Demographic data in research is typically presented in a dedicated table (often called “Table 1”) that summarizes participant characteristics, supported by a brief narrative description in the methods or results section. Getting this right involves choosing the correct summary statistics, using inclusive and precise language, organizing your table clearly, and protecting participant privacy when sample sizes are small.
What Goes in a Demographics Table
In most scientific papers, Table 1 provides an overview of demographic and baseline variables for the study population. The core variables almost always include age, sex or gender, and race or ethnicity. Depending on your field, you may also report body mass index, income or socioeconomic status, education level, marital status, geographic region, relevant medical conditions, or prior treatments.
When your study compares groups (treatment vs. control, for example), present these variables stratified by group in separate columns. This lets readers quickly judge whether the groups were balanced at baseline or whether differences might influence your results. A final column can show the overall sample, and some researchers include a column for the p-value of between-group comparisons, though this practice varies by journal.
Choosing the Right Summary Statistics
The type of variable determines how you summarize it. For continuous variables like age or income, check whether the distribution is roughly normal. If it is, report the mean and standard deviation. In your table header, label this as “Mean (SD),” and in running text write something like “the mean age was 46.5 (SD = 3.0).” If the distribution is skewed, as income data often is, report the median and interquartile range instead. Using the mean for a skewed variable will misrepresent your sample because a few extreme values pull the average away from where most participants actually fall.
For categorical variables like race, gender, or education level, report the count and percentage for each category. A common format is “n (%)” in the column header, with entries like “42 (35.0%)” in each cell. Always note the denominator if it differs from the total sample, such as when some participants declined to answer a particular question. Reporting the number of missing responses for each variable is equally important, because patterns of missingness can reveal bias in who provided data.
Writing About Demographics in Text
Your table does the heavy lifting, but the methods or results section should include a brief narrative summary highlighting the most important characteristics. This typically runs two to four sentences and covers sample size, age range, gender breakdown, and any demographic feature especially relevant to your research question. For instance, if you are studying a health disparity, the racial or ethnic composition of your sample deserves prominent mention in the text, not just a line buried in a table.
Go beyond age and gender when the context calls for it. Social class, immigration generation, citizenship status, religious affiliation, or majority/minority group distinctions can all shape the experiences and outcomes you are studying. Deciding which categories to report often depends on your research context. Consulting with experts familiar with the population can help you identify which social categories are most meaningful to assess and disclose.
Using Inclusive and Precise Language
How you label demographic categories matters as much as which ones you include. For gender, use phrasing like “45 participants identified as women, 38 as men, and 3 as non-binary” rather than reporting “males” and “females” unless you are specifically describing biological sex. If your study collected data on both sex assigned at birth and current gender identity, report them as separate variables. The CDC recommends distinguishing between sexual orientation, gender identity, and gender expression rather than collapsing them into a single item.
For race and ethnicity, capitalize group names (Black, White, Latino, Asian) and treat them as proper nouns. The term “Latinx” has been proposed as a gender-neutral alternative but remains debated, so consider your audience. If your research is federally funded in the United States, NIH recognizes these racial and ethnic categories: American Indian or Alaska Native, Asian, Black or African American, Hispanic or Latino, Middle Eastern or North African, and Native Hawaiian or Pacific Islander. Your study may need to align with these groupings for compliance, but you can always collect more granular data and report both.
Disaggregating broad categories is increasingly expected. Lumping all Asian participants together, for example, can mask enormous differences in health outcomes, socioeconomic status, and cultural background between people of Chinese, Indian, Filipino, or Vietnamese descent. Whenever your sample size allows, break broad groups into more specific subgroups. Report intersecting categories too: the experience of a young Black woman in your study may differ substantially from that of an older Black man, and collapsing both into “Black participants” hides that variation.
For studies conducted outside the United States, predetermined American labels may not fit. Consider letting participants describe themselves in their own terms and reporting those descriptions, with an explanation of the local context for international readers unfamiliar with it.
Visualizing Demographic Distributions
Tables are the default, but certain visualizations can make demographic patterns easier to grasp, especially in presentations or supplementary materials.
- Bar charts work well for categorical variables like race, education level, or employment status, where you want readers to compare the relative size of each group at a glance.
- Population pyramids are one of the most effective ways to visualize age and sex structure simultaneously. They display age groups on the vertical axis with bars extending left and right for each sex, making disparities immediately visible.
- Line charts suit continuous data tracked over time, such as poverty rates or population shifts across survey waves.
- Pictograms use small icons sized proportionally to represent counts, which can be effective for general audiences. For example, showing one figure per 10,000 people to illustrate a sex ratio disparity.
- Maps are useful when geographic variation in a demographic variable is central to your research, such as showing the percentage change in an aging population across states.
For most journal manuscripts, a well-constructed table remains the standard. Reserve visual formats for situations where the distribution pattern itself is a key finding or where you need to communicate with a non-specialist audience.
Protecting Privacy in Small Samples
When your sample is small, reporting exact demographic values can risk identifying individual participants, particularly in studies of rare conditions or specific communities. Several techniques can reduce this risk without stripping your data of meaning.
Generalization transforms specific values into broader ranges. Instead of reporting a participant’s exact age, you report a five-year age group (30 to 34, 35 to 39). Suppression goes further by removing or masking cells in your table when a category contains very few people. If only two participants identified as Native Hawaiian or Pacific Islander, you might combine smaller racial categories into an “Other” group, though this should be done carefully and with a note explaining why.
A more formal approach is the k-anonymity principle, which ensures that every combination of reported characteristics matches at least “k” individuals in the dataset. If k is set to 5, no disclosed record can correspond to fewer than five people, making it much harder to trace data back to a single person. For federally regulated health data, HIPAA’s Expert Determination method requires a qualified analyst to certify that the risk of re-identification is very small before the data can be shared.
In practice, the simplest steps are often enough for standard research papers: use age ranges instead of exact ages, combine rare categories, and avoid publishing unusual combinations of traits (such as a specific age, rare occupation, and small-town residence together) that could function as a fingerprint.
Explaining Sampling Decisions
Transparent reporting means going beyond what your sample looks like to explain how it got that way. If certain populations were excluded, state the reason. If you made deliberate efforts to diversify recruitment, such as using quota sampling to ensure enough participants from underrepresented groups or selecting from a larger pool based on diversity criteria, describe those efforts in your methods section.
Including representative population statistics alongside your sample demographics gives readers a benchmark. If your study sample is 82% White but the surrounding community is 60% White, that gap is worth flagging. It helps readers judge the generalizability of your findings without having to look up comparison data themselves. APA’s reporting standards for race, ethnicity, and culture encourage authors to address generalizability directly, reflecting on how the sample’s composition may shape what the results can and cannot tell us.