Temporal Data Structure Is Not Causal Identification: A Practical Framework for Appraising Causal Claims in Prevention and Population Health Research

Temporal data structure is not an identification strategy. In manuscript review and reporting, cross-sectional and longitudinal labels are sometimes used as shorthand for judging whether causal claims are appropriate. Whether a causal claim is credible depends on the effect being considered, what makes the comparison informative, what assumptions are required, and whether the analysis and supporting evidence address the main threats. We define temporal data structure as the temporal and sampling organization of measurement and distinguish it from causal identification, which concerns whether a causal estimand can be recovered under explicit assumptions. Longitudinal data can establish measurement order and show how outcomes, exposures, and other relevant factors change over time, but these advantages do not by themselves address confounding, selection, attrition, or measurement error. Conversely, cross-sectional or repeated cross-sectional data may contribute to causal inference when temporal ordering is established by design or substantive knowledge and identification arises from credible adjustment, assignment rules, thresholds, policy timing, or other design features. Using prevention examples, we show where causal leverage can come from, explain what repeated measurement adds and what it cannot establish by itself, and propose five plain-language questions for authors, reviewers, and editors. We also distinguish limitation language appropriate for associational analyses from qualified causal interpretations made under explicit assumptions. Temporal structure informs causal appraisal, but it does not determine it.