Candidate Evaluation Model
Also known as: Impression-Driven Evaluation Model, Online Processing Model, Candidate Trait Evaluation Model
A candidate evaluation model represents how voters form overall assessments of political candidates as a latent function of perceived traits (competence, leadership, integrity, empathy), partisanship, issue proximity, and affect. It spans the trait-based factor models of Kinder et al. (1980) and the online-processing tally model of Lodge, Steenbergen and Brau (1995), which describes evaluation as a running summary updated as information arrives.
Key highlights
- Formalizes candidate evaluation as a latent impression built from interpretable trait, party, and affect inputs.
- The online-tally account explains why evaluations persist even when specific facts are forgotten.
- Latent-variable measurement separates the true impression from noisy thermometer and trait indicators.
- Connects to vote-choice models, clarifying which candidate perceptions actually move votes.
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 a candidate evaluation model to explain how voters form and update overall assessments of candidates and to identify the relative weight of traits, partisanship, issues, and emotion. It is appropriate for studying vote choice, campaign effects, and the processing mode (online versus memory-based) underlying evaluations. Use trait-factor models to decompose the bases of evaluation and the online-tally model when studying dynamic updating; measure latent impressions with multiple indicators to separate signal from measurement error.
Strengths & limitations
- Formalizes candidate evaluation as a latent impression built from interpretable trait, party, and affect inputs.
- The online-tally account explains why evaluations persist even when specific facts are forgotten.
- Latent-variable measurement separates the true impression from noisy thermometer and trait indicators.
- Connects to vote-choice models, clarifying which candidate perceptions actually move votes.
- Distinguishing online from memory-based processing empirically is difficult and requires specialized designs.
- Trait dimensions and their weights can vary across candidates, offices, and elections, limiting generalization.
- Endogeneity: partisanship colors trait perceptions, so traits and party are hard to separate causally.
- Latent models require multiple indicators and assumptions (e.g., measurement invariance) that may not hold across groups.
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
What is the difference between online and memory-based evaluation?
Memory-based evaluation assumes voters retrieve specific considerations from memory and combine them when asked to judge a candidate. Online evaluation assumes voters update a running affective tally as information arrives and store that summary, discarding the details. The key empirical signature of online processing is that overall evaluations correlate poorly with the specific facts a voter can recall, because the tally outlives the memories.
Why model candidate evaluation as a latent variable?
Observed measures like a single feeling thermometer or trait rating contain measurement error and may each tap only part of the underlying impression. A latent-variable model uses multiple indicators (thermometers, trait batteries, like/dislike counts) to estimate the common impression they share, separating the true evaluation from noise and enabling cleaner tests of what drives it.
How do traits and partisanship interact in the model?
Partisanship shapes how voters perceive a candidate's traits (co-partisans look more competent and trustworthy) and also directly colors the overall evaluation. This makes traits and party endogenous: simply regressing evaluation on traits overstates trait effects because party drives both. Models that account for this, or experiments that manipulate information, are needed to estimate the independent contribution of traits.
Sources
- 1.Lodge, M., Steenbergen, M. R., & Brau, S. (1995). The responsive voter: Campaign information and the dynamics of candidate evaluation. American Political Science Review, 89(2), 309-326.
- 2.Kinder, D. R., Peters, M. D., Abelson, R. P., & Fiske, S. T. (1980). Presidential prototypes. Political Behavior, 2(4), 315-337.
You have read it. What now?
Cite this page
ScholarGate. (2026, June 22). Candidate Evaluation Model. ScholarGate. https://scholargate.app/political-psychology/candidate-evaluation-model