Pilot data from a survey experiment on affective polarization; illustrates experimental analysis with pre-post design

Pilot data from Cutler, Pietrka, and Rainey (2025), a preregistered replication of Ahler and Sood's (2018) survey experiment on affective polarization. Respondents are randomly assigned to one of three conditions: (1) a control, (2) an "ask" condition where they report their perceptions of the out-party's demographic composition, or (3) a "tell" condition where they report perceptions and then receive corrections.

Usage

asking_pilot

Format

A data frame with 33 variables. Variables are ordered as: respondent information, pre-treatment measures, treatment condition, then post-treatment outcomes.

out_party: respondent's out-party: "Democrat" or "Republican"
pre_out_party_ft: pre-treatment feeling thermometer rating of out-party supporters (0-100; higher = warmer)
pre_diff_ft: pre-treatment difference in feeling thermometers: in-party FT minus out-party FT (higher = more polarized)
pre_alt_aps: pre-treatment alternative three-item Affective Polarization Scale (sum of three items, each -3 to +3)
pre_political_sectarianism: pre-treatment political sectarianism composite (sum of 9 items, each 1-7; range 9-63)
pre_ps_othering: pre-treatment political sectarianism othering subdimension (sum of 3 items; range 3-21)
pre_ps_aversion: pre-treatment political sectarianism aversion subdimension (sum of 3 items; range 3-21)
pre_ps_moralization: pre-treatment political sectarianism moralization subdimension (sum of 3 items; range 3-21)
pre_ps_othering_1: political sectarianism othering item 1: "I don't feel distant from [out-party]" (1) to "I feel as if [out-party] and I are on separate planets" (7)
pre_ps_othering_2: political sectarianism othering item 2: "There are no ways in which I am different from [out-party]" (1) to "I am as different from [out-party] as can be" (7)
pre_ps_othering_3: political sectarianism othering item 3: "I don't have any difficulty seeing the world the way [out-party] do" (1) to "It's impossible for me to see the world the way [out-party] do" (7)
pre_ps_aversion_1: political sectarianism aversion item 1: "My feelings toward [out-party] aren't negative at all" (1) to "My feelings toward [out-party] are overwhelmingly negative" (7)
pre_ps_aversion_2: political sectarianism aversion item 2: "I don't hate [out-party] at all" (1) to "I have a fierce hatred for [out-party]" (7)
pre_ps_aversion_3: political sectarianism aversion item 3: "[Out-party] have no negative traits" (1) to "[Out-party] have every negative trait in the book" (7)
pre_ps_moralization_1: political sectarianism moralization item 1: "There is nothing immoral about [out-party]" (1) to "[Out-party] are completely immoral" (7)
pre_ps_moralization_2: political sectarianism moralization item 2: "[Out-party] are not evil in any way" (1) to "[Out-party] are evil in every way" (7)
pre_ps_moralization_3: political sectarianism moralization item 3: "[Out-party] do not lack integrity" (1) to "[Out-party] lack any shred of integrity" (7)
condition: experimental condition: "control", "ask", or "tell" (factor with control as reference level)
out_party_ft: post-treatment feeling thermometer rating of out-party supporters (0-100; higher = warmer)
diff_ft: post-treatment difference in feeling thermometers: in-party FT minus out-party FT (higher = more polarized)
aps: post-treatment Affective Polarization Scale composite (sum of 9 items; range -27 to +27)
aps_othering: post-treatment APS othering subdimension (sum of 3 items; range -9 to +9)
aps_aversion: post-treatment APS aversion subdimension (sum of 3 items; range -9 to +9)
aps_moralization: post-treatment APS moralization subdimension (sum of 3 items; range -9 to +9)
aps_othering_o1: APS othering item O1: "I feel as though [in-party] are very different from [out-party]" (-3 to +3)
aps_othering_o6: APS othering item O6: "[Out-party] live in a different world from us [in-party]" (-3 to +3)
aps_othering_o8: APS othering item O8: "[Out-party] act in ways that us [in-party] could never understand" (-3 to +3)
aps_aversion_a1: APS aversion item A1: "As a [in-party], I would not want to be friends with someone who was a [out-party]" (-3 to +3)
aps_aversion_a2: APS aversion item A2: "If I found out a friend of mine was a [out-party], I would want to stop spending time with them" (-3 to +3)
aps_aversion_a4: APS aversion item A4: "Although I do not agree with their political views, there are people I like who are [out-party]" (-3 to +3; reverse-coded)
aps_moralization_m2: APS moralization item M2: "My identity as a [in-party] is connected to my core moral beliefs" (-3 to +3)
aps_moralization_m4: APS moralization item M4: "My identity as a [in-party] reflects my beliefs about the difference between right and wrong" (-3 to +3)
aps_moralization_m6: APS moralization item M6: "My identity as a [in-party] is rooted in moral principles" (-3 to +3)

Details

The data include pre- and post-treatment measures of affective polarization. Pre-treatment measures include feeling thermometers, a three-item alternative Affective Polarization Scale (Campos and Federico 2026), and the nine-item political sectarianism scale (Finkel et al. 2024). Post-treatment measures include feeling thermometers and the nine-item Affective Polarization Scale (Campos and Federico 2026).

Pure independents are excluded from the data set.

The Affective Polarization Scale (APS) has three subdimensions, each with three items: othering (belief that partisans are fundamentally different), aversion (tendency to dislike and avoid out-partisans), and moralization (perception that one's partisan identity reflects moral values). APS items are coded from -3 (Strongly Disagree) to +3 (Strongly Agree). Item A4 is reverse-coded.

The political sectarianism scale shares the same three subdimensions but uses semantic differential items (1-7 bipolar scale) focused on attitudes toward out-partisans.

References

Cutler, Austin, Matthew Pietrka, and Carlisle Rainey. 2025. "Merely Asking: A Replication of Ahler and Sood (2018)." Working Paper.

Ahler, Douglas J. and Gaurav Sood. 2018. "The Parties in Our Heads: Misperceptions about Party Composition and Their Consequences." Journal of Politics 80(3): 964-981. doi:10.1086/697253

Campos, Nicolas and Christopher Federico. 2026. "A New Measure of Affective Polarization." American Political Science Review 120(1): 160-178. doi:10.1017/S0003055425000255

Finkel, Eli J., Alexander P. Landry, James N. Druckman, Jay J. Van Bavel, and Rick H. Hoyle. 2024. "Partisan Antipathy and the Erosion of Democratic Norms." Working Paper.

Examples


# load the pilot data
ap <- crdata::asking_pilot

# compare out-party feeling thermometer across conditions
tapply(ap$out_party_ft, ap$condition, mean, na.rm = TRUE)
#>  control      ask     tell 
#> 30.74324 23.98701 28.02740 

# estimate treatment effects with pre-post design
fit <- lm(out_party_ft ~ condition + pre_out_party_ft + pre_alt_aps, data = ap)
summary(fit)
#> 
#> Call:
#> lm(formula = out_party_ft ~ condition + pre_out_party_ft + pre_alt_aps, 
#>     data = ap)
#> 
#> Residuals:
#>     Min      1Q  Median      3Q     Max 
#> -42.145  -4.148  -1.715   1.966  93.543 
#> 
#> Coefficients:
#>                  Estimate Std. Error t value Pr(>|t|)    
#> (Intercept)       5.86730    1.81134   3.239  0.00139 ** 
#> conditionask     -3.31890    1.93521  -1.715  0.08776 .  
#> conditiontell    -1.29191    1.95725  -0.660  0.50991    
#> pre_out_party_ft  0.89113    0.03855  23.114  < 2e-16 ***
#> pre_alt_aps      -0.15057    0.27448  -0.549  0.58386    
#> ---
#> Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#> 
#> Residual standard error: 11.85 on 219 degrees of freedom
#>   (3 observations deleted due to missingness)
#> Multiple R-squared:  0.7639,	Adjusted R-squared:  0.7596 
#> F-statistic: 177.1 on 4 and 219 DF,  p-value: < 2.2e-16
#>