APA 7 Statistical Tables: Examples for t Tests, ANOVA, Correlation, and Regression
A good statistical table does more than arrange numbers. It lets a reader understand the comparison, estimate, uncertainty, and statistical conclusion without reconstructing the analysis from raw software output.
This guide shows how to turn common statistical results into concise APA 7-style tables. The examples are templates rather than universal journal rules, so always check the instructions for your target journal, university, or supervisor.
The anatomy of a clear APA statistical table
Most research tables need five elements:
- A table number in the order the table is mentioned in the text.
- A short, descriptive title that states what is being summarized.
- Column headings that identify every statistic and unit.
- A body containing only the precision needed to interpret the results.
- A note defining abbreviations, coding, confidence intervals, and significance markers.
Avoid copying a complete software output panel into a manuscript. Diagnostic output, intermediate calculations, and duplicated test statistics usually belong in your analysis record, not in the final table.
Example: independent-samples t test
For a two-group comparison, report the group summaries alongside the inferential result. Include the mean and standard deviation for each group, the mean difference or effect size, the test statistic, degrees of freedom, confidence interval, and exact p value when practical.
| Outcome | Group A, M (SD) | Group B, M (SD) | Mean difference [95% CI] | t(df) | p | Cohen's d |
|---|---|---|---|---|---|---|
| Example score | 24.60 (5.10) | 20.90 (4.80) | 3.70 [1.02, 6.38] | 2.76 (54) | .008 | 0.74 |
Note. Values are illustrative. Positive differences indicate higher scores in Group A. CI = confidence interval.
If equal variances are not defensible, label the result as Welch's t test and report its adjusted degrees of freedom. Do not silently switch tests.
Example: one-way ANOVA
An ANOVA table should make the source of variation visible and pair the omnibus test with an effect size. If the omnibus test is followed by pairwise comparisons, report the correction method and confidence intervals for those comparisons.
| Source | SS | df | MS | F | p | η² |
|---|---|---|---|---|---|---|
| Between groups | 182.40 | 2 | 91.20 | 6.31 | .003 | .18 |
| Within groups | 824.70 | 57 | 14.47 | |||
| Total | 1,007.10 | 59 |
Note. η² = eta squared. Values are illustrative.
The table does not replace the interpretation. In the manuscript text, state the practical direction of the finding and identify which groups differ when valid post hoc comparisons are available.
Example: correlation matrix
Correlation tables work best when variables are numbered and descriptive statistics appear in the same table. Include the correlation type in the note because Pearson, Spearman, and partial correlations answer different questions.
| Variable | M | SD | 1 | 2 | 3 |
|---|---|---|---|---|---|
| 1. Study time | 6.20 | 2.10 | NA | ||
| 2. Sleep quality | 3.80 | 0.90 | .24 | NA | |
| 3. Examination score | 74.30 | 8.70 | .46** | .31* | NA |
Note. Entries are Pearson correlations. p < .05. p < .01. Values are illustrative.
Significance stars can help scanning, but confidence intervals often communicate uncertainty more effectively. Use the convention preferred by your destination journal.
Example: multiple regression
Regression tables should distinguish unstandardized coefficients from standardized coefficients. Include standard errors or confidence intervals, and describe the outcome and reference categories in the note.
| Predictor | B | SE | β | 95% CI for B | t | p |
|---|---|---|---|---|---|---|
| Intercept | 51.20 | 4.80 | NA | [41.58, 60.82] | 10.67 | < .001 |
| Study time | 2.10 | 0.58 | .39 | [0.94, 3.26] | 3.62 | .001 |
| Sleep quality | 1.70 | 0.81 | .23 | [0.08, 3.32] | 2.10 | .041 |
Note. Outcome: examination score. R² = .29, adjusted R² = .26. Values are illustrative.
For categorical predictors, name the reference category. For logistic regression, replace B-only interpretation with odds ratios and confidence intervals where appropriate.
Precision and p-value rules
Consistency matters more than displaying every digit produced by software.
- Use the same number of decimal places for comparable statistics.
- Keep enough precision to avoid changing the interpretation.
- Report exact p values when practical, commonly to three decimals.
- Report very small values as p < .001 rather than p = .000.
- Do not use a leading zero for statistics that cannot exceed 1, such as p, r, and many effect sizes.
- Use a leading zero for values that can exceed 1, such as means and standard deviations.
A publication-ready workflow
Before exporting your table, verify the analysis, assumptions, variable labels, category coding, confidence level, and effect-size definition. Then remove columns that do not help the reader answer the research question.
DataStatPro's APA table generator is designed to shorten the formatting stage. You should still review the exported table against your study protocol and destination requirements. For the broader reporting workflow, see the publication-ready statistical reports guide.
Final checklist
- Can the table be understood without reading the software output?
- Are all abbreviations and reference categories defined?
- Are confidence intervals and effect sizes included where useful?
- Does the narrative text explain the result instead of repeating every cell?
- Are the title, numbering, notes, and precision consistent across the manuscript?
When those answers are yes, the table is doing its real job: making the evidence easier to inspect.
Frequently asked questions
What should an APA 7 statistical table include?
Include a table number, concise italicized title, clear row and column labels, the statistics needed to answer the research question, and notes that define abbreviations, reference categories, and significance conventions.
Should APA tables include vertical lines?
APA-style tables generally use minimal rules and avoid vertical lines. Always check the destination journal or institution because local requirements can differ.
How many decimal places should I report?
Use consistent precision for comparable statistics and retain enough digits to preserve interpretation. Exact p values are commonly reported to three decimals, with very small values shown as p < .001.
Should I put every software-output column in the table?
No. Include the estimates, uncertainty, tests, and effect sizes needed by the reader. Remove diagnostic or redundant columns that do not serve the research question.