Reporting Cross-Sectional and Survey Studies

Overview

Transparent reporting allows readers to understand what population was studied, how participants were selected, what was measured, how non-response and missing data were handled, and what the estimates mean. Reporting guidelines do not repair weak methods, but they expose decisions needed for appraisal and reproducibility.

Cross-sectional observational studies are commonly reported using the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement. Survey studies can also use the Consensus-Based Checklist for Reporting of Survey Studies (CROSS), and web surveys may use the Checklist for Reporting Results of Internet E-Surveys (CHERRIES).[1–4]

This lecture integrates those frameworks into a practical reporting workflow without treating checklists as substitutes for methodological judgement.

Why this matters

An article may say “a cross-sectional survey was conducted” yet omit:

• target population and sampling frame;
• dates and administration mode;
• questionnaire development;
• numbers eligible, invited, responding and analysed;
• response-rate definition;
• outcome denominator;
• weighting and survey design;
• missing-data handling;
• instrument scoring;
• ethical approval and consent; or
• the actual questionnaire.

Without these details, readers cannot assess selection, measurement, applicability or reproducibility.

Key concepts

Name the design accurately

Use a recognised design term in the title or abstract where possible: “cross-sectional study”, “population health survey”, “online staff survey” or “ecological study”. Avoid inflated labels such as prospective cross-sectional cohort unless the actual structure is explained.

Report what was planned and what occurred

The manuscript should distinguish protocol decisions from deviations. If sampling changed, recruitment was extended or scoring rules were modified, state when, why and how this may affect interpretation.

A reporting checklist is a minimum disclosure aid

STROBE covers title, abstract, introduction, methods, results, discussion and funding for observational studies.[1,2] CROSS adds survey-specific detail, and CHERRIES focuses on web-survey implementation and response metrics.[3,4]

Planning reporting before data collection

Reporting is easier when documentation begins during protocol development. Create a reporting map with:

• the applicable guideline;
• protocol section containing each item;
• data needed for the participant flow;
• tables and figures planned;
• instrument and codebook files;
• analysis outputs needed; and
• repository or supplement locations.

DECISION GUIDE

Selecting reporting guidance

• Use STROBE cross-sectional for observational cross-sectional studies.
• Use CROSS for survey studies, including web and non-web surveys.
• Add CHERRIES for internet e-surveys where its web-specific items apply.
• Add RECORD when routinely collected health data are central.
• Add topic-specific extensions when relevant, such as nutritional epidemiology or routinely collected data.
• Explain the design fully even when no checklist perfectly matches it.

Reporting the title, abstract and introduction

The title should identify the design without claiming causality. The abstract should summarise setting, participants, sampling, measurements, analysis, main estimates with uncertainty and principal limitations.

The introduction should establish the problem and evidence gap, then state specific objectives. Avoid presenting hypotheses discovered after analysis as though they were prespecified.

Reporting the Methods

PROTOCOL CHECKLIST

Essential Methods disclosures

• Design, setting, locations and field dates.
• Target and source populations, eligibility and sampling frame.
• Sampling or recruitment method and sample-size rationale.
• Invitation, reminders, consent and administration mode.
• Variables, operational definitions, instruments and data sources.
• Instrument development, translation, pretesting and scoring.
• Bias-reduction and quality-assurance procedures.
• Statistical methods, weights, strata, clusters and missing data.
• Ethical approval, privacy and data-governance arrangements.
• Protocol registration, deviations and availability of materials.

For analytical cross-sectional work, specify the estimand and covariate rationale. “Variables with p < 0.20 were entered” is not an adequate causal framework. For online surveys, report whether access was open or closed, how duplicates were prevented and how participation and completion metrics were calculated.[4]

From vague to reproducible Methods reporting

illustrative

Table 1. Reporting improvements

Vague statement	Information required
“Women with gestational diabetes were surveyed.”	Population, setting, delivery dates, eligibility, frame and sampling
“A validated questionnaire was used.”	Instrument name, language, version, evidence for intended use and scoring
“Reminders were sent.”	Number, timing, mode and whether reminders stopped after completion
“Logistic regression identified predictors.”	Outcome, estimand, model, covariates, reference groups, survey design and missing data
“The response rate was 72%.”	Numerator, denominator and named response-rate definition

Note: The table is illustrative.

Reporting instruments, files and reproducibility

Readers need access to the exact questionnaire version used, not only a citation to an earlier instrument. Provide the item wording, response options, routing, language, scoring rules and changes made during adaptation. When licensing prevents full reproduction, identify the instrument and version, describe permissions and provide the study-specific items and scoring details that can legally be shared.

A reproducibility package may contain:

• protocol and amendments;
• completed reporting checklists with page references;
• questionnaire or interview schedule;
• data dictionary and derived-variable rules;
• weighting specification;
• statistical analysis plan;
• code used to clean and analyse data;
• synthetic or de-identified example data; and
• a README describing software and execution order.

The survey literature has long emphasised complete reporting of design, conduct and analysis.[5] Standard disposition categories and outcome-rate definitions improve comparability across surveys.[6]

Designing the main tables and figures

A typical cross-sectional report may include:

1. participant-flow diagram;
2. table of sample characteristics with missing denominators;
3. primary prevalence estimates with confidence intervals;
4. main crude and adjusted contrasts;
5. sensitivity analyses; and
6. a figure showing subgroup estimates or survey flow where it adds information.

A “Table 1” should describe the sample rather than perform many significance tests between naturally occurring groups. Differences relevant to confounding should be judged using subject-matter knowledge and magnitude, not a collection of baseline p-values.

For complex surveys, table notes should state whether estimates are weighted, how standard errors were calculated and what unweighted counts support each estimate. Suppress or combine small cells according to governance rules without hiding analytically important missingness.

Reporting participant flow and response

Provide counts for:

• records or people assessed for eligibility;
• eligible;
• sampled or invited;
• undeliverable or unreachable;
• refused;
• partial responses;
• complete responses; and
• analysed for each main outcome.

When eligibility is unknown for non-contacts, state the outcome-rate definition and assumptions. A flow diagram is helpful when the pathway is complex. Report reasons for exclusion and the timing of exclusions.

Reporting Results

Begin with the participant flow, field period and sample characteristics. Provide unweighted counts and appropriate weighted estimates. Denominators should accompany all percentages when missingness varies.

Primary results should show:

• prevalence or mean estimates;
• confidence intervals;
• group contrasts and direction;
• adjusted estimates with named covariates;
• missing-data and sensitivity analyses; and
• survey-design or clustering information.

Avoid leading with p-values. Do not use “trend” for non-significant results or “independent predictor” without a clear predictive or causal framework.

Reporting Discussion and limitations

The discussion should answer the objective, compare findings with evidence and explain plausible mechanisms without exceeding the design. Limitations should describe the likely direction or consequence where possible.

For a cross-sectional study, discuss:

• temporality and reverse causation;
• prevalence-incidence or survivor bias;
• coverage and non-response;
• measurement and recall;
• residual confounding;
• missing data;
• setting-specific applicability; and
• any changes from the protocol.

APPRAISAL CHECKLIST

Final reporting audit

• Can the target population and participant pathway be reconstructed?
• Are numerators, denominators, dates and definitions explicit?
• Can readers understand and reproduce instrument scoring and exclusions?
• Are survey weights, clustering and missing data described?
• Do conclusions match cross-sectional temporality and the sampling basis?
• Are the questionnaire, protocol and analysis materials available where permitted?

Reporting null and uncertain findings

A confidence interval that includes no association does not prove equivalence or absence of a meaningful difference. Report the estimate and the range of values compatible with the data, then discuss whether the study was precise enough for the intended decision. Avoid rewriting the objective around whichever secondary result reached a conventional significance threshold.

Abstracts should preserve the same caution as the main text. Because many readers see only the abstract, it should identify the cross-sectional design and avoid causal verbs such as “led to”, “determined” or “reduced” unless the design and assumptions genuinely support them.

Ethical and governance considerations

Interface with statistical reporting

Statistical Methods should name the estimand, model and uncertainty procedure—not merely software. If weights are used, explain how they were constructed and whether variance estimation incorporated strata and clusters. Report software and version as implementation details after the method.

Tables should include reference categories, units, missing denominators and confidence intervals. Distinguish crude and adjusted estimates. Do not report more decimal places than the measurement and sample support.

Sharing survey instruments and analysis materialsMMed extension

A supplementary questionnaire, data dictionary, scoring syntax and analysis code improve reproducibility. Remove licensed content or obtain permission when required. If participant-level data cannot be shared, provide metadata, synthetic examples or a controlled-access process and explain the restriction.

Journal word limits do not justify omission of core methods. Use supplements and repositories for detailed instruments, codebooks, weighting specifications and analysis code, while keeping enough information in the main report for readers to understand the design and principal sources of bias.

Any substantive departure from the registered protocol should be identified and explained.

Worked example

WORKED EXAMPLE

Rewriting the postpartum study report

The revised report identifies a cross-sectional study of women who delivered after gestational diabetes at Hospital A in 2026. It describes the maternity-database frame, random sampling, eligibility, bilingual questionnaire development, record linkage, invitation sequence and final disposition.

The Results section shows 620 eligible women, 400 sampled, 18 undeliverable invitations, 27 refusals, 43 partial and 286 complete responses. Outcome-specific denominators are reported. The primary estimate is testing completion with a confidence interval. Reminder associations are reported as prevalence ratios with prespecified covariates, and current barriers are described as contemporaneous correlates. The questionnaire and codebook are provided as supplements.

Suggested protocol or reporting wording

Common mistakes

Knowledge check

Which guideline is the main starting point for an observational cross-sectional study?

Answer and reasoning

The STROBE cross-sectional checklist is the main starting point. Survey-specific guidance such as CROSS and web-specific guidance such as CHERRIES may be added according to the study procedures.

Why is ‘response rate 70%’ insufficient reporting?

Answer and reasoning

The reader needs the final disposition counts and the definition used, including how partial responses, ineligible cases and unknown eligibility were treated.

Should a manuscript report weighted percentages only?

Answer and reasoning

No. Weighted estimates describe the represented population, while unweighted counts show how many observations support each result. Both should be available where applicable.

Does adherence to STROBE prove that a cross-sectional study is valid?

Answer and reasoning

No. STROBE promotes transparent reporting. Validity depends on the actual design, sampling, measurement, conduct and analysis, which the reporting allows readers to appraise.

Key take-home messages

SUMMARY

Key take-home messages

• Use reporting guidance from protocol development, not only at submission.
• STROBE, CROSS and CHERRIES address overlapping but different reporting needs.
• Report population, sampling, fieldwork, instrument, response flow and analysis in reproducible detail.
• Present denominators, estimates and uncertainty before isolated p-values.
• Transparent reporting cannot repair bias, but it makes bias and applicability assessable.

References

1. von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP; STROBE Initiative. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. PLoS Med. 2007;4(10):e296. https://doi.org/10.1371/journal.pmed.0040296
2. Vandenbroucke JP, von Elm E, Altman DG, Gøtzsche PC, Mulrow CD, Pocock SJ, et al. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): explanation and elaboration. PLoS Med. 2007;4(10):e297. https://doi.org/10.1371/journal.pmed.0040297
3. Sharma A, Minh Duc NT, Luu Lam Thang T, Nam NH, Ng SJ, Abbas KS, et al. A consensus-based checklist for reporting of survey studies (CROSS). J Gen Intern Med. 2021;36(10):3179–3187. https://doi.org/10.1007/s11606-021-06737-1
4. Eysenbach G. Improving the quality of Web surveys: the Checklist for Reporting Results of Internet E-Surveys (CHERRIES). J Med Internet Res. 2004;6(3):e34. https://doi.org/10.2196/jmir.6.3.e34
5. Kelley K, Clark B, Brown V, Sitzia J. Good practice in the conduct and reporting of survey research. Int J Qual Health Care. 2003;15(3):261–266. https://doi.org/10.1093/intqhc/mzg031
6. American Association for Public Opinion Research. Standard definitions: final dispositions of case codes and outcome rates for surveys. 10th ed. Alexandria: AAPOR; 2023. https://aapor.org/wp-content/uploads/2023/05/Standards-Definitions-10th-edition.pdf