setwd("~/Documents/Dropbox/Research/Bernd") 
p<-read.csv ("measuring_TIF.csv", header=T, sep=",")
p$attn1[p$attn=="attention" |p$attn== "Attention" |p$attn== "ATTENTION"]<-1
p$attn1[p$attn!="attention" &p$attn!= "Attention" &p$attn!= "ATTENTION"]<-0
table(p$attn1)
## 
##   0   1 
##  59 344
#p<-subset(p, attn=="attention" |attn== "Attention" |attn== "ATTENTION")

MC check car

table(p$attncar)
## 
##   1   2   3   4 
## 254  19   7   6
p$attncar1[p$attncar==1]<-1
p$attncar1[p$attncar!=1]<-0
#p<-subset(p, attncar==1)

creating variables: the difference in willingness to try/buy driverless cars before vs. after facts, and trust in feelings + reverse coded trust in facts. then doing t tests and correlations.

p$cars1<-(p$Rcars1_1+p$Rcars1_2)/2
p$cars2<-(p$Rcars2_1+p$Rcars2_2)/2

#people are marginally more willing post-facts
t.test(p$cars1, p$cars2)
## 
##  Welch Two Sample t-test
## 
## data:  p$cars1 and p$cars2
## t = -1.6837, df = 569.76, p-value = 0.09279
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -9.8215752  0.7551417
## sample estimates:
## mean of x mean of y 
##  51.79021  56.32343
p$carsdiff<-p$cars2-p$cars1

#people rely more on feelings than on facts overall when reporting their willingness post-facts
t.test(p$carsmod_1, p$carsmod_2)
## 
##  Welch Two Sample t-test
## 
## data:  p$carsmod_1 and p$carsmod_2
## t = 2.3965, df = 561.21, p-value = 0.01688
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##   1.000033 10.086754
## sample estimates:
## mean of x mean of y 
##  66.13287  60.58947
p$carsmod<- p$carsmod_1 + (101-p$carsmod_2)

#so there's a negative correlation between trust in feelings and the size of the change in willingness pre to post fact. more trust in feelings, smaller change in willingness. 
cor.test(p$carsmod, p$carsdiff)
## 
##  Pearson's product-moment correlation
## 
## data:  p$carsmod and p$carsdiff
## t = -2.0432, df = 283, p-value = 0.04196
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.233484150 -0.004443013
## sample estimates:
##        cor 
## -0.1205678

The key relationship between reliance in feelings (carsmod) and the difference in attitudes pre/post fact holds even when controling for passing the attention checks.

summary(lm(carsdiff~carsmod + attn1 + attncar1, data=p))
## 
## Call:
## lm(formula = carsdiff ~ carsmod + attn1 + attncar1, data = p)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -66.836  -5.734  -1.926   3.395  93.766 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)   
## (Intercept)  0.75899    3.79303   0.200  0.84155   
## carsmod     -0.04145    0.01904  -2.176  0.03035 * 
## attn1        0.80457    2.50346   0.321  0.74816   
## attncar1     8.35721    2.81621   2.968  0.00326 **
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 14.78 on 281 degrees of freedom
##   (118 observations deleted due to missingness)
## Multiple R-squared:  0.04519,    Adjusted R-squared:  0.03499 
## F-statistic: 4.433 on 3 and 281 DF,  p-value: 0.004601

No correlation between reliance on feelings and confidence in decision strong correlation between reliance on feelings and belief that the facts are true

cor.test(p$carsmod, p$confcar_1)
## 
##  Pearson's product-moment correlation
## 
## data:  p$carsmod and p$confcar_1
## t = 0.040784, df = 278, p-value = 0.9675
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.1148083  0.1196332
## sample estimates:
##         cor 
## 0.002446044
cor.test(p$carsmod, p$believecar_1)
## 
##  Pearson's product-moment correlation
## 
## data:  p$carsmod and p$believecar_1
## t = -9.8815, df = 283, p-value < 2.2e-16
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.5880644 -0.4146997
## sample estimates:
##        cor 
## -0.5064831

MEAT

table(p$meatcheck)
## 
##   1   2   3   4 
##  29 215   7   6
p$meatcheck1[p$meatcheck==2]<-1
p$meatcheck1[p$meatcheck!=2]<-0

people are more willing to try this meat post facts. no difference in reliance on feelings vs. facts!

t.test(p$meat1_1, p$meat2_1)
## 
##  Welch Two Sample t-test
## 
## data:  p$meat1_1 and p$meat2_1
## t = -2.2176, df = 507.38, p-value = 0.02702
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -12.5358650  -0.7582526
## sample estimates:
## mean of x mean of y 
##  57.25882  63.90588
t.test(p$meatmod_1, p$meatmod_2)
## 
##  Welch Two Sample t-test
## 
## data:  p$meatmod_1 and p$meatmod_2
## t = 0.62897, df = 504.4, p-value = 0.5297
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -3.572218  6.936441
## sample estimates:
## mean of x mean of y 
##  62.80859  61.12648

no correlation here :(

p$meatdiff<-p$meat2_1-p$meat1_1
p$meatmod<-p$meatmod_1+(101-p$meatmod_2)

cor.test(p$meatmod, p$meatdiff)
## 
##  Pearson's product-moment correlation
## 
## data:  p$meatmod and p$meatdiff
## t = -0.18042, df = 249, p-value = 0.857
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.1350608  0.1125457
## sample estimates:
##         cor 
## -0.01143278

Interestingly, however, there IS a correlation between reliance on feelings and willingness to try the meat, poth pre- and post-facts

cor.test(p$meatmod, p$meat2_1)
## 
##  Pearson's product-moment correlation
## 
## data:  p$meatmod and p$meat2_1
## t = -10.591, df = 250, p-value < 2.2e-16
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.6363401 -0.4649309
## sample estimates:
##        cor 
## -0.5565296
cor.test(p$meatmod, p$meat1_1)
## 
##  Pearson's product-moment correlation
## 
## data:  p$meatmod and p$meat1_1
## t = -10.006, df = 249, p-value < 2.2e-16
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.6183491 -0.4409520
## sample estimates:
##        cor 
## -0.5355321

key relationship still isn’t there when controlling for attention check success

summary(lm(meatdiff~meatmod + attn1 + meatcheck1, data=p))
## 
## Call:
## lm(formula = meatdiff ~ meatmod + attn1 + meatcheck1, data = p)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -62.485  -8.296  -3.297   4.281  64.799 
## 
## Coefficients:
##              Estimate Std. Error t value Pr(>|t|)  
## (Intercept) -3.434714   3.636143  -0.945   0.3458  
## meatmod     -0.001215   0.018125  -0.067   0.9466  
## attn1        5.763243   2.552242   2.258   0.0248 *
## meatcheck1   6.091064   2.425843   2.511   0.0127 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 14.33 on 247 degrees of freedom
##   (152 observations deleted due to missingness)
## Multiple R-squared:  0.04529,    Adjusted R-squared:  0.03369 
## F-statistic: 3.906 on 3 and 247 DF,  p-value: 0.009434

TERRORISM

table(p$mc_terror)
## 
##   1   2   3   4 
## 236  12  10   3
p$terrorcheck[p$mc_terror==1]<-1
p$terrorcheck[p$mc_terror!=1]<-0

Much less concerned post facts, much more reliance on facts vs. feelings

p$terror1<-(p$terror_concern_1_1+p$terror_die_1_1)/2
p$terror2<-(p$terror_concern_2_1+p$terror_die_2_1)/2

t.test(p$terror1, p$terror2)
## 
##  Welch Two Sample t-test
## 
## data:  p$terror1 and p$terror2
## t = 4.0852, df = 518.98, p-value = 5.1e-05
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##   4.118053 11.747847
## sample estimates:
## mean of x mean of y 
##  33.14943  25.21648
t.test(p$terrormod_1, p$terrormod_2)
## 
##  Welch Two Sample t-test
## 
## data:  p$terrormod_1 and p$terrormod_2
## t = -7.345, df = 514.5, p-value = 8.115e-13
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -24.37736 -14.08871
## sample estimates:
## mean of x mean of y 
##  51.59004  70.82308

Expected correlation shows up again.

p$terrordiff<-p$terror1-p$terror2
p$terrormod<-p$terrormod_1+(101-p$terrormod_2)

cor.test(p$terrormod, p$terrordiff)
## 
##  Pearson's product-moment correlation
## 
## data:  p$terrormod and p$terrordiff
## t = -2.9858, df = 258, p-value = 0.0031
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.29779181 -0.06249484
## sample estimates:
##       cor 
## -0.182759

Holds controlling for attn checks.

summary(lm(terrordiff~terrormod + attn1 + terrorcheck, data=p))
## 
## Call:
## lm(formula = terrordiff ~ terrormod + attn1 + terrorcheck, data = p)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -39.633  -7.598  -3.108   6.029  51.684 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)   
## (Intercept)  9.86221    3.44198   2.865  0.00451 **
## terrormod   -0.04676    0.01624  -2.879  0.00432 **
## attn1       -2.90730    2.35779  -1.233  0.21868   
## terrorcheck  4.57706    2.80401   1.632  0.10384   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 12.99 on 256 degrees of freedom
##   (143 observations deleted due to missingness)
## Multiple R-squared:  0.04654,    Adjusted R-squared:  0.03537 
## F-statistic: 4.166 on 3 and 256 DF,  p-value: 0.006651
cor.test(p$believeterror_1, p$terrormod)
## 
##  Pearson's product-moment correlation
## 
## data:  p$believeterror_1 and p$terrormod
## t = -6.5385, df = 256, p-value = 3.35e-10
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.4783148 -0.2685695
## sample estimates:
##        cor 
## -0.3782871

Trust in institutions… more liberal people trust the govt less, academic research more, CDC more, no diff. for tech companies

cor.test(p$poli2_1, p$trust_1)
## 
##  Pearson's product-moment correlation
## 
## data:  p$poli2_1 and p$trust_1
## t = -1.7241, df = 398, p-value = 0.08547
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.18260968  0.01205389
## sample estimates:
##         cor 
## -0.08609959
cor.test(p$poli2_1, p$trust_3)
## 
##  Pearson's product-moment correlation
## 
## data:  p$poli2_1 and p$trust_3
## t = 6.2368, df = 398, p-value = 1.142e-09
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  0.2063664 0.3851636
## sample estimates:
##       cor 
## 0.2983805
cor.test(p$poli2_1, p$trust_4)
## 
##  Pearson's product-moment correlation
## 
## data:  p$poli2_1 and p$trust_4
## t = 5.943, df = 398, p-value = 6.117e-09
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  0.1928453 0.3731057
## sample estimates:
##       cor 
## 0.2854986
cor.test(p$poli2_1, p$trust_5)
## 
##  Pearson's product-moment correlation
## 
## data:  p$poli2_1 and p$trust_5
## t = 1.0076, df = 397, p-value = 0.3143
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.04790702  0.14794625
## sample estimates:
##        cor 
## 0.05050517
#republican/democrat t-tests
rep<-subset(p, party==1)
dem<-subset(p, party==2)
t.test(rep$trust_1, dem$trust_1)
## 
##  Welch Two Sample t-test
## 
## data:  rep$trust_1 and dem$trust_1
## t = 1.1475, df = 278.66, p-value = 0.2522
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -2.308800  8.762133
## sample estimates:
## mean of x mean of y 
##  39.70667  36.48000
t.test(rep$trust_3, dem$trust_3)
## 
##  Welch Two Sample t-test
## 
## data:  rep$trust_3 and dem$trust_3
## t = -4.8955, df = 276.63, p-value = 1.669e-06
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -17.627492  -7.516508
## sample estimates:
## mean of x mean of y 
##    59.020    71.592
t.test(rep$trust_4, dem$trust_4)
## 
##  Welch Two Sample t-test
## 
## data:  rep$trust_4 and dem$trust_4
## t = -5.3247, df = 293.7, p-value = 2.014e-07
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -20.246552  -9.318781
## sample estimates:
## mean of x mean of y 
##  54.89333  69.67600
t.test(rep$trust_5, dem$trust_5)
## 
##  Welch Two Sample t-test
## 
## data:  rep$trust_5 and dem$trust_5
## t = -0.14955, df = 283.84, p-value = 0.8812
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -5.343948  4.589264
## sample estimates:
## mean of x mean of y 
##  54.79866  55.17600

Effect of reliance on feelings on difference in attitudes towards terrorism pre/post facts holds after controlling for trust in CDC.

Also, trust in CDC correlates positively with the size of the difference pre/post facts, and negatively with trust in feelings

summary(lm(terrordiff~ terrormod + trust_4, data=p))
## 
## Call:
## lm(formula = terrordiff ~ terrormod + trust_4, data = p)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -36.581  -7.387  -2.518   6.212  50.013 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)  
## (Intercept)  7.00145    2.74356   2.552   0.0113 *
## terrormod   -0.04254    0.01647  -2.583   0.0103 *
## trust_4      0.06406    0.03044   2.104   0.0363 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 12.95 on 256 degrees of freedom
##   (144 observations deleted due to missingness)
## Multiple R-squared:  0.05114,    Adjusted R-squared:  0.04373 
## F-statistic: 6.899 on 2 and 256 DF,  p-value: 0.001207
cor.test(p$trust_4, p$terrordiff)
## 
##  Pearson's product-moment correlation
## 
## data:  p$trust_4 and p$terrordiff
## t = 2.9119, df = 258, p-value = 0.003906
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  0.05798543 0.29366178
## sample estimates:
##       cor 
## 0.1783808
cor.test(p$trust_4, p$terrormod)
## 
##  Pearson's product-moment correlation
## 
## data:  p$trust_4 and p$terrormod
## t = -3.2428, df = 257, p-value = 0.00134
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.31259995 -0.07826845
## sample estimates:
##        cor 
## -0.1982657

meat relationship doesn’t improve controlling for trust in academic research

summary(lm(meatdiff~meatmod + trust_3, data=p))
## 
## Call:
## lm(formula = meatdiff ~ meatmod + trust_3, data = p)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -66.481  -6.579  -4.767   3.496  66.780 
## 
## Coefficients:
##              Estimate Std. Error t value Pr(>|t|)  
## (Intercept)  6.237650   3.589733   1.738   0.0835 .
## meatmod     -0.003613   0.018895  -0.191   0.8485  
## trust_3      0.007530   0.037669   0.200   0.8417  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 14.64 on 246 degrees of freedom
##   (154 observations deleted due to missingness)
## Multiple R-squared:  0.0003822,  Adjusted R-squared:  -0.007745 
## F-statistic: 0.04703 on 2 and 246 DF,  p-value: 0.9541

main relationship for cars holds controlling for trust in academic research

summary(lm(carsdiff~carsmod + trust_3, data=p))
## 
## Call:
## lm(formula = carsdiff ~ carsmod + trust_3, data = p)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -74.247  -5.904  -1.454   3.249  92.933 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)  
## (Intercept)  4.19590    3.52098   1.192   0.2344  
## carsmod     -0.03442    0.01943  -1.772   0.0775 .
## trust_3      0.05973    0.03653   1.635   0.1031  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 14.94 on 281 degrees of freedom
##   (119 observations deleted due to missingness)
## Multiple R-squared:  0.0238, Adjusted R-squared:  0.01685 
## F-statistic: 3.425 on 2 and 281 DF,  p-value: 0.0339

other moderators - desire for control

p$cont<-(p$cont1_1+p$cont2_1)/2
summary(lm(carsdiff~carsmod * cont, data=p))
## 
## Call:
## lm(formula = carsdiff ~ carsmod * cont, data = p)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -74.364  -5.508  -2.132   3.410  93.430 
## 
## Coefficients:
##              Estimate Std. Error t value Pr(>|t|)
## (Intercept)   9.94024    9.20593   1.080    0.281
## carsmod      -0.07882    0.07796  -1.011    0.313
## cont         -0.25792    1.87479  -0.138    0.891
## carsmod:cont  0.00834    0.01584   0.527    0.599
## 
## Residual standard error: 14.99 on 281 degrees of freedom
##   (118 observations deleted due to missingness)
## Multiple R-squared:  0.01756,    Adjusted R-squared:  0.007068 
## F-statistic: 1.674 on 3 and 281 DF,  p-value: 0.1728
cor.test(p$cont, p$carsmod)
## 
##  Pearson's product-moment correlation
## 
## data:  p$cont and p$carsmod
## t = -0.26256, df = 283, p-value = 0.7931
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.1315539  0.1007644
## sample estimates:
##         cor 
## -0.01560539
summary(lm(meatdiff~meatmod * cont, data=p))
## 
## Call:
## lm(formula = meatdiff ~ meatmod * cont, data = p)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -66.276  -7.228  -3.994   4.417  68.345 
## 
## Coefficients:
##              Estimate Std. Error t value Pr(>|t|)  
## (Intercept)  -7.74952    8.52165  -0.909   0.3640  
## meatmod       0.10506    0.07482   1.404   0.1615  
## cont          3.19267    1.81509   1.759   0.0798 .
## meatmod:cont -0.02376    0.01575  -1.509   0.1326  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 14.57 on 247 degrees of freedom
##   (152 observations deleted due to missingness)
## Multiple R-squared:  0.01256,    Adjusted R-squared:  0.0005654 
## F-statistic: 1.047 on 3 and 247 DF,  p-value: 0.3723
cor.test(p$cont, p$meatmod)
## 
##  Pearson's product-moment correlation
## 
## data:  p$cont and p$meatmod
## t = 0.94618, df = 251, p-value = 0.345
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.06418398  0.18160862
## sample estimates:
##        cor 
## 0.05961589
summary(lm(terrordiff~terrormod * cont, data=p))
## 
## Call:
## lm(formula = terrordiff ~ terrormod * cont, data = p)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -39.631  -7.258  -3.032   5.806  51.608 
## 
## Coefficients:
##                  Estimate Std. Error t value Pr(>|t|)  
## (Intercept)    11.8109532  5.7026346   2.071   0.0393 *
## terrormod      -0.0518062  0.0545109  -0.950   0.3428  
## cont           -0.0413310  1.2023780  -0.034   0.9726  
## terrormod:cont  0.0007968  0.0116478   0.068   0.9455  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 13.08 on 256 degrees of freedom
##   (143 observations deleted due to missingness)
## Multiple R-squared:  0.03342,    Adjusted R-squared:  0.0221 
## F-statistic: 2.951 on 3 and 256 DF,  p-value: 0.03326
cor.test(p$cont, p$terrormod)
## 
##  Pearson's product-moment correlation
## 
## data:  p$cont and p$terrormod
## t = -0.67035, df = 258, p-value = 0.5032
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.16252712  0.08036342
## sample estimates:
##         cor 
## -0.04169791

reliance on feelings vs. facts for political and for tech decisions in general… works for meat!

summary(lm(carsdiff~carsmod + techmode_1, data=p))
## 
## Call:
## lm(formula = carsdiff ~ carsmod + techmode_1, data = p)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -73.942  -5.694  -1.841   2.944  93.707 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)
## (Intercept)  4.50801    5.06308   0.890    0.374
## carsmod     -0.03293    0.02044  -1.611    0.108
## techmode_1   0.65873    0.72020   0.915    0.361
## 
## Residual standard error: 14.97 on 282 degrees of freedom
##   (118 observations deleted due to missingness)
## Multiple R-squared:  0.01745,    Adjusted R-squared:  0.01048 
## F-statistic: 2.504 on 2 and 282 DF,  p-value: 0.08354
summary(lm(meatdiff~meatmod + techmode_1, data=p))
## 
## Call:
## lm(formula = meatdiff ~ meatmod + techmode_1, data = p)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -62.566  -7.441  -3.332   4.150  62.246 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)  
## (Intercept) -5.15291    5.16036  -0.999   0.3190  
## meatmod      0.01412    0.01953   0.723   0.4704  
## techmode_1   1.90937    0.75630   2.525   0.0122 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 14.45 on 248 degrees of freedom
##   (152 observations deleted due to missingness)
## Multiple R-squared:  0.02518,    Adjusted R-squared:  0.01732 
## F-statistic: 3.203 on 2 and 248 DF,  p-value: 0.04231
summary(lm(terrordiff~terrormod + polimode_1, data=p))
## 
## Call:
## lm(formula = terrordiff ~ terrormod + polimode_1, data = p)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -39.733  -7.244  -2.877   5.932  51.768 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)   
## (Intercept) 13.35590    4.17390   3.200  0.00155 **
## terrormod   -0.05007    0.01669  -3.000  0.00296 **
## polimode_1  -0.30360    0.68126  -0.446  0.65623   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 13.05 on 257 degrees of freedom
##   (143 observations deleted due to missingness)
## Multiple R-squared:  0.03415,    Adjusted R-squared:  0.02663 
## F-statistic: 4.543 on 2 and 257 DF,  p-value: 0.01151