To examine whether the covert mouse tracking measures also reflect a sense of conflict when disavowing party-aligned political conspiraices, we also evaluated participants’ confidence in the correctness of their given reponse to all statement types, as reported on a Likert scale of 1 to 5 (1= low confidence, 5 = high confidence). Specifically, we are interested in whether partisans report lower confidence to their disavowal of party-aligned political conspiracies (relative to other statement types), and/or higher confidence to their disavowal of political conspiracies that originate with the opposing party (relative to other statement types).
To evaluate these confidence scores, and the ways in which they might align with mouse-tracking’ latency and average deviation results, we ran identical mixed-effects models, but now with the Likert-scale confidence score entered as dependent variable. This means that non-belief responses to political and general conspiracy responses within each level of Republican or Democratic responders, testing the interactions between directness of statements and how political conspiracy statements were disavowed relative to a) each other, b) general knowledge statements as baseline, and c) general conspiracy items as baseline.
#### Read Data ####
fromprep = read.delim("/Users/nduran/Dropbox (ASU)/POLIBELIEFS/analysis/JESP/REDO/3_Ranalysis/Mouse-Traj-2GIT.csv", header = TRUE,sep=',')
Recoding:
fromprep <- within( fromprep, {
group.id <- subjnum ## subject number
item.id <- questionType ## item code
dependent.variable <- value ## action dynamic response variables
## Factor 1: Identification as Republican or Democrat (2-levels; between-subject)
political.identity <- poliID2
## Factor 2: Wording of item direct or indirect (2-levels; within-subject)
statement.direct <- explicit
## Factor 3: Item conspiracy about Bush, Obama, or general, or a general knowledge (4-levels; within-subject)
BO.GB.GK.GC <- resp_type
condition <- ifelse( BO.GB.GK.GC=="Gen Conspiracy", "Gen Conspiracy",
ifelse( BO.GB.GK.GC=="Right-wing", "Obama",
ifelse( BO.GB.GK.GC=="Left-wing", "Bush",
ifelse( BO.GB.GK.GC=="Gen Knowledge", "Gen Knowledge", NA )) ) )
} )
c_scale <- function(x) {
scale(x, scale = FALSE) }
fromprep <- within( fromprep, age.c <- c_scale(fromprep$age) )
Descriptives of main variables:
political.identity | condition | mean | se |
---|---|---|---|
Democrat | Bush | 4.041506 | 0.02387710 |
Democrat | Gen Conspiracy | 4.115098 | 0.02340956 |
Democrat | Gen Knowledge | 4.383957 | 0.01163638 |
Democrat | Obama | 4.167523 | 0.02148178 |
Republican | Bush | 4.453172 | 0.02941087 |
Republican | Gen Conspiracy | 4.231228 | 0.03572133 |
Republican | Gen Knowledge | 4.429036 | 0.01794082 |
Republican | Obama | 3.777589 | 0.04847632 |
Contrast structure:
nonbelief.data <- within( subset( fromprep, endorse!="C" ), {
## for within subjects analysis
republican <- ifelse( poliID2=="Republican", 0,
ifelse( poliID2=="Democrat", 1, NA ) )
## for within subjects analysis
democrat <- ifelse( poliID2=="Democrat", 0,
ifelse( poliID2=="Republican", 1, NA ) )
statement.direct.w <- ifelse( statement.direct=="Reject as False", 1/2,
ifelse( statement.direct=="Accept as True", -1/2, NA ) )
ObamaVsKnowledge <- ifelse( condition=="Bush", 0,
ifelse( condition=="Obama", 1/2,
ifelse( condition=="Gen Conspiracy", 0,
ifelse( condition=="Gen Knowledge", -1/2, NA ))))
BushVsKnowledge <- ifelse( condition=="Bush", 1/2,
ifelse( condition=="Obama", 0,
ifelse( condition=="Gen Conspiracy", 0,
ifelse( condition=="Gen Knowledge", -1/2, NA ))))
GeneralVsKnowledge <- ifelse( condition=="Bush", 0,
ifelse( condition=="Obama", 0,
ifelse( condition=="Gen Conspiracy", 1/2,
ifelse( condition=="Gen Knowledge", -1/2, NA ))))
} )
Republicans
republican.omnibus.nonbeliever.model <- lmer( dependent.variable ~ (1|group.id) + (1|item.id) + age.c + statement.direct.w * republican * (ObamaVsKnowledge + BushVsKnowledge + GeneralVsKnowledge), data=nonbelief.data, REML=FALSE, na.action = na.exclude, subset=variable=="choice" )
## Estimate
## (Intercept) 4.14255710
## age.c 0.01337005
## statement.direct.w -0.01524948
## republican -0.01546794
## ObamaVsKnowledge -0.97381184
## BushVsKnowledge 0.42182834
## GeneralVsKnowledge -0.03987610
## statement.direct.w:republican 0.10230144
## statement.direct.w:ObamaVsKnowledge -0.13460741
## statement.direct.w:BushVsKnowledge 0.11172117
## republican:ObamaVsKnowledge 0.97413110
## republican:BushVsKnowledge -0.77587413
## republican:GeneralVsKnowledge -0.21307830
## statement.direct.w:republican:ObamaVsKnowledge 0.29083832
## statement.direct.w:republican:BushVsKnowledge -0.12195486
## R2c
## 0.3149317
##// test for main effect in order to follow-up
republican.nonbeliever.main = update( republican.omnibus.nonbeliever.model, .~.-ObamaVsKnowledge-BushVsKnowledge-GeneralVsKnowledge)
## republican.nonbeliever.main
## df 2.000000e+00
## chi 2.983894e+01
## p 3.315552e-07
planCont = rbind(
# note that "Bush" items become the new baseline
"OvB" = c(0, 0, 0, 0, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, 0),
"OvK" = c(0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0),
"BvK" = c(0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0),
# note that "General Conspiracy" items become the new baseline
"OvGC" = c(0, 0, 0, 0, 1, 0, -1, 0, 0, 0, 0, 0, 0, 0, 0),
"BvGC" = c(0, 0, 0, 0, 0, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0))
# note that "General Conspiracy" items become the new baseline
# "OvB:Direct" = c(0, 0, 0, 0, 1, 0, -1, 0, 0, 0, 0, 0, 0, 1, -1),
# "OvK:Direct" = c(0, 0, 0, 0, 0, 1, -1, 0, 0, 0, 0, 0, 0, 1, 0),
# "BvK:Direct" = c(0, 0, 0, 0, 0, 1, -1, 0, 0, 0, 0, 0, 0, 0, 1))
## Estimate Std..Error t.value p.values DV
## OvB -1.3956402 0.3172819 -4.398739 1.088816e-05 republican/nonbelief
## OvK -0.9738118 0.1995765 -4.879390 1.064142e-06 republican/nonbelief
## BvK 0.4218283 0.1963049 2.148843 3.164686e-02 republican/nonbelief
## OvGC -0.9339357 0.3981288 -2.345813 1.898564e-02 republican/nonbelief
## BvGC 0.4617044 0.3946757 1.169832 2.420684e-01 republican/nonbelief
##// test for two-way interaction in order to follow-up
republican.nonbeliever.twoway = update( republican.omnibus.nonbeliever.model, .~.-statement.direct.w:ObamaVsKnowledge-statement.direct.w:BushVsKnowledge)
## republican.nonbeliever.twoway
## df 1.0000000
## chi 0.1507182
## p 0.6978499
Democrats
democrat.omnibus.nonbeliever.model <- lmer( dependent.variable ~ (1|group.id) + (1|item.id) + age.c + statement.direct.w * democrat * (ObamaVsKnowledge + BushVsKnowledge + GeneralVsKnowledge), data=nonbelief.data, REML=FALSE, na.action = na.exclude, subset=variable=="choice" )
## Estimate
## (Intercept) 4.1270891580
## age.c 0.0133700460
## statement.direct.w 0.0870519550
## democrat 0.0154679448
## ObamaVsKnowledge 0.0003192602
## BushVsKnowledge -0.3540457882
## GeneralVsKnowledge -0.2529544005
## statement.direct.w:democrat -0.1023014376
## statement.direct.w:ObamaVsKnowledge 0.1562309072
## statement.direct.w:BushVsKnowledge -0.0102336825
## democrat:ObamaVsKnowledge -0.9741311023
## democrat:BushVsKnowledge 0.7758741296
## democrat:GeneralVsKnowledge 0.2130783004
## statement.direct.w:democrat:ObamaVsKnowledge -0.2908383218
## statement.direct.w:democrat:BushVsKnowledge 0.1219548568
## R2c
## 0.3149316
##// test for main effect in order to follow-up
democrat.nonbeliever.main = update( democrat.omnibus.nonbeliever.model, .~.-ObamaVsKnowledge-BushVsKnowledge-GeneralVsKnowledge)
## democrat.nonbeliever.main
## df 2.000000000
## chi 14.964085199
## p 0.000563106
planCont = rbind(
# note that "Bush" items become the new baseline
"OvB" = c(0, 0, 0, 0, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0, 0),
"OvK" = c(0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0),
"BvK" = c(0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0),
# note that "General Conspiracy" items become the new baseline
"OvGC" = c(0, 0, 0, 0, 1, 0, -1, 0, 0, 0, 0, 0, 0, 0, 0),
"BvGC" = c(0, 0, 0, 0, 0, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0))
# note that "General Conspiracy" items become the new baseline
# "OvB:Direct" = c(0, 0, 0, 0, 1, 0, -1, 0, 0, 0, 0, 0, 0, 1, -1),
# "OvK:Direct" = c(0, 0, 0, 0, 0, 1, -1, 0, 0, 0, 0, 0, 0, 1, 0),
# "BvK:Direct" = c(0, 0, 0, 0, 0, 1, -1, 0, 0, 0, 0, 0, 0, 0, 1))
## Estimate Std..Error t.value p.values DV
## OvB 0.3543650484 0.3074936 1.152430599 0.24914417 democrat/nonbelief
## OvK 0.0003192602 0.1921718 0.001661327 0.99867445 democrat/nonbelief
## BvK -0.3540457882 0.1926749 -1.837529393 0.06613178 democrat/nonbelief
## OvGC 0.2532736607 0.3881890 0.652449380 0.51411132 democrat/nonbelief
## BvGC -0.1010913878 0.3886325 -0.260120803 0.79477059 democrat/nonbelief
##// test for two-way interaction in order to follow-up
democrat.nonbeliever.twoway = update( democrat.omnibus.nonbeliever.model, .~.-statement.direct.w:ObamaVsKnowledge-statement.direct.w:BushVsKnowledge)
## democrat.nonbeliever.twoway
## df 1.00000000
## chi 0.07324036
## p 0.78667595