Verbal IQ and Tolerance in the General Social Survey
Packages and Data Description and Formatting
library(pacman); p_load(haven, psych, dplyr, corrr, ggplot2, lavaan, sjmisc, specr, cowplot, ggpubr, pwr, ltm, polycor)Variables used:
Verbal Ability
- wordsum: a 10-item verbal IQ measure (higher = smarter)
Tolerance
- homosex: how wrong is homosexuality (positive)
- racmar: should racial intermarriage be outlawed (positive)
- racdin: would respondent object to opposite race at dinner (positive)
- racfew: would respondent object if a few children at school were opposite race (positive)
- racmost: would respondent object if most children at school were opposite race (positive)
- colath: should anti-religionist be allowed to teach (negative)
- colcom: should communist teacher be fired (positive)
- colhomo: should homosexual be allowed to teach (negative)
- colmil: should militarist be allowed to teach (negative)
- colmslm: should anti-American Muslim clergymen be allowed to teach (negative)
- colrac: should racist be allowed to teach (negative)
- libath: should anti-religious book be removed from public library (positive)
- libcom: should communist book be removed from public library (positive)
- libhomo: should homosexual book be removed from public library (positive)
- libmil: should militarist book be removed from public library (positive)
- libmslm: should anti-America Muslim clergymen book be removed from public library (positive)
- librac: should racist book be removed from public library (positive)
- spkath: should anti-religionist be allowed to make speech (negative)
- spkcom: should communist be allowed to make speech (negative)
- spkhomo: should homosexual be allowed to make speech (negative)
- spkmil: should militarist be allowed to make speech (negative)
- spkmslm: should anti-American Muslim clergymen be allowed to make speech (negative)
- spkrac: should racist be allowed to make speech (negative)
Though the initial codes above were sensible, for clarity, I reverse-coded all variables where negative relationships would mean a tolerance-wordsum association.
Controls
- sex
- race
- age
- year
- reg16: region
- res16: rural/urban
- mobile16: geographic mobility
Verbal ability being related to tolerance would be considered correlations that are in the direction indicated by the bracketed (positive) or (negative) notes above. Note that there is considerable attenuation because wordsum only goes 1-10; rescaling will not be granular because of this, but it will make means comparable to the typical IQ scale (mean = 100, sd = 15). The scaled sumscore is dubbed a “Wordsum IQ”. Wordsum represents a combination of general ability, g, verbal ability, specific skills, and error.
retain <- c("homosex", "racmar", "colath", "racdin", "racfew", "racmost", "colcom", "colhomo", "colmil", "colmslm", "colrac", "libath", "libcom", "libhomo", "libmil", "libmslm", "librac", "spkath", "spkcom", "spkhomo", "spkmil", "spkmslm", "spkrac", "wordsum", "sex", "race", "age", "year", "reg16", "res16", "mobile16", "degree", "polviews", "id")
data <- datafull[retain]; data <- zap_labels(data)
data$wordsum <- (scale(data$wordsum) * 15) + 100#Remove Values
data <- subset(data, homosex <= 4 | is.na(homosex)) #1 = Always wrong, 2 = almost always wrong, 3 = sometimes wrong, 4 = not wrong at all
data <- subset(data, racmar <= 2 | is.na(racmar)) #1 = Yes, 2 = no
data <- subset(data, racdin <= 3 | is.na(racdin)) #1 = Strongly object, 2 = mildly object, 3 = not object
data <- subset(data, racfew <= 2 | is.na(racfew)) #1 = Yes, object, 2 = no
data <- subset(data, racmost <= 2 | is.na(racmost)) #1 = Yes, object, 2 = no
data <- subset(data, colath = 4:5 | is.na(colath)) #4 = Allowed, 5 = not allowed
data <- subset(data, colcom = 4:5 | is.na(colcom)) #4 = Fired, 5 = not fired
data <- subset(data, colhomo = 4:5 | is.na(colhomo)) #4 = Allowed, 5 = not allowed
data <- subset(data, colmil = 4:5 | is.na(colmil)) #4 = Allowed, 5 = not allowed
data <- subset(data, colmslm = 4:5 | is.na(colmslm)) #4 = Yes, allowed, 5 = not allowed
data <- subset(data, colrac = 4:5 | is.na(colrac)) #4 = Allowed, 5 = not allowed
data <- subset(data, libath <= 2 | is.na(libath)) #1 = Remove, 2 = not remove
data <- subset(data, libcom <= 2 | is.na(libcom)) #1 = Remove, 2 = not remove
data <- subset(data, libhomo <= 2 | is.na(libhomo)) #1 = Remove, 2 = not remove
data <- subset(data, libmil <= 2 | is.na(libmil)) #1 = Remove, 2 = not remove
data <- subset(data, libmslm <= 2 | is.na(libmslm)) #1 = Remove, 2 = not remove
data <- subset(data, librac <= 2 | is.na(librac)) #1 = Remove, 2 = not remove
data <- subset(data, spkath <= 2 | is.na(spkath)) #1 = Allowed, 2 = not allowed
data <- subset(data, spkcom <= 2 | is.na(spkcom)) #1 = Allowed, 2 = not allowed
data <- subset(data, spkhomo <= 2 | is.na(spkhomo)) #1 = Allowed, 2 = not allowed
data <- subset(data, spkmil <= 2 | is.na(spkmil)) #1 = Allowed, 2 = not allowed
data <- subset(data, spkmslm <= 2 | is.na(spkmslm)) #1 = Yes, allowed, 2 = not allowed
data <- subset(data, spkrac <= 2 | is.na(spkrac)) #1 = Allowed, 2 = not allowed
data <- subset(data, degree <= 4 | is.na(degree)) #0 = Less than high school, 1 = high school, 2 = junior college, 3 = bachelor, 4 = graduate
data <- subset(data, polviews <= 7 | is.na(polviews)) #1 = Extremely liberal, 2 = liberal, 3 = slightly liberal, 4 = moderate, 5 = slightly conservative, 6 = conservative, 7 = extremely conservative
data$polcat <- rec(data$polviews, rec = "1,2,3 = 1; 4 = 2; 5, 6, 7 = 3")
#Recode and Create, for Logistic Regressions and Dichotomous Comparisons
data$homosexDI <- rec(data$homosex, rec = "1,2,3 = 0; 4 = 1") #Directionally Correct
data$racmar <- rec(data$racmar, rec = "1 = 0; 2 = 1") #Correct
data$racdinDI <- rec(data$racdin, rec = "1,2 = 0; 3 = 1") #Correct
data$racfew <- rec(data$racfew, rec = "1 = 0; 2 = 1") #Correct
data$racmost <- rec(data$racmost, rec = "1 = 0; 2 = 1") #Correct
data$colath <- rec(data$colath, rec = "4 = 1; 5 = 0") #Reversed
data$colcom <- rec(data$colcom, rec = "4 = 0; 5 = 1") #Correct
data$colhomo <- rec(data$colhomo, rec = "4 = 1; 5 = 0") #Reversed
data$colmil <- rec(data$colmil, rec = "4 = 1; 5 = 0") #Reversed
data$colmslm <- rec(data$colmslm, rec = "4 = 1; 5 = 0") #Reversed
data$colrac <- rec(data$colrac, rec = "4 = 1; 5 = 0") #Reversed
data$libath <- rec(data$libath, rec = "1 = 0; 2 = 1") #Correct
data$libcom <- rec(data$libcom, rec = "1 = 0; 2 = 1") #Correct
data$libhomo <- rec(data$libhomo, rec = "1 = 0; 2 = 1") #Correct
data$libmil <- rec(data$libmil, rec = "1 = 0; 2 = 1") #Correct
data$libmslm <- rec(data$libmslm, rec = "1 = 0; 2 = 1") #Correct
data$librac <- rec(data$librac, rec = "1 = 0; 2 = 1") #Correct
data$spkath <- rec(data$spkath, rec = "1 = 1; 2 = 0") #Reversed
data$spkcom <- rec(data$spkcom, rec = "1 = 1; 2 = 0") #Reversed
data$spkhomo <- rec(data$spkhomo, rec = "1 = 1; 2 = 0") #Reversed
data$spkmil <- rec(data$spkmil, rec = "1 = 1; 2 = 0") #Reversed
data$spkmslm <- rec(data$spkmslm, rec = "1 = 1; 2 = 0") #Reversed
data$spkrac <- rec(data$spkrac, rec = "1 = 1; 2 = 0") #Reversed
#Compute a Tolerance Sumscore
data$Tolerance = (scale(data$homosex) + scale(data$racmar) + scale(data$racdin) + scale(data$racfew) + scale(data$racmost) + scale(data$colath) + scale(data$colcom) + scale(data$colhomo) + scale(data$colmil) + scale(data$colrac) + scale(data$libath) + scale(data$libcom) + scale(data$libhomo) + scale(data$libmil) + scale(data$librac) + scale(data$spkath) + scale(data$spkcom) + scale(data$spkhomo) + scale(data$spkmil) + scale(data$spkrac))/20 #Muslim variables removed since they are only found in some years. Transformations not used for obvious reasons.Analyses
Rank Correlations between Wordsum IQ and Other Variables
data %>% correlate(method = "spearman") %>% focus(wordsum)## Warning in stats::cor(x = x, y = y, use = use, method = method): the standard
## deviation is zero
## Warning in stats::cor(x = x, y = y, use = use, method = method): the standard
## deviation is zero##
## Correlation method: 'spearman'
## Missing treated using: 'pairwise.complete.obs'#What is the largest wordsum-dependent variable complete case?
dataWF <- subset(data, !is.na(wordsum))
psych::describe(dataWF)The smallest test is with colmslm, with 4275 complete responses, and the largest is spkath with 18671 cases. This means that statistical power with an expected r of 0.072 ranges from
pwr.r.test(n = 4275, r = 0.072, sig.level = 0.01, alternative = "two.sided")##
## approximate correlation power calculation (arctangh transformation)
##
## n = 4275
## r = 0.072
## sig.level = 0.01
## power = 0.9837682
## alternative = two.sidedpwr.r.test(n = 18671, r = 0.072, sig.level = 0.01, alternative = "two.sided")##
## approximate correlation power calculation (arctangh transformation)
##
## n = 18671
## r = 0.072
## sig.level = 0.01
## power = 1
## alternative = two.sidedBut if we wanted to use our smallest sample at 80% power and p = 0.01, what’s the smallest correlation we could detect?
pwr.r.test(n = 4275, power = 0.8, sig.level = 0.01, alternative = "two.sided")##
## approximate correlation power calculation (arctangh transformation)
##
## n = 4275
## r = 0.05220613
## sig.level = 0.01
## power = 0.8
## alternative = two.sidedround(cor(data, method = "spearman", use = "pairwise.complete"), 3)## Warning in cor(data, method = "spearman", use = "pairwise.complete"): the
## standard deviation is zero
## Warning in cor(data, method = "spearman", use = "pairwise.complete"): the
## standard deviation is zero## homosex racmar colath racdin racfew racmost colcom colhomo colmil
## homosex 1.000 0.265 0.328 0.189 0.064 0.065 0.299 0.412 0.289
## racmar 0.265 1.000 0.301 0.415 0.194 0.138 0.284 0.377 0.229
## colath 0.328 0.301 1.000 0.222 0.086 0.082 0.477 0.455 0.567
## racdin 0.189 0.415 0.222 1.000 0.275 0.191 0.212 0.269 0.178
## racfew 0.064 0.194 0.086 0.275 1.000 NA 0.094 0.120 0.064
## racmost 0.065 0.138 0.082 0.191 NA 1.000 0.099 0.074 0.084
## colcom 0.299 0.284 0.477 0.212 0.094 0.099 1.000 0.395 0.473
## colhomo 0.412 0.377 0.455 0.269 0.120 0.074 0.395 1.000 0.399
## colmil 0.289 0.229 0.567 0.178 0.064 0.084 0.473 0.399 1.000
## colmslm 0.228 NA 0.425 NA NA NA 0.400 0.238 0.499
## colrac 0.159 0.099 0.546 0.067 0.013 0.060 0.332 0.272 0.506
## libath 0.319 0.327 0.448 0.218 0.105 0.028 0.372 0.383 0.328
## libcom 0.307 0.343 0.398 0.243 0.125 0.032 0.526 0.393 0.383
## libhomo 0.399 0.350 0.376 0.241 0.106 0.042 0.380 0.567 0.337
## libmil 0.293 0.277 0.373 0.195 0.078 0.011 0.403 0.350 0.505
## libmslm 0.276 NA 0.329 NA NA NA 0.387 0.236 0.366
## librac 0.212 0.175 0.316 0.105 0.033 -0.006 0.300 0.250 0.302
## spkath 0.278 0.324 0.517 0.215 0.112 0.038 0.338 0.386 0.340
## spkcom 0.296 0.344 0.430 0.258 0.132 0.061 0.505 0.390 0.404
## spkhomo 0.359 0.376 0.373 0.265 0.132 0.044 0.329 0.693 0.318
## spkmil 0.282 0.265 0.397 0.206 0.076 0.041 0.359 0.343 0.619
## spkmslm 0.267 NA 0.347 NA NA NA 0.366 0.244 0.384
## spkrac 0.163 0.149 0.349 0.091 0.031 0.020 0.255 0.230 0.325
## wordsum 0.242 0.216 0.239 0.196 0.116 -0.056 0.233 0.255 0.201
## sex 0.027 -0.025 -0.059 0.026 0.021 -0.025 -0.058 0.015 -0.039
## race -0.057 0.152 -0.041 0.125 0.002 0.226 0.001 0.002 -0.055
## age -0.166 -0.264 -0.232 -0.130 -0.064 -0.011 -0.186 -0.189 -0.194
## year 0.278 0.224 0.176 0.104 0.052 0.049 0.185 0.274 0.150
## reg16 -0.058 -0.056 -0.026 -0.067 -0.020 0.026 -0.053 -0.056 -0.011
## res16 0.186 0.205 0.154 0.133 0.039 -0.008 0.142 0.198 0.112
## mobile16 0.061 0.078 0.056 0.086 0.026 0.014 0.059 0.066 0.046
## degree 0.287 0.323 0.302 0.190 0.113 -0.031 0.281 0.318 0.252
## polviews -0.268 -0.092 -0.131 -0.100 -0.017 -0.062 -0.133 -0.138 -0.108
## id -0.029 0.002 0.003 -0.021 -0.005 0.029 0.009 0.000 0.003
## polcat -0.250 -0.081 -0.122 -0.092 -0.010 -0.058 -0.123 -0.126 -0.101
## homosexDI 0.886 0.210 0.289 0.154 0.056 0.076 0.263 0.350 0.259
## racdinDI 0.185 0.398 0.219 0.990 0.249 0.192 0.209 0.264 0.178
## Tolerance 0.546 0.453 0.722 0.307 NA 0.172 0.676 0.680 0.705
## colmslm colrac libath libcom libhomo libmil libmslm librac spkath
## homosex 0.228 0.159 0.319 0.307 0.399 0.293 0.276 0.212 0.278
## racmar NA 0.099 0.327 0.343 0.350 0.277 NA 0.175 0.324
## colath 0.425 0.546 0.448 0.398 0.376 0.373 0.329 0.316 0.517
## racdin NA 0.067 0.218 0.243 0.241 0.195 NA 0.105 0.215
## racfew NA 0.013 0.105 0.125 0.106 0.078 NA 0.033 0.112
## racmost NA 0.060 0.028 0.032 0.042 0.011 NA -0.006 0.038
## colcom 0.400 0.332 0.372 0.526 0.380 0.403 0.387 0.300 0.338
## colhomo 0.238 0.272 0.383 0.393 0.567 0.350 0.236 0.250 0.386
## colmil 0.499 0.506 0.328 0.383 0.337 0.505 0.366 0.302 0.340
## colmslm 1.000 0.473 0.286 0.336 0.242 0.318 0.565 0.322 0.288
## colrac 0.473 1.000 0.262 0.255 0.221 0.263 0.278 0.401 0.295
## libath 0.286 0.262 1.000 0.628 0.554 0.561 0.450 0.589 0.509
## libcom 0.336 0.255 0.628 1.000 0.579 0.652 0.530 0.551 0.431
## libhomo 0.242 0.221 0.554 0.579 1.000 0.551 0.398 0.445 0.407
## libmil 0.318 0.263 0.561 0.652 0.551 1.000 0.505 0.522 0.385
## libmslm 0.565 0.278 0.450 0.530 0.398 0.505 1.000 0.491 0.344
## librac 0.322 0.401 0.589 0.551 0.445 0.522 0.491 1.000 0.351
## spkath 0.288 0.295 0.509 0.431 0.407 0.385 0.344 0.351 1.000
## spkcom 0.388 0.299 0.445 0.601 0.424 0.447 0.453 0.362 0.567
## spkhomo 0.218 0.215 0.403 0.409 0.573 0.361 0.260 0.277 0.499
## spkmil 0.344 0.314 0.378 0.430 0.379 0.555 0.379 0.340 0.500
## spkmslm 0.673 0.319 0.342 0.400 0.291 0.376 0.631 0.361 0.402
## spkrac 0.363 0.546 0.328 0.318 0.276 0.314 0.368 0.470 0.505
## wordsum 0.256 0.152 0.283 0.311 0.274 0.280 0.306 0.229 0.284
## sex -0.090 -0.068 -0.048 -0.056 -0.006 -0.024 -0.071 -0.048 -0.075
## race -0.102 -0.082 -0.083 -0.072 -0.038 -0.091 -0.119 -0.121 -0.087
## age -0.031 -0.065 -0.169 -0.174 -0.176 -0.168 -0.027 -0.055 -0.185
## year 0.015 0.032 0.124 0.116 0.198 0.120 0.003 0.014 0.122
## reg16 0.015 -0.020 -0.048 -0.047 -0.051 -0.031 0.027 -0.038 -0.019
## res16 0.064 0.064 0.165 0.160 0.180 0.131 0.091 0.086 0.153
## mobile16 0.038 0.037 0.069 0.077 0.063 0.068 0.048 0.057 0.056
## degree 0.239 0.148 0.306 0.325 0.308 0.287 0.289 0.204 0.301
## polviews -0.125 -0.052 -0.111 -0.105 -0.123 -0.091 -0.128 -0.063 -0.080
## id -0.038 -0.018 -0.031 -0.029 -0.020 -0.028 -0.056 -0.036 -0.015
## polcat -0.113 -0.050 -0.101 -0.094 -0.111 -0.083 -0.118 -0.057 -0.073
## homosexDI 0.221 0.138 0.274 0.262 0.339 0.258 0.267 0.182 0.240
## racdinDI NA 0.070 0.212 0.236 0.234 0.192 NA 0.104 0.208
## Tolerance NA 0.615 0.724 0.740 0.722 0.721 NA 0.664 0.687
## spkcom spkhomo spkmil spkmslm spkrac wordsum sex race age
## homosex 0.296 0.359 0.282 0.267 0.163 0.242 0.027 -0.057 -0.166
## racmar 0.344 0.376 0.265 NA 0.149 0.216 -0.025 0.152 -0.264
## colath 0.430 0.373 0.397 0.347 0.349 0.239 -0.059 -0.041 -0.232
## racdin 0.258 0.265 0.206 NA 0.091 0.196 0.026 0.125 -0.130
## racfew 0.132 0.132 0.076 NA 0.031 0.116 0.021 0.002 -0.064
## racmost 0.061 0.044 0.041 NA 0.020 -0.056 -0.025 0.226 -0.011
## colcom 0.505 0.329 0.359 0.366 0.255 0.233 -0.058 0.001 -0.186
## colhomo 0.390 0.693 0.343 0.244 0.230 0.255 0.015 0.002 -0.189
## colmil 0.404 0.318 0.619 0.384 0.325 0.201 -0.039 -0.055 -0.194
## colmslm 0.388 0.218 0.344 0.673 0.363 0.256 -0.090 -0.102 -0.031
## colrac 0.299 0.215 0.314 0.319 0.546 0.152 -0.068 -0.082 -0.065
## libath 0.445 0.403 0.378 0.342 0.328 0.283 -0.048 -0.083 -0.169
## libcom 0.601 0.409 0.430 0.400 0.318 0.311 -0.056 -0.072 -0.174
## libhomo 0.424 0.573 0.379 0.291 0.276 0.274 -0.006 -0.038 -0.176
## libmil 0.447 0.361 0.555 0.376 0.314 0.280 -0.024 -0.091 -0.168
## libmslm 0.453 0.260 0.379 0.631 0.368 0.306 -0.071 -0.119 -0.027
## librac 0.362 0.277 0.340 0.361 0.470 0.229 -0.048 -0.121 -0.055
## spkath 0.567 0.499 0.500 0.402 0.505 0.284 -0.075 -0.087 -0.185
## spkcom 1.000 0.490 0.550 0.518 0.463 0.315 -0.086 -0.062 -0.160
## spkhomo 0.490 1.000 0.435 0.289 0.333 0.275 -0.002 -0.031 -0.151
## spkmil 0.550 0.435 1.000 0.451 0.460 0.255 -0.033 -0.083 -0.177
## spkmslm 0.518 0.289 0.451 1.000 0.496 0.291 -0.123 -0.123 -0.032
## spkrac 0.463 0.333 0.460 0.496 1.000 0.205 -0.101 -0.092 -0.068
## wordsum 0.315 0.275 0.255 0.291 0.205 1.000 0.021 -0.246 0.069
## sex -0.086 -0.002 -0.033 -0.123 -0.101 0.021 1.000 0.031 0.029
## race -0.062 -0.031 -0.083 -0.123 -0.092 -0.246 0.031 1.000 -0.102
## age -0.160 -0.151 -0.177 -0.032 -0.068 0.069 0.029 -0.102 1.000
## year 0.100 0.207 0.121 0.024 0.000 0.018 -0.001 0.122 0.081
## reg16 -0.032 -0.044 0.007 0.051 0.002 -0.050 0.005 0.021 -0.061
## res16 0.160 0.185 0.119 0.093 0.078 0.140 0.016 0.105 -0.119
## mobile16 0.075 0.062 0.050 0.044 0.038 0.120 -0.001 0.004 0.123
## degree 0.322 0.299 0.281 0.275 0.189 0.458 -0.034 -0.091 -0.107
## polviews -0.101 -0.108 -0.082 -0.129 -0.047 -0.031 -0.033 -0.092 0.118
## id -0.026 -0.016 -0.012 -0.050 -0.026 -0.058 -0.005 -0.023 0.037
## polcat -0.091 -0.097 -0.076 -0.118 -0.043 -0.022 -0.035 -0.091 0.112
## homosexDI 0.256 0.298 0.248 0.256 0.142 0.198 0.033 -0.041 -0.140
## racdinDI 0.251 0.256 0.203 NA 0.091 0.191 0.024 0.124 -0.127
## Tolerance 0.710 0.673 0.696 NA 0.619 0.430 -0.044 -0.077 -0.312
## year reg16 res16 mobile16 degree polviews id polcat homosexDI
## homosex 0.278 -0.058 0.186 0.061 0.287 -0.268 -0.029 -0.250 0.886
## racmar 0.224 -0.056 0.205 0.078 0.323 -0.092 0.002 -0.081 0.210
## colath 0.176 -0.026 0.154 0.056 0.302 -0.131 0.003 -0.122 0.289
## racdin 0.104 -0.067 0.133 0.086 0.190 -0.100 -0.021 -0.092 0.154
## racfew 0.052 -0.020 0.039 0.026 0.113 -0.017 -0.005 -0.010 0.056
## racmost 0.049 0.026 -0.008 0.014 -0.031 -0.062 0.029 -0.058 0.076
## colcom 0.185 -0.053 0.142 0.059 0.281 -0.133 0.009 -0.123 0.263
## colhomo 0.274 -0.056 0.198 0.066 0.318 -0.138 0.000 -0.126 0.350
## colmil 0.150 -0.011 0.112 0.046 0.252 -0.108 0.003 -0.101 0.259
## colmslm 0.015 0.015 0.064 0.038 0.239 -0.125 -0.038 -0.113 0.221
## colrac 0.032 -0.020 0.064 0.037 0.148 -0.052 -0.018 -0.050 0.138
## libath 0.124 -0.048 0.165 0.069 0.306 -0.111 -0.031 -0.101 0.274
## libcom 0.116 -0.047 0.160 0.077 0.325 -0.105 -0.029 -0.094 0.262
## libhomo 0.198 -0.051 0.180 0.063 0.308 -0.123 -0.020 -0.111 0.339
## libmil 0.120 -0.031 0.131 0.068 0.287 -0.091 -0.028 -0.083 0.258
## libmslm 0.003 0.027 0.091 0.048 0.289 -0.128 -0.056 -0.118 0.267
## librac 0.014 -0.038 0.086 0.057 0.204 -0.063 -0.036 -0.057 0.182
## spkath 0.122 -0.019 0.153 0.056 0.301 -0.080 -0.015 -0.073 0.240
## spkcom 0.100 -0.032 0.160 0.075 0.322 -0.101 -0.026 -0.091 0.256
## spkhomo 0.207 -0.044 0.185 0.062 0.299 -0.108 -0.016 -0.097 0.298
## spkmil 0.121 0.007 0.119 0.050 0.281 -0.082 -0.012 -0.076 0.248
## spkmslm 0.024 0.051 0.093 0.044 0.275 -0.129 -0.050 -0.118 0.256
## spkrac 0.000 0.002 0.078 0.038 0.189 -0.047 -0.026 -0.043 0.142
## wordsum 0.018 -0.050 0.140 0.120 0.458 -0.031 -0.058 -0.022 0.198
## sex -0.001 0.005 0.016 -0.001 -0.034 -0.033 -0.005 -0.035 0.033
## race 0.122 0.021 0.105 0.004 -0.091 -0.092 -0.023 -0.091 -0.041
## age 0.081 -0.061 -0.119 0.123 -0.107 0.118 0.037 0.112 -0.140
## year 1.000 0.008 0.084 0.047 0.222 0.007 0.274 0.005 0.294
## reg16 0.008 1.000 -0.091 -0.158 -0.054 0.039 0.196 0.034 -0.039
## res16 0.084 -0.091 1.000 0.017 0.170 -0.088 -0.194 -0.082 0.164
## mobile16 0.047 -0.158 0.017 1.000 0.149 -0.002 -0.034 -0.001 0.053
## degree 0.222 -0.054 0.170 0.149 1.000 -0.028 -0.008 -0.019 0.244
## polviews 0.007 0.039 -0.088 -0.002 -0.028 1.000 0.066 0.973 -0.243
## id 0.274 0.196 -0.194 -0.034 -0.008 0.066 1.000 0.060 -0.013
## polcat 0.005 0.034 -0.082 -0.001 -0.019 0.973 0.060 1.000 -0.226
## homosexDI 0.294 -0.039 0.164 0.053 0.244 -0.243 -0.013 -0.226 1.000
## racdinDI 0.103 -0.063 0.128 0.086 0.183 -0.100 -0.019 -0.093 0.153
## Tolerance -0.006 -0.112 0.236 0.098 0.436 -0.201 -0.109 -0.186 0.433
## racdinDI Tolerance
## homosex 0.185 0.546
## racmar 0.398 0.453
## colath 0.219 0.722
## racdin 0.990 0.307
## racfew 0.249 NA
## racmost 0.192 0.172
## colcom 0.209 0.676
## colhomo 0.264 0.680
## colmil 0.178 0.705
## colmslm NA NA
## colrac 0.070 0.615
## libath 0.212 0.724
## libcom 0.236 0.740
## libhomo 0.234 0.722
## libmil 0.192 0.721
## libmslm NA NA
## librac 0.104 0.664
## spkath 0.208 0.687
## spkcom 0.251 0.710
## spkhomo 0.256 0.673
## spkmil 0.203 0.696
## spkmslm NA NA
## spkrac 0.091 0.619
## wordsum 0.191 0.430
## sex 0.024 -0.044
## race 0.124 -0.077
## age -0.127 -0.312
## year 0.103 -0.006
## reg16 -0.063 -0.112
## res16 0.128 0.236
## mobile16 0.086 0.098
## degree 0.183 0.436
## polviews -0.100 -0.201
## id -0.019 -0.109
## polcat -0.093 -0.186
## homosexDI 0.153 0.433
## racdinDI 1.000 0.304
## Tolerance 0.304 1.000What if racmost is removed from the sumscore?
data$TolTest = (scale(data$homosex) + scale(data$racmar) + scale(data$racdin) + scale(data$racfew) + scale(data$colath) + scale(data$colcom) + scale(data$colhomo) + scale(data$colmil) + scale(data$colrac) + scale(data$libath) + scale(data$libcom) + scale(data$libhomo) + scale(data$libmil) + scale(data$librac) + scale(data$spkath) + scale(data$spkcom) + scale(data$spkhomo) + scale(data$spkmil) + scale(data$spkrac))/19
cor(data$TolTest, data$wordsum, method = "spearman", use = "complete.obs")## [,1]
## [1,] 0.4549631Biserial Correlations
First, the biserial correlations:
polyserial(data$wordsum, data$homosexDI)## [1] 0.2665036polyserial(data$wordsum, data$racmar)## [1] 0.2986141polyserial(data$wordsum, data$racdinDI)## [1] 0.2651545polyserial(data$wordsum, data$racfew)## [1] 0.2788201polyserial(data$wordsum, data$racmost)## [1] -0.07194818polyserial(data$wordsum, data$colath)## [1] 0.2965005polyserial(data$wordsum, data$colcom)## [1] 0.2901009polyserial(data$wordsum, data$colhomo)## [1] 0.3414164polyserial(data$wordsum, data$colmil)## [1] 0.2472782polyserial(data$wordsum, data$colmslm)## [1] 0.3299984polyserial(data$wordsum, data$colrac)## [1] 0.1882307polyserial(data$wordsum, data$libath)## [1] 0.3668813polyserial(data$wordsum, data$libcom)## [1] 0.3987142polyserial(data$wordsum, data$libhomo)## [1] 0.356505polyserial(data$wordsum, data$libmil)## [1] 0.3531073polyserial(data$wordsum, data$libmslm)## [1] 0.3812085polyserial(data$wordsum, data$librac)## [1] 0.2879645polyserial(data$wordsum, data$spkath)## [1] 0.3786574polyserial(data$wordsum, data$spkcom)## [1] 0.4015908polyserial(data$wordsum, data$spkhomo)## [1] 0.3884629polyserial(data$wordsum, data$spkmil)## [1] 0.3211459polyserial(data$wordsum, data$spkmslm)## [1] 0.3669985polyserial(data$wordsum, data$spkrac)## [1] 0.2636467Then, these are the point-biserial correlations. Note: I do not believe that tolerance is an A/B, pass/fail, yes/no question, and it is in fact artificially dichotomized for two questions, and those two questions show that intermediate tolerance levels have higher wordsum scores than lower levels, and the highest level of tolerance was in both cases even smarter, so there is some evidence against the use/interpretation of the point-biserial.
biserial.cor(data$wordsum, data$homosexDI, use = "complete.obs", level = 2)## [1] 0.1960655biserial.cor(data$wordsum, data$racmar, use = "complete.obs", level = 2)## [1] 0.2110091biserial.cor(data$wordsum, data$racdinDI, use = "complete.obs", level = 2)## [1] 0.1889019biserial.cor(data$wordsum, data$racfew, use = "complete.obs", level = 2)## [1] 0.1210425biserial.cor(data$wordsum, data$racmost, use = "complete.obs", level = 2)## [1] -0.05651882biserial.cor(data$wordsum, data$colath, use = "complete.obs", level = 2)## [1] 0.2359337biserial.cor(data$wordsum, data$colcom, use = "complete.obs", level = 2)## [1] 0.2305561biserial.cor(data$wordsum, data$colhomo, use = "complete.obs", level = 2)## [1] 0.2580286biserial.cor(data$wordsum, data$colmil, use = "complete.obs", level = 2)## [1] 0.1972631biserial.cor(data$wordsum, data$colmslm, use = "complete.obs", level = 2)## [1] 0.2541915biserial.cor(data$wordsum, data$colrac, use = "complete.obs", level = 2)## [1] 0.1498745biserial.cor(data$wordsum, data$libath, use = "complete.obs", level = 2)## [1] 0.2798046biserial.cor(data$wordsum, data$libcom, use = "complete.obs", level = 2)## [1] 0.3084304biserial.cor(data$wordsum, data$libhomo, use = "complete.obs", level = 2)## [1] 0.2729094biserial.cor(data$wordsum, data$libmil, use = "complete.obs", level = 2)## [1] 0.2742104biserial.cor(data$wordsum, data$libmslm, use = "complete.obs", level = 2)## [1] 0.3041215biserial.cor(data$wordsum, data$librac, use = "complete.obs", level = 2)## [1] 0.2242504biserial.cor(data$wordsum, data$spkath, use = "complete.obs", level = 2)## [1] 0.2836143biserial.cor(data$wordsum, data$spkcom, use = "complete.obs", level = 2)## [1] 0.312188biserial.cor(data$wordsum, data$spkhomo, use = "complete.obs", level = 2)## [1] 0.277902biserial.cor(data$wordsum, data$spkmil, use = "complete.obs", level = 2)## [1] 0.2508111biserial.cor(data$wordsum, data$spkmslm, use = "complete.obs", level = 2)## [1] 0.2915239biserial.cor(data$wordsum, data$spkrac, use = "complete.obs", level = 2)## [1] 0.2072019Bar Charts
Only the cumulative bar chart is used in the final paper.
HomosexD <- data %>%
group_by(homosex) %>%
summarise(
n = n(),
mean = mean(wordsum, na.rm = T),
sd = sd(wordsum, na.rm = T)) %>%
mutate(se = sd/sqrt(n)) %>%
mutate(ic = se * qt((1 - 0.05)/2 + 0.5, n - 1))
HD <- ggplot(HomosexD) +
geom_bar(aes(x = homosex, y = mean), stat = "identity", position = "dodge", fill = "steelblue", width = 0.9, alpha = 0.5) +
geom_errorbar(aes(x = homosex, ymin = mean - ic, ymax = mean + ic), width = 0.4, color = "orange", alpha = 0.9, size = 1) +
theme_bw() + xlab("") + ylab("") +
scale_x_continuous(breaks = c(1, 2, 3, 4), labels = c("AW", "AAW", "SW", "NWA")) +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1)); HDD <- HD + ylab("Wordsum IQ") + xlab("Position on Homosexuality") + scale_x_continuous(breaks = c(1, 2, 3, 4), labels = c("Always Wrong", "Almost Always Wrong", "Sometimes Wrong", "Not Wrong at All")); HDD## Scale for 'x' is already present. Adding another scale for 'x', which will
## replace the existing scale.## Warning: Removed 1 rows containing missing values (geom_bar).
homosexDID <- data %>%
group_by(homosexDI) %>%
summarise(
n = n(),
mean = mean(wordsum, na.rm = T),
sd = sd(wordsum, na.rm = T)) %>%
mutate(se = sd/sqrt(n)) %>%
mutate(ic = se * qt((1 - 0.05)/2 + 0.5, n - 1))
HDI <- ggplot(homosexDID) +
geom_bar(aes(x = homosexDI, y = mean), stat = "identity", position = "dodge", fill = "steelblue", width = 0.9, alpha = 0.5) +
geom_errorbar(aes(x = homosexDI, ymin = mean - ic, ymax = mean + ic), width = 0.4, color = "orange", alpha = 0.9, size = 1) +
theme_bw() + ylab("") + xlab("") +
scale_x_continuous(breaks = c(0, 1), labels = c("Wrong", "Not Wrong")) +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1)); HDID <- HDI + ylab("Wordsum IQ") + xlab("Position on Homosexuality"); HDID## Warning: Removed 1 rows containing missing values (geom_bar).
racmarD <- data %>%
group_by(racmar) %>%
summarise(
n = n(),
mean = mean(wordsum, na.rm = T),
sd = sd(wordsum, na.rm = T)) %>%
mutate(se = sd/sqrt(n)) %>%
mutate(ic = se * qt((1 - 0.05)/2 + 0.5, n - 1))
RD <- ggplot(racmarD) +
geom_bar(aes(x = racmar, y = mean), stat = "identity", position = "dodge", fill = "lightgreen", width = 0.9, alpha = 0.5) +
geom_errorbar(aes(x = racmar, ymin = mean - ic, ymax = mean + ic), width = 0.4, color = "violet", alpha = 0.9, size = 1) +
theme_bw() + ylab("") + xlab("") +
scale_x_continuous(breaks = c(0, 1), labels = c("Support", "Oppose")) +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1)); RDD <- RD + ylab("Wordsum IQ") + xlab("Position on Laws Against Racial Intermarriage"); RDD## Warning: Removed 1 rows containing missing values (geom_bar).
colathD <- data %>%
group_by(colath) %>%
summarise(
n = n(),
mean = mean(wordsum, na.rm = T),
sd = sd(wordsum, na.rm = T)) %>%
mutate(se = sd/sqrt(n)) %>%
mutate(ic = se * qt((1 - 0.05)/2 + 0.5, n - 1))
CA <- ggplot(colathD) +
geom_bar(aes(x = colath, y = mean), stat = "identity", position = "dodge", fill = "orangered", width = 0.9, alpha = 0.5) +
geom_errorbar(aes(x = colath, ymin = mean - ic, ymax = mean + ic), width = 0.4, color = "steelblue", alpha = 0.9, size = 1) +
theme_bw() + ylab("") + xlab("") +
scale_x_continuous(breaks = c(0, 1), labels = c("Disallow", "Allow")) +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1)); CAD <- CA + ylab("Wordsum IQ") + xlab("Position on Whether Anti-Religionists Should be Allowed to Teach"); CAD## Warning: Removed 1 rows containing missing values (geom_bar).
colcomD <- data %>%
group_by(colcom) %>%
summarise(
n = n(),
mean = mean(wordsum, na.rm = T),
sd = sd(wordsum, na.rm = T)) %>%
mutate(se = sd/sqrt(n)) %>%
mutate(ic = se * qt((1 - 0.05)/2 + 0.5, n - 1))
CC <- ggplot(colcomD) +
geom_bar(aes(x = colcom, y = mean), stat = "identity", position = "dodge", fill = "orangered", width = 0.9, alpha = 0.5) +
geom_errorbar(aes(x = colcom, ymin = mean - ic, ymax = mean + ic), width = 0.4, color = "steelblue", alpha = 0.9, size = 1) +
theme_bw() + ylab("") + xlab("") +
scale_x_continuous(breaks = c(0, 1), labels = c("Fire", "Do Not Fire")) +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1)); CCD <- CC + ylab("Wordsum IQ") + xlab("Position on Whether Communists Should be Allowed to Teach")
colhomoD <- data %>%
group_by(colhomo) %>%
summarise(
n = n(),
mean = mean(wordsum, na.rm = T),
sd = sd(wordsum, na.rm = T)) %>%
mutate(se = sd/sqrt(n)) %>%
mutate(ic = se * qt((1 - 0.05)/2 + 0.5, n - 1))
CH <- ggplot(colhomoD) +
geom_bar(aes(x = colhomo, y = mean), stat = "identity", position = "dodge", fill = "orangered", width = 0.9, alpha = 0.5) +
geom_errorbar(aes(x = colhomo, ymin = mean - ic, ymax = mean + ic), width = 0.4, color = "steelblue", alpha = 0.9, size = 1) +
theme_bw() + ylab("") + xlab("") +
scale_x_continuous(breaks = c(0, 1), labels = c("Disallow", "Allow")) +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1)); CHD <- CH + ylab("Wordsum IQ") + xlab("Position on Whether Homosexuals Should be Allowed to Teach"); CHD## Warning: Removed 1 rows containing missing values (geom_bar).
colmilD <- data %>%
group_by(colmil) %>%
summarise(
n = n(),
mean = mean(wordsum, na.rm = T),
sd = sd(wordsum, na.rm = T)) %>%
mutate(se = sd/sqrt(n)) %>%
mutate(ic = se * qt((1 - 0.05)/2 + 0.5, n - 1))
CM <- ggplot(colmilD) +
geom_bar(aes(x = colmil, y = mean), stat = "identity", position = "dodge", fill = "orangered", width = 0.9, alpha = 0.5) +
geom_errorbar(aes(x = colmil, ymin = mean - ic, ymax = mean + ic), width = 0.4, color = "steelblue", alpha = 0.9, size = 1) +
theme_bw() + ylab("") + xlab("") +
scale_x_continuous(breaks = c(0, 1), labels = c("Disallow", "Allow")) +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1)); CMD <- CM + ylab("Wordsum IQ") + xlab("Position on Whether Militarists Should be Allowed to Teach"); CMD## Warning: Removed 1 rows containing missing values (geom_bar).
colmslmD <- data %>%
group_by(colmslm) %>%
summarise(
n = n(),
mean = mean(wordsum, na.rm = T),
sd = sd(wordsum, na.rm = T)) %>%
mutate(se = sd/sqrt(n)) %>%
mutate(ic = se * qt((1 - 0.05)/2 + 0.5, n - 1))
CU <- ggplot(colmslmD) +
geom_bar(aes(x = colmslm, y = mean), stat = "identity", position = "dodge", fill = "orangered", width = 0.9, alpha = 0.5) +
geom_errorbar(aes(x = colmslm, ymin = mean - ic, ymax = mean + ic), width = 0.4, color = "steelblue", alpha = 0.9, size = 1) +
theme_bw() + ylab("") + xlab("") +
scale_x_continuous(breaks = c(0, 1), labels = c("Disallow", "Allow")) +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1)); CUD <- CU + ylab("Wordsum IQ") + xlab("Position on Whether Anti-American Muslim Clerics Should be Allowed to Teach"); CUD ## Warning: Removed 1 rows containing missing values (geom_bar).
colracD <- data %>%
group_by(colrac) %>%
summarise(
n = n(),
mean = mean(wordsum, na.rm = T),
sd = sd(wordsum, na.rm = T)) %>%
mutate(se = sd/sqrt(n)) %>%
mutate(ic = se * qt((1 - 0.05)/2 + 0.5, n - 1))
CR <- ggplot(colracD) +
geom_bar(aes(x = colrac, y = mean), stat = "identity", position = "dodge", fill = "orangered", width = 0.9, alpha = 0.5) +
geom_errorbar(aes(x = colrac, ymin = mean - ic, ymax = mean + ic), width = 0.4, color = "steelblue", alpha = 0.9, size = 1) +
theme_bw() + ylab("") + xlab("") +
scale_x_continuous(breaks = c(0, 1), labels = c("Disallow", "Allow")) +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1)); CRD <- CR + ylab("Wordsum IQ") + xlab("Position on Whether Racists Should be Allowed to Teach"); CRD## Warning: Removed 1 rows containing missing values (geom_bar).
racdinD <- data %>%
group_by(racdin) %>%
summarise(
n = n(),
mean = mean(wordsum, na.rm = T),
sd = sd(wordsum, na.rm = T)) %>%
mutate(se = sd/sqrt(n)) %>%
mutate(ic = se * qt((1 - 0.05)/2 + 0.5, n - 1))
CRA <- ggplot(racdinD) +
geom_bar(aes(x = racdin, y = mean), stat = "identity", position = "dodge", fill = "gold", width = 0.9, alpha = 0.5) +
geom_errorbar(aes(x = racdin, ymin = mean - ic, ymax = mean + ic), width = 0.4, color = "steelblue", alpha = 0.9, size = 1) +
theme_bw() + ylab("") + xlab("") +
scale_x_continuous(breaks = c(1, 2, 3), labels = c("Strongly Object", "Mildly Object", "Not Object")) +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1)); CRAD <- CRA + ylab("Wordsum IQ") + xlab("If Respondent would Object to a Member of Another Race at Dinner"); CRAD## Warning: Removed 1 rows containing missing values (geom_bar).
racdinDID <- data %>%
group_by(racdinDI) %>%
summarise(
n = n(),
mean = mean(wordsum, na.rm = T),
sd = sd(wordsum, na.rm = T)) %>%
mutate(se = sd/sqrt(n)) %>%
mutate(ic = se * qt((1 - 0.05)/2 + 0.5, n - 1))
CRD <- ggplot(racdinDID) +
geom_bar(aes(x = racdinDI, y = mean), stat = "identity", position = "dodge", fill = "gold", width = 0.9, alpha = 0.5) +
geom_errorbar(aes(x = racdinDI, ymin = mean - ic, ymax = mean + ic), width = 0.4, color = "steelblue", alpha = 0.9, size = 1) +
theme_bw() + ylab("") + xlab("") +
scale_x_continuous(breaks = c(0, 1), labels = c("Object", "Not Object")) +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1)); CRDD <- CRD + ylab("Wordsum IQ") + xlab("If Respondent would Object to a Member of Another Race at Dinner"); CRDD## Warning: Removed 1 rows containing missing values (geom_bar).
racfewD <- data %>%
group_by(racfew) %>%
summarise(
n = n(),
mean = mean(wordsum, na.rm = T),
sd = sd(wordsum, na.rm = T)) %>%
mutate(se = sd/sqrt(n)) %>%
mutate(ic = se * qt((1 - 0.05)/2 + 0.5, n - 1))
RF <- ggplot(racfewD) +
geom_bar(aes(x = racfew, y = mean), stat = "identity", position = "dodge", fill = "indianred4", width = 0.9, alpha = 0.5) +
geom_errorbar(aes(x = racfew, ymin = mean - ic, ymax = mean + ic), width = 0.4, color = "snow4", alpha = 0.9, size = 1) +
theme_bw() + ylab("") + xlab("") +
scale_x_continuous(breaks = c(0, 1), labels = c("Yes, Object", "Do not Object")) +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1)); RFD <- RF + ylab("Wordsum IQ") + xlab("If Respondent would Object to Sending their Children to School with a Few Members of Other Races"); RFD## Warning: Removed 1 rows containing missing values (geom_bar).
racmostD <- data %>%
group_by(racmost) %>%
summarise(
n = n(),
mean = mean(wordsum, na.rm = T),
sd = sd(wordsum, na.rm = T)) %>%
mutate(se = sd/sqrt(n)) %>%
mutate(ic = se * qt((1 - 0.05)/2 + 0.5, n - 1))
RM <- ggplot(racmostD) +
geom_bar(aes(x = racmost, y = mean), stat = "identity", position = "dodge", fill = "indianred4", width = 0.9, alpha = 0.5) +
geom_errorbar(aes(x = racmost, ymin = mean - ic, ymax = mean + ic), width = 0.4, color = "snow4", alpha = 0.9, size = 1) +
theme_bw() + ylab("") + xlab("") +
scale_x_continuous(breaks = c(0, 1), labels = c("Yes, Object", "Do not Object")) +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1)); RMD <- RM + ylab("Wordsum IQ") + xlab("If Respondent would Object to Sending their Children to a School that is Mostly Racially Different"); RMD## Warning: Removed 1 rows containing missing values (geom_bar).
libathD <- data %>%
group_by(libath) %>%
summarise(
n = n(),
mean = mean(wordsum, na.rm = T),
sd = sd(wordsum, na.rm = T)) %>%
mutate(se = sd/sqrt(n)) %>%
mutate(ic = se * qt((1 - 0.05)/2 + 0.5, n - 1))
LA <- ggplot(libathD) +
geom_bar(aes(x = libath, y = mean), stat = "identity", position = "dodge", fill = "#cc6699", width = 0.9, alpha = 0.5) +
geom_errorbar(aes(x = libath, ymin = mean - ic, ymax = mean + ic), width = 0.4, color = "#3399ff", alpha = 0.9, size = 1) +
theme_bw() + ylab("") + xlab("") +
scale_x_continuous(breaks = c(0, 1), labels = c("Remove", "Do not Remove")) +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1)); LAD <- LA + ylab("Wordsum IQ") + xlab("Position on if Anti-Religious Books should remain in Public Libraries"); LAD## Warning: Removed 1 rows containing missing values (geom_bar).
libcomD <- data %>%
group_by(libcom) %>%
summarise(
n = n(),
mean = mean(wordsum, na.rm = T),
sd = sd(wordsum, na.rm = T)) %>%
mutate(se = sd/sqrt(n)) %>%
mutate(ic = se * qt((1 - 0.05)/2 + 0.5, n - 1))
LC <- ggplot(libcomD) +
geom_bar(aes(x = libcom, y = mean), stat = "identity", position = "dodge", fill = "#cc6699", width = 0.9, alpha = 0.5) +
geom_errorbar(aes(x = libcom, ymin = mean - ic, ymax = mean + ic), width = 0.4, color = "#3399ff", alpha = 0.9, size = 1) +
theme_bw() + ylab("") + xlab("") +
scale_x_continuous(breaks = c(0, 1), labels = c("Remove", "Do not Remove")) +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1)); LCD <- LC + ylab("Wordsum IQ") + xlab("Position on if Communisted-authored Books should remain in Public Libraries"); LCD## Warning: Removed 1 rows containing missing values (geom_bar).
libhomoD <- data %>%
group_by(libhomo) %>%
summarise(
n = n(),
mean = mean(wordsum, na.rm = T),
sd = sd(wordsum, na.rm = T)) %>%
mutate(se = sd/sqrt(n)) %>%
mutate(ic = se * qt((1 - 0.05)/2 + 0.5, n - 1))
LH <- ggplot(libhomoD) +
geom_bar(aes(x = libhomo, y = mean), stat = "identity", position = "dodge", fill = "#cc6699", width = 0.9, alpha = 0.5) +
geom_errorbar(aes(x = libhomo, ymin = mean - ic, ymax = mean + ic), width = 0.4, color = "#3399ff", alpha = 0.9, size = 1) +
theme_bw() + ylab("") + xlab("") +
scale_x_continuous(breaks = c(0, 1), labels = c("Remove", "Do not Remove")) +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1)); LHD <- LH + ylab("Wordsum IQ") + xlab("Position on if Pro-Homosexual Books should remain in Public Libraries"); LHD## Warning: Removed 1 rows containing missing values (geom_bar).
libmilD <- data %>%
group_by(libmil) %>%
summarise(
n = n(),
mean = mean(wordsum, na.rm = T),
sd = sd(wordsum, na.rm = T)) %>%
mutate(se = sd/sqrt(n)) %>%
mutate(ic = se * qt((1 - 0.05)/2 + 0.5, n - 1))
LM <- ggplot(libmilD) +
geom_bar(aes(x = libmil, y = mean), stat = "identity", position = "dodge", fill = "#cc6699", width = 0.9, alpha = 0.5) +
geom_errorbar(aes(x = libmil, ymin = mean - ic, ymax = mean + ic), width = 0.4, color = "#3399ff", alpha = 0.9, size = 1) +
theme_bw() + ylab("") + xlab("") +
scale_x_continuous(breaks = c(0, 1), labels = c("Remove", "Do not Remove")) +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1)); LMD <- LM + ylab("Wordsum IQ") + xlab("Position on if Militarist Books should remain in Public Libraries"); LMD## Warning: Removed 1 rows containing missing values (geom_bar).
libmslmD <- data %>%
group_by(libmslm) %>%
summarise(
n = n(),
mean = mean(wordsum, na.rm = T),
sd = sd(wordsum, na.rm = T)) %>%
mutate(se = sd/sqrt(n)) %>%
mutate(ic = se * qt((1 - 0.05)/2 + 0.5, n - 1))
LU <- ggplot(libmslmD) +
geom_bar(aes(x = libmslm, y = mean), stat = "identity", position = "dodge", fill = "#cc6699", width = 0.9, alpha = 0.5) +
geom_errorbar(aes(x = libmslm, ymin = mean - ic, ymax = mean + ic), width = 0.4, color = "#3399ff", alpha = 0.9, size = 1) +
theme_bw() + ylab("") + xlab("") +
scale_x_continuous(breaks = c(0, 1), labels = c("Remove", "Do not Remove")) +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1)); LUD <- LU + ylab("Wordsum IQ") + xlab("Position on if Anti-American Muslim Clergyman Books should remain in Public Libraries"); LUD## Warning: Removed 1 rows containing missing values (geom_bar).
libracD <- data %>%
group_by(librac) %>%
summarise(
n = n(),
mean = mean(wordsum, na.rm = T),
sd = sd(wordsum, na.rm = T)) %>%
mutate(se = sd/sqrt(n)) %>%
mutate(ic = se * qt((1 - 0.05)/2 + 0.5, n - 1))
LR <- ggplot(libracD) +
geom_bar(aes(x = librac, y = mean), stat = "identity", position = "dodge", fill = "#cc6699", width = 0.9, alpha = 0.5) +
geom_errorbar(aes(x = librac, ymin = mean - ic, ymax = mean + ic), width = 0.4, color = "#3399ff", alpha = 0.9, size = 1) +
theme_bw() + ylab("") + xlab("") +
scale_x_continuous(breaks = c(0, 1), labels = c("Remove", "Do not Remove")) +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1)); LRD <- LR + ylab("Wordsum IQ") + xlab("Position on if Racist Books should remain in Public Libraries"); LRD## Warning: Removed 1 rows containing missing values (geom_bar).
spkathD <- data %>%
group_by(spkath) %>%
summarise(
n = n(),
mean = mean(wordsum, na.rm = T),
sd = sd(wordsum, na.rm = T)) %>%
mutate(se = sd/sqrt(n)) %>%
mutate(ic = se * qt((1 - 0.05)/2 + 0.5, n - 1))
SA <- ggplot(spkathD) +
geom_bar(aes(x = spkath, y = mean), stat = "identity", position = "dodge", fill = "#009999", width = 0.9, alpha = 0.5) +
geom_errorbar(aes(x = spkath, ymin = mean - ic, ymax = mean + ic), width = 0.4, color = "#666699", alpha = 0.9, size = 1) +
theme_bw() + ylab("") + xlab("") +
scale_x_continuous(breaks = c(0, 1), labels = c("Disallowed", "Allowed")) +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1)); SAD <- SA + ylab("Wordsum IQ") + xlab("Position on if Anti-Religious Speakers should be Allowed to make a Public Speech"); SAD## Warning: Removed 1 rows containing missing values (geom_bar).
spkcomD <- data %>%
group_by(spkcom) %>%
summarise(
n = n(),
mean = mean(wordsum, na.rm = T),
sd = sd(wordsum, na.rm = T)) %>%
mutate(se = sd/sqrt(n)) %>%
mutate(ic = se * qt((1 - 0.05)/2 + 0.5, n - 1))
SC <- ggplot(spkcomD) +
geom_bar(aes(x = spkcom, y = mean), stat = "identity", position = "dodge", fill = "#009999", width = 0.9, alpha = 0.5) +
geom_errorbar(aes(x = spkcom, ymin = mean - ic, ymax = mean + ic), width = 0.4, color = "#666699", alpha = 0.9, size = 1) +
theme_bw() + ylab("") + xlab("") +
scale_x_continuous(breaks = c(0, 1), labels = c("Disallowed", "Allowed")) +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1)); SCD <- SC + ylab("Wordsum IQ") + xlab("Position on if Communist Speakers should be Allowed to make a Public Speech"); SCD## Warning: Removed 1 rows containing missing values (geom_bar).
spkhomoD <- data %>%
group_by(spkhomo) %>%
summarise(
n = n(),
mean = mean(wordsum, na.rm = T),
sd = sd(wordsum, na.rm = T)) %>%
mutate(se = sd/sqrt(n)) %>%
mutate(ic = se * qt((1 - 0.05)/2 + 0.5, n - 1))
SH <- ggplot(spkhomoD) +
geom_bar(aes(x = spkhomo, y = mean), stat = "identity", position = "dodge", fill = "#009999", width = 0.9, alpha = 0.5) +
geom_errorbar(aes(x = spkhomo, ymin = mean - ic, ymax = mean + ic), width = 0.4, color = "#666699", alpha = 0.9, size = 1) +
theme_bw() + ylab("") + xlab("") +
scale_x_continuous(breaks = c(0, 1), labels = c("Disallowed", "Allowed")) +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1)); SHD <- SH + ylab("Wordsum IQ") + xlab("Position on if Pro-Homosexual Speakers should be Allowed to make a Public Speech"); SHD## Warning: Removed 1 rows containing missing values (geom_bar).
spkmilD <- data %>%
group_by(spkmil) %>%
summarise(
n = n(),
mean = mean(wordsum, na.rm = T),
sd = sd(wordsum, na.rm = T)) %>%
mutate(se = sd/sqrt(n)) %>%
mutate(ic = se * qt((1 - 0.05)/2 + 0.5, n - 1))
SM <- ggplot(spkmilD) +
geom_bar(aes(x = spkmil, y = mean), stat = "identity", position = "dodge", fill = "#009999", width = 0.9, alpha = 0.5) +
geom_errorbar(aes(x = spkmil, ymin = mean - ic, ymax = mean + ic), width = 0.4, color = "#666699", alpha = 0.9, size = 1) +
theme_bw() + ylab("") + xlab("") +
scale_x_continuous(breaks = c(0, 1), labels = c("Disallowed", "Allowed")) +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1)); SMD <- SM + ylab("Wordsum IQ") + xlab("Position on if Militarist Speakers should be Allowed to make a Public Speech"); SMD## Warning: Removed 1 rows containing missing values (geom_bar).
spkmslmD <- data %>%
group_by(spkmslm) %>%
summarise(
n = n(),
mean = mean(wordsum, na.rm = T),
sd = sd(wordsum, na.rm = T)) %>%
mutate(se = sd/sqrt(n)) %>%
mutate(ic = se * qt((1 - 0.05)/2 + 0.5, n - 1))
SU <- ggplot(spkmslmD) +
geom_bar(aes(x = spkmslm, y = mean), stat = "identity", position = "dodge", fill = "#009999", width = 0.9, alpha = 0.5) +
geom_errorbar(aes(x = spkmslm, ymin = mean - ic, ymax = mean + ic), width = 0.4, color = "#666699", alpha = 0.9, size = 1) +
theme_bw() + ylab("") + xlab("") +
scale_x_continuous(breaks = c(0, 1), labels = c("Disallowed", "Allowed")) +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1)); SUD <- SU + ylab("Wordsum IQ") + xlab("Position on if Anti-American Muslim Clergyman Speakers should be Allowed to make a Public Speech"); SUD## Warning: Removed 1 rows containing missing values (geom_bar).
spkracD <- data %>%
group_by(spkrac) %>%
summarise(
n = n(),
mean = mean(wordsum, na.rm = T),
sd = sd(wordsum, na.rm = T)) %>%
mutate(se = sd/sqrt(n)) %>%
mutate(ic = se * qt((1 - 0.05)/2 + 0.5, n - 1))
SR <- ggplot(spkracD) +
geom_bar(aes(x = spkrac, y = mean), stat = "identity", position = "dodge", fill = "#009999", width = 0.9, alpha = 0.5) +
geom_errorbar(aes(x = spkrac, ymin = mean - ic, ymax = mean + ic), width = 0.4, color = "#666699", alpha = 0.9, size = 1) +
theme_bw() + ylab("") + xlab("") +
scale_x_continuous(breaks = c(0, 1), labels = c("Disallowed", "Allowed")) +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1)); SRD <- SR + ylab("Wordsum IQ") + xlab("Position on if Racist Speakers should be Allowed to make a Public Speech"); SRD## Warning: Removed 1 rows containing missing values (geom_bar).
ggarrange(HD, HDI, RD, CA, CC, CH, CM, CU, CR, CRA , CRD, RF, RM, LA, LC, LH, LM, LU, LR, SA, SC, SH, SM, SU, SR)## Warning: Removed 1 rows containing missing values (geom_bar).
## Warning: Removed 1 rows containing missing values (geom_bar).
## Warning: Removed 1 rows containing missing values (geom_bar).
## Warning: Removed 1 rows containing missing values (geom_bar).
## Warning: Removed 1 rows containing missing values (geom_bar).
## Warning: Removed 1 rows containing missing values (geom_bar).
## Warning: Removed 1 rows containing missing values (geom_bar).
## Warning: Removed 1 rows containing missing values (geom_bar).
## Warning: Removed 1 rows containing missing values (geom_bar).
## Warning: Removed 1 rows containing missing values (geom_bar).
## Warning: Removed 1 rows containing missing values (geom_bar).
## Warning: Removed 1 rows containing missing values (geom_bar).
## Warning: Removed 1 rows containing missing values (geom_bar).
## Warning: Removed 1 rows containing missing values (geom_bar).
## Warning: Removed 1 rows containing missing values (geom_bar).
## Warning: Removed 1 rows containing missing values (geom_bar).
## Warning: Removed 1 rows containing missing values (geom_bar).
## Warning: Removed 1 rows containing missing values (geom_bar).
## Warning: Removed 1 rows containing missing values (geom_bar).
## Warning: Removed 1 rows containing missing values (geom_bar).
## Warning: Removed 1 rows containing missing values (geom_bar).
## Warning: Removed 1 rows containing missing values (geom_bar).
## Warning: Removed 1 rows containing missing values (geom_bar).
## Warning: Removed 1 rows containing missing values (geom_bar).
## Warning: Removed 1 rows containing missing values (geom_bar).
Wordsum Differences by Tolerance
psych::cohen.d(data$wordsum, data$homosexDI)$cohen.d[,2]## effect
## 0.4598247psych::cohen.d(data$wordsum, data$racmar)$cohen.d[,2]## effect
## 0.5311738psych::cohen.d(data$wordsum, data$racdinDI)$cohen.d[,2]## effect
## 0.4669406psych::cohen.d(data$wordsum, data$racfew)$cohen.d[,2]## effect
## 0.6338781psych::cohen.d(data$wordsum, data$racmost)$cohen.d[,2]## effect
## -0.1163406psych::cohen.d(data$wordsum, data$colath)$cohen.d[,2]## effect
## 0.4878822psych::cohen.d(data$wordsum, data$colcom)$cohen.d[,2]## effect
## 0.4771611psych::cohen.d(data$wordsum, data$colhomo)$cohen.d[,2]## effect
## 0.5867238psych::cohen.d(data$wordsum, data$colmil)$cohen.d[,2]## effect
## 0.4025638psych::cohen.d(data$wordsum, data$colmslm)$cohen.d[,2]## effect
## 0.5588219psych::cohen.d(data$wordsum, data$colrac)$cohen.d[,2]## effect
## 0.304279psych::cohen.d(data$wordsum, data$libath)$cohen.d[,2]## effect
## 0.6303764psych::cohen.d(data$wordsum, data$libcom)$cohen.d[,2]## effect
## 0.6843602psych::cohen.d(data$wordsum, data$libhomo)$cohen.d[,2]## effect
## 0.6096453psych::cohen.d(data$wordsum, data$libmil)$cohen.d[,2]## effect
## 0.5978203psych::cohen.d(data$wordsum, data$libmslm)$cohen.d[,2]## effect
## 0.6386289psych::cohen.d(data$wordsum, data$librac)$cohen.d[,2]## effect
## 0.4801116psych::cohen.d(data$wordsum, data$spkath)$cohen.d[,2]## effect
## 0.6597864psych::cohen.d(data$wordsum, data$spkcom)$cohen.d[,2]## effect
## 0.6877138psych::cohen.d(data$wordsum, data$spkhomo)$cohen.d[,2]## effect
## 0.6973724psych::cohen.d(data$wordsum, data$spkmil)$cohen.d[,2]## effect
## 0.5378884psych::cohen.d(data$wordsum, data$spkmslm)$cohen.d[,2]## effect
## 0.6142829psych::cohen.d(data$wordsum, data$spkrac)$cohen.d[,2]## effect
## 0.4349318Verbal IQ and a Tolerance Sumscore
ggplot(data, aes(x = Tolerance, y = wordsum)) + geom_point(color = "#cc6699", alpha = 0.9, position = "jitter") +
geom_smooth(method = "gam", color = "#3399ff") +
geom_smooth(method = "lm", color = "orangered", formula = "y ~ x") +
geom_smooth(method = "loess", color = "gold") +
theme_bw() +
ylab("Wordsum IQ") + xlab("Tolerance") +
theme(axis.title.y = element_text(vjust = 3), axis.title.x = element_text(vjust = -1))## `geom_smooth()` using formula 'y ~ s(x, bs = "cs")'## Warning: Removed 63707 rows containing non-finite values (stat_smooth).
## Warning: Removed 63707 rows containing non-finite values (stat_smooth).## `geom_smooth()` using formula 'y ~ x'## Warning: Removed 63707 rows containing non-finite values (stat_smooth).## Warning: Removed 63707 rows containing missing values (geom_point).
Specification Curve Analysis
Linear models - not admissible. Political views had to be trichotomized because analyses were not possible with all seven categories, regardless of whether they were used for subsetting (as is best) or a continuous control. This is unfortunate, but a reality required to conduct estimation, and it should provide asymptotically very similar results regardless. Even more unfortunately, the trichotomized political views variable, polcat, had to be used as a control rather than for subsetting, lest estimation with it become impossible.
ControlSet <- run_specs(df = data,
y = c("homosex", "racmar", "colath", "racdin", "racfew", "racmost", "colcom", "colhomo", "colmil", "colmslm", "colrac", "libcom", "libhomo", "libmil", "libmslm", "librac", "spkath", "spkcom", "spkhomo", "spkmil", "spkmslm", "spkrac"),
x = c("wordsum"),
model = c("glm"),
controls = c("age", "year", "reg16", "res16", "mobile16", "degree", "polcat"),
subsets = list(sex = unique(data$sex),
race = unique(data$race)))
head(ControlSet)plot_specs(ControlSet, choices = c("x", "y", "controls", "subsets"))
pc <- plot_curve(ControlSet,
ci = T,
ribbon = F) +
geom_hline(yintercept = 0,
linetype = "dashed",
color = "black") +
labs(x = "", y = "Unstandardized Beta Coefficient") +
theme_half_open()
pcr <- plot_curve(ControlSet,
ci = F,
ribbon = T) +
geom_hline(yintercept = 0,
linetype = "dashed",
color = "black") +
labs(x = "", y = "Unstandardized Beta Coefficient") +
theme_half_open()
pcc <- plot_choices(ControlSet,
choices = c("x", "y", "controls", "subsets")) +
labs(x = "Ranked Specifications") +
theme_half_open() +
theme(strip.text.x = element_blank())
ph <- plot_samplesizes(ControlSet) +
labs(y = "Sample Size") +
theme_half_open()
plot_grid(pc, pcc, ph, ncol = 1, align = "v", rel_heights = c(1.5, 2, 0.8), axis = "rbl")
plot_grid(pcr, ph, ncol = 1, align = "v", rel_heights = c(1.5, 0.6), axis = "rbl")
Ordinal logistic models.
I did not run these yet, since specr does not want to take them via polr(). It is hard to imagine the results will differ much and I can do them by hand if really necessary.
Binomial logistic models.
BinomialGLM = function(formula, data){
glm(formula = formula, data = data, family = "binomial")}
BControlSet <- run_specs(df = data,
y = c("homosexDI", "racmar", "colath", "racdinDI", "racfew", "racmost", "colcom", "colhomo", "colmil", "colmslm", "colrac", "libcom", "libhomo", "libmil", "libmslm", "librac", "spkath", "spkcom", "spkhomo", "spkmil", "spkmslm", "spkrac"),
x = c("wordsum"),
model = c("BinomialGLM"),
controls = c("age", "year", "reg16", "res16", "mobile16", "degree", "polcat"),
subsets = list(sex = unique(data$sex),
race = unique(data$race)))## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurred
## Warning: glm.fit: fitted probabilities numerically 0 or 1 occurredhead(BControlSet)plot_specs(BControlSet, choices = c("x", "y", "controls", "subsets"))
pc <- plot_curve(BControlSet,
ci = T,
ribbon = F) +
geom_hline(yintercept = 0,
linetype = "dashed",
color = "black") +
labs(x = "", y = "Unstandardized Beta Coefficient") +
theme_half_open()
pcr <- plot_curve(BControlSet,
ci = F,
ribbon = T) +
geom_hline(yintercept = 0,
linetype = "dashed",
color = "black") +
labs(x = "", y = "Unstandardized Beta Coefficient") +
theme_half_open()
pcc <- plot_choices(BControlSet,
choices = c("x", "y", "controls", "subsets")) +
labs(x = "Ranked Specifications") +
theme_half_open() +
theme(strip.text.x = element_blank())
ph <- plot_samplesizes(BControlSet) +
labs(y = "Sample Size") +
theme_half_open()
plot_grid(pc, pcc, ph, ncol = 1, align = "v", rel_heights = c(1.5, 2, 0.8), axis = "rbl")
plot_grid(pcr, ph, ncol = 1, align = "v", rel_heights = c(1.5, 0.6), axis = "rbl")
Assessing Moderation of the Wordsum-Tolerance Relationship by Sex and Race
list(
summary(lm(racmost ~ wordsum, data)),
summary(lm(racmost ~ wordsum + as.factor(sex), data)),
summary(lm(racmost ~ wordsum * as.factor(sex), data)),
summary(lm(racmost ~ wordsum + as.factor(race), data)),
summary(lm(racmost ~ wordsum * as.factor(race), data))) %>%
print() ## [[1]]
##
## Call:
## lm(formula = racmost ~ wordsum, data = data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.6928 -0.6031 0.3585 0.3841 0.4354
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.7963515 0.0340579 23.382 < 2e-16 ***
## wordsum -0.0018042 0.0003351 -5.384 7.46e-08 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.4859 on 9046 degrees of freedom
## (55684 observations deleted due to missingness)
## Multiple R-squared: 0.003194, Adjusted R-squared: 0.003084
## F-statistic: 28.99 on 1 and 9046 DF, p-value: 7.462e-08
##
##
## [[2]]
##
## Call:
## lm(formula = racmost ~ wordsum + as.factor(sex), data = data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.7023 -0.6005 0.3613 0.3917 0.4426
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.8052027 0.0344480 23.374 < 2e-16 ***
## wordsum -0.0017924 0.0003351 -5.348 9.09e-08 ***
## as.factor(sex)2 -0.0175887 0.0103202 -1.704 0.0884 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.4858 on 9045 degrees of freedom
## (55684 observations deleted due to missingness)
## Multiple R-squared: 0.003514, Adjusted R-squared: 0.003294
## F-statistic: 15.95 on 2 and 9045 DF, p-value: 1.217e-07
##
##
## [[3]]
##
## Call:
## lm(formula = racmost ~ wordsum * as.factor(sex), data = data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.7268 -0.6091 0.3624 0.3821 0.4694
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.6810371 0.0512318 13.293 < 2e-16 ***
## wordsum -0.0005523 0.0005057 -1.092 0.27484
## as.factor(sex)2 0.2042197 0.0685528 2.979 0.00290 **
## wordsum:as.factor(sex)2 -0.0022092 0.0006750 -3.273 0.00107 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.4855 on 9044 degrees of freedom
## (55684 observations deleted due to missingness)
## Multiple R-squared: 0.004693, Adjusted R-squared: 0.004363
## F-statistic: 14.22 on 3 and 9044 DF, p-value: 3.07e-09
##
##
## [[4]]
##
## Call:
## lm(formula = racmost ~ wordsum + as.factor(race), data = data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.8804 -0.5635 0.1249 0.4365 0.4405
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 5.541e-01 3.498e-02 15.842 < 2e-16 ***
## wordsum 9.351e-05 3.378e-04 0.277 0.782
## as.factor(race)2 3.143e-01 1.458e-02 21.554 < 2e-16 ***
## as.factor(race)3 2.006e-01 2.992e-02 6.704 2.15e-11 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.4733 on 9044 degrees of freedom
## (55684 observations deleted due to missingness)
## Multiple R-squared: 0.05409, Adjusted R-squared: 0.05377
## F-statistic: 172.4 on 3 and 9044 DF, p-value: < 2.2e-16
##
##
## [[5]]
##
## Call:
## lm(formula = racmost ~ wordsum * as.factor(race), data = data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.9261 -0.5630 0.1449 0.4370 0.4497
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.5331598 0.0382645 13.934 < 2e-16 ***
## wordsum 0.0002982 0.0003703 0.805 0.4206
## as.factor(race)2 0.4748966 0.0949093 5.004 5.73e-07 ***
## as.factor(race)3 0.1523598 0.1689113 0.902 0.3671
## wordsum:as.factor(race)2 -0.0017260 0.0010060 -1.716 0.0863 .
## wordsum:as.factor(race)3 0.0005388 0.0017799 0.303 0.7621
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.4733 on 9042 degrees of freedom
## (55684 observations deleted due to missingness)
## Multiple R-squared: 0.05442, Adjusted R-squared: 0.05389
## F-statistic: 104.1 on 5 and 9042 DF, p-value: < 2.2e-16