Feeling Thermometer Analysis
Also known as: Feeling Thermometer, Affect Thermometer, Thermometer Rating Scale
The feeling thermometer is a survey instrument that asks respondents to rate their warmth or favorability toward a person, group, or institution on a 0-to-100 scale, where 0 is very cold/unfavorable, 100 is very warm/favorable, and 50 is neutral. Introduced in the American National Election Studies in the 1960s, it is the standard measure of political affect, and its analysis underpins candidate evaluation, group affect, and affective-polarization research.
Key highlights
- Provides a fine-grained, intuitive continuous measure of affect that respondents find easy to use.
- Carried consistently in major surveys for over half a century, enabling long time series and comparative analysis.
- Supports many downstream analyses: difference scores, spatial models, and polarization indices.
- Flexible across targets, allowing parallel measurement of warmth toward candidates, parties, and social groups.
Intuition
This section is available to Pro members. Upgrade to Pro
How it works
This section is available to Pro members. Upgrade to Pro
When to use it
Use feeling-thermometer analysis whenever you need a continuous measure of political affect toward candidates, parties, groups, or institutions, and especially when comparing warmth across many targets or computing affect differences. It is appropriate for candidate-evaluation models, group-affect studies, and affective-polarization indices. Apply within-respondent standardization or difference scores to remove scale-use artifacts, and treat the bounded, heaped distribution carefully rather than assuming normality.
Strengths & limitations
- Provides a fine-grained, intuitive continuous measure of affect that respondents find easy to use.
- Carried consistently in major surveys for over half a century, enabling long time series and comparative analysis.
- Supports many downstream analyses: difference scores, spatial models, and polarization indices.
- Flexible across targets, allowing parallel measurement of warmth toward candidates, parties, and social groups.
- Ratings are contaminated by individual scale-use differences (level and spread) that require statistical correction.
- Responses heap heavily at round numbers and at 50, 85, and 100, violating interval-scale and normality assumptions.
- 'Don't know' and refusal to rate are nonrandom and can bias group comparisons if mishandled.
- The same numeric rating can mean different things to different respondents, threatening cross-person comparability.
Common pitfalls
This section is available to Pro members. Upgrade to Pro
Applications
This section is available to Pro members. Upgrade to Pro
Frequently asked
Why standardize thermometer ratings within respondent?
People use the 0-100 scale differently: some rate every target warmly, others coldly, and they differ in how much they spread their ratings. Raw scores therefore mix genuine affect with personal scale-use style. Subtracting each respondent's own mean (and optionally dividing by their standard deviation) isolates how warm they feel toward a target relative to their other ratings, making cross-person comparisons valid.
What causes the heaping at 50, 85, and 100?
Fifty serves as a neutral anchor that respondents select when indifferent or unsure, producing a spike. Round and culturally salient numbers like 85 and 100 attract responses because the scale invites approximate, intuitive judgments. This heaping means the data are not smoothly continuous, so analysts should inspect the distribution and use methods robust to it rather than assuming normality.
How are thermometers used to measure affective polarization?
For each partisan, analysts subtract the warmth rating of the out-party from that of the in-party. A larger difference indicates stronger affective polarization. Aggregating these differences over time tracks the rise of partisan animosity, and decomposing them shows whether the trend is driven by warmer in-party or colder out-party feelings.
Sources
- 1.Wilcox, C., Sigelman, L., & Cook, E. (1989). Some like it hot: Individual differences in responses to group feeling thermometers. Public Opinion Quarterly, 53(2), 246-257.
- 2.Weisberg, H. F., & Miller, A. H. (1980). Evaluation of the feeling thermometer: A report to the National Election Study Board. NES Technical Report. American National Election Studies.
You have read it. What now?
Cite this page
ScholarGate. (2026, June 22). Feeling Thermometer Analysis. ScholarGate. https://scholargate.app/political-psychology/feeling-thermometer-analysis