How to Make an APA 7 Correlation Table (With Example)

Once you correlate more than two or three variables, you need a table. Software prints a full matrix with significance values and sample sizes stacked under every coefficient, and pasting that into a thesis is a quick way to get a table sent back. This guide shows the layout that works in APA 7, builds one from a small dataset you can reproduce, and lists the errors reviewers flag.

What you need before you build the table

For every variable in the matrix, pull these from your output:

1. The mean (M) and standard deviation (SD).
2. The Pearson correlation (r) with every other variable.
3. The exact p value for each correlation, so you can mark or report significance.
4. The sample size (N), and whether it is the same for every pair. With pairwise deletion it may not be.

If your variables are ordinal or clearly non-linear, use Spearman correlations instead and say so in the title or note; the layout stays the same. See the Spearman guide.

The anatomy of an APA correlation table

Number and title above the table. APA 7 puts the table number in bold on the first line ("Table 1") and the title on the line below it, in italic title case. Nothing about the table goes underneath it except the note.

Numbered variables. List the variables down the first column with a number in front of each ("1. Study hours per week"). The correlation columns are then headed only by those numbers (1, 2, 3, 4). This keeps the table narrow and avoids repeating long names across the top.

M and SD columns. Put the descriptives next to the variable names, before the correlation columns, so the reader sees each variable's scale before its relationships.

Lower triangle only. A correlation matrix is symmetric: the correlation of 2 with 3 is the same as 3 with 2. Fill in the cells below the diagonal and leave the cells above it empty. Printing both halves doubles the numbers without adding any information.

The diagonal. Each diagonal cell is a variable correlated with itself, which is always 1, so it carries no information. Most tables mark it with a dash or leave it blank. In this guide we use a plain hyphen. Whichever you pick, use it consistently across every table in the paper. If the last column holds nothing but a diagonal mark, you may drop it.

No leading zero on r. A correlation cannot exceed 1 in absolute value, so APA drops the leading zero: .50, not 0.50. Keep the minus sign for negative correlations. M and SD can exceed 1, so they keep theirs (0.82).

Two decimals for r. Report every coefficient to two decimals, and give M and SD the same number of decimals within a column.

Significance in the note. Mark significant correlations with asterisks and explain them in a probability note under the table. The usual convention is one asterisk for p < .05 and two for p < .01. Alternatively, drop the asterisks and give exact p values in the note.

The note. It starts with the italic word Note., gives N and anything the reader needs to read the table, then ends with the probability note. APA 7 orders notes as general, then specific, then probability.

Horizontal rules only. No vertical lines and no grid.

A worked example

Twenty students (fictional data) reported their weekly study hours, their average sleep per night, and a test anxiety score (10 to 50), and later sat an exam scored 0 to 100. The raw data, one row per student:

study,sleep,anxiety,score
15,8.0,25,76
17,7.0,26,75
12,6.5,43,69
13,5.5,25,68
10,7.5,28,84
8,8.5,20,77
5,7.0,43,57
8,6.5,31,71
11,6.5,31,63
9,7.0,22,82
11,8.0,27,52
12,7.5,33,61
5,7.0,45,54
13,8.5,24,73
2,6.5,33,61
19,8.5,23,80
15,6.5,23,75
7,6.5,28,56
10,7.5,29,70
6,7.5,33,68

Pearson correlations (two-tailed) on those 20 rows give this table:

Table 1

Means, Standard Deviations, and Correlations for Study Habits and Exam Performance

Note. N = 20. Data are fictional. Test anxiety was scored from 10 to 50 and exam score from 0 to 100.

*p < .05. **p < .01.

If you prefer exact values to asterisks, the note becomes:

Note. N = 20. Two-tailed p values: study hours and test anxiety, p = .024; study hours and exam score, p = .025; test anxiety and exam score, p = .005; all other correlations, p > .15.

Point to the table in the text and report the correlation that matters for your question:

As Table 1 shows, students with higher test anxiety scored lower on the exam, r(18) = -.60, p = .005.

boltPaste your data and get this table built for you: the Excel builder is in the Rescue Vault, $27 →

The degrees of freedom are N minus 2. The Pearson correlation guide covers the single-correlation sentence in detail, including its confidence interval.

Common errors in correlation tables

These are the errors reviewers flag most often in theses and manuscripts:

• Pasting the software matrix. SPSS-style output with "Sig. (2-tailed)" and N under every coefficient is not an APA table. Keep the coefficients; move N and significance to the note.
• Filling both triangles. Every coefficient then appears twice. Keep the lower triangle.
• A leading zero on r. 0.50 should be .50. The reverse error also happens: dropping the zero from an SD of 0.82.
• Asterisks with no probability note, or a note whose thresholds do not match the asterisks in the table.
• Missing N. If pairwise deletion left different sample sizes for different pairs, give the range in the note (for example, N = 112 to 120).
• Variable names repeated across the top. Number the rows and use the numbers as column headers.
• Title under the table, or not italic. The bold number and italic title go above.
• Mixed decimals, such as .5 next to .28. Use two decimals throughout.
• Writing p = .000 in the note. Report p < .001.
• Reading every asterisk as a finding. Six correlations means six tests. With a large matrix, some coefficients will cross .05 by chance, so interpret the pattern your hypotheses predicted and say if you corrected for multiple comparisons.

Before you report: did the correlations' assumptions hold?

Each Pearson coefficient assumes a linear relationship, two continuous variables, and no single extreme case driving the result. A table hides all of that, so look at a scatterplot for every pair you discuss in the text. The assumptions guide explains the checks, and the effect size guide explains why r doubles as its own effect size. If your table sits next to group comparisons, the Cohen's d guide covers those effect sizes.

Let KyroStat build the table for you

Retyping a matrix by hand is where transposed digits and stray leading zeros creep in, and errors in reported statistics are common. KyroStat computes the correlations on your data with real Python or R, lays them out as a lower-triangle APA 7 table with numbered variables, M and SD columns and a finished note, and gives you the code behind it. Upload your spreadsheet, and the table is done in seconds.

Frequently asked questions

Should the diagonal of a correlation table show 1.00? It can, but the value is always 1 and carries no information, so most tables use a dash or a blank instead. Pick one and use it in every table.

Do I report the upper or lower triangle? Conventionally the lower triangle. Some authors use the upper triangle for a second group or for another statistic, and then the note must say so.

Where does N go in a correlation table? In the note, for example "N = 20." If it differs across pairs, give the range.

What is a Korrelationsmatrix in APA format? Korrelationsmatrix is German for correlation matrix. Theses written in German that follow APA style use the same layout: numbered variables, M and SD, the lower triangle, and a note.

Which related guides might I need? Often these: APA table format, descriptive statistics and partial correlation.