Cross-Sectional Study

A cross-sectional study is one of the three main observational study designs used in epidemiological and social-science research — alongside the cohort study and the case-control study — and it is defined by a single feature: every variable of interest, whether exposure, outcome, or covariate, is measured at one point in time, in one data-collection event, on a sample selected by criteria unrelated to exposure or outcome status. That single-timepoint structure is a design decision made at the protocol stage, and it shapes what a manuscript can legitimately claim from the resulting data — which is why reviewers and editors read the methods section closely for exactly this term.

What makes a study cross-sectional

Three conditions distinguish a cross-sectional design from the other two main observational designs:

• Single time point. Exposure and outcome (or whatever variables are being studied) are captured at the same visit, survey administration, or data extract — not measured at one time and followed up at a later time.
• Selection independent of exposure or outcome. Participants are enrolled based on membership in the population of interest (e.g., current patients at a clinic, respondents to a national survey), not because they already have a given exposure or a given outcome. This is the feature that separates a cross-sectional design from a cohort study (which selects on exposure) or a case-control study (which selects on outcome).
• No temporal sequence to exploit. Because exposure and outcome are recorded simultaneously, the design cannot show which came first. A cross-sectional study can report that two variables co-occur or are statistically associated; it cannot, by design, establish that one preceded or caused the other.

The STROBE (STrengthening the Reporting of OBservational studies in Epidemiology) reporting guideline treats cohort, case-control, and cross-sectional studies together as “the three main analytical designs … used in observational research,” and provides a dedicated reporting checklist for cross-sectional studies specifically, reflecting how established the term is as a formal design category rather than a loose descriptor.

Cross-sectional vs. longitudinal and cohort designs

The distinction most manuscripts need to get right in the methods section is cross-sectional vs. longitudinal:

• A cross-sectional study measures a sample once. It answers “what proportion of this population currently has X” or “is X associated with Y in this population right now” — questions about prevalence and association at a moment in time.
• A longitudinal study (including the cohort design) measures the same or a comparable population at two or more time points, which allows it to track change over time and, when exposure is measured before outcome, to support a stronger case for temporal sequence.

Repeated cross-sectional studies — the same question asked of a fresh, independently drawn sample at each round, as in many recurring national surveys — are sometimes confused with a true longitudinal design. They are not the same: repeated cross-sectional data can show that a population-level prevalence has changed over time, but because the same individuals are not tracked across rounds, it cannot show individual-level change or establish a within-person time sequence. A manuscript describing repeated cross-sectional survey rounds should say so explicitly rather than let “repeated” imply the individual-level tracking that only a true longitudinal or cohort design provides.

How to describe it in a methods section

A methods section that identifies a study as cross-sectional should typically state, in plain terms: (1) the single point or narrow window in which data were collected, (2) the selection criteria for the sample (and confirm they are independent of exposure/outcome status), and (3) the resulting scope of causal claim the paper will make — association and prevalence, not causation or incidence. Reviewers commonly flag manuscripts that use causal language (“X leads to Y,” “X reduces Y”) to describe findings from a cross-sectional dataset; the design supports “X was associated with Y,” not a causal claim, and stating the design explicitly and early helps set that expectation for the reader before the results are presented.

Strengths and limitations

Cross-sectional designs are comparatively fast and inexpensive to conduct because they require only one round of data collection, and they are well suited to estimating the prevalence of a condition, exposure, or attitude in a defined population, and to generating hypotheses about associations that a later cohort or experimental study could test more rigorously. Their central limitation is the one built into the definition: without a temporal sequence, they are vulnerable to reverse causation (does the exposure cause the outcome, or does the outcome change behavior in a way that produces the exposure?) and cannot separate that possibility from a genuine causal relationship using the data alone.

Related terms

See Research Study Types for the full observational-vs-experimental design taxonomy that cross-sectional, cohort, and case-control studies all sit within, and Defining the Population for how a manuscript should specify the sample a cross-sectional study’s findings apply to. For related sample-size and clinical-study reporting concepts, see Clinical Study and Non-Interventional Study.