401 0.689
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## aveCueF -0.571ggpredict(m, c("aveCueF")) %>% plot()
6.7.1 Moderated by valence
m <- lmer(selfResp ~ aveCueF * valence +
( aveCueF * valence | subID) + (1 | trait), data = reValDf3, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ aveCueF * valence + (aveCueF * valence | subID) +
## (1 | trait)
## Data: reValDf3
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 108569.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.3578 -0.6237 0.0110 0.6215 4.2488
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 0.3736 0.6112
## subID (Intercept) 1.1717 1.0825
## aveCueF 3.2975 1.8159 -0.75
## valencePos 2.6325 1.6225 -0.77 0.51
## aveCueF:valencePos 6.5287 2.5551 0.56 -0.77 -0.71
## Residual 1.8733 1.3687
## Number of obs: 30669, groups: trait, 210; subID, 195
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 3.14006 0.10728 335.58046 29.270 <2e-16 ***
## aveCueF 0.05993 0.15532 193.80432 0.386 0.700
## valencePos 1.97389 0.15652 325.51656 12.611 <2e-16 ***
## aveCueF:valencePos -0.22230 0.21888 191.52478 -1.016 0.311
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) aveCuF vlncPs
## aveCueF -0.674
## valencePos -0.738 0.465
## avCF:vlncPs 0.495 -0.752 -0.654ggpredict(m, c("aveCueF","valence")) %>% plot()
6.8 Combined
m <- lmer(selfResp ~ aveCueF * condition +
( aveCueF | subID) + (1 | trait), data = reValDf3, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ aveCueF * condition + (aveCueF | subID) + (1 | trait)
## Data: reValDf3
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 113044.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.5351 -0.6942 0.0042 0.6967 3.6901
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 1.2350 1.1113
## subID (Intercept) 0.5511 0.7424
## aveCueF 1.6268 1.2754 -0.88
## Residual 2.2169 1.4889
## Number of obs: 30669, groups: trait, 210; subID, 195
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.13395 0.11758 380.66279 35.158 <2e-16 ***
## aveCueF -0.03883 0.15888 190.62815 -0.244 0.807
## conditionRepublican -0.01353 0.12525 190.60048 -0.108 0.914
## aveCueF:conditionRepublican -0.01095 0.22291 190.36504 -0.049 0.961
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) aveCuF cndtnR
## aveCueF -0.686
## cndtnRpblcn -0.540 0.644
## avCF:cndtnR 0.489 -0.713 -0.905ggpredict(m, c("aveCueF","condition")) %>% plot()
6.8.1 Moderated by valence
m <- lmer(selfResp ~ aveCueF * valence * condition +
( aveCueF * valence | subID) + (1 | trait), data = reValDf3, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ aveCueF * valence * condition + (aveCueF * valence |
## subID) + (1 | trait)
## Data: reValDf3
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 108576.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.3573 -0.6232 0.0109 0.6218 4.2472
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 0.3737 0.6113
## subID (Intercept) 1.1796 1.0861
## aveCueF 3.3208 1.8223 -0.76
## valencePos 2.6497 1.6278 -0.77 0.51
## aveCueF:valencePos 6.5784 2.5648 0.56 -0.77 -0.72
## Residual 1.8733 1.3687
## Number of obs: 30669, groups: trait, 210; subID, 195
##
## Fixed effects:
## Estimate Std. Error df t value
## (Intercept) 3.15198 0.14059 270.81726 22.419
## aveCueF 0.09560 0.22196 193.23337 0.431
## valencePos 1.94429 0.20627 263.04820 9.426
## conditionRepublican -0.02355 0.17870 191.14344 -0.132
## aveCueF:valencePos -0.24269 0.31305 191.07823 -0.775
## aveCueF:conditionRepublican -0.07005 0.31141 192.81715 -0.225
## valencePos:conditionRepublican 0.05839 0.26447 190.60128 0.221
## aveCueF:valencePos:conditionRepublican 0.04006 0.43942 191.18131 0.091
## Pr(>|t|)
## (Intercept) <2e-16 ***
## aveCueF 0.667
## valencePos <2e-16 ***
## conditionRepublican 0.895
## aveCueF:valencePos 0.439
## aveCueF:conditionRepublican 0.822
## valencePos:conditionRepublican 0.826
## aveCueF:valencePos:conditionRepublican 0.927
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) aveCuF vlncPs cndtnR avCF:P avCF:R vlnP:R
## aveCueF -0.734
## valencePos -0.744 0.504
## cndtnRpblcn -0.645 0.577 0.489
## avCF:vlncPs 0.539 -0.752 -0.709 -0.424
## avCF:cndtnR 0.523 -0.713 -0.359 -0.811 0.536
## vlncPs:cndR 0.485 -0.393 -0.650 -0.752 0.553 0.551
## avCF:vlnP:R -0.384 0.536 0.505 0.595 -0.713 -0.752 -0.777ggpredict(m, c("aveCueF","valence","condition")) %>% plot()
6.8.2 Ingroup vs. Outgroup
m <- lmer(selfResp ~ aveCueF * ingroup +
( aveCueF | subID) + (1 | traits), data = reValDf, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ aveCueF * ingroup + (aveCueF | subID) + (1 | traits)
## Data: reValDf
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 214065.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.5432 -0.6920 0.0194 0.7041 3.4869
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.6893 0.8302
## aveCueF 2.0150 1.4195 -0.89
## traits (Intercept) 1.0024 1.0012
## Residual 2.1130 1.4536
## Number of obs: 58974, groups: subID, 376; traits, 257
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.18116 0.08468 571.29988 49.377 <2e-16 ***
## aveCueF 0.22339 0.10078 364.42329 2.216 0.0273 *
## ingroupOut -0.03038 0.11097 359.92013 -0.274 0.7844
## aveCueF:ingroupOut -0.26734 0.19519 356.39919 -1.370 0.1717
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) aveCuF ingrpO
## aveCueF -0.615
## ingroupOut -0.347 0.470
## avCF:ngrpOt 0.317 -0.516 -0.910ggpredict(m, c("aveCueF","ingroup")) %>% plot()
6.8.3 Ingroup vs. Outgroup moderated by valence
m <- lmer(selfResp ~ aveCueF * ingroup * valence +
( aveCueF + valence | subID) + (1 | traits), data = reValDf, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ aveCueF * ingroup * valence + (aveCueF + valence |
## subID) + (1 | traits)
## Data: reValDf
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 206681.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.3609 -0.6358 0.0224 0.6450 4.3225
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 3.3806 1.8386
## aveCueF 1.8401 1.3565 -0.38
## valence 1.1285 1.0623 -0.90 -0.01
## traits (Intercept) 0.3458 0.5881
## Residual 1.8320 1.3535
## Number of obs: 58974, groups: subID, 376; traits, 257
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 1.689e+00 1.830e-01 8.459e+02 9.233 <2e-16
## aveCueF 4.430e-01 1.799e-01 4.110e+03 2.462 0.0138
## ingroupOut -5.777e-01 2.679e-01 7.508e+02 -2.157 0.0313
## valence 1.611e+00 1.108e-01 8.825e+02 14.537 <2e-16
## aveCueF:ingroupOut -1.925e-01 3.449e-01 3.864e+03 -0.558 0.5767
## aveCueF:valence -1.439e-01 1.013e-01 5.842e+04 -1.421 0.1553
## ingroupOut:valence 3.713e-01 1.604e-01 9.468e+02 2.314 0.0209
## aveCueF:ingroupOut:valence -6.513e-02 1.939e-01 5.835e+04 -0.336 0.7370
##
## (Intercept) ***
## aveCueF *
## ingroupOut *
## valence ***
## aveCueF:ingroupOut
## aveCueF:valence
## ingroupOut:valence *
## aveCueF:ingroupOut:valence
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) aveCuF ingrpO valenc avCF:O avCF:v ingrO:
## aveCueF -0.496
## ingroupOut -0.390 0.340
## valence -0.932 0.385 0.353
## avCF:ngrpOt 0.259 -0.521 -0.650 -0.201
## aveCuF:vlnc 0.417 -0.845 -0.287 -0.458 0.441
## ingrpOt:vln 0.357 -0.268 -0.917 -0.384 0.509 0.320
## avCF:ngrpO: -0.218 0.442 0.546 0.240 -0.842 -0.523 -0.609ggpredict(m, c("aveCueF","valence","ingroup")) %>% plot()
7 Question 9: Does final value estimate predict re-evaluations?
7.1 Study 1
m <- lmer(selfResp ~ valEstF +
( valEstF | subID) + (1 | trait), data = reValDf1, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ valEstF + (valEstF | subID) + (1 | trait)
## Data: reValDf1
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 33340.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.0875 -0.5762 0.0710 0.6437 4.2275
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 0.290 0.5385
## subID (Intercept) 1.923 1.3866
## valEstF 8.097 2.8455 -0.88
## Residual 1.629 1.2762
## Number of obs: 9781, groups: trait, 149; subID, 76
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 2.1760 0.2053 160.3426 10.60 <2e-16 ***
## valEstF 6.4781 0.4754 182.6982 13.63 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## valEstF -0.900ggpredict(m, c("valEstF")) %>% plot()
m <- lmer(selfResp ~ SV_F + desirability +
( SV_F + desirability | subID) + (1 | trait), data = reValDf1, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ SV_F + desirability + (SV_F + desirability | subID) +
## (1 | trait)
## Data: reValDf1
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 33228.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.2385 -0.5805 0.0634 0.6414 4.3023
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 0.34890 0.5907
## subID (Intercept) 1.84504 1.3583
## SV_F 18.33953 4.2825 -0.25
## desirability 0.07857 0.2803 -0.81 -0.29
## Residual 1.59769 1.2640
## Number of obs: 9781, groups: trait, 149; subID, 76
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 1.51407 0.23063 191.89346 6.565 4.74e-10 ***
## SV_F 10.55323 1.13063 418.54893 9.334 < 2e-16 ***
## desirability 0.33274 0.04765 187.01602 6.983 4.89e-11 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) SV_F
## SV_F -0.409
## desirabilty -0.642 -0.378ggpredict(m, c("SV_F")) %>% plot()
7.2 Study 2
m <- lmer(selfResp ~ valEstF +
( valEstF | subID) + (1 | trait), data = reValDf2, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ valEstF + (valEstF | subID) + (1 | trait)
## Data: reValDf2
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 58397.7
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.8939 -0.6329 0.0301 0.6649 3.8414
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 0.3217 0.5672
## subID (Intercept) 1.4330 1.1971
## valEstF 13.8465 3.7211 -0.55
## Residual 1.7168 1.3103
## Number of obs: 16905, groups: trait, 210; subID, 105
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 1.5510 0.1489 211.2656 10.42 <2e-16 ***
## valEstF 9.9293 0.4940 163.3564 20.10 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## valEstF -0.691ggpredict(m, c("valEstF")) %>% plot()
m <- lmer(selfResp ~ SV_F + desirability +
( SV_F + desirability | subID) + (1 | trait), data = reValDf2, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ SV_F + desirability + (SV_F + desirability | subID) +
## (1 | trait)
## Data: reValDf2
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 57837.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.1718 -0.6367 0.0231 0.6552 3.8248
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 0.65349 0.8084
## subID (Intercept) 0.87598 0.9359
## SV_F 7.15172 2.6743 -0.14
## desirability 0.04389 0.2095 -0.79 -0.39
## Residual 1.66002 1.2884
## Number of obs: 16905, groups: trait, 210; subID, 105
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 3.708e-01 2.040e-01 3.075e+02 1.818 0.07008 .
## SV_F 2.359e+01 9.247e-01 1.081e+03 25.511 < 2e-16 ***
## desirability 1.389e-01 4.263e-02 2.747e+02 3.258 0.00126 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) SV_F
## SV_F -0.397
## desirabilty -0.707 -0.289ggpredict(m, c("SV_F")) %>% plot()
7.3 Study 3
m <- lmer(selfResp ~ valEstF +
( valEstF | subID) + (1 | trait), data = reValDf3, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ valEstF + (valEstF | subID) + (1 | trait)
## Data: reValDf3
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 108763.8
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.3820 -0.6333 0.0083 0.6282 4.3181
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 0.2487 0.4987
## subID (Intercept) 2.5164 1.5863
## valEstF 17.3264 4.1625 -0.80
## Residual 1.9004 1.3786
## Number of obs: 30669, groups: trait, 210; subID, 195
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.9471 0.1395 367.6710 6.789 4.55e-11 ***
## valEstF 8.8962 0.3703 325.8350 24.023 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## valEstF -0.834ggpredict(m, c("valEstF")) %>% plot()
m <- lmer(selfResp ~ SV_F + desirability +
( SV_F + desirability | subID) + (1 | trait), data = reValDf3, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ SV_F + desirability + (SV_F + desirability | subID) +
## (1 | trait)
## Data: reValDf3
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 107447.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.4068 -0.6215 0.0069 0.6244 4.4215
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 0.5687 0.7541
## subID (Intercept) 1.8572 1.3628
## SV_F 15.0848 3.8839 -0.05
## desirability 0.1036 0.3218 -0.87 -0.38
## Residual 1.8214 1.3496
## Number of obs: 30669, groups: trait, 210; subID, 195
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) -0.30119 0.18856 336.78271 -1.597 0.111
## SV_F 22.34699 0.76047 1767.84315 29.386 < 2e-16 ***
## desirability 0.29008 0.04195 346.66181 6.915 2.26e-11 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) SV_F
## SV_F -0.275
## desirabilty -0.776 -0.316ggpredict(m, c("SV_F")) %>% plot()
7.4 Combined
m <- lmer(selfResp ~ valEstF * condition +
( valEstF | subID) + (1 | traits), data = reValDf, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ valEstF * condition + (valEstF | subID) + (1 | traits)
## Data: reValDf
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 200470.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.4910 -0.6287 0.0249 0.6444 4.3879
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 2.1084 1.4520
## valEstF 14.0568 3.7492 -0.76
## traits (Intercept) 0.2499 0.4999
## Residual 1.8171 1.3480
## Number of obs: 57355, groups: subID, 376; traits, 257
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 8.310e-01 1.619e-01 4.852e+02 5.133 4.12e-07 ***
## valEstF 9.149e+00 4.203e-01 4.073e+02 21.770 < 2e-16 ***
## conditionAsian 9.525e-01 2.103e-01 3.704e+02 4.529 7.99e-06 ***
## conditionDemocrat 4.196e-04 2.143e-01 3.655e+02 0.002 0.998439
## conditionLatino 8.748e-01 2.308e-01 3.826e+02 3.790 0.000175 ***
## valEstF:conditionAsian -3.673e-01 5.642e-01 3.554e+02 -0.651 0.515510
## valEstF:conditionDemocrat 8.464e-03 5.618e-01 3.264e+02 0.015 0.987990
## valEstF:conditionLatino -1.474e+00 6.020e-01 3.382e+02 -2.449 0.014814 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) vlEstF cndtnA cndtnD cndtnL vlEF:A vlEF:D
## valEstF -0.782
## conditinAsn -0.661 0.513
## condtnDmcrt -0.655 0.510 0.502
## conditinLtn -0.608 0.473 0.466 0.457
## vlEstF:cndA 0.470 -0.618 -0.769 -0.377 -0.350
## vlEstF:cndD 0.505 -0.658 -0.386 -0.772 -0.352 0.487
## vlEstF:cndL 0.472 -0.617 -0.361 -0.354 -0.776 0.454 0.457ggpredict(m, c("valEstF","condition")) %>% plot()
m <- lmer(selfResp ~ SV_F * condition + desirability * condition +
( SV_F + desirability | subID) + (1 | traits), data = reValDf, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ SV_F * condition + desirability * condition + (SV_F +
## desirability | subID) + (1 | traits)
## Data: reValDf
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 198397.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.4999 -0.6260 0.0230 0.6378 4.4755
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 1.61983 1.2727
## SV_F 13.97091 3.7378 -0.13
## desirability 0.08364 0.2892 -0.85 -0.34
## traits (Intercept) 0.53567 0.7319
## Residual 1.75475 1.3247
## Number of obs: 57355, groups: subID, 376; traits, 257
##
## Fixed effects:
## Estimate Std. Error df t value
## (Intercept) -3.468e-01 1.957e-01 5.751e+02 -1.772
## SV_F 2.138e+01 7.052e-01 1.136e+03 30.314
## conditionAsian 7.928e-01 1.889e-01 3.681e+02 4.197
## conditionDemocrat -2.700e-02 1.924e-01 3.627e+02 -0.140
## conditionLatino 9.945e-01 2.091e-01 3.921e+02 4.757
## desirability 3.388e-01 4.278e-02 5.949e+02 7.920
## SV_F:conditionAsian -1.930e-02 7.329e-01 3.838e+02 -0.026
## SV_F:conditionDemocrat 2.093e-01 7.495e-01 3.827e+02 0.279
## SV_F:conditionLatino -2.640e+00 9.187e-01 6.625e+02 -2.873
## conditionAsian:desirability -1.361e-01 4.234e-02 3.705e+02 -3.214
## conditionDemocrat:desirability 3.014e-03 4.340e-02 3.740e+02 0.069
## conditionLatino:desirability -1.168e-01 4.817e-02 4.401e+02 -2.425
## Pr(>|t|)
## (Intercept) 0.07700 .
## SV_F < 2e-16 ***
## conditionAsian 3.39e-05 ***
## conditionDemocrat 0.88850
## conditionLatino 2.76e-06 ***
## desirability 1.17e-14 ***
## SV_F:conditionAsian 0.97900
## SV_F:conditionDemocrat 0.78023
## SV_F:conditionLatino 0.00419 **
## conditionAsian:desirability 0.00142 **
## conditionDemocrat:desirability 0.94467
## conditionLatino:desirability 0.01572 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) SV_F cndtnA cndtnD cndtnL dsrblt SV_F:A SV_F:D SV_F:L
## SV_F -0.245
## conditinAsn -0.483 0.093
## condtnDmcrt -0.481 0.114 0.501
## conditinLtn -0.451 0.128 0.460 0.452
## desirabilty -0.807 -0.293 0.405 0.392 0.355
## SV_F:cndtnA 0.112 -0.531 -0.243 -0.116 -0.107 0.184
## SV_F:cndtnD 0.112 -0.527 -0.115 -0.233 -0.104 0.183 0.504
## SV_F:cndtnL 0.094 -0.499 -0.089 -0.092 -0.260 0.194 0.410 0.404
## cndtnAsn:ds 0.380 0.223 -0.784 -0.395 -0.363 -0.519 -0.349 -0.184 -0.154
## cndtnDmcrt: 0.377 0.197 -0.393 -0.783 -0.354 -0.502 -0.183 -0.359 -0.148
## cndtnLtn:ds 0.353 0.180 -0.355 -0.347 -0.726 -0.474 -0.164 -0.162 -0.421
## cndtA: cndtD:
## SV_F
## conditinAsn
## condtnDmcrt
## conditinLtn
## desirabilty
## SV_F:cndtnA
## SV_F:cndtnD
## SV_F:cndtnL
## cndtnAsn:ds
## cndtnDmcrt: 0.506
## cndtnLtn:ds 0.456 0.445ggpredict(m, c("SV_F","condition")) %>% plot()
7.4.1 Ingroup vs. Outgroup
m <- lmer(selfResp ~ valEstF * ingroup +
( valEstF | subID) + (1 | traits), data = reValDf, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ valEstF * ingroup + (valEstF | subID) + (1 | traits)
## Data: reValDf
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 200511
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.4902 -0.6284 0.0249 0.6455 4.3822
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 2.2509 1.5003
## valEstF 14.2245 3.7715 -0.76
## traits (Intercept) 0.2493 0.4993
## Residual 1.8173 1.3481
## Number of obs: 57355, groups: subID, 376; traits, 257
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 1.4562 0.1104 692.6299 13.189 < 2e-16 ***
## valEstF 8.5140 0.2905 613.5862 29.304 < 2e-16 ***
## ingroupOut -0.6042 0.1810 370.6886 -3.338 0.000931 ***
## valEstF:ingroupOut 0.5711 0.4637 333.7180 1.232 0.218989
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) vlEstF ingrpO
## valEstF -0.792
## ingroupOut -0.435 0.329
## vlEstF:ngrO 0.354 -0.452 -0.767ggpredict(m, c("valEstF","ingroup")) %>% plot()
m <- lmer(selfResp ~ SV_F * ingroup + desirability * ingroup +
( SV_F + desirability | subID) + (1 | traits), data = reValDf, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: selfResp ~ SV_F * ingroup + desirability * ingroup + (SV_F +
## desirability | subID) + (1 | traits)
## Data: reValDf
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 198423.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.5025 -0.6258 0.0228 0.6383 4.4694
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 1.76180 1.3273
## SV_F 14.50395 3.8084 -0.16
## desirability 0.08693 0.2948 -0.85 -0.30
## traits (Intercept) 0.53800 0.7335
## Residual 1.75466 1.3246
## Number of obs: 57355, groups: subID, 376; traits, 257
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.21904 0.16576 427.67036 1.321 0.187067
## SV_F 20.65930 0.55237 2567.00048 37.401 < 2e-16 ***
## ingroupOut -0.55322 0.16376 369.57117 -3.378 0.000807 ***
## desirability 0.25762 0.03497 409.85356 7.368 9.63e-13 ***
## SV_F:ingroupOut 0.46417 0.62760 396.41785 0.740 0.459991
## ingroupOut:desirability 0.08755 0.03626 380.80038 2.415 0.016220 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) SV_F ingrpO dsrblt SV_F:O
## SV_F -0.279
## ingroupOut -0.267 0.096
## desirabilty -0.824 -0.222 0.212
## SV_F:ngrpOt 0.071 -0.293 -0.252 0.083
## ingrpOt:dsr 0.205 0.088 -0.787 -0.266 -0.338ggpredict(m, c("SV_F","ingroup")) %>% plot()
8 Question 10: Does the effect of category representativeness on re-evaluations depend on centrality?
8.1 Study 1
m <- lmer(scale(selfResp) ~ scale(SV_F)*scale(outDegree) + scale(desirability) +
( scale(SV_F) + scale(outDegree) | subID) + (1 | trait), data = reValDf1, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula:
## scale(selfResp) ~ scale(SV_F) * scale(outDegree) + scale(desirability) +
## (scale(SV_F) + scale(outDegree) | subID) + (1 | trait)
## Data: reValDf1
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 22365.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.2024 -0.5955 0.0725 0.6588 3.9132
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 0.103775 0.32214
## subID (Intercept) 0.055005 0.23453
## scale(SV_F) 0.033231 0.18229 0.00
## scale(outDegree) 0.005011 0.07079 -0.27 -0.86
## Residual 0.532639 0.72982
## Number of obs: 9781, groups: trait, 149; subID, 76
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.082278 0.039046 184.536140 2.107 0.0364
## scale(SV_F) 0.233092 0.030595 199.896502 7.619 1e-12
## scale(outDegree) -0.031167 0.030299 178.846041 -1.029 0.3050
## scale(desirability) 0.313127 0.029823 173.382857 10.499 <2e-16
## scale(SV_F):scale(outDegree) 0.004535 0.014950 912.257836 0.303 0.7617
##
## (Intercept) *
## scale(SV_F) ***
## scale(outDegree)
## scale(desirability) ***
## scale(SV_F):scale(outDegree)
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) sc(SV_F) scl(D) scl(d)
## scale(SV_F) 0.035
## scale(tDgr) -0.040 -0.382
## scl(dsrblt) -0.005 -0.266 0.101
## s(SV_F):(D) -0.173 -0.049 -0.098 -0.049ggpredict(m, c("SV_F", "outDegree" )) %>% plot()
m <- lmer(scale(selfResp) ~ scale(SV_F)*scale(inDegree) + scale(desirability) +
( scale(SV_F) + scale(inDegree) | subID) + (1 | trait), data = reValDf1, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))## boundary (singular) fit: see ?isSingularsummary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula:
## scale(selfResp) ~ scale(SV_F) * scale(inDegree) + scale(desirability) +
## (scale(SV_F) + scale(inDegree) | subID) + (1 | trait)
## Data: reValDf1
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 22305.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.1323 -0.5960 0.0675 0.6530 4.0441
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 0.093813 0.30629
## subID (Intercept) 0.054938 0.23439
## scale(SV_F) 0.036283 0.19048 -0.09
## scale(inDegree) 0.007811 0.08838 0.16 -1.00
## Residual 0.530592 0.72842
## Number of obs: 9781, groups: trait, 149; subID, 76
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.09161 0.03818 183.74820 2.400 0.017406
## scale(SV_F) 0.24021 0.03086 190.98529 7.783 4.35e-13
## scale(inDegree) -0.11234 0.02967 201.90514 -3.786 0.000202
## scale(desirability) 0.30466 0.02881 188.82469 10.574 < 2e-16
## scale(SV_F):scale(inDegree) -0.01787 0.01643 1490.30498 -1.087 0.277036
##
## (Intercept) *
## scale(SV_F) ***
## scale(inDegree) ***
## scale(desirability) ***
## scale(SV_F):scale(inDegree)
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) sc(SV_F) scl(D) scl(d)
## scale(SV_F) -0.014
## scale(nDgr) 0.040 -0.462
## scl(dsrblt) -0.011 -0.279 0.158
## s(SV_F):(D) -0.179 -0.025 -0.073 -0.024
## optimizer (bobyqa) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingularggpredict(m, c("SV_F","inDegree")) %>% plot()
8.2 Study 2
m <- lmer(scale(selfResp) ~ scale(SV_F)*scale(outDegree) + scale(desirability) +
( scale(SV_F) + scale(outDegree) | subID) + (1 | trait), data = reValDf2, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula:
## scale(selfResp) ~ scale(SV_F) * scale(outDegree) + scale(desirability) +
## (scale(SV_F) + scale(outDegree) | subID) + (1 | trait)
## Data: reValDf2
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 40775.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.2763 -0.6407 0.0254 0.6647 3.5941
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 0.185643 0.43086
## subID (Intercept) 0.045980 0.21443
## scale(SV_F) 0.017397 0.13190 -0.11
## scale(outDegree) 0.003026 0.05501 -0.04 -0.96
## Residual 0.611362 0.78190
## Number of obs: 16905, groups: trait, 210; subID, 105
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) -0.01572 0.03736 286.88163 -0.421 0.67426
## scale(SV_F) 0.58504 0.01989 267.96506 29.410 < 2e-16
## scale(outDegree) -0.28136 0.03181 216.79032 -8.846 3.23e-16
## scale(desirability) 0.09700 0.03119 205.80457 3.110 0.00214
## scale(SV_F):scale(outDegree) 0.03611 0.01215 4379.68562 2.971 0.00299
##
## (Intercept)
## scale(SV_F) ***
## scale(outDegree) ***
## scale(desirability) **
## scale(SV_F):scale(outDegree) **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) sc(SV_F) scl(D) scl(d)
## scale(SV_F) -0.029
## scale(tDgr) 0.000 -0.280
## scl(dsrblt) 0.009 -0.154 0.019
## s(SV_F):(D) -0.164 -0.063 -0.034 -0.054ggpredict(m, c("SV_F", "outDegree" )) %>% plot()
m <- lmer(scale(selfResp) ~ scale(SV_F)*scale(inDegree) + scale(desirability) +
( scale(SV_F) + scale(inDegree) | subID) + (1 | trait), data = reValDf2, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))## boundary (singular) fit: see ?isSingularsummary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula:
## scale(selfResp) ~ scale(SV_F) * scale(inDegree) + scale(desirability) +
## (scale(SV_F) + scale(inDegree) | subID) + (1 | trait)
## Data: reValDf2
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 40732
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.2855 -0.6395 0.0264 0.6633 3.7830
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 0.18066 0.42504
## subID (Intercept) 0.04610 0.21472
## scale(SV_F) 0.01929 0.13889 -0.15
## scale(inDegree) 0.00515 0.07176 0.11 -1.00
## Residual 0.60966 0.78080
## Number of obs: 16905, groups: trait, 210; subID, 105
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) -7.591e-03 3.698e-02 2.884e+02 -0.205 0.83751
## scale(SV_F) 5.811e-01 2.019e-02 2.643e+02 28.778 < 2e-16
## scale(inDegree) -2.954e-01 3.148e-02 2.256e+02 -9.384 < 2e-16
## scale(desirability) 9.971e-02 3.077e-02 2.078e+02 3.241 0.00139
## scale(SV_F):scale(inDegree) 1.728e-02 1.216e-02 7.541e+03 1.422 0.15521
##
## (Intercept)
## scale(SV_F) ***
## scale(inDegree) ***
## scale(desirability) **
## scale(SV_F):scale(inDegree)
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) sc(SV_F) scl(D) scl(d)
## scale(SV_F) -0.051
## scale(nDgr) 0.022 -0.307
## scl(dsrblt) -0.001 -0.158 0.028
## s(SV_F):(D) -0.153 -0.043 -0.036 0.006
## optimizer (bobyqa) convergence code: 0 (OK)
## boundary (singular) fit: see ?isSingularggpredict(m, c("SV_F","inDegree")) %>% plot()
8.3 Study 3
m <- lmer(scale(selfResp) ~ scale(SV_F)*scale(outDegree) + scale(desirability) +
( scale(SV_F) + scale(outDegree) | subID) + (1 | trait), data = reValDf3, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula:
## scale(selfResp) ~ scale(SV_F) * scale(outDegree) + scale(desirability) +
## (scale(SV_F) + scale(outDegree) | subID) + (1 | trait)
## Data: reValDf3
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 70047.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.9596 -0.6457 0.0035 0.6480 3.9375
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 0.134593 0.36687
## subID (Intercept) 0.036187 0.19023
## scale(SV_F) 0.051512 0.22696 -0.06
## scale(outDegree) 0.008657 0.09304 -0.04 -0.98
## Residual 0.543110 0.73696
## Number of obs: 30669, groups: trait, 210; subID, 195
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 1.763e-03 2.930e-02 2.934e+02 0.060 0.952
## scale(SV_F) 5.020e-01 2.009e-02 3.372e+02 24.992 < 2e-16
## scale(outDegree) -1.815e-01 2.697e-02 2.298e+02 -6.731 1.33e-10
## scale(desirability) 2.384e-01 2.652e-02 2.140e+02 8.989 < 2e-16
## scale(SV_F):scale(outDegree) -5.460e-03 8.273e-03 9.493e+03 -0.660 0.509
##
## (Intercept)
## scale(SV_F) ***
## scale(outDegree) ***
## scale(desirability) ***
## scale(SV_F):scale(outDegree)
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) sc(SV_F) scl(D) scl(d)
## scale(SV_F) -0.016
## scale(tDgr) -0.002 -0.306
## scl(dsrblt) 0.006 -0.133 0.015
## s(SV_F):(D) -0.129 -0.057 -0.029 -0.038ggpredict(m, c("SV_F", "outDegree" )) %>% plot()
m <- lmer(scale(selfResp) ~ scale(SV_F)*scale(inDegree) + scale(desirability) +
( scale(SV_F) + scale(inDegree) | subID) + (1 | trait), data = reValDf3, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula:
## scale(selfResp) ~ scale(SV_F) * scale(inDegree) + scale(desirability) +
## (scale(SV_F) + scale(inDegree) | subID) + (1 | trait)
## Data: reValDf3
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 70012.7
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.9564 -0.6463 0.0047 0.6448 3.7467
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## trait (Intercept) 0.113866 0.33744
## subID (Intercept) 0.036128 0.19007
## scale(SV_F) 0.049646 0.22281 -0.08
## scale(inDegree) 0.008513 0.09227 0.04 -0.98
## Residual 0.543060 0.73693
## Number of obs: 30669, groups: trait, 210; subID, 195
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 1.145e-03 2.753e-02 3.124e+02 0.042 0.967
## scale(SV_F) 4.936e-01 1.968e-02 3.389e+02 25.079 <2e-16
## scale(inDegree) -2.231e-01 2.505e-02 2.376e+02 -8.909 <2e-16
## scale(desirability) 2.399e-01 2.454e-02 2.215e+02 9.774 <2e-16
## scale(SV_F):scale(inDegree) -6.387e-03 8.385e-03 1.168e+04 -0.762 0.446
##
## (Intercept)
## scale(SV_F) ***
## scale(inDegree) ***
## scale(desirability) ***
## scale(SV_F):scale(inDegree)
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) sc(SV_F) scl(D) scl(d)
## scale(SV_F) -0.025
## scale(nDgr) 0.007 -0.319
## scl(dsrblt) 0.000 -0.145 0.030
## s(SV_F):(D) -0.128 -0.047 -0.013 0.008ggpredict(m, c("SV_F","inDegree")) %>% plot()
8.4 Combined
m <- lmer(scale(selfResp) ~ scale(SV_F)*scale(outDegree)*condition + scale(desirability) +
( scale(SV_F) + scale(outDegree) | subID) + (1 | traits), data = reValDf, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: scale(selfResp) ~ scale(SV_F) * scale(outDegree) * condition +
## scale(desirability) + (scale(SV_F) + scale(outDegree) | subID) +
## (1 | traits)
## Data: reValDf
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 132787.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.1407 -0.6449 0.0228 0.6569 3.9924
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.042040 0.20504
## scale(SV_F) 0.040629 0.20157 -0.02
## scale(outDegree) 0.006841 0.08271 -0.12 -0.97
## traits (Intercept) 0.144138 0.37966
## Residual 0.563943 0.75096
## Number of obs: 57355, groups: subID, 376; traits, 257
##
## Fixed effects:
## Estimate Std. Error df
## (Intercept) -2.944e-02 3.239e-02 5.525e+02
## scale(SV_F) 5.202e-01 2.298e-02 4.405e+02
## scale(outDegree) -1.767e-01 2.661e-02 3.406e+02
## conditionAsian 1.150e-01 3.013e-02 3.773e+02
## conditionDemocrat 2.336e-03 3.080e-02 3.772e+02
## conditionLatino 8.102e-02 3.386e-02 4.286e+02
## scale(desirability) 2.213e-01 2.452e-02 2.503e+02
## scale(SV_F):scale(outDegree) -9.579e-03 7.826e-03 3.295e+04
## scale(SV_F):conditionAsian -1.327e-02 3.055e-02 3.702e+02
## scale(SV_F):conditionDemocrat 6.969e-02 3.100e-02 3.595e+02
## scale(SV_F):conditionLatino -1.331e-01 3.351e-02 3.798e+02
## scale(outDegree):conditionAsian -2.775e-02 1.536e-02 3.708e+02
## scale(outDegree):conditionDemocrat -3.708e-02 1.564e-02 3.650e+02
## scale(outDegree):conditionLatino 5.842e-02 1.874e-02 5.840e+02
## scale(SV_F):scale(outDegree):conditionAsian 2.208e-03 7.842e-03 5.625e+04
## scale(SV_F):scale(outDegree):conditionDemocrat 2.526e-02 7.614e-03 5.598e+04
## scale(SV_F):scale(outDegree):conditionLatino -9.727e-03 8.875e-03 5.355e+04
## t value Pr(>|t|)
## (Intercept) -0.909 0.363874
## scale(SV_F) 22.631 < 2e-16 ***
## scale(outDegree) -6.640 1.24e-10 ***
## conditionAsian 3.818 0.000157 ***
## conditionDemocrat 0.076 0.939580
## conditionLatino 2.393 0.017151 *
## scale(desirability) 9.028 < 2e-16 ***
## scale(SV_F):scale(outDegree) -1.224 0.220934
## scale(SV_F):conditionAsian -0.434 0.664425
## scale(SV_F):conditionDemocrat 2.248 0.025176 *
## scale(SV_F):conditionLatino -3.974 8.47e-05 ***
## scale(outDegree):conditionAsian -1.806 0.071689 .
## scale(outDegree):conditionDemocrat -2.371 0.018276 *
## scale(outDegree):conditionLatino 3.117 0.001916 **
## scale(SV_F):scale(outDegree):conditionAsian 0.282 0.778265
## scale(SV_F):scale(outDegree):conditionDemocrat 3.318 0.000908 ***
## scale(SV_F):scale(outDegree):conditionLatino -1.096 0.273088
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1##
## Correlation matrix not shown by default, as p = 17 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need itggpredict(m, c("SV_F", "outDegree" ,"condition")) %>% plot()
m <- lmer(scale(selfResp) ~ scale(SV_F)*scale(inDegree)*condition + scale(desirability) +
( scale(SV_F) + scale(inDegree) | subID) + (1 | traits), data = reValDf, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: scale(selfResp) ~ scale(SV_F) * scale(inDegree) * condition +
## scale(desirability) + (scale(SV_F) + scale(inDegree) | subID) +
## (1 | traits)
## Data: reValDf
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 132682.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.0965 -0.6461 0.0196 0.6546 3.8627
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.042229 0.20550
## scale(SV_F) 0.040140 0.20035 -0.07
## scale(inDegree) 0.007474 0.08645 0.06 -0.99
## traits (Intercept) 0.116925 0.34194
## Residual 0.563483 0.75066
## Number of obs: 57355, groups: subID, 376; traits, 257
##
## Fixed effects:
## Estimate Std. Error df
## (Intercept) -2.334e-02 3.071e-02 5.934e+02
## scale(SV_F) 5.194e-01 2.274e-02 4.420e+02
## scale(inDegree) -2.475e-01 2.407e-02 3.858e+02
## conditionAsian 1.041e-01 3.022e-02 3.788e+02
## conditionDemocrat -2.889e-03 3.089e-02 3.784e+02
## conditionLatino 8.586e-02 3.397e-02 4.308e+02
## scale(desirability) 2.118e-01 2.225e-02 2.605e+02
## scale(SV_F):scale(inDegree) -2.178e-02 8.255e-03 3.856e+04
## scale(SV_F):conditionAsian -3.576e-02 3.024e-02 3.711e+02
## scale(SV_F):conditionDemocrat 5.928e-02 3.070e-02 3.609e+02
## scale(SV_F):conditionLatino -1.340e-01 3.320e-02 3.826e+02
## scale(inDegree):conditionAsian 1.218e-02 1.549e-02 3.577e+02
## scale(inDegree):conditionDemocrat -1.778e-02 1.577e-02 3.521e+02
## scale(inDegree):conditionLatino 6.815e-02 1.979e-02 6.743e+02
## scale(SV_F):scale(inDegree):conditionAsian 2.466e-02 9.119e-03 5.499e+04
## scale(SV_F):scale(inDegree):conditionDemocrat 3.693e-02 8.759e-03 5.481e+04
## scale(SV_F):scale(inDegree):conditionLatino -1.009e-02 1.075e-02 5.526e+04
## t value Pr(>|t|)
## (Intercept) -0.760 0.447495
## scale(SV_F) 22.842 < 2e-16 ***
## scale(inDegree) -10.283 < 2e-16 ***
## conditionAsian 3.444 0.000638 ***
## conditionDemocrat -0.094 0.925540
## conditionLatino 2.527 0.011848 *
## scale(desirability) 9.520 < 2e-16 ***
## scale(SV_F):scale(inDegree) -2.639 0.008318 **
## scale(SV_F):conditionAsian -1.183 0.237757
## scale(SV_F):conditionDemocrat 1.931 0.054238 .
## scale(SV_F):conditionLatino -4.036 6.58e-05 ***
## scale(inDegree):conditionAsian 0.787 0.432012
## scale(inDegree):conditionDemocrat -1.128 0.260264
## scale(inDegree):conditionLatino 3.444 0.000608 ***
## scale(SV_F):scale(inDegree):conditionAsian 2.704 0.006844 **
## scale(SV_F):scale(inDegree):conditionDemocrat 4.217 2.48e-05 ***
## scale(SV_F):scale(inDegree):conditionLatino -0.939 0.347840
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1##
## Correlation matrix not shown by default, as p = 17 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need itggpredict(m, c("SV_F","inDegree", "condition")) %>% plot()
8.4.1 Ingroup vs. Outgroup
m <- lmer( scale(selfResp) ~ scale(SV_F)*ingroup*scale(outDegree) + scale(desirability) +
( scale(SV_F) + scale(outDegree) | subID) + (1 | traits), data = reValDf, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: scale(selfResp) ~ scale(SV_F) * ingroup * scale(outDegree) +
## scale(desirability) + (scale(SV_F) + scale(outDegree) | subID) +
## (1 | traits)
## Data: reValDf
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 132804.8
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.1364 -0.6458 0.0220 0.6565 3.9947
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.043168 0.2078
## scale(SV_F) 0.044906 0.2119 -0.05
## scale(outDegree) 0.007534 0.0868 -0.11 -0.96
## traits (Intercept) 0.143185 0.3784
## Residual 0.564096 0.7511
## Number of obs: 57355, groups: subID, 376; traits, 257
##
## Fixed effects:
## Estimate Std. Error df
## (Intercept) 3.341e-02 2.723e-02 3.760e+02
## scale(SV_F) 5.027e-01 1.558e-02 6.072e+02
## ingroupOut -6.407e-02 2.558e-02 3.845e+02
## scale(outDegree) -1.864e-01 2.500e-02 2.775e+02
## scale(desirability) 2.210e-01 2.444e-02 2.510e+02
## scale(SV_F):ingroupOut 1.540e-02 2.655e-02 3.679e+02
## scale(SV_F):scale(outDegree) 1.603e-03 6.436e-03 1.431e+04
## ingroupOut:scale(outDegree) 1.254e-02 1.328e-02 3.734e+02
## scale(SV_F):ingroupOut:scale(outDegree) -9.209e-03 6.304e-03 5.509e+04
## t value Pr(>|t|)
## (Intercept) 1.227 0.2207
## scale(SV_F) 32.274 < 2e-16 ***
## ingroupOut -2.505 0.0127 *
## scale(outDegree) -7.454 1.16e-12 ***
## scale(desirability) 9.045 < 2e-16 ***
## scale(SV_F):ingroupOut 0.580 0.5623
## scale(SV_F):scale(outDegree) 0.249 0.8033
## ingroupOut:scale(outDegree) 0.944 0.3455
## scale(SV_F):ingroupOut:scale(outDegree) -1.461 0.1441
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) sc(SV_F) ingrpO scl(D) scl(d) sc(SV_F):O s(SV_F):( iO:(D)
## scale(SV_F) -0.009
## ingroupOut -0.244 0.038
## scale(tDgr) 0.024 -0.256 0.003
## scl(dsrblt) -0.061 -0.080 -0.002 0.006
## scl(SV_F):O 0.013 -0.455 -0.043 0.115 0.000
## s(SV_F):(D) -0.108 -0.052 0.027 -0.045 -0.026 -0.016
## ingrpOt:(D) 0.011 0.366 -0.075 -0.139 -0.001 -0.808 0.066
## s(SV_F):O:( 0.031 -0.030 -0.111 0.026 0.003 0.042 -0.280 -0.103ggpredict(m, c("SV_F", "outDegree" ,"ingroup")) %>% plot()
m <- lmer(scale(selfResp) ~ scale(SV_F)*scale(inDegree)*ingroup + scale(desirability) +
( scale(SV_F) + scale(inDegree) | subID) + (1 | traits), data = reValDf, control = lmerControl(optimizer = "bobyqa",
optCtrl = list(maxfun = 100000)))
summary(m)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula:
## scale(selfResp) ~ scale(SV_F) * scale(inDegree) * ingroup + scale(desirability) +
## (scale(SV_F) + scale(inDegree) | subID) + (1 | traits)
## Data: reValDf
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+05))
##
## REML criterion at convergence: 132694
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.0867 -0.6463 0.0190 0.6547 3.9063
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## subID (Intercept) 0.043354 0.20822
## scale(SV_F) 0.044108 0.21002 -0.10
## scale(inDegree) 0.007925 0.08902 0.08 -0.99
## traits (Intercept) 0.117222 0.34238
## Residual 0.563627 0.75075
## Number of obs: 57355, groups: subID, 376; traits, 257
##
## Fixed effects:
## Estimate Std. Error df t value
## (Intercept) 3.833e-02 2.529e-02 4.126e+02 1.516
## scale(SV_F) 4.892e-01 1.537e-02 6.094e+02 31.824
## scale(inDegree) -2.327e-01 2.227e-02 2.958e+02 -10.450
## ingroupOut -5.955e-02 2.565e-02 3.857e+02 -2.322
## scale(desirability) 2.114e-01 2.227e-02 2.605e+02 9.492
## scale(SV_F):scale(inDegree) -3.609e-03 6.674e-03 2.301e+04 -0.541
## scale(SV_F):ingroupOut 2.922e-02 2.623e-02 3.688e+02 1.114
## scale(inDegree):ingroupOut -1.340e-02 1.327e-02 3.633e+02 -1.009
## scale(SV_F):scale(inDegree):ingroupOut -2.132e-02 7.286e-03 5.415e+04 -2.926
## Pr(>|t|)
## (Intercept) 0.13030
## scale(SV_F) < 2e-16 ***
## scale(inDegree) < 2e-16 ***
## ingroupOut 0.02078 *
## scale(desirability) < 2e-16 ***
## scale(SV_F):scale(inDegree) 0.58871
## scale(SV_F):ingroupOut 0.26595
## scale(inDegree):ingroupOut 0.31343
## scale(SV_F):scale(inDegree):ingroupOut 0.00344 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) sc(SV_F) scl(D) ingrpO scl(d) sc(SV_F):(D) s(SV_F):O s(D):O
## scale(SV_F) -0.028
## scale(nDgr) 0.016 -0.290
## ingroupOut -0.264 0.057 -0.016
## scl(dsrblt) -0.063 -0.094 0.071 -0.003
## sc(SV_F):(D) -0.112 -0.039 -0.032 0.037 0.006
## scl(SV_F):O 0.025 -0.455 0.133 -0.087 0.001 -0.018
## scl(nDgr):O -0.023 0.378 -0.155 0.063 -0.003 0.063 -0.833
## s(SV_F):(D): 0.035 -0.026 0.023 -0.119 0.001 -0.320 0.045 -0.067ggpredict(m, c("SV_F","inDegree", "ingroup")) %>% plot()