How to Report Descriptive Statistics: Tables and Examples

Report the valid sample size, counts and percentages for categories, and a suitable centre and spread for numerical variables. Use mean and SD when they serve your reporting purpose; consider median and IQR when skew or extreme observations make them more informative. State units, scale direction and missing-data denominators.

Start with a compact table, then explain the values that matter to the research question. A descriptive statistics table describes the observations you analysed; it does not establish a causal effect or make a sample representative.

For spread, NIST explains why tail values affect SD and why IQR can be more stable. This is a reporting decision, not a rule that every continuous variable must have one particular summary.

From five observations to a reporting paragraph

Invented teaching data, not study findings: five students report weekly study hours of 2, 4, 4, 5 and 10. Three study full-time and two part-time. There are no missing values in this example.

Note. Percentages use all five observations. Study mode is categorical, so its codes are not averaged. SD describes variation among observations; it is not the standard error of the mean. The two study-mode rows describe one variable.

Reporting paragraph: “In this five-person example, weekly study time averaged 5.0 hours (SD = 3.0), with observed values from 2 to 10 hours. Three students studied full-time (60%) and two part-time (40%).”

The mean is 25/5 = 5. The squared deviations sum to 36, giving a sample variance of 36/(5−1) = 9 and SD = 3. With real data, inspect the distribution before choosing the summary. These five observations cannot establish what students generally do.

All other numerical examples below are also invented for teaching. Replace them with checked output from your own analysis.

In this guide

• What are descriptive statistics in research?
• What descriptive statistics should you report for your variables?
• How do you build a descriptive statistics table that readers can use?
• How should mean and standard deviation reporting look in a paper?
• How do descriptive statistics change across disciplines and data types?
• What mistakes do students commonly make when reporting descriptive statistics?
• How do you connect descriptive statistics to later statistical tests?
• What should you check before moving on from descriptive statistics?

What are descriptive statistics in research?

Descriptive statistics in research are numbers that summarise the observed data without making claims beyond the sample. They describe patterns such as typical values, spread, counts, percentages, and ranges. They do not test whether a relationship is statistically significant or whether one group differs from another in the wider population.

Describing before explaining

Descriptive statistics are summaries of what is present in the dataset. They answer questions such as “How many participants were in each group?”, “What was the average score?”, and “How spread out were the responses?” In a student paper, they usually appear at the start of the results section, often before inferential statistics.

Inferential statistics use sample data and assumptions to estimate or test claims about a broader population. A t-test or an inferential regression analysis can serve that purpose. A correlation coefficient can also simply describe an association in the observed data; inference requires the relevant inferential procedure and assumptions. Descriptive summaries give those analyses context.

For example, a psychology paper on stress and sleep quality might report the mean stress score, standard deviation, minimum, maximum, and sample size before testing whether stress predicts sleep quality. Without those descriptive values, the later regression result has less context.

What descriptive statistics can and cannot claim

Descriptive results can say, “Participants reported a mean sleep quality score of 6.2 on a 10-point scale.” They cannot say, by themselves, “Stress causes poor sleep” or “students generally have poor sleep quality.” The first statement describes the sample; the second and third move into explanation or generalisation.

A useful way to think about descriptive statistics is that they make your dataset visible. They show whether the sample is balanced, whether scale responses cluster tightly or vary widely, and whether any values look unusual. If you have not yet defined your variables clearly, start with the measurement logic first; the guide to defining variables in quantitative research helps connect constructs to measurable indicators before analysis.

What descriptive statistics should you report for your variables?

Match the summary to the measurement level, distribution and question. Counts and percentages describe categories. Mean and SD or median and IQR describe different aspects of numerical observations. A single ordinal item and a multi-item composite need separate consideration; do not treat every variable with numeric codes as continuous.

Match the statistic to the variable type

Categorical variables place observations into groups, such as gender category, study mode, employment status, programme type, or treatment group. Report the frequency, which is the number of cases in each category, and the percentage, which is the proportion of the sample in that category.

Numerical measurements include age, income and reaction time; some other numerical variables, such as test scores or event counts, are discrete. Where appropriate, report the mean, the arithmetic average, and standard deviation, a measure of spread in the original units. Check whether the median and IQR better communicate the distribution. Minimum and maximum show the observed limits, not how values are distributed between them.

Ordinal variables have ordered categories, but the distance between categories may not be equal. Likert-type items are a common example. Some courses allow mean and standard deviation for summed or averaged Likert scales, especially when several items form a scale; single ordinal items are often clearer with frequencies, percentages, medians, or category distributions.

Common reporting choices by variable

The guiding question is not “How many numbers can I report?” but “Which numbers help the reader understand the variable?” If your assignment brief asks for specific outputs, convert that requirement into a plan rather than copying every software table; the guide on turning assignment brief requirements into a paper plan is useful at that stage.

What to include for sample characteristics

Sample characteristics usually describe the participants, cases, documents, organisations, or observations used in the study. In undergraduate and master’s papers, these may include age, gender category, study level, employment status, health condition, school type, company size, or other context variables.

Do not treat every demographic variable as equally central. A paper on workplace burnout may need role type, weekly hours, and industry more than it needs every background category collected in the survey. A nursing paper on medication adherence among elderly patients discharged to home care may report age, living arrangement, number of prescribed medications, and support at home because those features help readers understand adherence behaviour.

How do you build a descriptive statistics table that readers can use?

A descriptive statistics table should group related variables, use consistent decimal places, name each statistic clearly, and avoid unnecessary software output. Readers should be able to see the sample size, central tendency, spread, and category distribution without decoding abbreviations. A good table supports the text rather than replacing it.

Table structure that works in student papers

A basic descriptive statistics table usually has one row per variable and columns for the statistics that fit those variables. For continuous variables, common columns are N, M, SD, minimum, and maximum. For categorical variables, use category rows with n and %. Mixing both types in one table is acceptable if the layout is readable.

A simple structure might look like this:

Note. n = valid observations; M = mean; SD = standard deviation; Min and Max = observed limits. These are invented examples, not the five-person dataset above. Choose this format only where mean and SD suit the data.

For categorical variables, a separate table may be cleaner:

A descriptive statistics table becomes confusing when it mixes raw software labels, unexplained codes, and too many decimal places. If the variable was coded as 1 = full-time and 2 = part-time, report the category names, not the numeric code mean.

A practical table-building process

Use a short process before writing the table into your paper:

1. List every variable used in the research question, hypotheses, or sample description.
2. Mark each variable as categorical, ordinal, continuous, or scale-based.
3. Choose counts and percentages for categories.
4. Inspect numerical distributions and justify mean/SD, median/IQR or another appropriate summary.
5. Check whether missing data means N differs across variables.
6. Round values consistently, usually to one or two decimal places.
7. Add a note if higher scores, scale ranges, or coding decisions need explanation.

Table notes and readability

Table notes are not decorative. They prevent misreading. If a scale runs from 1 to 5, state what higher values mean. If percentages are based on valid responses rather than the full sample, say so. If some variables have lower N because of missing responses, do not hide that difference.

A concise note might read: “Higher academic self-efficacy scores indicate greater confidence in completing academic tasks; possible scores range from 1 to 5.” That sentence gives the reader enough information to interpret the mean without searching the methodology chapter.

How should mean and standard deviation reporting look in a paper?

Name the variable, give the mean and SD consistently, and explain units or scale direction. For an invented example: “Academic self-efficacy scores averaged 3.71 (SD = 0.62) on a 1–5 scale, where higher scores indicated greater self-efficacy.” A label such as “moderate” requires a justified interpretation rule; the scale endpoints alone do not establish it.

What the mean and SD tell the reader

The mean is the average score. It gives a quick sense of the typical value in the sample, especially for approximately continuous or scale-based variables. The standard deviation shows spread: a small SD means scores cluster near the mean, while a large SD means responses vary more widely.

In an invented satisfaction example, two groups might have the same mean of 5.8 on a 1–7 scale but SDs of 0.6 and 1.9. Their means match, while the second has greater spread by that measure. Inspect the distributions and the instrument's scoring guidance before labelling either group's satisfaction “high” or “positive”.

Mean and standard deviation reporting also needs the scale range. “M = 3.4” means little unless the reader knows whether the scale runs from 1 to 4, 1 to 5, 1 to 7, or 0 to 100.

Weak versus stronger reporting

These rewrites describe the supplied summaries without inferring a category majority, a scale cut-off or a population pattern.

Rounding and formatting

Most student papers use one or two decimal places for means and standard deviations. More precision rarely improves interpretation. If your survey scale has whole-number response options from 1 to 5, reporting M = 3.916472 adds noise rather than accuracy.

Use the same format throughout the paper. Do not switch between “mean,” “Mean,” “M,” and “average” without reason. Many style guides accept italic statistical symbols, but formatting expectations vary by institution, so check your course handbook.

How do descriptive statistics change across disciplines and data types?

Descriptive statistics change because fields measure different kinds of variables and answer different kinds of questions. Psychology often reports scale means and reliability-related context; health sciences may report clinical ranges, patient characteristics, and adherence categories; education and business papers often combine survey scales with group characteristics. The statistics must fit both the data type and the research aim.

Social sciences and psychology example

For a psychology assignment, distinguish a single questionnaire item from a scored multi-item scale. State the instrument's possible range and direction; apply descriptive labels only when its documented scoring guidance supports them. A summary does not establish a diagnosis or a causal effect.

Health sciences or nursing example

A health-sciences assignment may need age, medication counts and categorical support measures. Choose summaries for each variable separately, and use clinical categories only when a specified instrument or protocol defines them. Descriptive differences alone do not demonstrate intervention effectiveness.

Education and business examples

Education and business assignments can mix numerical scores with course, role or department categories. Use a readable combination of numerical summaries and category counts. A department coded 1, 2 or 3 is still a category; averaging those codes does not describe department membership.

What mistakes do students commonly make when reporting descriptive statistics?

Students often make mistakes by treating all variables as if they need the same statistics, copying raw software output, or interpreting descriptive values as proof of a hypothesis. The fix is to match each statistic to the variable type, explain scale direction, and keep claims descriptive until inferential tests are reported. Clear reporting prevents readers from questioning the basic credibility of the results section.

Mistakes that distort the data

1. Averaging category codes
   Student example: “The mean gender was 1.42, showing that most participants were female.”
   Correction: Report category counts and percentages instead: “Female: 70 (58.3%); male: 45 (37.5%); non-binary/prefer to self-describe: 5 (4.2%),” if those were the actual categories used.

2. Reporting a mean without the scale direction
   Student example: “The mean attitude score was 4.1, which is good.”
   Correction: State the scale range and what higher scores mean: “Attitude scores averaged 4.1 (SD = 0.7) on a 1–5 scale, where higher scores indicated more positive attitudes.”

3. Claiming significance from descriptive differences
   Student example: “Students who worked part-time had lower grades, so employment negatively affected performance.”
   Correction: Describe the observed pattern first, then use an appropriate test if your design supports it: “The part-time employed group had a lower mean grade than the non-employed group; an independent-samples t-test was used to assess whether the difference was statistically significant.”

4. Copying too many software decimals
   Student example: “The average satisfaction score was 3.843721 and the standard deviation was 0.927364.”
   Correction: Round consistently: “Satisfaction scores averaged 3.84 (SD = 0.93).”

5. Ignoring missing data in the table
   Student example: “N = 150” at the top of the table, even though only 132 students answered the income question.
   Correction: Report variable-specific N where needed, or add a table note: “N varies because some participants did not answer all items.”

Mistakes that weaken the writing

A results section can be numerically correct and still hard to read. Students often write a sentence for every row in the descriptive statistics table, which creates repetition. The text should pull out patterns that matter for the research question, while the table holds the full set of values.

Another common problem is evaluative wording. Words such as “good,” “bad,” “excellent,” or “poor” need criteria. If you write “the mean score was good,” the reader may ask, “Good compared with what?” Use neutral description unless the scale manual, course framework, or prior literature gives a threshold.

Before and after revision

Weak: “The descriptive statistics show that students used social media a lot and were anxious. The mean was 4.6 and the standard deviation was 1.2.”

Stronger: “Students reported an average of 4.6 hours of social media use per day (SD = 1.2). Anxiety scores averaged 18.4 (SD = 5.0) on a 7–28 scale, where higher scores indicated more frequent anxiety symptoms.”

The stronger version separates the two variables, gives units, reports scale direction, and avoids interpreting the pattern as causal.

How do you connect descriptive statistics to later statistical tests?

Descriptive statistics connect to later tests by showing whether the data are suitable for the planned comparison, association, or model. They help readers see group sizes, variable spread, possible floor or ceiling effects, and missing data before inferential results appear. This connection makes the results section feel planned rather than assembled from software output.

Use descriptives as a bridge, not a substitute

If your research question asks whether study hours predict exam score, descriptive statistics first show the mean and spread for both variables. The later correlation or regression then tests the association. If your research question compares two teaching methods, descriptive statistics first show group sizes, mean scores, and standard deviations for each group. The later t-test or ANOVA evaluates whether the observed difference is unlikely to be due to sampling variation.

Descriptive statistics also help justify why a test was chosen. Unequal group sizes, skewed variables, or ordinal outcomes may affect the analysis plan. If you are unsure which test matches your variables and question, the guide to choosing a statistical test for student research gives a decision structure that connects measurement level, groups, and hypotheses.

Connect to hypotheses without overclaiming

Suppose a master’s education paper tests the hypothesis that students receiving weekly formative feedback report higher engagement than students receiving feedback only at midterm. A descriptive paragraph might state that the weekly-feedback group had a higher mean engagement score. The inferential test then evaluates whether that difference is statistically supported.

The sequence matters:

1. Describe the sample and variables.
2. Report group or variable descriptives.
3. Identify the observed pattern.
4. Report the inferential test.
5. Interpret the result in relation to the research question.

This order prevents a common problem: announcing that a hypothesis is supported before giving the reader the descriptive pattern or test result. For quantitative projects, the descriptive section is the first evidence layer, not the final argument.

What should you check before moving on from descriptive statistics?

Before leaving descriptive statistics, check that every reported number has a purpose, matches the variable type, and can be understood without opening your raw dataset. Your table and text should agree, use consistent rounding, and explain any scale or coding decisions that affect interpretation. If the reader can see who was studied, what was measured, and how the values are distributed, the section is doing its job.

Final checks for tables and text

Read your descriptive statistics table as if you had not collected the data. Can you tell what each variable means? Can you tell whether a higher score is positive, negative, more frequent, more severe, or simply larger? Can you tell whether percentages use the full sample or only valid responses?

Then compare the table with the prose. If the table reports M = 3.71 and the text says M = 3.17, fix the inconsistency before doing anything else. Small numerical errors are easy to miss, but they damage confidence in the analysis.

Before you move on: descriptive statistics checklist

• Each variable is classified as categorical, ordinal, continuous, or scale-based.
• Categorical variables are reported with counts and percentages, not code averages.
• Continuous or scale variables include mean and standard deviation where appropriate.
• Medians or category distributions are used when the mean would mislead the reader.
• Every table has clear variable names rather than raw software labels.
• Scale ranges and score direction are explained where needed.
• N is reported clearly, especially when missing data differ by variable.
• Decimal places are consistent across similar statistics.
• The text describes key patterns instead of repeating every table row.
• Descriptive claims are not presented as proof of causation or statistical significance.
• The descriptive results connect logically to the research question, hypotheses, and planned tests.