new_articles_rep_analyses_9_4_2020_DG
Dani Grant
9/4/2020
- Contrast Codes
- Condition Differences
- 1a. Are there difference between the conditions for perceived expression of gratitude in Marel?
- 1b. What are the simple effects for no fear control vs. gratitude?
- 2. Are there differences between conditions for willingness to acceptance Marel?
- 3. Are there differences between conditions for believing Marel is trustworthy?
- 4. Are there differences between condition for feeling of closeness to Marel?
- 5. Are there differences between condition for oppisition to immigration?
- 6. Are there condition differences for opposition to speedy eligibility for government services?
- 7. Are there condition differences for willingness to help with resume?
- 8. Are there condition differencs for liking Marel?
- 9. Are there condition differences for wanting to befriend Marel?
- 10. Are there condition differences for perceived cooperativeness in Marel?
- 11. Are there condition differences in williness to donate?
- Differences between political parties
- Gratitude as a predictor
Call libraries
Contrast Codes
#control = -.5; gratitude = .5
d$CvG <-
(d$condition == 'controlFear')*(-.5) +
(d$condition == 'controlNoFear')*(-.5) +
(d$condition == 'gratFear')*(.5) +
(d$condition == 'gratNoFear')*(.5)
#Fear expressed = -.5; no fear expressed = .5
d$FvNF <-
(d$condition =='controlFear')*(-.5) +
(d$condition =='controlNoFear')*(.5) +
(d$condition =='gratFear')*(-.5) +
(d$condition =='gratNoFear')*(.5)
#Contrast Codes for immigrant friend
#-1/2 & 1 = no; +1/2 & 0 = yes
d$immFriYes_.5 <- as.factor(d$immFRI)
d$immFriYes_.5 <- (d$immFRI == '0')*(.5) + (d$immFRI == '1')*(-.5)Condition Differences
1a. Are there difference between the conditions for perceived expression of gratitude in Marel?
summary(model<-lmer(gratSCALE ~ CvG*FvNF + (1|region), data = d))## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: gratSCALE ~ CvG * FvNF + (1 | region)
## Data: d
##
## REML criterion at convergence: 983.2
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -5.9248 -0.4107 0.4049 0.7080 0.9644
##
## Random effects:
## Groups Name Variance Std.Dev.
## region (Intercept) 0.01297 0.1139
## Residual 0.79906 0.8939
## Number of obs: 372, groups: region, 6
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 6.37767 0.06696 4.03392 95.244 6.48e-08 ***
## CvG 0.09814 0.09424 365.45451 1.041 0.2984
## FvNF 0.19441 0.09359 367.46583 2.077 0.0385 *
## CvG:FvNF 0.29498 0.18711 367.13142 1.577 0.1158
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) CvG FvNF
## CvG 0.048
## FvNF -0.006 0.071
## CvG:FvNF 0.042 0.014 0.074There is a significant effect on reader’s perception of gratitude for fear expression in Marel’s story. When Marel DOES NOT express fear (M=6.46), participants rate him as being .19 points more grateful compared to when he expresses fear of living in the Honduras (M=6.29), b=.19, t(367.47)=2.08, p=.036.
p1 <- plot_model(model, type = "pred", terms = c("FvNF"), legend.title = "FvNF")
p1 + labs(title = " ",
x = "fear (-.5) vs no fear (.5)",
y = "rating of percieved gratitudes") + theme_sjplot()
d$fear <- ifelse(d$FvNF == -.5, "fear", "no fear")
d %>% ggplot(aes(condition, gratSCALE, fill = fear)) +
stat_summary(fun.data = mean_sdl, geom = "bar", position = "dodge") + theme_minimal() + coord_cartesian(ylim = c(6, 7))
1b. What are the simple effects for no fear control vs. gratitude?
d$NF_0 <-
(d$condition == 'controlFear')*(1) +
(d$condition == 'controlNoFear')*(0) +
(d$condition == 'gratFear')*(1) +
(d$condition == 'gratNoFear')*(0)
summary(m1b<-lmer(gratSCALE ~ CvG*NF_0 + (1|region), data = d))## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: gratSCALE ~ CvG * NF_0 + (1 | region)
## Data: d
##
## REML criterion at convergence: 983.2
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -5.9248 -0.4107 0.4049 0.7080 0.9644
##
## Random effects:
## Groups Name Variance Std.Dev.
## region (Intercept) 0.01297 0.1139
## Residual 0.79906 0.8939
## Number of obs: 372, groups: region, 6
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 6.47488 0.08145 9.04307 79.498 3.53e-14 ***
## CvG 0.24563 0.13374 364.52662 1.837 0.0671 .
## NF_0 -0.19441 0.09359 367.46583 -2.077 0.0385 *
## CvG:NF_0 -0.29498 0.18711 367.13142 -1.577 0.1158
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) CvG NF_0
## CvG 0.110
## NF_0 -0.569 -0.102
## CvG:NF_0 -0.077 -0.710 0.074mean(d$gratSCALE[d$NF_0 == 0 & d$condition == 'controlNoFear'])## [1] 6.364486mean(d$gratSCALE[d$NF_0 == 0 & d$condition == 'gratNoFear'])## [1] 6.595833There are marginally signficant effects for control vs. gratitude within the no fear condition. Specifically, readers perceive Marel as having .25 points more gratitude when he expresses gratitude in his story (M=6.60) compared to the control (M=6.36), t(364.53)=1.84, p=.067. ***
2. Are there differences between conditions for willingness to acceptance Marel?
summary(m2 <- lm(acceptSCALE ~ CvG*FvNF, data = d))##
## Call:
## lm(formula = acceptSCALE ~ CvG * FvNF, data = d)
##
## Residuals:
## Min 1Q Median 3Q Max
## -4.6500 -0.6683 0.3500 1.0616 2.0145
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 5.28902 0.08365 63.224 <2e-16 ***
## CvG 0.05746 0.16731 0.343 0.7315
## FvNF 0.54236 0.16731 3.242 0.0013 **
## CvG:FvNF 0.24427 0.33462 0.730 0.4659
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.605 on 368 degrees of freedom
## Multiple R-squared: 0.02841, Adjusted R-squared: 0.02049
## F-statistic: 3.587 on 3 and 368 DF, p-value: 0.01397There is a significant effect of fear expression on reader’s willingness to accept Marel into their community.When Marel DOES NOT expresses fear living in the Honduras, participants rate themselves as being .54 points more willing to accept Marel into their community (M=5.55) compared to when he DOES express fear (M=5.01), b=.54, t(368)=3.24, p=.001. ***
d %>% ggplot(aes(condition, acceptSCALE, fill = fear)) +
stat_summary(fun.data = mean_sdl, geom = "bar", position = "dodge") + theme_minimal() + coord_cartesian(ylim = c(3, 7))
p2 <- plot_model(m2, type = "pred", terms = c("FvNF"), legend.title = "FvNF")
p2 + labs(title = " ",
x = "fear (-.5) vs no fear (.5)",
y = "rating for willingness to accept Marel") + theme_sjplot()
3. Are there differences between conditions for believing Marel is trustworthy?
mcSummary(m3 <-lm(trustworthy ~ CvG*FvNF, data = d)) ## lm(formula = trustworthy ~ CvG * FvNF, data = d)
##
## Omnibus ANOVA
## SS df MS EtaSq F p
## Model 23.758 3 7.919 0.042 5.44 0.001
## Error 535.758 368 1.456
## Corr Total 559.516 371 1.508
##
## RMSE AdjEtaSq
## 1.207 0.035
##
## Coefficients
## Est StErr t SSR(3) EtaSq tol CI_2.5 CI_97.5 p
## (Intercept) 5.409 0.063 86.014 10770.969 0.953 NA 5.286 5.533 0.000
## CvG 0.063 0.126 0.502 0.367 0.001 0.995 -0.184 0.311 0.616
## FvNF 0.495 0.126 3.936 22.549 0.040 0.990 0.248 0.742 0.000
## CvG:FvNF 0.296 0.252 1.176 2.013 0.004 0.994 -0.199 0.791 0.240mean(d$trustworthy[d$FvNF == -.5])## [1] 5.162162mean(d$trustworthy[d$FvNF == .5])## [1] 5.641711Yes, there is a significant effect of fear expression on whether readers think Marel is trustworthy. Specifically, participants rate marel as .50 points more trustworthy when his story DOES NOT expressess fear (M=5.64) compared to when he does express fear (M=5.16), t(368)=3.94,PRE=.04, p < .001. ***
d %>% ggplot(aes(condition, trustworthy, fill = fear)) +
stat_summary(fun.data = mean_sdl, geom = "bar", position = "dodge") + theme_minimal() + coord_cartesian(ylim = c(4, 6))
p3 <- plot_model(m3, type = "pred", terms = c("FvNF"), legend.title = "FvNF")
p3 + labs(title = " ",
x = "fear (-.5) vs no fear (.5)",
y = "rating for Marel's trustworthiness") + theme_sjplot()
4. Are there differences between condition for feeling of closeness to Marel?
mcSummary(m4<-lm(close_1 ~ CvG*FvNF, data = d))## lm(formula = close_1 ~ CvG * FvNF, data = d)
##
## Omnibus ANOVA
## SS df MS EtaSq F p
## Model 7.508 3 2.503 0.007 0.805 0.492
## Error 1143.941 368 3.109
## Corr Total 1151.449 371 3.104
##
## RMSE AdjEtaSq
## 1.763 -0.002
##
## Coefficients
## Est StErr t SSR(3) EtaSq tol CI_2.5 CI_97.5 p
## (Intercept) 3.855 0.092 41.953 5471.118 0.827 NA 3.675 4.036 0.000
## CvG 0.096 0.184 0.524 0.855 0.001 0.995 -0.265 0.458 0.600
## FvNF 0.217 0.184 1.183 4.352 0.004 0.990 -0.144 0.579 0.237
## CvG:FvNF -0.299 0.368 -0.814 2.058 0.002 0.994 -1.022 0.424 0.416mean(d$close_1[d$FvNF == -.5])## [1] 3.745946mean(d$close_1[d$FvNF == .5])## [1] 3.967914There are no effects for fear expression on closeness, p=.237. ***
d %>% ggplot(aes(condition, close_1, fill = fear)) +
stat_summary(fun.data = mean_sdl, geom = "bar", position = "dodge") + theme_minimal() + coord_cartesian(ylim = c(3, 4.5))
5. Are there differences between condition for oppisition to immigration?
mcSummary(m5<-lm(OppositionToImm ~ CvG*FvNF, data = d))## lm(formula = OppositionToImm ~ CvG * FvNF, data = d)
##
## Omnibus ANOVA
## SS df MS EtaSq F p
## Model 5.457 3 1.819 0.011 1.317 0.268
## Error 508.188 368 1.381
## Corr Total 513.645 371 1.384
##
## RMSE AdjEtaSq
## 1.175 0.003
##
## Coefficients
## Est StErr t SSR(3) EtaSq tol CI_2.5 CI_97.5 p
## (Intercept) 2.789 0.061 45.540 2863.882 0.849 NA 2.669 2.910 0.000
## CvG -0.059 0.123 -0.483 0.322 0.001 0.995 -0.300 0.182 0.630
## FvNF -0.237 0.123 -1.936 5.173 0.010 0.990 -0.478 0.004 0.054
## CvG:FvNF 0.035 0.245 0.143 0.028 0.000 0.994 -0.447 0.517 0.886mean(d$OppositionToImm[d$FvNF == -.5]) #fear## [1] 2.908108mean(d$OppositionToImm[d$FvNF == .5]) #no fear## [1] 2.673797There is a marginally significant effect of fear expression on whether readers oppose immigration (1 = increase immigration a lot, 7 = decrease immigration a lot). Specifically, there is a .24 point rating decrease in opposition to immigration from fear expression (M=2.91) to no fear expression (M=2.67), t(368)=-1.94,PRE=.01, p=.054. ***
d %>% ggplot(aes(condition, OppositionToImm, fill = fear)) +
stat_summary(fun.data = mean_sdl, geom = "bar", position = "dodge") + theme_minimal() + coord_cartesian(ylim = c(2, 4))
p5 <- plot_model(m5, type = "pred", terms = c("FvNF"), legend.title = "FvNF")
p5 + labs(title = " ",
x = "fear (-.5) vs no fear (.5)",
y = "rating opposition to immigration") + theme_sjplot()
6. Are there condition differences for opposition to speedy eligibility for government services?
mcSummary(m6<-lm(OppositionToElig ~ CvG*FvNF, data = d))## lm(formula = OppositionToElig ~ CvG * FvNF, data = d)
##
## Omnibus ANOVA
## SS df MS EtaSq F p
## Model 3.774 3 1.258 0.008 0.948 0.418
## Error 488.428 368 1.327
## Corr Total 492.202 371 1.327
##
## RMSE AdjEtaSq
## 1.152 0
##
## Coefficients
## Est StErr t SSR(3) EtaSq tol CI_2.5 CI_97.5 p
## (Intercept) 2.313 0.06 38.526 1969.949 0.801 NA 2.195 2.431 0.000
## CvG -0.136 0.12 -1.128 1.690 0.003 0.995 -0.372 0.101 0.260
## FvNF 0.088 0.12 0.730 0.707 0.001 0.990 -0.148 0.324 0.466
## CvG:FvNF 0.242 0.24 1.008 1.348 0.003 0.994 -0.230 0.714 0.314There are no condition effects on opposition/support for speedy eligibility for government services. ***
7. Are there condition differences for willingness to help with resume?
mcSummary(lm(resume ~ CvG*FvNF, data = d))## lm(formula = resume ~ CvG * FvNF, data = d)
##
## Omnibus ANOVA
## SS df MS EtaSq F p
## Model 0.922 3 0.307 0.002 0.215 0.886
## Error 526.981 368 1.432
## Corr Total 527.903 371 1.423
##
## RMSE AdjEtaSq
## 1.197 -0.006
##
## Coefficients
## Est StErr t SSR(3) EtaSq tol CI_2.5 CI_97.5 p
## (Intercept) 1.980 0.062 31.749 1443.458 0.733 NA 1.858 2.103 0.000
## CvG -0.003 0.125 -0.024 0.001 0.000 0.995 -0.248 0.242 0.981
## FvNF -0.008 0.125 -0.060 0.005 0.000 0.990 -0.253 0.238 0.952
## CvG:FvNF -0.200 0.249 -0.802 0.922 0.002 0.994 -0.691 0.290 0.423there are no significant effects on willingness to help write resume for Marel. ***
8. Are there condition differencs for liking Marel?
mcSummary(m8<-lm(likeSCALE ~ CvG*FvNF, data = d))## lm(formula = likeSCALE ~ CvG * FvNF, data = d)
##
## Omnibus ANOVA
## SS df MS EtaSq F p
## Model 13.699 3 4.566 0.021 2.576 0.054
## Error 652.398 368 1.773
## Corr Total 666.096 371 1.795
##
## RMSE AdjEtaSq
## 1.331 0.013
##
## Coefficients
## Est StErr t SSR(3) EtaSq tol CI_2.5 CI_97.5 p
## (Intercept) 5.013 0.069 72.240 9251.571 0.934 NA 4.877 5.150 0.000
## CvG 0.154 0.139 1.111 2.187 0.003 0.995 -0.119 0.427 0.267
## FvNF 0.361 0.139 2.604 12.023 0.018 0.990 0.089 0.634 0.010
## CvG:FvNF 0.132 0.278 0.475 0.399 0.001 0.994 -0.414 0.678 0.635mean(d$likeSCALE[d$FvNF == -.5])## [1] 4.832432mean(d$likeSCALE[d$FvNF == .5])## [1] 5.178253There is a significant effect for fear condition on liking Marel. Specifically, readers like Marel less when he expresses fear (M=4.83) in his story compared to when he does not express fear (M=5.18) by .36 points. ***
d %>% ggplot(aes(condition, likeSCALE, fill = fear)) +
stat_summary(fun.data = mean_sdl, geom = "bar", position = "dodge") + theme_minimal() + coord_cartesian(ylim = c(3, 6))
p8 <- plot_model(m8, type = "pred", terms = c("FvNF"), legend.title = "FvNF")
p8 + labs(title = " ",
x = "fear (-.5) vs no fear (.5)",
y = "rating for liking Marel") + theme_sjplot()
9. Are there condition differences for wanting to befriend Marel?
mcSummary(m9<-lm(befriendSCALE ~ CvG*FvNF, data = d))## lm(formula = befriendSCALE ~ CvG * FvNF, data = d)
##
## Omnibus ANOVA
## SS df MS EtaSq F p
## Model 9.952 3 3.317 0.013 1.572 0.196
## Error 776.599 368 2.110
## Corr Total 786.550 371 2.120
##
## RMSE AdjEtaSq
## 1.453 0.005
##
## Coefficients
## Est StErr t SSR(3) EtaSq tol CI_2.5 CI_97.5 p
## (Intercept) 5.344 0.076 70.583 10513.478 0.931 NA 5.195 5.493 0.000
## CvG 0.082 0.151 0.544 0.625 0.001 0.995 -0.215 0.380 0.587
## FvNF 0.321 0.151 2.118 9.463 0.012 0.990 0.023 0.618 0.035
## CvG:FvNF -0.034 0.303 -0.111 0.026 0.000 0.994 -0.629 0.562 0.911mean(d$befriendSCALE[d$FvNF == -.5])## [1] 5.183784mean(d$befriendSCALE[d$FvNF == .5])## [1] 5.5There is a significant effect for fear condition on willingness to befriend Marel. Specifically, readers are less willing to befriend Marel when he expresses fear (M=5.18) in his story compared to when he does not express fear (M=5.50) by .32 points, PRE = .012, p = .035. ***
d %>% ggplot(aes(condition, befriendSCALE, fill = fear)) +
stat_summary(fun.data = mean_sdl, geom = "bar", position = "dodge") + theme_minimal() + coord_cartesian(ylim = c(4, 6))
p9 <- plot_model(m9, type = "pred", terms = c("FvNF"), legend.title = "FvNF")
p9 + labs(title = " ",
x = "fear (-.5) vs no fear (.5)",
y = "rating for willingness to befriend Marel") + theme_sjplot()
10. Are there condition differences for perceived cooperativeness in Marel?
mcSummary(m10<-lm(cooperative ~ CvG*FvNF, data = d))## lm(formula = cooperative ~ CvG * FvNF, data = d)
##
## Omnibus ANOVA
## SS df MS EtaSq F p
## Model 8.354 3 2.785 0.02 2.447 0.063
## Error 418.708 368 1.138
## Corr Total 427.062 371 1.151
##
## RMSE AdjEtaSq
## 1.067 0.012
##
## Coefficients
## Est StErr t SSR(3) EtaSq tol CI_2.5 CI_97.5 p
## (Intercept) 5.801 0.056 104.339 12386.714 0.967 NA 5.692 5.910 0.000
## CvG 0.155 0.111 1.390 2.197 0.005 0.995 -0.064 0.373 0.165
## FvNF 0.261 0.111 2.348 6.272 0.015 0.990 0.042 0.480 0.019
## CvG:FvNF 0.165 0.222 0.741 0.625 0.001 0.994 -0.273 0.602 0.459mean(d$cooperative[d$FvNF == -.5])## [1] 5.67027mean(d$cooperative[d$FvNF == .5])## [1] 5.914439There is a significant effect for fear on perception of Marel’s cooperativeness. Readers perceive Marel as less cooperative (M=5.67) when he expresses fear in his story compared to when he does not express any fear (M=5.91), PRE =.015, p = .019. ***
d %>% ggplot(aes(condition, cooperative, fill = fear)) +
stat_summary(fun.data = mean_sdl, geom = "bar", position = "dodge") + theme_minimal() + coord_cartesian(ylim = c(5, 7))
p10 <- plot_model(m10, type = "pred", terms = c("FvNF"), legend.title = "FvNF")
p10 + labs(title = " ",
x = "fear (-.5) vs no fear (.5)",
y = "rating for Marel's perceived cooperativeness") + theme_sjplot()
11. Are there condition differences in williness to donate?
mcSummary(donate<-lm(donate ~ CvG*FvNF, data = d))## lm(formula = donate ~ CvG * FvNF, data = d)
##
## Omnibus ANOVA
## SS df MS EtaSq F p
## Model 0.040 3 0.013 0.001 0.084 0.969
## Error 51.821 326 0.159
## Corr Total 51.862 329 0.158
##
## RMSE AdjEtaSq
## 0.399 -0.008
##
## Coefficients
## Est StErr t SSR(3) EtaSq tol CI_2.5 CI_97.5 p
## (Intercept) 0.252 0.022 11.410 20.696 0.285 NA 0.208 0.295 0.000
## CvG -0.014 0.044 -0.325 0.017 0.000 0.998 -0.101 0.072 0.746
## FvNF 0.016 0.044 0.358 0.020 0.000 0.991 -0.071 0.103 0.720
## CvG:FvNF 0.010 0.088 0.116 0.002 0.000 0.992 -0.163 0.184 0.908There are no condition effects for willingness to donate money. ***
Differences between political parties
0. Make new dataset and codes for party
#contrast codes for party
d$DvR <-
(d$party=="Democrat")*(-.5) +
(d$party=="Independent")*(0) +
(d$party=="Republican")*(.5)
d$IvDR<- ifelse(d$party=="Independent", (-2/3), (1/3))
d$IndCode <- ifelse(d$party == "Independent" | is.na(d$party), 1, 0)
#new dataset without true independents
dNoI <- filter(d, IndCode == 0)
dNoI$dem_0 <- (0)*(dNoI$party=="Democrat") + (1)*(dNoI$party=="Republican")
dNoI$rep_0 <- (1)*(dNoI$party=="Democrat") + (0)*(dNoI$party=="Republican")1a. gratitudeScale ~ Grat x fear x party
#looking at simple effects for democrats
mcSummary(m1p<-lm(gratSCALE ~ CvG*FvNF*DvR, data = dNoI))## lm(formula = gratSCALE ~ CvG * FvNF * DvR, data = dNoI)
##
## Omnibus ANOVA
## SS df MS EtaSq F p
## Model 9.786 7 1.398 0.047 1.968 0.059
## Error 200.276 282 0.710
## Corr Total 210.062 289 0.727
##
## RMSE AdjEtaSq
## 0.843 0.023
##
## Coefficients
## Est StErr t SSR(3) EtaSq tol CI_2.5 CI_97.5 p
## (Intercept) 6.376 0.056 114.710 9345.008 0.979 NA 6.266 6.485 0.000
## CvG 0.015 0.111 0.134 0.013 0.000 0.793 -0.204 0.234 0.893
## FvNF 0.145 0.111 1.300 1.200 0.006 0.793 -0.074 0.363 0.195
## DvR -0.223 0.111 -2.004 2.853 0.014 0.915 -0.442 -0.004 0.046
## CvG:FvNF 0.076 0.222 0.343 0.083 0.000 0.794 -0.361 0.514 0.732
## CvG:DvR 0.237 0.222 1.066 0.807 0.004 0.795 -0.201 0.675 0.287
## FvNF:DvR -0.068 0.222 -0.307 0.067 0.000 0.810 -0.506 0.369 0.759
## CvG:FvNF:DvR -0.797 0.445 -1.793 2.282 0.011 0.796 -1.672 0.078 0.074dNoI %>% ggplot(aes(condition, gratSCALE, fill = party)) +
stat_summary(fun.data = mean_sdl, geom = "bar", position = "dodge") + theme_minimal() + coord_cartesian(ylim = c(6, 7))
1b. interaction between fear X gratitude for democrats
#looking at simple effects for democrats
mcSummary(m2p<-lm(likeSCALE ~ CvG*FvNF*dem_0, data = dNoI))## lm(formula = likeSCALE ~ CvG * FvNF * dem_0, data = dNoI)
##
## Omnibus ANOVA
## SS df MS EtaSq F p
## Model 37.079 7 5.297 0.074 3.219 0.003
## Error 464.019 282 1.645
## Corr Total 501.097 289 1.734
##
## RMSE AdjEtaSq
## 1.283 0.051
##
## Coefficients
## Est StErr t SSR(3) EtaSq tol CI_2.5 CI_97.5 p
## (Intercept) 5.259 0.092 57.319 5406.156 0.921 NA 5.079 5.440 0.000
## CvG -0.017 0.184 -0.095 0.015 0.000 0.674 -0.379 0.344 0.924
## FvNF 0.276 0.184 1.506 3.733 0.008 0.674 -0.085 0.638 0.133
## dem_0 -0.539 0.169 -3.188 16.724 0.035 0.915 -0.872 -0.206 0.002
## CvG:FvNF -0.035 0.367 -0.095 0.015 0.000 0.675 -0.757 0.687 0.924
## CvG:dem_0 0.365 0.338 1.078 1.913 0.004 0.627 -0.301 1.031 0.282
## FvNF:dem_0 0.284 0.338 0.838 1.156 0.002 0.634 -0.382 0.950 0.403
## CvG:FvNF:dem_0 0.007 0.677 0.010 0.000 0.000 0.629 -1.325 1.339 0.9921c. interaction between fear x grat for republicans
#looking at simple effects for republicans
mcSummary(likingR<-lm(likeSCALE ~ CvG*FvNF*rep_0, data = dNoI))## lm(formula = likeSCALE ~ CvG * FvNF * rep_0, data = dNoI)
##
## Omnibus ANOVA
## SS df MS EtaSq F p
## Model 37.079 7 5.297 0.074 3.219 0.003
## Error 464.019 282 1.645
## Corr Total 501.097 289 1.734
##
## RMSE AdjEtaSq
## 1.283 0.051
##
## Coefficients
## Est StErr t SSR(3) EtaSq tol CI_2.5 CI_97.5 p
## (Intercept) 4.720 0.142 33.201 1813.825 0.796 NA 4.440 5.000 0.000
## CvG 0.347 0.284 1.222 2.456 0.005 0.281 -0.212 0.907 0.223
## FvNF 0.560 0.284 1.970 6.384 0.014 0.281 0.000 1.120 0.050
## rep_0 0.539 0.169 3.188 16.724 0.035 0.915 0.206 0.872 0.002
## CvG:FvNF -0.028 0.569 -0.049 0.004 0.000 0.281 -1.147 1.091 0.961
## CvG:rep_0 -0.365 0.338 -1.078 1.913 0.004 0.290 -1.031 0.301 0.282
## FvNF:rep_0 -0.284 0.338 -0.838 1.156 0.002 0.293 -0.950 0.382 0.403
## CvG:FvNF:rep_0 -0.007 0.677 -0.010 0.000 0.000 0.290 -1.339 1.325 0.9922a. befriend ~ grat X fear X part
#looking at simple effects for republicans
mcSummary(befriending<-lm(befriendSCALE ~ CvG*FvNF*DvR, data = dNoI))## lm(formula = befriendSCALE ~ CvG * FvNF * DvR, data = dNoI)
##
## Omnibus ANOVA
## SS df MS EtaSq F p
## Model 39.684 7 5.669 0.069 2.968 0.005
## Error 538.713 282 1.910
## Corr Total 578.397 289 2.001
##
## RMSE AdjEtaSq
## 1.382 0.045
##
## Coefficients
## Est StErr t SSR(3) EtaSq tol CI_2.5 CI_97.5 p
## (Intercept) 5.293 0.091 58.069 6441.608 0.923 NA 5.114 5.473 0.000
## CvG -0.065 0.182 -0.357 0.243 0.000 0.793 -0.424 0.294 0.722
## FvNF 0.291 0.182 1.595 4.859 0.009 0.793 -0.068 0.650 0.112
## DvR -0.647 0.182 -3.548 24.045 0.043 0.915 -1.006 -0.288 0.000
## CvG:FvNF -0.068 0.365 -0.186 0.066 0.000 0.794 -0.786 0.650 0.852
## CvG:DvR 0.288 0.365 0.789 1.188 0.002 0.795 -0.430 1.005 0.431
## FvNF:DvR 0.455 0.365 1.249 2.979 0.006 0.810 -0.262 1.173 0.213
## CvG:FvNF:DvR -0.346 0.729 -0.474 0.430 0.001 0.796 -1.781 1.089 0.636dNoI %>% ggplot(aes(condition, befriendSCALE, fill = party)) +
stat_summary(fun.data = mean_sdl, geom = "bar", position = "dodge") + theme_minimal() + coord_cartesian(ylim = c(4, 6))
2b. befriend ~ simple effects for dems
#looking at simple effects for democrats
trustD<-lm(trustworthy ~ CvG*FvNF*dem_0, data = dNoI)
mcSummary(trustD)## lm(formula = trustworthy ~ CvG * FvNF * dem_0, data = dNoI)
##
## Omnibus ANOVA
## SS df MS EtaSq F p
## Model 34.424 7 4.918 0.095 4.205 0
## Error 329.796 282 1.169
## Corr Total 364.221 289 1.260
##
## RMSE AdjEtaSq
## 1.081 0.072
##
## Coefficients
## Est StErr t SSR(3) EtaSq tol CI_2.5 CI_97.5 p
## (Intercept) 5.637 0.077 72.874 6210.732 0.950 NA 5.485 5.789 0.000
## CvG -0.132 0.155 -0.851 0.847 0.003 0.674 -0.436 0.173 0.395
## FvNF 0.343 0.155 2.214 5.735 0.017 0.674 0.038 0.647 0.028
## dem_0 -0.503 0.143 -3.529 14.562 0.042 0.915 -0.784 -0.223 0.000
## CvG:FvNF 0.328 0.309 1.059 1.311 0.004 0.675 -0.282 0.937 0.291
## CvG:dem_0 0.261 0.285 0.913 0.975 0.003 0.627 -0.301 0.822 0.362
## FvNF:dem_0 0.140 0.285 0.490 0.280 0.001 0.634 -0.422 0.701 0.625
## CvG:FvNF:dem_0 -0.085 0.571 -0.149 0.026 0.000 0.629 -1.208 1.038 0.8812c. befriend ~ simple effects for reps
#looking at simple effects for republicans
mcSummary(trustR<-lm(trustworthy ~ CvG*FvNF*rep_0, data = dNoI))## lm(formula = trustworthy ~ CvG * FvNF * rep_0, data = dNoI)
##
## Omnibus ANOVA
## SS df MS EtaSq F p
## Model 34.424 7 4.918 0.095 4.205 0
## Error 329.796 282 1.169
## Corr Total 364.221 289 1.260
##
## RMSE AdjEtaSq
## 1.081 0.072
##
## Coefficients
## Est StErr t SSR(3) EtaSq tol CI_2.5 CI_97.5 p
## (Intercept) 5.134 0.120 42.835 2145.862 0.867 NA 4.898 5.370 0.000
## CvG 0.129 0.240 0.537 0.338 0.001 0.281 -0.343 0.601 0.591
## FvNF 0.482 0.240 2.012 4.734 0.014 0.281 0.010 0.954 0.045
## rep_0 0.503 0.143 3.529 14.562 0.042 0.915 0.223 0.784 0.000
## CvG:FvNF 0.242 0.479 0.505 0.299 0.001 0.281 -0.701 1.186 0.614
## CvG:rep_0 -0.261 0.285 -0.913 0.975 0.003 0.290 -0.822 0.301 0.362
## FvNF:rep_0 -0.140 0.285 -0.490 0.280 0.001 0.293 -0.701 0.422 0.625
## CvG:FvNF:rep_0 0.085 0.571 0.149 0.026 0.000 0.290 -1.038 1.208 0.8813a. acceptance Marel ~ grat X fear X party
#looking at simple effects for republicans
mcSummary(accept<-lm(acceptSCALE ~ CvG*FvNF*DvR, data = dNoI))## lm(formula = acceptSCALE ~ CvG * FvNF * DvR, data = dNoI)
##
## Omnibus ANOVA
## SS df MS EtaSq F p
## Model 49.797 7 7.114 0.067 2.906 0.006
## Error 690.268 282 2.448
## Corr Total 740.065 289 2.561
##
## RMSE AdjEtaSq
## 1.565 0.044
##
## Coefficients
## Est StErr t SSR(3) EtaSq tol CI_2.5 CI_97.5 p
## (Intercept) 5.239 0.103 50.772 6309.903 0.901 NA 5.036 5.442 0.000
## CvG -0.131 0.206 -0.637 0.994 0.001 0.793 -0.538 0.275 0.525
## FvNF 0.433 0.206 2.098 10.777 0.015 0.793 0.027 0.839 0.037
## DvR -0.613 0.206 -2.970 21.594 0.030 0.915 -1.019 -0.207 0.003
## CvG:FvNF 0.251 0.413 0.608 0.905 0.001 0.794 -0.561 1.063 0.544
## CvG:DvR 0.144 0.413 0.349 0.299 0.000 0.795 -0.668 0.957 0.727
## FvNF:DvR 0.321 0.413 0.777 1.478 0.002 0.810 -0.492 1.133 0.438
## CvG:FvNF:DvR -0.987 0.825 -1.195 3.497 0.005 0.796 -2.612 0.638 0.233dNoI %>% ggplot(aes(condition, acceptSCALE, fill = party)) +
stat_summary(fun.data = mean_sdl, geom = "bar", position = "dodge") + theme_minimal() + coord_cartesian(ylim = c(4, 6))
4a. cooperation ~ grat x fear x party
#looking at simple effects for republicans
mcSummary(cooperation<-lm(cooperative ~ CvG*FvNF*DvR, data = dNoI))## lm(formula = cooperative ~ CvG * FvNF * DvR, data = dNoI)
##
## Omnibus ANOVA
## SS df MS EtaSq F p
## Model 12.904 7 1.843 0.045 1.89 0.071
## Error 275.113 282 0.976
## Corr Total 288.017 289 0.997
##
## RMSE AdjEtaSq
## 0.988 0.021
##
## Coefficients
## Est StErr t SSR(3) EtaSq tol CI_2.5 CI_97.5 p
## (Intercept) 5.827 0.065 89.457 7807.075 0.966 NA 5.699 5.956 0.000
## CvG 0.274 0.130 2.106 4.329 0.015 0.793 0.018 0.531 0.036
## FvNF 0.226 0.130 1.731 2.925 0.011 0.793 -0.031 0.482 0.084
## DvR -0.112 0.130 -0.859 0.720 0.003 0.915 -0.368 0.145 0.391
## CvG:FvNF 0.067 0.261 0.259 0.065 0.000 0.794 -0.446 0.580 0.796
## CvG:DvR 0.463 0.261 1.777 3.079 0.011 0.795 -0.050 0.976 0.077
## FvNF:DvR -0.037 0.261 -0.142 0.020 0.000 0.810 -0.550 0.476 0.887
## CvG:FvNF:DvR -0.492 0.521 -0.944 0.869 0.003 0.796 -1.518 0.534 0.346dNoI %>% ggplot(aes(condition, cooperative, fill = party)) +
stat_summary(fun.data = mean_sdl, geom = "bar", position = "dodge") + theme_minimal() + coord_cartesian(ylim = c(5, 6.5))
4b. cooperation ~ grat x fear x dem_0
#looking at simple effects for republicans
mcSummary(cooperation<-lm(cooperative ~ CvG*FvNF*dem_0, data = dNoI))## lm(formula = cooperative ~ CvG * FvNF * dem_0, data = dNoI)
##
## Omnibus ANOVA
## SS df MS EtaSq F p
## Model 12.904 7 1.843 0.045 1.89 0.071
## Error 275.113 282 0.976
## Corr Total 288.017 289 0.997
##
## RMSE AdjEtaSq
## 0.988 0.021
##
## Coefficients
## Est StErr t SSR(3) EtaSq tol CI_2.5 CI_97.5 p
## (Intercept) 5.883 0.071 83.273 6765.086 0.961 NA 5.744 6.022 0.000
## CvG 0.043 0.141 0.304 0.090 0.000 0.674 -0.235 0.321 0.761
## FvNF 0.244 0.141 1.727 2.911 0.010 0.674 -0.034 0.522 0.085
## dem_0 -0.112 0.130 -0.859 0.720 0.003 0.915 -0.368 0.145 0.391
## CvG:FvNF 0.313 0.283 1.108 1.199 0.004 0.675 -0.243 0.870 0.269
## CvG:dem_0 0.463 0.261 1.777 3.079 0.011 0.627 -0.050 0.976 0.077
## FvNF:dem_0 -0.037 0.261 -0.142 0.020 0.000 0.634 -0.550 0.476 0.887
## CvG:FvNF:dem_0 -0.492 0.521 -0.944 0.869 0.003 0.629 -1.518 0.534 0.3464c. cooperation ~ grat x fear x rep_0
#looking at simple effects for republicans
mcSummary(cooperation<-lm(cooperative ~ CvG*FvNF*DvR, data = dNoI))## lm(formula = cooperative ~ CvG * FvNF * DvR, data = dNoI)
##
## Omnibus ANOVA
## SS df MS EtaSq F p
## Model 12.904 7 1.843 0.045 1.89 0.071
## Error 275.113 282 0.976
## Corr Total 288.017 289 0.997
##
## RMSE AdjEtaSq
## 0.988 0.021
##
## Coefficients
## Est StErr t SSR(3) EtaSq tol CI_2.5 CI_97.5 p
## (Intercept) 5.827 0.065 89.457 7807.075 0.966 NA 5.699 5.956 0.000
## CvG 0.274 0.130 2.106 4.329 0.015 0.793 0.018 0.531 0.036
## FvNF 0.226 0.130 1.731 2.925 0.011 0.793 -0.031 0.482 0.084
## DvR -0.112 0.130 -0.859 0.720 0.003 0.915 -0.368 0.145 0.391
## CvG:FvNF 0.067 0.261 0.259 0.065 0.000 0.794 -0.446 0.580 0.796
## CvG:DvR 0.463 0.261 1.777 3.079 0.011 0.795 -0.050 0.976 0.077
## FvNF:DvR -0.037 0.261 -0.142 0.020 0.000 0.810 -0.550 0.476 0.887
## CvG:FvNF:DvR -0.492 0.521 -0.944 0.869 0.003 0.796 -1.518 0.534 0.3466a. Opposition to immigration ~ grat X Fear X party
#looking at simple effects for democrats
mcSummary(OppositionToImm<-lm(OppositionToImm ~ CvG*FvNF*DvR, data = dNoI))## lm(formula = OppositionToImm ~ CvG * FvNF * DvR, data = dNoI)
##
## Omnibus ANOVA
## SS df MS EtaSq F p
## Model 62.357 7 8.908 0.159 7.592 0
## Error 330.888 282 1.173
## Corr Total 393.245 289 1.361
##
## RMSE AdjEtaSq
## 1.083 0.138
##
## Coefficients
## Est StErr t SSR(3) EtaSq tol CI_2.5 CI_97.5 p
## (Intercept) 2.872 0.071 40.203 1896.440 0.851 NA 2.731 3.013 0.000
## CvG 0.010 0.143 0.071 0.006 0.000 0.793 -0.271 0.291 0.943
## FvNF -0.228 0.143 -1.597 2.992 0.009 0.793 -0.509 0.053 0.111
## DvR 0.919 0.143 6.435 48.589 0.128 0.915 0.638 1.201 0.000
## CvG:FvNF 0.158 0.286 0.553 0.359 0.001 0.794 -0.404 0.721 0.581
## CvG:DvR 0.020 0.286 0.070 0.006 0.000 0.795 -0.542 0.583 0.944
## FvNF:DvR -0.288 0.286 -1.007 1.190 0.004 0.810 -0.850 0.275 0.315
## CvG:FvNF:DvR -0.230 0.572 -0.403 0.190 0.001 0.796 -1.355 0.895 0.688dNoI %>% ggplot(aes(condition, OppositionToImm, fill = party)) +
stat_summary(fun.data = mean_sdl, geom = "bar", position = "dodge") + theme_minimal() + coord_cartesian(ylim = c(2, 4))
7a. opposition to eligibity for gov’t services
#looking at simple effects for democrats
mcSummary(OppositionToElig<-lm(OppositionToElig ~ CvG*FvNF*DvR, data = dNoI))## lm(formula = OppositionToElig ~ CvG * FvNF * DvR, data = dNoI)
##
## Omnibus ANOVA
## SS df MS EtaSq F p
## Model 30.946 7 4.421 0.087 3.857 0.001
## Error 323.223 282 1.146
## Corr Total 354.169 289 1.225
##
## RMSE AdjEtaSq
## 1.071 0.065
##
## Coefficients
## Est StErr t SSR(3) EtaSq tol CI_2.5 CI_97.5 p
## (Intercept) 2.320 0.071 32.854 1237.149 0.793 NA 2.181 2.459 0.000
## CvG -0.127 0.141 -0.901 0.930 0.003 0.793 -0.405 0.151 0.368
## FvNF 0.075 0.141 0.528 0.319 0.001 0.793 -0.203 0.352 0.598
## DvR 0.653 0.141 4.622 24.490 0.070 0.915 0.375 0.931 0.000
## CvG:FvNF 0.274 0.282 0.969 1.077 0.003 0.794 -0.282 0.830 0.333
## CvG:DvR -0.198 0.282 -0.701 0.563 0.002 0.795 -0.754 0.358 0.484
## FvNF:DvR -0.234 0.282 -0.827 0.784 0.002 0.810 -0.790 0.322 0.409
## CvG:FvNF:DvR 0.024 0.565 0.043 0.002 0.000 0.796 -1.088 1.136 0.966dNoI %>% ggplot(aes(condition, OppositionToElig, fill = party)) +
stat_summary(fun.data = mean_sdl, geom = "bar", position = "dodge") + theme_minimal() + coord_cartesian(ylim = c(1, 4))
Gratitude as a predictor
1. hispanic threat ~ gratitude X party
summary(m1b<-lmer(hispThreatSCALE ~ gratSCALE*DvR + (1|region), data = dNoI))## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: hispThreatSCALE ~ gratSCALE * DvR + (1 | region)
## Data: dNoI
##
## REML criterion at convergence: 791.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.0973 -0.6691 -0.1507 0.5281 3.8334
##
## Random effects:
## Groups Name Variance Std.Dev.
## region (Intercept) 0.02102 0.1450
## Residual 0.86102 0.9279
## Number of obs: 290, groups: region, 6
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 8.46967 0.44309 276.73730 19.115 < 2e-16 ***
## gratSCALE -0.56310 0.06863 284.78356 -8.205 8.02e-15 ***
## DvR -0.54175 0.88354 285.98225 -0.613 0.540
## gratSCALE:DvR 0.30644 0.13788 285.95476 2.223 0.027 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) grSCALE DvR
## gratSCALE -0.981
## DvR 0.289 -0.301
## grtSCALE:DR -0.302 0.321 -0.991p2 <- plot_model(m1b, type = "pred", terms = c("gratSCALE", "DvR"), legend.title = "condition")
p2 + labs(title = "hispThrear ~ gratitude X DvR",
x = "gratitude score",
y = "hispanic threat score") +
scale_color_discrete(labels = c("dem", "rep")) +
theme_sjplot()## Scale for 'colour' is already present. Adding another scale for 'colour',
## which will replace the existing scale.
2. patient ~ grat X party
summary(m2<-lmer(patient ~ gratSCALE*DvR + (1|region), data = dNoI))## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: patient ~ gratSCALE * DvR + (1 | region)
## Data: dNoI
##
## REML criterion at convergence: 868.7
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.8811 -0.7655 0.1404 0.7738 2.1473
##
## Random effects:
## Groups Name Variance Std.Dev.
## region (Intercept) 0.003066 0.05537
## Residual 1.140071 1.06774
## Number of obs: 290, groups: region, 6
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 2.46314 0.50393 281.92862 4.888 1.71e-06 ***
## gratSCALE 0.46266 0.07868 285.94964 5.880 1.14e-08 ***
## DvR 2.67332 1.00818 280.58658 2.652 0.00847 **
## gratSCALE:DvR -0.42132 0.15754 283.30783 -2.674 0.00792 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) grSCALE DvR
## gratSCALE -0.990
## DvR 0.292 -0.305
## grtSCALE:DR -0.306 0.325 -0.991p2 <- plot_model(m2, type = "pred", terms = c("gratSCALE", "DvR"), legend.title = "condition")
p2 + labs(title = "pateint ~ gratitude X DvR",
x = "gratitude score",
y = "patience score") +
scale_color_discrete(labels = c("dem", "rep")) +
theme_sjplot()## Scale for 'colour' is already present. Adding another scale for 'colour',
## which will replace the existing scale.
3. oppToImm ~ grat X party
summary(m3<-lmer(OppositionToImm ~ gratSCALE*DvR + (1|region), data = dNoI))## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: OppositionToImm ~ gratSCALE * DvR + (1 | region)
## Data: dNoI
##
## REML criterion at convergence: 871.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.4332 -0.4717 -0.1025 0.5629 2.4500
##
## Random effects:
## Groups Name Variance Std.Dev.
## region (Intercept) 0.02134 0.1461
## Residual 1.14068 1.0680
## Number of obs: 290, groups: region, 6
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 3.86177 0.50865 277.59555 7.592 4.81e-13 ***
## gratSCALE -0.15554 0.07894 284.90955 -1.970 0.0498 *
## DvR 2.03010 1.01566 285.73984 1.999 0.0466 *
## gratSCALE:DvR -0.17667 0.15853 285.99406 -1.114 0.2660
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) grSCALE DvR
## gratSCALE -0.984
## DvR 0.289 -0.301
## grtSCALE:DR -0.303 0.322 -0.991p2 <- plot_model(m3, type = "pred", terms = c("gratSCALE", "DvR"), legend.title = "condition")
p2 + labs(title = "opposition to immigration ~ gratitude X DvR",
x = "gratitude score",
y = "opposition to immigration score") +
scale_color_discrete(labels = c("dem", "rep")) +
theme_sjplot()## Scale for 'colour' is already present. Adding another scale for 'colour',
## which will replace the existing scale.
4. opposition to eligibility for gov’t assistance ~ grat X party
summary(m4<-lmer(OppositionToElig ~ gratSCALE*DvR + (1|region), data = dNoI))## boundary (singular) fit: see ?isSingular## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: OppositionToElig ~ gratSCALE * DvR + (1 | region)
## Data: dNoI
##
## REML criterion at convergence: 859.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.90509 -0.90411 -0.00463 0.89486 1.94897
##
## Random effects:
## Groups Name Variance Std.Dev.
## region (Intercept) 0.000 0.000
## Residual 1.106 1.051
## Number of obs: 290, groups: region, 6
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 3.86945 0.49522 286.00000 7.814 1.07e-13 ***
## gratSCALE -0.24416 0.07741 286.00000 -3.154 0.00178 **
## DvR 2.32005 0.99044 286.00000 2.342 0.01984 *
## gratSCALE:DvR -0.27153 0.15483 286.00000 -1.754 0.08053 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) grSCALE DvR
## gratSCALE -0.991
## DvR 0.293 -0.306
## grtSCALE:DR -0.306 0.326 -0.991
## convergence code: 0
## boundary (singular) fit: see ?isSingularp2 <- plot_model(m4, type = "pred", terms = c("gratSCALE", "DvR"), legend.title = "condition")
p2 + labs(title = "opposition to immigration ~ gratitude X DvR",
x = "gratitude score",
y = "opp to elig for gov't assistance score") +
scale_color_discrete(labels = c("dem", "rep")) +
theme_sjplot()## Scale for 'colour' is already present. Adding another scale for 'colour',
## which will replace the existing scale.