Home Statistical Dictionary Cross-Sectional Survey
Cross-Sectional Survey
A survey design that collects data from a sample of the population at a single point in time, providing a snapshot of the population's characteristics but no direct information about how those characteristics change over time.
In Plain English
A cross-sectional survey is a single snapshot: you sample the population once, ask your questions once, and that's it, no follow-up, no repeated waves. It's the most common and generally cheapest and fastest survey design, well suited for measuring how common something is at a given moment, an opinion, a behavior, a demographic characteristic, but it has a fundamental limitation, because you only ever observe each person once, you can't directly track how any individual's situation changes over time, and you generally can't establish that one thing caused another based on a cross-sectional survey alone, since you don't know which variable came first.
Definition
A cross-sectional survey collects data from a sample of the population at a single point in time (or over a short field period treated as a single wave), producing a snapshot estimate of the prevalence of characteristics, behaviors, or opinions in the population at that moment, without repeated observation of the same individuals over time. Cross-sectional surveys are well suited for descriptive population-level estimation (prevalence rates, demographic distributions, opinion snapshots) and are generally faster and less expensive than longitudinal designs, but they cannot directly measure individual-level change over time, and any association observed between two variables in a cross-sectional survey cannot establish temporal order or rule out reverse causation, a key limitation relative to cohort studies or panel surveys, which explicitly track the same units across multiple time points and can more directly establish the sequence of events.
Properties
- Cross-sectional surveys cannot establish temporal precedence between two associated variables, a well-documented association observed in cross-sectional data, say between a behavior and an outcome, could reflect the behavior causing the outcome, the outcome causing the behavior, or some third variable driving both, the survey's single-point-in-time design provides no way to distinguish between these possibilities.
- Repeated cross-sectional surveys, the same questions fielded to a fresh, independently drawn sample at each time point, can track how population-level prevalence changes over time even though no individual is followed, this is distinct from a panel or longitudinal survey, which tracks the same specific individuals across waves and can additionally measure individual-level change.
- Cross-sectional surveys are generally the fastest and least expensive survey design to field, since there is no need to track, re-contact, or retain the same respondents over multiple waves, this practical advantage is exactly why cross-sectional designs remain the default choice for most opinion polling and prevalence-estimation surveys where individual-level change is not the research question.
At a Glance