Psychometric validity evidence for a measure of trust in U.S. public health actors for vaccine information.

Résumé

PURPOSE: Generate psychometric validity evidence for a novel 4-item measure of trust in public health actors for vaccine information for use in U.S.-population survey research.METHODS: We analyzed data from 7394 respondents in U.S.-nationally representative KFF public opinion surveys over five time points (2020-2025) regarding their level of trust in FDA, CDC, their local public health department, and their own doctor to provide reliable vaccine information. We used classical test theory to estimate internal consistency reliability and parallel analysis to evaluate dimensionality of the proposed measure. We fit an Item Response Theory graded response model to discern item and scale properties, and estimated differential test functioning (DTF) by political identity and time point. We estimated associations between scale scores and vaccine-related outcomes, and compared scale scores between groups of differing partisanship and vaccination status.RESULTS: The 4-item scale had high internal consistency (Cronbach's alpha = 0.84) and a single underlying dimension ("trust in public health actors for vaccine information"). IRT suggested items provided sufficient information (a parameters: 1.55-4.04) across a wide range of latent trust levels (b parameters: -2.62-0.69). We did not find substantively meaningful DTF by political partisanship or survey time point at the scale level. Higher scale scores were positively associated with vaccine uptake and pro-vaccine attitudes, including COVID, flu, and MMR vaccine confidence (all p