Key points: (i) Overall effect, (ii) Cross country generalization, not limited to Fox

1. import data sets

DESCRIPTIVE STATS

1. demographics

a. sample size

table(!is.na(d1$participant))

## 
## TRUE 
## 4836

table(!is.na(d1$participant[d1$country_factor == "US"]))

## 
## TRUE 
## 3316

table(!is.na(d1$participant[d1$country_factor == "UK"]))

## 
## TRUE 
## 1520

b. age

describe(d1$age)

describe(d1$age[d1$country_factor == "US"])

describe(d1$age[d1$country_factor == "UK"])

c. gender

(t <- table(d1$gender_factor))

## 
## custom female   male 
##     16   2480   2053

round(prop.table(t), 3)

## 
## custom female   male 
##  0.004  0.545  0.451

(t <- table(d1$gender_factor[d1$country_factor == "US"]))

## 
## custom female   male 
##      6   1564   1477

round(prop.table(t), 3)

## 
## custom female   male 
##  0.002  0.513  0.485

(t <- table(d1$gender_factor[d1$country_factor == "UK"]))

## 
## custom female   male 
##     10    916    576

round(prop.table(t), 3)

## 
## custom female   male 
##  0.007  0.610  0.383

d. party identity

(t <- table(d1$party_factor))

## 
##    Democrat Independent  Republican 
##        2268         916        1470

round(prop.table(t), 3)

## 
##    Democrat Independent  Republican 
##       0.487       0.197       0.316

(t <- table(d1$party_factor[d1$country_factor == "US"]))

## 
##    Democrat Independent  Republican 
##        1402         627        1118

round(prop.table(t), 3)

## 
##    Democrat Independent  Republican 
##       0.446       0.199       0.355

(t <- table(d1$party_factor[d1$country_factor == "UK"]))

## 
##    Democrat Independent  Republican 
##         866         289         352

round(prop.table(t), 3)

## 
##    Democrat Independent  Republican 
##       0.575       0.192       0.234

e. education

describe(d1$education)

describe(d1$education[d1$country_factor == "US"])

describe(d1$education[d1$country_factor == "UK"])

2. variables of interest

a. media exposure

Not included here, media exposure asks participants to “Consider each of the media sources below. In general, how much do you get news about Covid-19 from each source?” on a scale from 1 (Not at all) to 3 (Somewhat) to 5 (A great deal).

b. analytical media index

The first step to creating the analytical media index is multiplying individual analytic thinking scores for each media outlet by participant rated exposure to that outlet, Then taking the proportion of each of these products (i.e., dividing by the 12 possible US outlets to be exposed to, or 8 outlets for the UK)

foxAnalyticalIndex = (foxNewsExposure x foxAnalyticalScore) (foxAnalyticalIndex + cnnAnalyticalIndex + msnbcAnalyticalIndex + …) / 12 total outlets

describe(d1$index_ANexp)

describe(d1$index_ANexp[d1$country_factor == "US"])

describe(d1$index_ANexp[d1$country_factor == "UK"])

c. symbolic ideology

symbolic ideology is an average of three items. They ask “How liberal/conservative…” (1) in general, (2) on social issues, and (3) on economic issues. Participants answered on a scale from -3 (Very liberal) to 0 (Moderate) to +3 (Very conservative).

print("US")

## [1] "US"

psych::alpha (data.frame(
  d1$symbolic_beliefs_1[d1$country_factor == "US"], 
  d1$symbolic_beliefs_2[d1$country_factor == "US"], 
  d1$symbolic_beliefs_3[d1$country_factor == "US"]), cumulative = F, na.rm = T, delete = T) #alpha = .94

## 
## Reliability analysis   
## Call: psych::alpha(x = data.frame(d1$symbolic_beliefs_1[d1$country_factor == 
##     "US"], d1$symbolic_beliefs_2[d1$country_factor == "US"], 
##     d1$symbolic_beliefs_3[d1$country_factor == "US"]), cumulative = F, 
##     na.rm = T, delete = T)
## 
##   raw_alpha std.alpha G6(smc) average_r S/N    ase  mean  sd median_r
##       0.94      0.94    0.92      0.85  17 0.0017 0.089 1.6     0.86
## 
##  lower alpha upper     95% confidence boundaries
## 0.94 0.94 0.95 
## 
##  Reliability if an item is dropped:
##                                                  raw_alpha std.alpha G6(smc)
## d1.symbolic_beliefs_1.d1.country_factor.....US..      0.89      0.89    0.80
## d1.symbolic_beliefs_2.d1.country_factor.....US..      0.92      0.92    0.86
## d1.symbolic_beliefs_3.d1.country_factor.....US..      0.94      0.94    0.89
##                                                  average_r S/N alpha se var.r
## d1.symbolic_beliefs_1.d1.country_factor.....US..      0.80   8   0.0039    NA
## d1.symbolic_beliefs_2.d1.country_factor.....US..      0.86  12   0.0027    NA
## d1.symbolic_beliefs_3.d1.country_factor.....US..      0.89  16   0.0020    NA
##                                                  med.r
## d1.symbolic_beliefs_1.d1.country_factor.....US..  0.80
## d1.symbolic_beliefs_2.d1.country_factor.....US..  0.86
## d1.symbolic_beliefs_3.d1.country_factor.....US..  0.89
## 
##  Item statistics 
##                                                     n raw.r std.r r.cor r.drop
## d1.symbolic_beliefs_1.d1.country_factor.....US.. 3162  0.97  0.97  0.95   0.92
## d1.symbolic_beliefs_2.d1.country_factor.....US.. 3163  0.95  0.95  0.91   0.88
## d1.symbolic_beliefs_3.d1.country_factor.....US.. 3163  0.93  0.93  0.88   0.85
##                                                    mean  sd
## d1.symbolic_beliefs_1.d1.country_factor.....US..  0.060 1.7
## d1.symbolic_beliefs_2.d1.country_factor.....US.. -0.067 1.7
## d1.symbolic_beliefs_3.d1.country_factor.....US..  0.272 1.7
## 
## Non missing response frequency for each item
##                                                    -3   -2   -1    0    1    2
## d1.symbolic_beliefs_1.d1.country_factor.....US.. 0.08 0.13 0.09 0.38 0.09 0.13
## d1.symbolic_beliefs_2.d1.country_factor.....US.. 0.10 0.14 0.12 0.33 0.10 0.13
## d1.symbolic_beliefs_3.d1.country_factor.....US.. 0.06 0.11 0.09 0.35 0.13 0.16
##                                                     3 miss
## d1.symbolic_beliefs_1.d1.country_factor.....US.. 0.09 0.05
## d1.symbolic_beliefs_2.d1.country_factor.....US.. 0.09 0.05
## d1.symbolic_beliefs_3.d1.country_factor.....US.. 0.11 0.05

print("UK")

## [1] "UK"

psych::alpha (data.frame(
  d1$symbolic_beliefs_1[d1$country_factor == "UK"], 
  d1$symbolic_beliefs_2[d1$country_factor == "UK"], 
  d1$symbolic_beliefs_3[d1$country_factor == "UK"]), cumulative = F, na.rm = T, delete = T)

## 
## Reliability analysis   
## Call: psych::alpha(x = data.frame(d1$symbolic_beliefs_1[d1$country_factor == 
##     "UK"], d1$symbolic_beliefs_2[d1$country_factor == "UK"], 
##     d1$symbolic_beliefs_3[d1$country_factor == "UK"]), cumulative = F, 
##     na.rm = T, delete = T)
## 
##   raw_alpha std.alpha G6(smc) average_r S/N    ase  mean  sd median_r
##       0.92      0.92     0.9      0.79  12 0.0037 -0.61 1.4     0.81
## 
##  lower alpha upper     95% confidence boundaries
## 0.91 0.92 0.93 
## 
##  Reliability if an item is dropped:
##                                                  raw_alpha std.alpha G6(smc)
## d1.symbolic_beliefs_1.d1.country_factor.....UK..      0.83      0.83    0.72
## d1.symbolic_beliefs_2.d1.country_factor.....UK..      0.90      0.90    0.81
## d1.symbolic_beliefs_3.d1.country_factor.....UK..      0.92      0.92    0.85
##                                                  average_r  S/N alpha se var.r
## d1.symbolic_beliefs_1.d1.country_factor.....UK..      0.72  5.0   0.0085    NA
## d1.symbolic_beliefs_2.d1.country_factor.....UK..      0.81  8.7   0.0053    NA
## d1.symbolic_beliefs_3.d1.country_factor.....UK..      0.85 11.6   0.0041    NA
##                                                  med.r
## d1.symbolic_beliefs_1.d1.country_factor.....UK..  0.72
## d1.symbolic_beliefs_2.d1.country_factor.....UK..  0.81
## d1.symbolic_beliefs_3.d1.country_factor.....UK..  0.85
## 
##  Item statistics 
##                                                     n raw.r std.r r.cor r.drop
## d1.symbolic_beliefs_1.d1.country_factor.....UK.. 1509  0.96  0.96  0.94   0.90
## d1.symbolic_beliefs_2.d1.country_factor.....UK.. 1509  0.92  0.92  0.87   0.82
## d1.symbolic_beliefs_3.d1.country_factor.....UK.. 1509  0.91  0.91  0.83   0.79
##                                                   mean  sd
## d1.symbolic_beliefs_1.d1.country_factor.....UK.. -0.64 1.4
## d1.symbolic_beliefs_2.d1.country_factor.....UK.. -0.84 1.5
## d1.symbolic_beliefs_3.d1.country_factor.....UK.. -0.35 1.5
## 
## Non missing response frequency for each item
##                                                    -3   -2   -1    0    1    2
## d1.symbolic_beliefs_1.d1.country_factor.....UK.. 0.09 0.24 0.19 0.27 0.12 0.07
## d1.symbolic_beliefs_2.d1.country_factor.....UK.. 0.13 0.25 0.20 0.24 0.10 0.06
## d1.symbolic_beliefs_3.d1.country_factor.....UK.. 0.07 0.18 0.19 0.29 0.16 0.09
##                                                     3 miss
## d1.symbolic_beliefs_1.d1.country_factor.....UK.. 0.01 0.01
## d1.symbolic_beliefs_2.d1.country_factor.....UK.. 0.01 0.01
## d1.symbolic_beliefs_3.d1.country_factor.....UK.. 0.02 0.01

describe(d1$ideology)

describe(d1$ideology[d1$country_factor == "US"])

describe(d1$ideology[d1$country_factor == "UK"])

d. average trust in experts

trust in experts is an average of 3-items asking about how much participants trust experts, medicine, economics, public health, and science. The trust response scale ranges from -3 (strongly distrust) to 0 (Neither trust nor distrust) to (Strongly trust)

describe(d1$trustExpert)

describe(d1$trustExpert[d1$country_factor == "US"], na.rm = T)

describe(d1$trustExpert[d1$country_factor == "UK"], na.rm = T)

3. LIWC media measures

a. Analytical thinking scales

Media outlet LIWC analytic thinking scores captures the degree to which people use words that suggest formal, logical, and hierarchical thinking patterns. Column two is raw scores, and column three is standardized scores.

i. US

ii. UK

4. outcome measures

Participants were asked “would you get a Covid-19 vaccine?” and answered on a scale from -3 (Definitely would not get it) to 0 (Undecided) to +3 (Definitely would get it).

describe(d1$vaxxAttitudes)

describe(d1$vaxxAttitudes[d1$country_factor == "US"])

describe(d1$vaxxAttitudes[d1$country_factor == "UK"])

5. pairwise correlation table for IVs, DVs, and covariates

i. US

i. UK

6. Analyses for Study 1 US + UK (July-August November)

USvUK (US = -.5, UK = +.5)

a. vaxxAttitudes ~ analticalIndex * ideology + USvUK + age + education

summary(m.s1.a.robust <- lm(vaxxAttitudes ~ index_ANexp.c * ideology.c + USvUK + 
                              age.c + education.c, data = d1))

## 
## Call:
## lm(formula = vaxxAttitudes ~ index_ANexp.c * ideology.c + USvUK + 
##     age.c + education.c, data = d1)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -5.7081 -1.3369  0.3965  1.5363  4.3853 
## 
## Coefficients:
##                            Estimate Std. Error t value Pr(>|t|)    
## (Intercept)               1.0251836  0.0315810  32.462  < 2e-16 ***
## index_ANexp.c             0.0051306  0.0004295  11.944  < 2e-16 ***
## ideology.c               -0.2307232  0.0196799 -11.724  < 2e-16 ***
## USvUK                     0.9023665  0.0690258  13.073  < 2e-16 ***
## age.c                     0.0215370  0.0020142  10.693  < 2e-16 ***
## education.c               0.0525331  0.0110634   4.748 2.12e-06 ***
## index_ANexp.c:ideology.c  0.0012543  0.0002619   4.788 1.74e-06 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.952 on 4462 degrees of freedom
##   (367 observations deleted due to missingness)
## Multiple R-squared:  0.1207, Adjusted R-squared:  0.1196 
## F-statistic: 102.1 on 6 and 4462 DF,  p-value: < 2.2e-16

tab_model(m.s1.a.robust,
          show.df = T, 
          show.ci = .95,
          show.se = T,
          show.stat = T,
          string.stat = "t",
          string.se="SE",
          string.est = "Est",
          digits = 3)

	vaxxAttitudes
Predictors	Est	SE	CI	t	p	df
(Intercept)	1.025	0.032	0.963 – 1.087	32.462	<0.001	4462.000
index_ANexp.c	0.005	0.000	0.004 – 0.006	11.944	<0.001	4462.000
ideology.c	-0.231	0.020	-0.269 – -0.192	-11.724	<0.001	4462.000
USvUK	0.902	0.069	0.767 – 1.038	13.073	<0.001	4462.000
age.c	0.022	0.002	0.018 – 0.025	10.693	<0.001	4462.000
education.c	0.053	0.011	0.031 – 0.074	4.748	<0.001	4462.000
index_ANexp.c * ideology.c	0.001	0.000	0.001 – 0.002	4.788	<0.001	4462.000
Observations	4469
R² / R² adjusted	0.121 / 0.120

b. vaxxAttitudes ~ analticalIndex * ideology * USvUK + age + education

summary(m.s1.b.robust <- lm(vaxxAttitudes ~ index_ANexp.c * ideology.c * USvUK + 
                              age.c + education.c, data = d1))

## 
## Call:
## lm(formula = vaxxAttitudes ~ index_ANexp.c * ideology.c * USvUK + 
##     age.c + education.c, data = d1)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -5.3370 -1.3492  0.4032  1.5282  4.4067 
## 
## Coefficients:
##                                  Estimate Std. Error t value Pr(>|t|)    
## (Intercept)                     1.0197428  0.0330645  30.841  < 2e-16 ***
## index_ANexp.c                   0.0047055  0.0004899   9.605  < 2e-16 ***
## ideology.c                     -0.2257272  0.0222675 -10.137  < 2e-16 ***
## USvUK                           0.8962025  0.0700062  12.802  < 2e-16 ***
## age.c                           0.0213948  0.0020185  10.599  < 2e-16 ***
## education.c                     0.0531431  0.0110745   4.799 1.65e-06 ***
## index_ANexp.c:ideology.c        0.0011419  0.0003204   3.564 0.000369 ***
## index_ANexp.c:USvUK            -0.0018075  0.0009783  -1.848 0.064729 .  
## ideology.c:USvUK                0.0064325  0.0439408   0.146 0.883620    
## index_ANexp.c:ideology.c:USvUK -0.0001690  0.0006397  -0.264 0.791593    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.952 on 4459 degrees of freedom
##   (367 observations deleted due to missingness)
## Multiple R-squared:  0.1214, Adjusted R-squared:  0.1197 
## F-statistic: 68.48 on 9 and 4459 DF,  p-value: < 2.2e-16

tab_model(m.s1.a.robust, m.s1.b.robust,
          show.df = T, 
          show.ci = .95,
          show.se = T,
          show.stat = T,
          string.stat = "t",
          string.se="SE",
          string.est = "Est",
          digits = 3)

	vaxxAttitudes						vaxxAttitudes
Predictors	Est	SE	CI	t	p	df	Est	SE	CI	t	p	df
(Intercept)	1.025	0.032	0.963 – 1.087	32.462	<0.001	4462.000	1.020	0.033	0.955 – 1.085	30.841	<0.001	4459.000
index_ANexp.c	0.005	0.000	0.004 – 0.006	11.944	<0.001	4462.000	0.005	0.000	0.004 – 0.006	9.605	<0.001	4459.000
ideology.c	-0.231	0.020	-0.269 – -0.192	-11.724	<0.001	4462.000	-0.226	0.022	-0.269 – -0.182	-10.137	<0.001	4459.000
USvUK	0.902	0.069	0.767 – 1.038	13.073	<0.001	4462.000	0.896	0.070	0.759 – 1.033	12.802	<0.001	4459.000
age.c	0.022	0.002	0.018 – 0.025	10.693	<0.001	4462.000	0.021	0.002	0.017 – 0.025	10.599	<0.001	4459.000
education.c	0.053	0.011	0.031 – 0.074	4.748	<0.001	4462.000	0.053	0.011	0.031 – 0.075	4.799	<0.001	4459.000
index_ANexp.c * ideology.c	0.001	0.000	0.001 – 0.002	4.788	<0.001	4462.000	0.001	0.000	0.001 – 0.002	3.564	<0.001	4459.000
index_ANexp.c * USvUK							-0.002	0.001	-0.004 – 0.000	-1.848	0.065	4459.000
ideology.c * USvUK							0.006	0.044	-0.080 – 0.093	0.146	0.884	4459.000
(index_ANexp.c ideology.c) USvUK							-0.000	0.001	-0.001 – 0.001	-0.264	0.792	4459.000
Observations	4469						4469
R² / R² adjusted	0.121 / 0.120						0.121 / 0.120

c. mediation path: analyticalIndex -> trustInExperts -> vaccineAttitudes

- no interaction

summary(m.xy <- lm(vaxxAttitudes ~ index_ANexp.z + 
                         ideology.z + USvUK + age.z + education.z, data = d1))

## 
## Call:
## lm(formula = vaxxAttitudes ~ index_ANexp.z + ideology.z + USvUK + 
##     age.z + education.z, data = d1)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -5.6783 -1.3178  0.3948  1.5544  4.2232 
## 
## Coefficients:
##               Estimate Std. Error t value Pr(>|t|)    
## (Intercept)    1.01028    0.03150  32.068  < 2e-16 ***
## index_ANexp.z  0.35835    0.02986  11.999  < 2e-16 ***
## ideology.z    -0.35660    0.03066 -11.632  < 2e-16 ***
## USvUK          0.93011    0.06895  13.490  < 2e-16 ***
## age.z          0.33344    0.03230  10.325  < 2e-16 ***
## education.z    0.14015    0.03005   4.663  3.2e-06 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.957 on 4463 degrees of freedom
##   (367 observations deleted due to missingness)
## Multiple R-squared:  0.1162, Adjusted R-squared:  0.1152 
## F-statistic: 117.4 on 5 and 4463 DF,  p-value: < 2.2e-16

summary(m.xm <- lm(trustExpert ~ index_ANexp.z + 
                         ideology.z + USvUK + age.z + education.z, data = d1))

## 
## Call:
## lm(formula = trustExpert ~ index_ANexp.z + ideology.z + USvUK + 
##     age.z + education.z, data = d1)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -5.6925 -0.6256  0.2825  0.9497  3.0967 
## 
## Coefficients:
##               Estimate Std. Error t value Pr(>|t|)    
## (Intercept)    1.69460    0.02198  77.107  < 2e-16 ***
## index_ANexp.z  0.19312    0.02083   9.270  < 2e-16 ***
## ideology.z    -0.31151    0.02139 -14.566  < 2e-16 ***
## USvUK          0.58358    0.04810  12.133  < 2e-16 ***
## age.z          0.14460    0.02253   6.418 1.52e-10 ***
## education.z    0.13956    0.02097   6.657 3.14e-11 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.365 on 4463 degrees of freedom
##   (367 observations deleted due to missingness)
## Multiple R-squared:  0.1213, Adjusted R-squared:  0.1203 
## F-statistic: 123.2 on 5 and 4463 DF,  p-value: < 2.2e-16

summary(m.xmy <- lm(vaxxAttitudes ~ trustExpert.z + index_ANexp.z + ideology.z + 
                      USvUK + age.z + education.z, data = d1))

## 
## Call:
## lm(formula = vaxxAttitudes ~ trustExpert.z + index_ANexp.z + 
##     ideology.z + USvUK + age.z + education.z, data = d1)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -5.7735 -1.1918  0.3258  1.3076  5.4251 
## 
## Coefficients:
##               Estimate Std. Error t value Pr(>|t|)    
## (Intercept)    0.93912    0.02938  31.966  < 2e-16 ***
## trustExpert.z  0.78471    0.02939  26.702  < 2e-16 ***
## index_ANexp.z  0.25559    0.02800   9.128  < 2e-16 ***
## ideology.z    -0.19085    0.02914  -6.550 6.42e-11 ***
## USvUK          0.61960    0.06508   9.521  < 2e-16 ***
## age.z          0.25650    0.03013   8.513  < 2e-16 ***
## education.z    0.06589    0.02805   2.349   0.0189 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.817 on 4462 degrees of freedom
##   (367 observations deleted due to missingness)
## Multiple R-squared:  0.238,  Adjusted R-squared:  0.237 
## F-statistic: 232.3 on 6 and 4462 DF,  p-value: < 2.2e-16

tab_model(m.xy, m.xm, m.xmy,
          show.df = T, 
          show.ci = .95,
          show.se = T,
          show.stat = T,
          string.stat = "t",
          string.se="SE",
          string.est = "Est",
          digits = 3)

	vaxxAttitudes						trustExpert						vaxxAttitudes
Predictors	Est	SE	CI	t	p	df	Est	SE	CI	t	p	df	Est	SE	CI	t	p	df
(Intercept)	1.010	0.032	0.949 – 1.072	32.068	<0.001	4463.000	1.695	0.022	1.652 – 1.738	77.107	<0.001	4463.000	0.939	0.029	0.882 – 0.997	31.966	<0.001	4462.000
index_ANexp.z	0.358	0.030	0.300 – 0.417	11.999	<0.001	4463.000	0.193	0.021	0.152 – 0.234	9.270	<0.001	4463.000	0.256	0.028	0.201 – 0.310	9.128	<0.001	4462.000
ideology.z	-0.357	0.031	-0.417 – -0.296	-11.632	<0.001	4463.000	-0.312	0.021	-0.353 – -0.270	-14.566	<0.001	4463.000	-0.191	0.029	-0.248 – -0.134	-6.550	<0.001	4462.000
USvUK	0.930	0.069	0.795 – 1.065	13.490	<0.001	4463.000	0.584	0.048	0.489 – 0.678	12.133	<0.001	4463.000	0.620	0.065	0.492 – 0.747	9.521	<0.001	4462.000
age.z	0.333	0.032	0.270 – 0.397	10.325	<0.001	4463.000	0.145	0.023	0.100 – 0.189	6.418	<0.001	4463.000	0.257	0.030	0.197 – 0.316	8.513	<0.001	4462.000
education.z	0.140	0.030	0.081 – 0.199	4.663	<0.001	4463.000	0.140	0.021	0.098 – 0.181	6.657	<0.001	4463.000	0.066	0.028	0.011 – 0.121	2.349	0.019	4462.000
trustExpert.z													0.785	0.029	0.727 – 0.842	26.702	<0.001	4462.000
Observations	4469						4469						4469
R² / R² adjusted	0.116 / 0.115						0.121 / 0.120						0.238 / 0.237

- interaction between moderator and ideology

summary(m.xy <- lm(vaxxAttitudes ~ index_ANexp.c * ideology.c + 
                     USvUK + age.c + education.c, data = d1))

## 
## Call:
## lm(formula = vaxxAttitudes ~ index_ANexp.c * ideology.c + USvUK + 
##     age.c + education.c, data = d1)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -5.7081 -1.3369  0.3965  1.5363  4.3853 
## 
## Coefficients:
##                            Estimate Std. Error t value Pr(>|t|)    
## (Intercept)               1.0251836  0.0315810  32.462  < 2e-16 ***
## index_ANexp.c             0.0051306  0.0004295  11.944  < 2e-16 ***
## ideology.c               -0.2307232  0.0196799 -11.724  < 2e-16 ***
## USvUK                     0.9023665  0.0690258  13.073  < 2e-16 ***
## age.c                     0.0215370  0.0020142  10.693  < 2e-16 ***
## education.c               0.0525331  0.0110634   4.748 2.12e-06 ***
## index_ANexp.c:ideology.c  0.0012543  0.0002619   4.788 1.74e-06 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.952 on 4462 degrees of freedom
##   (367 observations deleted due to missingness)
## Multiple R-squared:  0.1207, Adjusted R-squared:  0.1196 
## F-statistic: 102.1 on 6 and 4462 DF,  p-value: < 2.2e-16

summary(m.xm <- lm(trustExpert ~ index_ANexp.c * ideology.c +
                     USvUK + age.c + education.c, data = d1))

## 
## Call:
## lm(formula = trustExpert ~ index_ANexp.c * ideology.c + USvUK + 
##     age.c + education.c, data = d1)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -5.3331 -0.6186  0.2866  0.9480  3.2123 
## 
## Coefficients:
##                            Estimate Std. Error t value Pr(>|t|)    
## (Intercept)               1.7052211  0.0220280  77.411  < 2e-16 ***
## index_ANexp.c             0.0027588  0.0002996   9.208  < 2e-16 ***
## ideology.c               -0.2013451  0.0137269 -14.668  < 2e-16 ***
## USvUK                     0.5637911  0.0481461  11.710  < 2e-16 ***
## age.c                     0.0095502  0.0014049   6.798 1.20e-11 ***
## education.c               0.0520792  0.0077169   6.749 1.68e-11 ***
## index_ANexp.c:ideology.c  0.0008944  0.0001827   4.895 1.02e-06 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.361 on 4462 degrees of freedom
##   (367 observations deleted due to missingness)
## Multiple R-squared:  0.126,  Adjusted R-squared:  0.1248 
## F-statistic: 107.2 on 6 and 4462 DF,  p-value: < 2.2e-16

summary(m.xmy <- lm(vaxxAttitudes ~ (trustExpert.c + index_ANexp.c) * ideology.c + 
                      USvUK + age.c + education.c, data = d1))

## 
## Call:
## lm(formula = vaxxAttitudes ~ (trustExpert.c + index_ANexp.c) * 
##     ideology.c + USvUK + age.c + education.c, data = d1)
## 
## Residuals:
##    Min     1Q Median     3Q    Max 
## -5.671 -1.200  0.325  1.302  5.446 
## 
## Coefficients:
##                            Estimate Std. Error t value Pr(>|t|)    
## (Intercept)               0.9454010  0.0301096  31.399  < 2e-16 ***
## trustExpert.c             0.5286189  0.0200586  26.354  < 2e-16 ***
## index_ANexp.c             0.0036938  0.0004044   9.135  < 2e-16 ***
## ideology.c               -0.1237554  0.0187824  -6.589 4.95e-11 ***
## USvUK                     0.6054085  0.0651792   9.288  < 2e-16 ***
## age.c                     0.0165169  0.0018831   8.771  < 2e-16 ***
## education.c               0.0247892  0.0103518   2.395  0.01668 *  
## trustExpert.c:ideology.c -0.0071300  0.0116994  -0.609  0.54227    
## index_ANexp.c:ideology.c  0.0008140  0.0002497   3.260  0.00112 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.815 on 4460 degrees of freedom
##   (367 observations deleted due to missingness)
## Multiple R-squared:  0.2398, Adjusted R-squared:  0.2384 
## F-statistic: 175.9 on 8 and 4460 DF,  p-value: < 2.2e-16

tab_model(m.xy, m.xm, m.xmy,
          show.df = T, 
          show.ci = .95,
          show.se = T,
          show.stat = T,
          string.stat = "t",
          string.se="SE",
          string.est = "Est",
          digits = 3)

	vaxxAttitudes						trustExpert						vaxxAttitudes
Predictors	Est	SE	CI	t	p	df	Est	SE	CI	t	p	df	Est	SE	CI	t	p	df
(Intercept)	1.025	0.032	0.963 – 1.087	32.462	<0.001	4462.000	1.705	0.022	1.662 – 1.748	77.411	<0.001	4462.000	0.945	0.030	0.886 – 1.004	31.399	<0.001	4460.000
index_ANexp.c	0.005	0.000	0.004 – 0.006	11.944	<0.001	4462.000	0.003	0.000	0.002 – 0.003	9.208	<0.001	4462.000	0.004	0.000	0.003 – 0.004	9.135	<0.001	4460.000
ideology.c	-0.231	0.020	-0.269 – -0.192	-11.724	<0.001	4462.000	-0.201	0.014	-0.228 – -0.174	-14.668	<0.001	4462.000	-0.124	0.019	-0.161 – -0.087	-6.589	<0.001	4460.000
USvUK	0.902	0.069	0.767 – 1.038	13.073	<0.001	4462.000	0.564	0.048	0.469 – 0.658	11.710	<0.001	4462.000	0.605	0.065	0.478 – 0.733	9.288	<0.001	4460.000
age.c	0.022	0.002	0.018 – 0.025	10.693	<0.001	4462.000	0.010	0.001	0.007 – 0.012	6.798	<0.001	4462.000	0.017	0.002	0.013 – 0.020	8.771	<0.001	4460.000
education.c	0.053	0.011	0.031 – 0.074	4.748	<0.001	4462.000	0.052	0.008	0.037 – 0.067	6.749	<0.001	4462.000	0.025	0.010	0.004 – 0.045	2.395	0.017	4460.000
index_ANexp.c * ideology.c	0.001	0.000	0.001 – 0.002	4.788	<0.001	4462.000	0.001	0.000	0.001 – 0.001	4.895	<0.001	4462.000	0.001	0.000	0.000 – 0.001	3.260	0.001	4460.000
trustExpert.c													0.529	0.020	0.489 – 0.568	26.354	<0.001	4460.000
trustExpert.c * ideology.c													-0.007	0.012	-0.030 – 0.016	-0.609	0.542	4460.000
Observations	4469						4469						4469
R² / R² adjusted	0.121 / 0.120						0.126 / 0.125						0.240 / 0.238

7. Figures

d.us <- d1[d1$country_factor == "US",]
d.us$ideology.c <- d.us$ideology - mean(d.us$ideology, na.rm = T) 
d.us$index_ANexp.c <- d.us$index_ANexp - mean(d.us$index_ANexp, na.rm = T) 
d.us$age.c <- d.us$age - mean(d.us$age, na.rm = T) 
d.us$education.c <- d.us$education - mean(d.us$education, na.rm = T) 

#run model
m.US <- lm(vaxxAttitudes ~ ideology.c * index_ANexp.c + 
             age.c + education.c, data = d.us)

#create plot
p <- plot_model(m.US, type = "pred", 
                terms = c("index_ANexp.c", "ideology.c [-1.59, 1.59]")) +
  ggtitle("") + 
  ylab("") +
  xlab("") +
  theme_minimal() +
  theme(panel.grid.major = element_blank(), 
        panel.grid.minor = element_blank(),
        panel.background = element_blank(), 
        axis.line = element_line(colour = "black"),
        legend.position = c(.5, .2), 
        legend.direction = "horizontal", 
        legend.box = "horizontal",
        legend.background = element_rect(fill = "white", color = "white"),
        legend.title = element_blank())

p.US <-  p + scale_color_manual(
      labels = c("Liberal (-1 SD)", "Conservative (+1 SD)"),
      values = c("blue", "red")) +
  scale_fill_manual(values = c("blue", "red")) +
  scale_x_continuous(breaks = c(-100, -50, 0, 50, 100, 150, 200, 240)) + #-87.82    235.66  
  scale_y_continuous(breaks = c(-3, -2, -1, 0, 1, 2, 3), 
                     limits = c(-3, 3),
                     labels = c("Definitely
                                would not", 
                                "Probably
                                would not",
                                "Leaning toward
                                would not",
                                "Undecided",
                                "Leaning toward
                                would",
                                "Probably
                                would",
                                "Definitely
                                would"))

## Scale for 'colour' is already present. Adding another scale for 'colour',
## which will replace the existing scale.

## Scale for 'y' is already present. Adding another scale for 'y', which will
## replace the existing scale.

p.US

d.uk <- d1[d1$country_factor == "UK",]
d.uk$ideology.c <- d.uk$ideology - mean(d.uk$ideology, na.rm = T) 
d.uk$index_ANexp.c <- d.uk$index_ANexp - mean(d.uk$index_ANexp, na.rm = T) 
d.uk$age.c <- d.uk$age - mean(d.uk$age, na.rm = T) 
d.uk$education.c <- d.uk$education - mean(d.uk$education, na.rm = T) 

#run model
m.UK <- lm(vaxxAttitudes ~ ideology.c * index_ANexp.c + 
             age.c + education.c, data = d1[d1$country_factor == "UK",])

#create plot
p <- plot_model(m.UK, type = "pred", 
                terms = c("index_ANexp.c", "ideology.c [-1.35, 1.35]")) +
  ggtitle("") + 
  ylab("") +
  xlab("") +
  theme_minimal() +
  theme(panel.grid.major = element_blank(), 
        panel.grid.minor = element_blank(),
        panel.background = element_blank(), 
        axis.line = element_line(colour = "black"),
        legend.position = "none", 
        legend.background = element_rect(fill = "white", color = "white"),
        legend.title = element_blank())

p.UK <-  p + scale_color_manual(labels = NULL, 
                       values = c("blue", "red")) +
  scale_fill_manual(values = c("blue", "red")) +
  scale_x_continuous(breaks = c(-100, -50, 0, 50, 100, 150, 200, 250, 270),
                     limits = c(-90, 270),
) + #-75.14,    284.47
  scale_y_continuous(breaks = c(-3, -2, -1, 0, 1, 2, 3), 
                     limits = c(-3, 3),
                     labels = c("Definitely
                                would not", 
                                "Probably
                                would not",
                                "Leaning toward
                                would not",
                                "Undecided",
                                "Leaning toward
                                would",
                                "Probably
                                would",
                                "Definitely
                                would"))

## Scale for 'colour' is already present. Adding another scale for 'colour',
## which will replace the existing scale.

## Scale for 'y' is already present. Adding another scale for 'y', which will
## replace the existing scale.

p.UK

## Warning: Removed 4 row(s) containing missing values (geom_path).

p <- ggarrange(p.US, p.UK, 
          labels = c("United States: Willingness to get vaccinated ", 
                     "United Kingdom: Willingness to get vaccinated"),
          ncol = 2, nrow = 1, align = "hv",
          common.legend = T)

## Warning: Removed 4 row(s) containing missing values (geom_path).

Covid Media Manuscript Analyses for Study 1 (US vs. UK)

Dani Grant

1. import data sets

DESCRIPTIVE STATS

1. demographics

a. sample size

b. age

c. gender

d. party identity

e. education

2. variables of interest

a. media exposure

b. analytical media index

c. symbolic ideology

d. average trust in experts

3. LIWC media measures

a. Analytical thinking scales

i. US

ii. UK

4. outcome measures

5. pairwise correlation table for IVs, DVs, and covariates

i. US

i. UK

6. Analyses for Study 1 US + UK (July-August November)

a. vaxxAttitudes ~ analticalIndex * ideology + USvUK + age + education

b. vaxxAttitudes ~ analticalIndex * ideology * USvUK + age + education

c. mediation path: analyticalIndex -> trustInExperts -> vaccineAttitudes

- no interaction

- interaction between moderator and ideology

7. Figures