How do measurement methods (feeling thermometers vs. trait attributions vs. experiments) change findings about partisan hostility?
This fact-check may be outdated. Consider refreshing it to get the most current information.
Executive summary
Different tools for measuring partisan hostility—feeling thermometers, trait attributions, and behavioral experiments—tend to tell overlapping but not identical stories: thermometers are broad and sensitive to context or priming, trait ratings are more precise about negative stereotyping of supporters, and experiments reveal how survey wording, targets, and incentives change apparent levels of hostility and sometimes expose measurement-driven inflation or attenuation of partisan animus [1] [2] [3].
1. Feeling thermometers: a blunt, sensitive gauge that captures mood and elites as well as voters
Feeling thermometers ask respondents to place parties or actors on a warm–cold scale and reliably register generalized sentiment, making them widely used across studies of affective polarization [1] [4]; yet they are also highly sensitive to survey context, mode, and whom respondents are thinking about—party organizations, leaders, or ordinary voters—which can inflate apparent hostility when political content or elite exemplars are salient [5] [6]. Researchers find thermometers correlate with other measures of out-party affect but show more variation and are susceptible to priming and mode effects that can make polarization look larger than it might be in behavioral terms [2] [5] [3].
2. Trait attributions: sharper measurement of stereotyping and interpersonal hostility
Asking respondents to rate the traits of opposing partisans (honesty, intelligence, hypocrisy, etc.) narrows the focus to stereotyping and interpersonal evaluations and often produces less variance than thermometers—interpreted by some authors as greater precision [7] [2]. Trait measures can diverge from thermometers on causal tests: for example, emotional manipulations sometimes change thermometer scores but not trait attributions, and in some cases fear reduced trait-based affective polarization even when thermometers did not register change [7]. This suggests trait ratings capture a more cognitive, stereotype-based component of hostility rather than broad affect alone [7] [1].
3. Experiments and behavioral tasks: revealing causal drivers and correcting measurement illusions
Experimental approaches—ranging from survey experiments that vary targets and informational primes to behavioral games and the Equality Equivalency Test—show that much of what looks like entrenched hostility depends on framing, perceived extremity, and the salience of elite cues [3] [8]. Controlled manipulations can make partisan animus rise or fall: for instance, experimentally varying the portrayed engagement and extremity of out-partisans reduces hostile ratings, implying canonical measures overstate dislike when respondents rely on stereotypes of extreme, visible exemplars [3]. Behavioral experiments add another layer: observable choices and allocations sometimes show less bias than thermometer differentials, indicating attitudinal measures do not always translate into discriminatory behavior [8] [3].
4. Why the measures diverge—and what that implies about "hostility"
The divergence among methods stems from different targets (parties vs. elites vs. ordinary supporters), different psychological processes (affect vs. stereotyping vs. motivated behavior), and measurement artifacts (survey mode, sampling of strong partisans, priming) that shape responses [6] [5] [9]. Thermometers capture broad affect—including reactions to elites and media exemplars—so they are useful for tracking mood and societal trends but risk conflating elite-driven outrage with interpersonal animus; trait ratings better isolate stereotypes about ordinary supporters; experiments show how fragile apparent hostility can be when key cues are manipulated [6] [9] [3].
5. Competing interpretations, hidden incentives, and research agendas
Scholars disagree about whether the apparent rise in affective polarization is a measurement artifact or a substantive social trend: some work argues canonical survey instruments overstate hostility by priming extreme exemplars (the "illusion" argument) while other longitudinal analyses find consistent increases across thermometer, trait, and affect batteries, especially when face-to-face or behavioral measures are considered [3] [9]. Methodological choices reflect disciplinary priorities—survey researchers prize comparability and breadth, experimentalists prioritize causal leverage, and behavioral work focuses on real-world consequences—so claims about dangerous polarization often depend on which instrument an author favors [3] [8] [1].
6. Practical takeaway for interpreting research on partisan hostility
Read findings with an eye to measurement: warm–cold thermometers are efficient for broad surveillance but sensitive to framing and may overstate outgroup hate when elites are salient; trait measures give a clearer window into stereotyping of supporters; and experiments/behavioral tests are essential to assess causality and whether attitudes predict discriminatory behavior. No single instrument is definitive; triangulation across thermometers, trait attributions, and behavioral experiments produces the most reliable picture of when partisan hostility is real, contextual, or instrument-driven [2] [3] [8].