1. means polarization

describe(ds$media_1, na.rm = T)  #NTTimes = 3.78

##    vars  n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 18 3.78 2.76      4    3.88 2.97  -2   8    10 -0.1    -0.68 0.65

describe(ds$media_2, na.rm = T)  #WSJ = -1.38

##    vars  n  mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 16 -1.38 4.94     -2   -1.57 4.45  -8   8    16 0.76    -0.69 1.23

describe(ds$media_3, na.rm = T)  #WashPost = 1.88

##    vars  n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 16 1.88 4.22      2    1.86 2.97  -6  10    16 0.39    -0.34 1.06

describe(ds$media_4, na.rm = T)  #USA today = 1.67

##    vars  n mean  sd median trimmed  mad min max range skew kurtosis  se
## X1    1 12 1.67 3.8      0     1.2 1.48  -2  10    12 1.09    -0.25 1.1

describe(ds$media_5, na.rm = T)  #Fox News = -6.78

##    vars  n  mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 18 -6.78 4.76     -8   -7.25 1.48 -10   4    14 1.53     0.64 1.12

describe(ds$media_6, na.rm = T)  #CNN = 5.56

##    vars  n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 18 5.56 2.91      6    5.56 2.97   1  10     9 0.09     -1.3 0.69

describe(ds$media_7, na.rm = T)  #MSNBC = 5.17

##    vars  n mean   sd median trimmed  mad min max range  skew kurtosis   se
## X1    1 18 5.17 3.76      6    5.31 5.19  -2  10    12 -0.32    -1.25 0.89

describe(ds$media_8, na.rm = T)  #Yahoo = 1.92

##    vars  n mean   sd median trimmed  mad min max range skew kurtosis  se
## X1    1 12 1.92 3.82      0     1.5 2.22  -2  10    12 0.92    -0.58 1.1

describe(ds$media_9, na.rm = T)  #HuffPost = 5.56

##    vars  n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 16 5.56 3.16      6    5.79 2.97  -2  10    12 -0.9     0.02 0.79

describe(ds$media_10, na.rm = T) #AOL = 2

##    vars n mean   sd median trimmed mad min max range skew kurtosis   se
## X1    1 7    2 4.32      0       2   0  -2  10    12 0.85    -1.08 1.63

describe(ds$media_11, na.rm = T) #NPR = 3.5

##    vars  n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 14  3.5 3.67      2    3.25 2.97   0  10    10 0.51    -1.53 0.98

describe(ds$media_12, na.rm = T) #ABC = 0.81

##    vars  n mean   sd median trimmed  mad min max range  skew kurtosis  se
## X1    1 16 0.81 2.01      1    0.79 1.48  -2   4     6 -0.13     -1.2 0.5

describe(ds$media_13, na.rm = T) #NBC = 2.18

##    vars  n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 17 2.18 3.38      2    2.07 1.48  -4  10    14 0.62     0.13 0.82

describe(ds$media_14, na.rm = T) #CBS = 3.24

##    vars  n mean  sd median trimmed  mad min max range skew kurtosis   se
## X1    1 17 3.24 3.4      2    3.13 2.97  -2  10    12 0.59    -0.94 0.82

describe(ds$media_15, na.rm = T) #PBS = 1.38

##    vars  n mean   sd median trimmed mad min max range skew kurtosis   se
## X1    1 16 1.38 2.13      0    1.14   0   0   6     6 1.26     0.05 0.53

2. interrate reliability

kripp.alpha(as.matrix(ds), method = c("interval"))

##  Krippendorff's alpha
## 
##  Subjects = 16 
##    Raters = 18 
##     alpha = 0.466

3. create media polarization index

describe media index

describe(d$mediaIndex, na.rm = T)

##    vars    n mean   sd median trimmed  mad  min   max range skew kurtosis   se
## X1    1 1207 5.03 2.36   5.08    5.01 2.45 0.13 11.75 11.62 0.06    -0.51 0.07

hist(d$mediaIndex)

4. analyses

a. own vote ~ index * partyID

Yes, there is a sig interaction between party ID and media type watched in effect on confidence in own vote counted.

simple effects for media index:

Dems: B = .18, p < .001(M = .30)

Reps: B = .09, p < .001 (M = -.24)

Inds: B .24, p < .001 (M = -.65)

ma <- lm(ownvote.c ~  mediaIndex.c * (pDem_Rep + pInd_Not), data = d)

tab_model(ma)

	ownvote.c
Predictors	Estimates	CI	p
(Intercept)	-0.19	-0.28 – -0.11	<0.001
mediaIndex.c	0.17	0.13 – 0.21	<0.001
pDem_Rep	-0.54	-0.72 – -0.35	<0.001
pInd_Not	0.68	0.48 – 0.87	<0.001
mediaIndex.c * pDem_Rep	-0.09	-0.17 – -0.01	0.023
mediaIndex.c * pInd_Not	-0.10	-0.20 – 0.00	0.055
Observations	1202
R² / R² adjusted	0.202 / 0.198

summary(ma)

## 
## Call:
## lm(formula = ownvote.c ~ mediaIndex.c * (pDem_Rep + pInd_Not), 
##     data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.4141 -0.9061  0.1746  0.8278  2.8186 
## 
## Coefficients:
##                       Estimate Std. Error t value Pr(>|t|)    
## (Intercept)           -0.19255    0.04340  -4.436 9.99e-06 ***
## mediaIndex.c           0.17151    0.02041   8.402  < 2e-16 ***
## pDem_Rep              -0.53610    0.09598  -5.585 2.88e-08 ***
## pInd_Not               0.67660    0.10053   6.730 2.62e-11 ***
## mediaIndex.c:pDem_Rep -0.08963    0.03943  -2.273   0.0232 *  
## mediaIndex.c:pInd_Not -0.09837    0.05112  -1.924   0.0545 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.207 on 1196 degrees of freedom
##   (34 observations deleted due to missingness)
## Multiple R-squared:  0.2018, Adjusted R-squared:  0.1984 
## F-statistic: 60.46 on 5 and 1196 DF,  p-value: < 2.2e-16

maa <- lm(ownvote.c ~  mediaIndex.c * (party_factor), data = d)

plot_model(maa, type = "pred", terms = c("mediaIndex.c", "party_factor"))

i. simple effects for dems

mai <- lm(ownvote.c ~  mediaIndex.c * (pDemR + pDemI), data = d)
tab_model(mai)

	ownvote.c
Predictors	Estimates	CI	p
(Intercept)	0.30	0.17 – 0.43	<0.001
mediaIndex.c	0.18	0.13 – 0.24	<0.001
pDemR	-0.54	-0.72 – -0.35	<0.001
pDemI	-0.94	-1.16 – -0.73	<0.001
mediaIndex.c * pDemR	-0.09	-0.17 – -0.01	0.023
mediaIndex.c * pDemI	0.05	-0.06 – 0.16	0.334
Observations	1202
R² / R² adjusted	0.202 / 0.198

summary(mai)

## 
## Call:
## lm(formula = ownvote.c ~ mediaIndex.c * (pDemR + pDemI), data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.4141 -0.9061  0.1746  0.8278  2.8186 
## 
## Coefficients:
##                    Estimate Std. Error t value Pr(>|t|)    
## (Intercept)         0.29878    0.06543   4.567 5.47e-06 ***
## mediaIndex.c        0.18387    0.02917   6.304 4.07e-10 ***
## pDemR              -0.53610    0.09598  -5.585 2.88e-08 ***
## pDemI              -0.94465    0.10993  -8.593  < 2e-16 ***
## mediaIndex.c:pDemR -0.08963    0.03943  -2.273   0.0232 *  
## mediaIndex.c:pDemI  0.05356    0.05545   0.966   0.3343    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.207 on 1196 degrees of freedom
##   (34 observations deleted due to missingness)
## Multiple R-squared:  0.2018, Adjusted R-squared:  0.1984 
## F-statistic: 60.46 on 5 and 1196 DF,  p-value: < 2.2e-16

ii. simple effects for reps

maii <- lm(ownvote.c ~  mediaIndex.c * (pRepD + pRepI), data = d)
tab_model(maii)

	ownvote.c
Predictors	Estimates	CI	p
(Intercept)	-0.24	-0.38 – -0.10	0.001
mediaIndex.c	0.09	0.04 – 0.15	<0.001
pRepD	0.54	0.35 – 0.72	<0.001
pRepI	-0.41	-0.63 – -0.19	<0.001
mediaIndex.c * pRepD	0.09	0.01 – 0.17	0.023
mediaIndex.c * pRepI	0.14	0.04 – 0.25	0.008
Observations	1202
R² / R² adjusted	0.202 / 0.198

summary(maii)

## 
## Call:
## lm(formula = ownvote.c ~ mediaIndex.c * (pRepD + pRepI), data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.4141 -0.9061  0.1746  0.8278  2.8186 
## 
## Coefficients:
##                    Estimate Std. Error t value Pr(>|t|)    
## (Intercept)        -0.23732    0.07023  -3.379 0.000750 ***
## mediaIndex.c        0.09424    0.02654   3.551 0.000398 ***
## pRepD               0.53610    0.09598   5.585 2.88e-08 ***
## pRepI              -0.40854    0.11285  -3.620 0.000307 ***
## mediaIndex.c:pRepD  0.08963    0.03943   2.273 0.023204 *  
## mediaIndex.c:pRepI  0.14318    0.05412   2.646 0.008254 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.207 on 1196 degrees of freedom
##   (34 observations deleted due to missingness)
## Multiple R-squared:  0.2018, Adjusted R-squared:  0.1984 
## F-statistic: 60.46 on 5 and 1196 DF,  p-value: < 2.2e-16

iii. simple effects for indep

maiii <- lm(ownvote.c ~  mediaIndex.c * (pIndD + pIndR), data = d)
tab_model(maiii)

	ownvote.c
Predictors	Estimates	CI	p
(Intercept)	-0.65	-0.82 – -0.47	<0.001
mediaIndex.c	0.24	0.14 – 0.33	<0.001
pIndD	0.94	0.73 – 1.16	<0.001
pIndR	0.41	0.19 – 0.63	<0.001
mediaIndex.c * pIndD	-0.05	-0.16 – 0.06	0.334
mediaIndex.c * pIndR	-0.14	-0.25 – -0.04	0.008
Observations	1202
R² / R² adjusted	0.202 / 0.198

summary(maiii)

## 
## Call:
## lm(formula = ownvote.c ~ mediaIndex.c * (pIndD + pIndR), data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.4141 -0.9061  0.1746  0.8278  2.8186 
## 
## Coefficients:
##                    Estimate Std. Error t value Pr(>|t|)    
## (Intercept)        -0.64587    0.08834  -7.311 4.82e-13 ***
## mediaIndex.c        0.23742    0.04716   5.034 5.54e-07 ***
## pIndD               0.94465    0.10993   8.593  < 2e-16 ***
## pIndR               0.40854    0.11285   3.620 0.000307 ***
## mediaIndex.c:pIndD -0.05356    0.05545  -0.966 0.334320    
## mediaIndex.c:pIndR -0.14318    0.05412  -2.646 0.008254 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.207 on 1196 degrees of freedom
##   (34 observations deleted due to missingness)
## Multiple R-squared:  0.2018, Adjusted R-squared:  0.1984 
## F-statistic: 60.46 on 5 and 1196 DF,  p-value: < 2.2e-16

b. national vote ~ index * partyID

There is again a significant interaction for Party ID and mediaIndex for effect on trust in national vote counted. Slope steepness move from Reps, Dems, to Indeps.

Simple effects for mediaIndex:

Dems: B = .17, p < .001 (M = .56)

Reps: B = .26, p < .001 (M = -.40)

Ind: B = .34, p < .001 (M = -.33)

mb <- lm(overallvote.c ~  mediaIndex.c * (pDem_Rep + pInd_Not), data = d)
summary(mb)

## 
## Call:
## lm(formula = overallvote.c ~ mediaIndex.c * (pDem_Rep + pInd_Not), 
##     data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.3221 -0.8638  0.0721  0.8326  3.3625 
## 
## Coefficients:
##                       Estimate Std. Error t value Pr(>|t|)    
## (Intercept)           -0.05265    0.04027  -1.307   0.1913    
## mediaIndex.c           0.25740    0.01899  13.553  < 2e-16 ***
## pDem_Rep              -0.95450    0.08897 -10.728  < 2e-16 ***
## pInd_Not               0.40725    0.09334   4.363 1.39e-05 ***
## mediaIndex.c:pDem_Rep  0.09057    0.03655   2.478   0.0134 *  
## mediaIndex.c:pInd_Not -0.11847    0.04764  -2.487   0.0130 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.119 on 1195 degrees of freedom
##   (35 observations deleted due to missingness)
## Multiple R-squared:  0.3894, Adjusted R-squared:  0.3868 
## F-statistic: 152.4 on 5 and 1195 DF,  p-value: < 2.2e-16

tab_model(mb)

	overallvote.c
Predictors	Estimates	CI	p
(Intercept)	-0.05	-0.13 – 0.03	0.191
mediaIndex.c	0.26	0.22 – 0.29	<0.001
pDem_Rep	-0.95	-1.13 – -0.78	<0.001
pInd_Not	0.41	0.22 – 0.59	<0.001
mediaIndex.c * pDem_Rep	0.09	0.02 – 0.16	0.013
mediaIndex.c * pInd_Not	-0.12	-0.21 – -0.03	0.013
Observations	1201
R² / R² adjusted	0.389 / 0.387

maa <- lm(overallvote.c ~  mediaIndex.c * (party_factor), data = d)

plot_model(maa, type = "pred", terms = c("mediaIndex.c", "party_factor"))

i. simple effects for dems

mbi <- lm(overallvote.c ~  mediaIndex.c * (pDemR + pDemI), data = d)
summary(mbi)

## 
## Call:
## lm(formula = overallvote.c ~ mediaIndex.c * (pDemR + pDemI), 
##     data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.3221 -0.8638  0.0721  0.8326  3.3625 
## 
## Coefficients:
##                    Estimate Std. Error t value Pr(>|t|)    
## (Intercept)         0.55900    0.06065   9.217  < 2e-16 ***
## mediaIndex.c        0.17302    0.02704   6.400 2.23e-10 ***
## pDemR              -0.95450    0.08897 -10.728  < 2e-16 ***
## pDemI              -0.88450    0.10204  -8.669  < 2e-16 ***
## mediaIndex.c:pDemR  0.09057    0.03655   2.478  0.01335 *  
## mediaIndex.c:pDemI  0.16375    0.05164   3.171  0.00156 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.119 on 1195 degrees of freedom
##   (35 observations deleted due to missingness)
## Multiple R-squared:  0.3894, Adjusted R-squared:  0.3868 
## F-statistic: 152.4 on 5 and 1195 DF,  p-value: < 2.2e-16

tab_model(mbi)

	overallvote.c
Predictors	Estimates	CI	p
(Intercept)	0.56	0.44 – 0.68	<0.001
mediaIndex.c	0.17	0.12 – 0.23	<0.001
pDemR	-0.95	-1.13 – -0.78	<0.001
pDemI	-0.88	-1.08 – -0.68	<0.001
mediaIndex.c * pDemR	0.09	0.02 – 0.16	0.013
mediaIndex.c * pDemI	0.16	0.06 – 0.27	0.002
Observations	1201
R² / R² adjusted	0.389 / 0.387

ii. simple effects for reps

mbii <- lm(overallvote.c ~  mediaIndex.c * (pRepD + pRepI), data = d)
summary(mbii)

## 
## Call:
## lm(formula = overallvote.c ~ mediaIndex.c * (pRepD + pRepI), 
##     data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.3221 -0.8638  0.0721  0.8326  3.3625 
## 
## Coefficients:
##                    Estimate Std. Error t value Pr(>|t|)    
## (Intercept)        -0.39551    0.06510  -6.075 1.66e-09 ***
## mediaIndex.c        0.26360    0.02460  10.716  < 2e-16 ***
## pRepD               0.95450    0.08897  10.728  < 2e-16 ***
## pRepI               0.07000    0.10474   0.668   0.5040    
## mediaIndex.c:pRepD -0.09057    0.03655  -2.478   0.0134 *  
## mediaIndex.c:pRepI  0.07318    0.05040   1.452   0.1468    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.119 on 1195 degrees of freedom
##   (35 observations deleted due to missingness)
## Multiple R-squared:  0.3894, Adjusted R-squared:  0.3868 
## F-statistic: 152.4 on 5 and 1195 DF,  p-value: < 2.2e-16

tab_model(mbii)

	overallvote.c
Predictors	Estimates	CI	p
(Intercept)	-0.40	-0.52 – -0.27	<0.001
mediaIndex.c	0.26	0.22 – 0.31	<0.001
pRepD	0.95	0.78 – 1.13	<0.001
pRepI	0.07	-0.14 – 0.28	0.504
mediaIndex.c * pRepD	-0.09	-0.16 – -0.02	0.013
mediaIndex.c * pRepI	0.07	-0.03 – 0.17	0.147
Observations	1201
R² / R² adjusted	0.389 / 0.387

iii. simple effects for indep

mbiii <- lm(overallvote.c ~  mediaIndex.c * (pIndD + pIndR), data = d)
summary(mbiii)

## 
## Call:
## lm(formula = overallvote.c ~ mediaIndex.c * (pIndD + pIndR), 
##     data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.3221 -0.8638  0.0721  0.8326  3.3625 
## 
## Coefficients:
##                    Estimate Std. Error t value Pr(>|t|)    
## (Intercept)        -0.32550    0.08205  -3.967 7.71e-05 ***
## mediaIndex.c        0.33678    0.04399   7.655 3.96e-14 ***
## pIndD               0.88450    0.10204   8.669  < 2e-16 ***
## pIndR              -0.07000    0.10474  -0.668  0.50403    
## mediaIndex.c:pIndD -0.16375    0.05164  -3.171  0.00156 ** 
## mediaIndex.c:pIndR -0.07318    0.05040  -1.452  0.14679    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.119 on 1195 degrees of freedom
##   (35 observations deleted due to missingness)
## Multiple R-squared:  0.3894, Adjusted R-squared:  0.3868 
## F-statistic: 152.4 on 5 and 1195 DF,  p-value: < 2.2e-16

tab_model(mbiii)

	overallvote.c
Predictors	Estimates	CI	p
(Intercept)	-0.33	-0.49 – -0.16	<0.001
mediaIndex.c	0.34	0.25 – 0.42	<0.001
pIndD	0.88	0.68 – 1.08	<0.001
pIndR	-0.07	-0.28 – 0.14	0.504
mediaIndex.c * pIndD	-0.16	-0.27 – -0.06	0.002
mediaIndex.c * pIndR	-0.07	-0.17 – 0.03	0.147
Observations	1201
R² / R² adjusted	0.389 / 0.387

c. voteLegit ~ index * partyID

there is no sig. interaction between DvR and MediaIndex, but there is for IvDR and mediaIndex - the steepness of the slope is smaller for DR than for Indep.

Simple effects of mediaIndex:

Dems: B = .18, p < .001 (M = .43)

Reps: B = .18, p < .001 (M = -.32)

Ind: B = .28, p < .001 (M = -.48)

mc <- lm(voteLegit.c ~  mediaIndex.c * (pDem_Rep + pInd_Not), data = d)
tab_model(mc)

	voteLegit.c
Predictors	Estimates	CI	p
(Intercept)	-0.12	-0.20 – -0.05	0.002
mediaIndex.c	0.21	0.18 – 0.25	<0.001
pDem_Rep	-0.75	-0.91 – -0.58	<0.001
pInd_Not	0.54	0.36 – 0.71	<0.001
mediaIndex.c * pDem_Rep	0.00	-0.07 – 0.07	0.989
mediaIndex.c * pInd_Not	-0.11	-0.20 – -0.02	0.021
Observations	1201
R² / R² adjusted	0.319 / 0.316

summary(mc)

## 
## Call:
## lm(formula = voteLegit.c ~ mediaIndex.c * (pDem_Rep + pInd_Not), 
##     data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.3393 -0.7781  0.1391  0.7875  2.8425 
## 
## Coefficients:
##                         Estimate Std. Error t value Pr(>|t|)    
## (Intercept)           -0.1226780  0.0386485  -3.174  0.00154 ** 
## mediaIndex.c           0.2134940  0.0182274  11.713  < 2e-16 ***
## pDem_Rep              -0.7453025  0.0853933  -8.728  < 2e-16 ***
## pInd_Not               0.5387997  0.0895802   6.015 2.39e-09 ***
## mediaIndex.c:pDem_Rep  0.0004742  0.0350811   0.014  0.98922    
## mediaIndex.c:pInd_Not -0.1054936  0.0457196  -2.307  0.02120 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.074 on 1195 degrees of freedom
##   (35 observations deleted due to missingness)
## Multiple R-squared:  0.3186, Adjusted R-squared:  0.3157 
## F-statistic: 111.7 on 5 and 1195 DF,  p-value: < 2.2e-16

maa <- lm(voteLegit.c ~  mediaIndex.c * (party_factor), data = d)

plot_model(maa, type = "pred", terms = c("mediaIndex.c", "party_factor"))

i. simple effects for dems

mci <- lm(voteLegit.c ~  mediaIndex.c * (pDemR + pDemI), data = d)
tab_model(mci)

	voteLegit.c
Predictors	Estimates	CI	p
(Intercept)	0.43	0.31 – 0.54	<0.001
mediaIndex.c	0.18	0.13 – 0.23	<0.001
pDemR	-0.75	-0.91 – -0.58	<0.001
pDemI	-0.91	-1.10 – -0.72	<0.001
mediaIndex.c * pDemR	0.00	-0.07 – 0.07	0.989
mediaIndex.c * pDemI	0.11	0.01 – 0.20	0.033
Observations	1201
R² / R² adjusted	0.319 / 0.316

summary(mci)

## 
## Call:
## lm(formula = voteLegit.c ~ mediaIndex.c * (pDemR + pDemI), data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.3393 -0.7781  0.1391  0.7875  2.8425 
## 
## Coefficients:
##                      Estimate Std. Error t value Pr(>|t|)    
## (Intercept)         0.4277772  0.0582096   7.349 3.69e-13 ***
## mediaIndex.c        0.1784440  0.0259480   6.877 9.83e-12 ***
## pDemR              -0.7453025  0.0853933  -8.728  < 2e-16 ***
## pDemI              -0.9114510  0.0979284  -9.307  < 2e-16 ***
## mediaIndex.c:pDemR  0.0004742  0.0350811   0.014   0.9892    
## mediaIndex.c:pDemI  0.1057307  0.0495571   2.134   0.0331 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.074 on 1195 degrees of freedom
##   (35 observations deleted due to missingness)
## Multiple R-squared:  0.3186, Adjusted R-squared:  0.3157 
## F-statistic: 111.7 on 5 and 1195 DF,  p-value: < 2.2e-16

ii. simple effects for reps

mcii <- lm(voteLegit.c ~  mediaIndex.c * (pRepD + pRepI), data = d)
tab_model(mcii)

	voteLegit.c
Predictors	Estimates	CI	p
(Intercept)	-0.32	-0.44 – -0.19	<0.001
mediaIndex.c	0.18	0.13 – 0.23	<0.001
pRepD	0.75	0.58 – 0.91	<0.001
pRepI	-0.17	-0.36 – 0.03	0.099
mediaIndex.c * pRepD	-0.00	-0.07 – 0.07	0.989
mediaIndex.c * pRepI	0.11	0.01 – 0.20	0.030
Observations	1201
R² / R² adjusted	0.319 / 0.316

summary(mcii)

## 
## Call:
## lm(formula = voteLegit.c ~ mediaIndex.c * (pRepD + pRepI), data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.3393 -0.7781  0.1391  0.7875  2.8425 
## 
## Coefficients:
##                      Estimate Std. Error t value Pr(>|t|)    
## (Intercept)        -0.3175253  0.0624793  -5.082 4.33e-07 ***
## mediaIndex.c        0.1789183  0.0236090   7.578 6.99e-14 ***
## pRepD               0.7453025  0.0853933   8.728  < 2e-16 ***
## pRepI              -0.1661485  0.1005250  -1.653   0.0986 .  
## mediaIndex.c:pRepD -0.0004742  0.0350811  -0.014   0.9892    
## mediaIndex.c:pRepI  0.1052565  0.0483735   2.176   0.0298 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.074 on 1195 degrees of freedom
##   (35 observations deleted due to missingness)
## Multiple R-squared:  0.3186, Adjusted R-squared:  0.3157 
## F-statistic: 111.7 on 5 and 1195 DF,  p-value: < 2.2e-16

iii. simple effects for indep

mciii <- lm(voteLegit.c ~  mediaIndex.c * (pIndD + pIndR), data = d)
tab_model(mciii)

	voteLegit.c
Predictors	Estimates	CI	p
(Intercept)	-0.48	-0.64 – -0.33	<0.001
mediaIndex.c	0.28	0.20 – 0.37	<0.001
pIndD	0.91	0.72 – 1.10	<0.001
pIndR	0.17	-0.03 – 0.36	0.099
mediaIndex.c * pIndD	-0.11	-0.20 – -0.01	0.033
mediaIndex.c * pIndR	-0.11	-0.20 – -0.01	0.030
Observations	1201
R² / R² adjusted	0.319 / 0.316

summary(mciii)

## 
## Call:
## lm(formula = voteLegit.c ~ mediaIndex.c * (pIndD + pIndR), data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.3393 -0.7781  0.1391  0.7875  2.8425 
## 
## Coefficients:
##                    Estimate Std. Error t value Pr(>|t|)    
## (Intercept)        -0.48367    0.07875  -6.142 1.11e-09 ***
## mediaIndex.c        0.28417    0.04222   6.731 2.62e-11 ***
## pIndD               0.91145    0.09793   9.307  < 2e-16 ***
## pIndR               0.16615    0.10053   1.653   0.0986 .  
## mediaIndex.c:pIndD -0.10573    0.04956  -2.134   0.0331 *  
## mediaIndex.c:pIndR -0.10526    0.04837  -2.176   0.0298 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.074 on 1195 degrees of freedom
##   (35 observations deleted due to missingness)
## Multiple R-squared:  0.3186, Adjusted R-squared:  0.3157 
## F-statistic: 111.7 on 5 and 1195 DF,  p-value: < 2.2e-16

d. risk severity ~ index * partyID

There is an interactin between mediaIndex and DvR but not IvDR. The slope of Reps is steeper than for Dems – the more liberal media they watch the more risk they see covid has to economy and themselves.

Simple effect for mediaIndex:

Dem: B = .10, p < .001 (M = 4.37)

Rep: B = .22, p < .001 (M = 4.07)

Ind: B = .18, p < .001 (M = 4.05)

#risk3 = severity of health consequences of covid
#risk4 = severity of economic consequences of covid
#risk5 = severity of personal economic consequences of covid

psych::alpha(d[,c("risk3",
                  "risk4",
                  "risk5"
                  )])

## 
## Reliability analysis   
## Call: psych::alpha(x = d[, c("risk3", "risk4", "risk5")])
## 
##   raw_alpha std.alpha G6(smc) average_r S/N   ase mean  sd median_r
##       0.65      0.67    0.58       0.4   2 0.017  4.1 1.2     0.38
## 
##  lower alpha upper     95% confidence boundaries
## 0.62 0.65 0.69 
## 
##  Reliability if an item is dropped:
##       raw_alpha std.alpha G6(smc) average_r S/N alpha se var.r med.r
## risk3      0.53      0.55    0.38      0.38 1.2    0.026    NA  0.38
## risk4      0.50      0.50    0.34      0.34 1.0    0.028    NA  0.34
## risk5      0.65      0.66    0.49      0.49 1.9    0.020    NA  0.49
## 
##  Item statistics 
##          n raw.r std.r r.cor r.drop mean  sd
## risk3 1226  0.78  0.79  0.62   0.48  4.5 1.6
## risk4 1226  0.77  0.80  0.65   0.52  4.9 1.4
## risk5 1227  0.77  0.74  0.50   0.41  3.0 1.8
## 
## Non missing response frequency for each item
##          1    2    3    4    5    6    7 miss
## risk3 0.05 0.09 0.11 0.17 0.27 0.20 0.11 0.01
## risk4 0.02 0.04 0.08 0.19 0.30 0.26 0.11 0.01
## risk5 0.29 0.21 0.12 0.17 0.09 0.07 0.04 0.01

#alpha = .65

d$riskSeverity <- (d$risk3 + d$risk4 + d$risk5)/3

psych::describeBy(d$risk3, list(d$party_factor))

## 
##  Descriptive statistics by group 
## : Democrat
##    vars   n mean   sd median trimmed  mad min max range  skew kurtosis   se
## X1    1 564 5.29 1.22      5    5.38 1.48   1   7     6 -0.65     0.38 0.05
## ------------------------------------------------------------ 
## : Republican
##    vars   n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 466 3.76 1.66      4    3.76 1.48   1   7     6 0.04    -0.89 0.08
## ------------------------------------------------------------ 
## : Independent
##    vars   n mean   sd median trimmed  mad min max range  skew kurtosis   se
## X1    1 194 4.24 1.66      4    4.29 1.48   1   7     6 -0.25     -0.7 0.12

psych::describeBy(d$risk4, list(d$party_factor))

## 
##  Descriptive statistics by group 
## : Democrat
##    vars   n mean  sd median trimmed  mad min max range  skew kurtosis   se
## X1    1 564 5.16 1.2      5    5.22 1.48   1   7     6 -0.64     0.52 0.05
## ------------------------------------------------------------ 
## : Republican
##    vars   n mean   sd median trimmed  mad min max range  skew kurtosis   se
## X1    1 466 4.73 1.48      5    4.81 1.48   1   7     6 -0.49    -0.35 0.07
## ------------------------------------------------------------ 
## : Independent
##    vars   n mean   sd median trimmed  mad min max range  skew kurtosis   se
## X1    1 194 4.68 1.56      5    4.79 1.48   1   7     6 -0.59     0.01 0.11

psych::describeBy(d$risk5, list(d$party_factor))

## 
##  Descriptive statistics by group 
## : Democrat
##    vars   n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 565 3.14 1.82      3    2.98 1.48   1   7     6  0.5    -0.81 0.08
## ------------------------------------------------------------ 
## : Republican
##    vars   n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 466 2.67 1.71      2    2.45 1.48   1   7     6 0.79    -0.43 0.08
## ------------------------------------------------------------ 
## : Independent
##    vars   n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 194 3.09 1.91      3    2.92 2.97   1   7     6 0.49    -0.96 0.14

describe(d$riskSeverity)

##    vars    n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 1226 4.14 1.25      4    4.12 0.99   1   7     6 0.04        0 0.04

psych::describeBy(d$riskSeverity, list(d$party_factor))

## 
##  Descriptive statistics by group 
## : Democrat
##    vars   n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 564 4.53 1.08   4.33    4.51 0.99   1   7     6 0.13     0.19 0.05
## ------------------------------------------------------------ 
## : Republican
##    vars   n mean   sd median trimmed  mad min max range skew kurtosis   se
## X1    1 466 3.72 1.23   3.67    3.68 0.99   1   7     6  0.3     0.01 0.06
## ------------------------------------------------------------ 
## : Independent
##    vars   n mean   sd median trimmed  mad min max range skew kurtosis  se
## X1    1 194 4.01 1.37      4       4 0.99   1   7     6    0    -0.02 0.1

md <- lm(riskSeverity ~  mediaIndex.c * (pDem_Rep + pInd_Not), data = d)
tab_model(md)

	riskSeverity
Predictors	Estimates	CI	p
(Intercept)	4.17	4.09 – 4.25	<0.001
mediaIndex.c	0.17	0.13 – 0.21	<0.001
pDem_Rep	-0.29	-0.47 – -0.12	0.001
pInd_Not	0.17	-0.01 – 0.36	0.065
mediaIndex.c * pDem_Rep	0.12	0.04 – 0.19	0.002
mediaIndex.c * pInd_Not	-0.02	-0.12 – 0.07	0.619
Observations	1205
R² / R² adjusted	0.164 / 0.160

summary(md)

## 
## Call:
## lm(formula = riskSeverity ~ mediaIndex.c * (pDem_Rep + pInd_Not), 
##     data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.8442 -0.7581 -0.0771  0.6807  3.9349 
## 
## Coefficients:
##                       Estimate Std. Error t value Pr(>|t|)    
## (Intercept)            4.16619    0.04086 101.956  < 2e-16 ***
## mediaIndex.c           0.17068    0.01918   8.896  < 2e-16 ***
## pDem_Rep              -0.29493    0.09056  -3.257  0.00116 ** 
## pInd_Not               0.17435    0.09451   1.845  0.06531 .  
## mediaIndex.c:pDem_Rep  0.11599    0.03721   3.117  0.00187 ** 
## mediaIndex.c:pInd_Not -0.02385    0.04795  -0.497  0.61893    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.138 on 1199 degrees of freedom
##   (31 observations deleted due to missingness)
## Multiple R-squared:  0.1636, Adjusted R-squared:  0.1602 
## F-statistic: 46.92 on 5 and 1199 DF,  p-value: < 2.2e-16

maa <- lm(riskSeverity ~  mediaIndex.c * (party_factor), data = d)

plot_model(maa, type = "pred", terms = c("mediaIndex.c", "party_factor"))

i. simple effects for dems

mdi <- lm(riskSeverity ~  mediaIndex.c * (pDemR + pDemI), data = d)
tab_model(mdi)

	riskSeverity
Predictors	Estimates	CI	p
(Intercept)	4.37	4.25 – 4.49	<0.001
mediaIndex.c	0.10	0.05 – 0.16	<0.001
pDemR	-0.29	-0.47 – -0.12	0.001
pDemI	-0.32	-0.52 – -0.12	0.002
mediaIndex.c * pDemR	0.12	0.04 – 0.19	0.002
mediaIndex.c * pDemI	0.08	-0.02 – 0.18	0.116
Observations	1205
R² / R² adjusted	0.164 / 0.160

summary(mdi)

## 
## Call:
## lm(formula = riskSeverity ~ mediaIndex.c * (pDemR + pDemI), data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.8442 -0.7581 -0.0771  0.6807  3.9349 
## 
## Coefficients:
##                    Estimate Std. Error t value Pr(>|t|)    
## (Intercept)         4.37118    0.06177  70.769  < 2e-16 ***
## mediaIndex.c        0.10481    0.02755   3.804 0.000149 ***
## pDemR              -0.29493    0.09056  -3.257 0.001159 ** 
## pDemI              -0.32181    0.10342  -3.112 0.001905 ** 
## mediaIndex.c:pDemR  0.11599    0.03721   3.117 0.001868 ** 
## mediaIndex.c:pDemI  0.08185    0.05208   1.572 0.116268    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.138 on 1199 degrees of freedom
##   (31 observations deleted due to missingness)
## Multiple R-squared:  0.1636, Adjusted R-squared:  0.1602 
## F-statistic: 46.92 on 5 and 1199 DF,  p-value: < 2.2e-16

ii. simple effects for reps

mdii <- lm(riskSeverity ~  mediaIndex.c * (pRepD + pRepI), data = d)
tab_model(mdii)

	riskSeverity
Predictors	Estimates	CI	p
(Intercept)	4.08	3.95 – 4.21	<0.001
mediaIndex.c	0.22	0.17 – 0.27	<0.001
pRepD	0.29	0.12 – 0.47	0.001
pRepI	-0.03	-0.24 – 0.18	0.800
mediaIndex.c * pRepD	-0.12	-0.19 – -0.04	0.002
mediaIndex.c * pRepI	-0.03	-0.13 – 0.07	0.501
Observations	1205
R² / R² adjusted	0.164 / 0.160

summary(mdii)

## 
## Call:
## lm(formula = riskSeverity ~ mediaIndex.c * (pRepD + pRepI), data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.8442 -0.7581 -0.0771  0.6807  3.9349 
## 
## Coefficients:
##                    Estimate Std. Error t value Pr(>|t|)    
## (Intercept)         4.07626    0.06623  61.549  < 2e-16 ***
## mediaIndex.c        0.22080    0.02501   8.828  < 2e-16 ***
## pRepD               0.29493    0.09056   3.257  0.00116 ** 
## pRepI              -0.02688    0.10615  -0.253  0.80011    
## mediaIndex.c:pRepD -0.11599    0.03721  -3.117  0.00187 ** 
## mediaIndex.c:pRepI -0.03414    0.05078  -0.672  0.50148    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.138 on 1199 degrees of freedom
##   (31 observations deleted due to missingness)
## Multiple R-squared:  0.1636, Adjusted R-squared:  0.1602 
## F-statistic: 46.92 on 5 and 1199 DF,  p-value: < 2.2e-16

iii. simple effects for indep

mdiii <- lm(riskSeverity ~  mediaIndex.c * (pIndD + pIndR), data = d)
tab_model(mdiii)

	riskSeverity
Predictors	Estimates	CI	p
(Intercept)	4.05	3.89 – 4.21	<0.001
mediaIndex.c	0.19	0.10 – 0.27	<0.001
pIndD	0.32	0.12 – 0.52	0.002
pIndR	0.03	-0.18 – 0.24	0.800
mediaIndex.c * pIndD	-0.08	-0.18 – 0.02	0.116
mediaIndex.c * pIndR	0.03	-0.07 – 0.13	0.501
Observations	1205
R² / R² adjusted	0.164 / 0.160

summary(mdiii)

## 
## Call:
## lm(formula = riskSeverity ~ mediaIndex.c * (pIndD + pIndR), data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.8442 -0.7581 -0.0771  0.6807  3.9349 
## 
## Coefficients:
##                    Estimate Std. Error t value Pr(>|t|)    
## (Intercept)         4.04937    0.08295  48.815  < 2e-16 ***
## mediaIndex.c        0.18666    0.04419   4.224 2.58e-05 ***
## pIndD               0.32181    0.10342   3.112   0.0019 ** 
## pIndR               0.02688    0.10615   0.253   0.8001    
## mediaIndex.c:pIndD -0.08185    0.05208  -1.572   0.1163    
## mediaIndex.c:pIndR  0.03414    0.05078   0.672   0.5015    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.138 on 1199 degrees of freedom
##   (31 observations deleted due to missingness)
## Multiple R-squared:  0.1636, Adjusted R-squared:  0.1602 
## F-statistic: 46.92 on 5 and 1199 DF,  p-value: < 2.2e-16

e. worst of Covid behind/ahead ~ index * partyID

There is a sig interaction between DvR and mediaIndex (not IvDR though) on whether risk of covid is ahead or behind us. Reps think the worst is ahead more when they watch more liberal tv compared with Dems.

Simple effects for media index:

Dem: B = .10, p =.007 (M = 1.13)

Rep: B = .31, p < .001 (M = .59)

Ind: B = .19, p = .001 (M = .93)

# risk6 = worst of covid ahead or behind?
# (-3 = completely behind, 0 = in the worst, +3 = completely ahead)

describe(d$risk6)

##    vars    n mean   sd median trimmed  mad min max range  skew kurtosis   se
## X1    1 1227 0.76 1.63      1    0.88 1.48  -3   3     6 -0.46    -0.42 0.05

psych::describeBy(d$risk6, list(d$party_factor))

## 
##  Descriptive statistics by group 
## : Democrat
##    vars   n mean  sd median trimmed  mad min max range  skew kurtosis   se
## X1    1 565 1.27 1.3      1    1.36 1.48  -3   3     6 -0.47    -0.17 0.05
## ------------------------------------------------------------ 
## : Republican
##    vars   n mean   sd median trimmed  mad min max range  skew kurtosis   se
## X1    1 466 0.09 1.76      0    0.11 1.48  -3   3     6 -0.07    -0.87 0.08
## ------------------------------------------------------------ 
## : Independent
##    vars   n mean   sd median trimmed  mad min max range  skew kurtosis   se
## X1    1 194 0.89 1.59      1    1.01 1.48  -3   3     6 -0.43    -0.35 0.11

me <- lm(risk6 ~  mediaIndex.c * (pDem_Rep + pInd_Not), data = d)
tab_model(me)

	risk6
Predictors	Estimates	CI	p
(Intercept)	0.88	0.78 – 0.99	<0.001
mediaIndex.c	0.20	0.15 – 0.25	<0.001
pDem_Rep	-0.54	-0.77 – -0.31	<0.001
pInd_Not	-0.07	-0.31 – 0.17	0.580
mediaIndex.c * pDem_Rep	0.22	0.12 – 0.31	<0.001
mediaIndex.c * pInd_Not	0.01	-0.11 – 0.13	0.835
Observations	1206
R² / R² adjusted	0.186 / 0.183

summary(me)

## 
## Call:
## lm(formula = risk6 ~ mediaIndex.c * (pDem_Rep + pInd_Not), data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -5.2012 -1.1139 -0.0428  0.9612  3.9182 
## 
## Coefficients:
##                       Estimate Std. Error t value Pr(>|t|)    
## (Intercept)            0.88396    0.05291  16.706  < 2e-16 ***
## mediaIndex.c           0.20101    0.02484   8.091 1.44e-15 ***
## pDem_Rep              -0.54306    0.11716  -4.635 3.96e-06 ***
## pInd_Not              -0.06774    0.12245  -0.553    0.580    
## mediaIndex.c:pDem_Rep  0.21718    0.04812   4.514 7.00e-06 ***
## mediaIndex.c:pInd_Not  0.01295    0.06213   0.208    0.835    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.475 on 1200 degrees of freedom
##   (30 observations deleted due to missingness)
## Multiple R-squared:  0.1861, Adjusted R-squared:  0.1828 
## F-statistic: 54.89 on 5 and 1200 DF,  p-value: < 2.2e-16

maa <- lm(risk6 ~  mediaIndex.c * (party_factor), data = d)

plot_model(maa, type = "pred", terms = c("mediaIndex.c", "party_factor"))

i. simple effects for dems

mei <- lm(risk6 ~  mediaIndex.c * (pDemR + pDemI), data = d)
tab_model(mei)

	risk6
Predictors	Estimates	CI	p
(Intercept)	1.13	0.98 – 1.29	<0.001
mediaIndex.c	0.10	0.03 – 0.17	0.007
pDemR	-0.54	-0.77 – -0.31	<0.001
pDemI	-0.20	-0.47 – 0.06	0.128
mediaIndex.c * pDemR	0.22	0.12 – 0.31	<0.001
mediaIndex.c * pDemI	0.10	-0.04 – 0.23	0.156
Observations	1206
R² / R² adjusted	0.186 / 0.183

summary(mei)

## 
## Call:
## lm(formula = risk6 ~ mediaIndex.c * (pDemR + pDemI), data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -5.2012 -1.1139 -0.0428  0.9612  3.9182 
## 
## Coefficients:
##                    Estimate Std. Error t value Pr(>|t|)    
## (Intercept)         1.13314    0.07973  14.212  < 2e-16 ***
## mediaIndex.c        0.09669    0.03555   2.719  0.00663 ** 
## pDemR              -0.54306    0.11716  -4.635 3.96e-06 ***
## pDemI              -0.20379    0.13386  -1.522  0.12817    
## mediaIndex.c:pDemR  0.21718    0.04812   4.514 7.00e-06 ***
## mediaIndex.c:pDemI  0.09564    0.06742   1.419  0.15630    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.475 on 1200 degrees of freedom
##   (30 observations deleted due to missingness)
## Multiple R-squared:  0.1861, Adjusted R-squared:  0.1828 
## F-statistic: 54.89 on 5 and 1200 DF,  p-value: < 2.2e-16

ii. simple effects for reps

meii <- lm(risk6 ~  mediaIndex.c * (pRepD + pRepI), data = d)
tab_model(meii)

	risk6
Predictors	Estimates	CI	p
(Intercept)	0.59	0.42 – 0.76	<0.001
mediaIndex.c	0.31	0.25 – 0.38	<0.001
pRepD	0.54	0.31 – 0.77	<0.001
pRepI	0.34	0.07 – 0.61	0.014
mediaIndex.c * pRepD	-0.22	-0.31 – -0.12	<0.001
mediaIndex.c * pRepI	-0.12	-0.25 – 0.01	0.065
Observations	1206
R² / R² adjusted	0.186 / 0.183

summary(meii)

## 
## Call:
## lm(formula = risk6 ~ mediaIndex.c * (pRepD + pRepI), data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -5.2012 -1.1139 -0.0428  0.9612  3.9182 
## 
## Coefficients:
##                    Estimate Std. Error t value Pr(>|t|)    
## (Intercept)         0.59007    0.08585   6.873 1.00e-11 ***
## mediaIndex.c        0.31387    0.03242   9.681  < 2e-16 ***
## pRepD               0.54306    0.11716   4.635 3.96e-06 ***
## pRepI               0.33927    0.13759   2.466   0.0138 *  
## mediaIndex.c:pRepD -0.21718    0.04812  -4.514 7.00e-06 ***
## mediaIndex.c:pRepI -0.12154    0.06582  -1.847   0.0651 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.475 on 1200 degrees of freedom
##   (30 observations deleted due to missingness)
## Multiple R-squared:  0.1861, Adjusted R-squared:  0.1828 
## F-statistic: 54.89 on 5 and 1200 DF,  p-value: < 2.2e-16

iii. simple effects for indep

meiii <- lm(risk6 ~  mediaIndex.c * (pIndD + pIndR), data = d)
tab_model(meiii)

	risk6
Predictors	Estimates	CI	p
(Intercept)	0.93	0.72 – 1.14	<0.001
mediaIndex.c	0.19	0.08 – 0.30	0.001
pIndD	0.20	-0.06 – 0.47	0.128
pIndR	-0.34	-0.61 – -0.07	0.014
mediaIndex.c * pIndD	-0.10	-0.23 – 0.04	0.156
mediaIndex.c * pIndR	0.12	-0.01 – 0.25	0.065
Observations	1206
R² / R² adjusted	0.186 / 0.183

summary(meiii)

## 
## Call:
## lm(formula = risk6 ~ mediaIndex.c * (pIndD + pIndR), data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -5.2012 -1.1139 -0.0428  0.9612  3.9182 
## 
## Coefficients:
##                    Estimate Std. Error t value Pr(>|t|)    
## (Intercept)         0.92934    0.10753   8.643  < 2e-16 ***
## mediaIndex.c        0.19233    0.05728   3.357 0.000811 ***
## pIndD               0.20379    0.13386   1.522 0.128172    
## pIndR              -0.33927    0.13759  -2.466 0.013811 *  
## mediaIndex.c:pIndD -0.09564    0.06742  -1.419 0.156301    
## mediaIndex.c:pIndR  0.12154    0.06582   1.847 0.065059 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.475 on 1200 degrees of freedom
##   (30 observations deleted due to missingness)
## Multiple R-squared:  0.1861, Adjusted R-squared:  0.1828 
## F-statistic: 54.89 on 5 and 1200 DF,  p-value: < 2.2e-16

f. vaxxAttitudes ~ index * partyID

There is a sig interacton betwen DvR and mediaIndex on willingness to get vaxxed, but not int for IvDR and mediaIndex. As Dems and Inds watch more liberal media they are more willing to get vaxxed, but there is no change in Reps based on media.

Simple Effects for mediaIndex:

Dems: B = .32, p < .001 (M = .34)

Reps: B = .05, p = .268 (M = .24)

Ind: B = .26, p = .001, (M =.14)

# vaxxAttitudes = would you get a covid vaxx?
# (-3 = def not, 0 = undecided, +3 = def would)

describe(d$vaxxAttitudes)

##    vars    n mean   sd median trimmed  mad min max range  skew kurtosis   se
## X1    1 1217  0.4 2.14      0     0.5 2.97  -3   3     6 -0.32    -1.19 0.06

psych::describeBy(d$vaxxAttitude, list(d$party_factor, d$election_timing))

## 
##  Descriptive statistics by group 
## : Democrat
## : Pre-election
##   vars* n* mean* sd* median* trimmed* mad* min* max* range* skew* kurtosis* se*
## 1    NA NA    NA  NA      NA       NA   NA   NA   NA     NA    NA        NA  NA
## ------------------------------------------------------------ 
## : Republican
## : Pre-election
##   vars* n* mean* sd* median* trimmed* mad* min* max* range* skew* kurtosis* se*
## 1    NA NA    NA  NA      NA       NA   NA   NA   NA     NA    NA        NA  NA
## ------------------------------------------------------------ 
## : Independent
## : Pre-election
##   vars* n* mean* sd* median* trimmed* mad* min* max* range* skew* kurtosis* se*
## 1    NA NA    NA  NA      NA       NA   NA   NA   NA     NA    NA        NA  NA
## ------------------------------------------------------------ 
## : Democrat
## : During-election
##    vars   n mean   sd median trimmed  mad min max range  skew kurtosis   se
## X1    1 251 0.94 1.91      1    1.15 1.48  -3   3     6 -0.62    -0.67 0.12
## ------------------------------------------------------------ 
## : Republican
## : During-election
##    vars   n mean   sd median trimmed  mad min max range  skew kurtosis   se
## X1    1 238 0.27 2.35      0    0.34 4.45  -3   3     6 -0.23    -1.47 0.15
## ------------------------------------------------------------ 
## : Independent
## : During-election
##    vars  n  mean   sd median trimmed  mad min max range skew kurtosis  se
## X1    1 87 -0.16 1.82      0    -0.2 1.48  -3   3     6 0.01    -0.72 0.2
## ------------------------------------------------------------ 
## : Democrat
## : Post-election
##    vars   n mean   sd median trimmed  mad min max range  skew kurtosis   se
## X1    1 309 0.68 2.02      1    0.84 1.48  -3   3     6 -0.54    -0.86 0.11
## ------------------------------------------------------------ 
## : Republican
## : Post-election
##    vars   n mean   sd median trimmed  mad min max range  skew kurtosis   se
## X1    1 227 0.08 2.32      0     0.1 4.45  -3   3     6 -0.11    -1.49 0.15
## ------------------------------------------------------------ 
## : Independent
## : Post-election
##    vars   n  mean   sd median trimmed  mad min max range skew kurtosis  se
## X1    1 104 -0.24 2.01      0    -0.3 2.97  -3   3     6 0.07     -1.1 0.2

mf <- lm(vaxxAttitudes ~  mediaIndex.c * (pDem_Rep + pInd_Not), data = d)
summary(mf)

## 
## Call:
## lm(formula = vaxxAttitudes ~ mediaIndex.c * (pDem_Rep + pInd_Not), 
##     data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -4.7877 -1.7114  0.0725  1.7848  3.6912 
## 
## Coefficients:
##                       Estimate Std. Error t value Pr(>|t|)    
## (Intercept)            0.14902    0.07424   2.007   0.0449 *  
## mediaIndex.c           0.21198    0.03485   6.083 1.59e-09 ***
## pDem_Rep              -0.09926    0.16441  -0.604   0.5461    
## pInd_Not               0.43422    0.17177   2.528   0.0116 *  
## mediaIndex.c:pDem_Rep -0.27155    0.06751  -4.022 6.12e-05 ***
## mediaIndex.c:pInd_Not -0.07821    0.08715  -0.897   0.3697    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.069 on 1198 degrees of freedom
##   (32 observations deleted due to missingness)
## Multiple R-squared:  0.07324,    Adjusted R-squared:  0.06938 
## F-statistic: 18.94 on 5 and 1198 DF,  p-value: < 2.2e-16

tab_model(mf)

	vaxxAttitudes
Predictors	Estimates	CI	p
(Intercept)	0.15	0.00 – 0.29	0.045
mediaIndex.c	0.21	0.14 – 0.28	<0.001
pDem_Rep	-0.10	-0.42 – 0.22	0.546
pInd_Not	0.43	0.10 – 0.77	0.012
mediaIndex.c * pDem_Rep	-0.27	-0.40 – -0.14	<0.001
mediaIndex.c * pInd_Not	-0.08	-0.25 – 0.09	0.370
Observations	1204
R² / R² adjusted	0.073 / 0.069

maa <- lm(vaxxAttitudes ~  mediaIndex.c * (party_factor), data = d)

plot_model(maa, type = "pred", terms = c("mediaIndex.c", "party_factor"))

i. simple effects for dems

mei <- lm(vaxxAttitudes ~  mediaIndex.c * (pDemR + pDemI), data = d)
summary(mei)

## 
## Call:
## lm(formula = vaxxAttitudes ~ mediaIndex.c * (pDemR + pDemI), 
##     data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -4.7877 -1.7114  0.0725  1.7848  3.6912 
## 
## Coefficients:
##                    Estimate Std. Error t value Pr(>|t|)    
## (Intercept)         0.34195    0.11193   3.055   0.0023 ** 
## mediaIndex.c        0.32195    0.04987   6.455 1.57e-10 ***
## pDemR              -0.09926    0.16441  -0.604   0.5461    
## pDemI              -0.48385    0.18782  -2.576   0.0101 *  
## mediaIndex.c:pDemR -0.27155    0.06751  -4.022 6.12e-05 ***
## mediaIndex.c:pDemI -0.05756    0.09457  -0.609   0.5428    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.069 on 1198 degrees of freedom
##   (32 observations deleted due to missingness)
## Multiple R-squared:  0.07324,    Adjusted R-squared:  0.06938 
## F-statistic: 18.94 on 5 and 1198 DF,  p-value: < 2.2e-16

tab_model(mei)

	vaxxAttitudes
Predictors	Estimates	CI	p
(Intercept)	0.34	0.12 – 0.56	0.002
mediaIndex.c	0.32	0.22 – 0.42	<0.001
pDemR	-0.10	-0.42 – 0.22	0.546
pDemI	-0.48	-0.85 – -0.12	0.010
mediaIndex.c * pDemR	-0.27	-0.40 – -0.14	<0.001
mediaIndex.c * pDemI	-0.06	-0.24 – 0.13	0.543
Observations	1204
R² / R² adjusted	0.073 / 0.069

ii. simple effects for reps

meii <- lm(vaxxAttitudes ~  mediaIndex.c * (pRepD + pRepI), data = d)
summary(meii)

## 
## Call:
## lm(formula = vaxxAttitudes ~ mediaIndex.c * (pRepD + pRepI), 
##     data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -4.7877 -1.7114  0.0725  1.7848  3.6912 
## 
## Coefficients:
##                    Estimate Std. Error t value Pr(>|t|)    
## (Intercept)         0.24269    0.12043   2.015   0.0441 *  
## mediaIndex.c        0.05040    0.04550   1.108   0.2682    
## pRepD               0.09926    0.16441   0.604   0.5461    
## pRepI              -0.38459    0.19301  -1.993   0.0465 *  
## mediaIndex.c:pRepD  0.27155    0.06751   4.022 6.12e-05 ***
## mediaIndex.c:pRepI  0.21399    0.09234   2.318   0.0206 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.069 on 1198 degrees of freedom
##   (32 observations deleted due to missingness)
## Multiple R-squared:  0.07324,    Adjusted R-squared:  0.06938 
## F-statistic: 18.94 on 5 and 1198 DF,  p-value: < 2.2e-16

tab_model(meii)

	vaxxAttitudes
Predictors	Estimates	CI	p
(Intercept)	0.24	0.01 – 0.48	0.044
mediaIndex.c	0.05	-0.04 – 0.14	0.268
pRepD	0.10	-0.22 – 0.42	0.546
pRepI	-0.38	-0.76 – -0.01	0.047
mediaIndex.c * pRepD	0.27	0.14 – 0.40	<0.001
mediaIndex.c * pRepI	0.21	0.03 – 0.40	0.021
Observations	1204
R² / R² adjusted	0.073 / 0.069

iii. simple effects for indep

meiii <- lm(vaxxAttitudes ~  mediaIndex.c * (pIndD + pIndR), data = d)
summary(meiii)

## 
## Call:
## lm(formula = vaxxAttitudes ~ mediaIndex.c * (pIndD + pIndR), 
##     data = d)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -4.7877 -1.7114  0.0725  1.7848  3.6912 
## 
## Coefficients:
##                    Estimate Std. Error t value Pr(>|t|)   
## (Intercept)        -0.14190    0.15082  -0.941  0.34696   
## mediaIndex.c        0.26439    0.08035   3.290  0.00103 **
## pIndD               0.48385    0.18782   2.576  0.01011 * 
## pIndR               0.38459    0.19301   1.993  0.04653 * 
## mediaIndex.c:pIndD  0.05756    0.09457   0.609  0.54285   
## mediaIndex.c:pIndR -0.21399    0.09234  -2.318  0.02064 * 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.069 on 1198 degrees of freedom
##   (32 observations deleted due to missingness)
## Multiple R-squared:  0.07324,    Adjusted R-squared:  0.06938 
## F-statistic: 18.94 on 5 and 1198 DF,  p-value: < 2.2e-16

tab_model(meiii)

	vaxxAttitudes
Predictors	Estimates	CI	p
(Intercept)	-0.14	-0.44 – 0.15	0.347
mediaIndex.c	0.26	0.11 – 0.42	0.001
pIndD	0.48	0.12 – 0.85	0.010
pIndR	0.38	0.01 – 0.76	0.047
mediaIndex.c * pIndD	0.06	-0.13 – 0.24	0.543
mediaIndex.c * pIndR	-0.21	-0.40 – -0.03	0.021
Observations	1204
R² / R² adjusted	0.073 / 0.069

US_media_analyses

Dani Grant

6/10/2021

1. means polarization

2. interrate reliability

3. create media polarization index

describe media index

4. analyses

a. own vote ~ index * partyID

i. simple effects for dems

ii. simple effects for reps

iii. simple effects for indep

b. national vote ~ index * partyID

i. simple effects for dems

ii. simple effects for reps

iii. simple effects for indep

c. voteLegit ~ index * partyID

i. simple effects for dems

ii. simple effects for reps

iii. simple effects for indep

d. risk severity ~ index * partyID

i. simple effects for dems

ii. simple effects for reps

iii. simple effects for indep

e. worst of Covid behind/ahead ~ index * partyID

i. simple effects for dems

ii. simple effects for reps

iii. simple effects for indep

f. vaxxAttitudes ~ index * partyID

i. simple effects for dems

ii. simple effects for reps

iii. simple effects for indep