StatFas.com

Correlogram Explained: How to Make, Read and Report a Correlation Matrix Plot

By StatFas · Updated 2026-11-09

A correlogram shows the correlation between every pair of variables as a colored matrix, so blocks of related variables stand out immediately. This guide explains how to read one, which coefficient to use, why you must adjust for multiple testing, and how to present the result as an APA7 correlation table.

Open the free Correlogram GeneratorStep-by-step tutorial (StatFas, SPSS, R, Python)

What is a correlogram?

A correlogram (corrgram, correlation matrix plot, correlation heatmap) replaces the numbers of a correlation matrix with visual marks: the colour shows direction and the size of the circle or the depth of the colour shows strength (Friendly, 2002). With ten variables there are 45 pairs; the picture lets you see the structure in seconds, where a table of 45 coefficients would take minutes.

It is used at the start of an analysis to screen variables, to find redundant measurements before regression or PCA, and in papers to present a full correlation matrix.

How to read a correlogram

As a rough guide to size, |r| near .10, .30 and .50 are often called small, medium and large (Cohen, 1988), but interpretation depends on your field.

Pearson, Spearman or Kendall?

Pearson measures linear association and suits roughly normal, continuous variables without extreme outliers. Spearman (rho) works on ranks and suits skewed data, outliers and ordinal variables. Kendall (tau) also uses ranks and is preferred for small samples with many ties. The Correlogram Generator lets you choose and includes a distribution check (normality, skewness) so you can justify the choice. Read more in Pearson, Spearman and Kendall explained and normality tests explained.

Adjust p-values for multiple testing

With k variables you run k(k − 1)/2 tests. With 10 variables that is 45 tests, so around two significant results are expected by chance alone at α = .05. Two common remedies are the Holm procedure (Holm, 1979), which controls the family-wise error rate and is uniformly more powerful than the Bonferroni correction, and the Benjamini–Hochberg false discovery rate (1995), which is less strict and common in genomics. The generator reports both the raw and the adjusted p-values and the number of pairs tested.

Also report 95% confidence intervals: they show how precisely each correlation is estimated, which a p-value does not.

Step by step: make a correlogram in StatFas

  1. Open the Correlogram Generator.
  2. Load the example data, paste a table, or upload a CSV, Excel or SPSS file with three or more numeric columns. Missing values are handled by pairwise deletion.
  3. Choose the coefficient (Pearson, Spearman or Kendall tau-b), the p-value adjustment (Holm, Benjamini–Hochberg, Bonferroni or none), the significance level, whether to order variables by clustering, how to draw the upper and lower triangles (circles, squares, ellipses, numbers or blank) and how to treat non-significant cells (cross, show or hide).
  4. Run the analysis to get the matrix, confidence intervals, adjusted p-values, descriptive statistics and the APA7 table.
  5. Style the figure, then log in (free) to download it, the tables and a Word report.

The tutorial repeats the plot in SPSS, R, Python and Excel.

How to report it in APA7

An APA7 correlation table lists variables down the side, numbers across the top, correlations in the cells (without a leading zero), means and standard deviations, and a note defining significance and the correction. In the text, report the key pairs with r, degrees of freedom, 95% CI and the adjusted p. Example wording (illustrative numbers):

Pearson correlations among ten variables are shown in Figure 1. After Holm adjustment, interleukin-6 correlated positively with tumour necrosis factor, r(88) = .69, 95% CI [.56, .79], padj < .001, and with C-reactive protein, r(88) = .73, 95% CI [.62, .82], padj < .001. Nine of 45 correlations were significant after adjustment.

Common mistakes to avoid

Frequently asked questions

What is the difference between a correlogram and a heatmap?

A correlogram is a heatmap specialised for a correlation matrix, often with circles or significance marks. See also the heatmap maker for general matrices.

Which p-value correction should I use?

Holm controls the chance of any false positive and is a safe default. Benjamini–Hochberg (FDR) controls the proportion of false positives and suits exploratory screening with many variables.

How should I handle missing values?

Pairwise deletion uses all available data for each pair, but different pairs may then use different n. Report the range of n, or impute or remove incomplete cases if you need a single n.

Can I make a correlogram for free?

Yes. The StatFas generator is free to run in your browser; a free account is needed only to download results.

Try it now. Run the analysis in your browser with the free Correlogram Generator. No account is needed to analyze; a free StatFas account is needed only to download figures, tables and the Word report.

Related tools, guides and tutorials

References

  1. American Psychological Association. (2020). Publication manual of the American Psychological Association (7th ed.).
  2. Benjamini, Y., & Hochberg, Y. (1995). Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society B, 57(1), 289–300.
  3. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Lawrence Erlbaum Associates.
  4. Friendly, M. (2002). Corrgrams: Exploratory displays for correlation matrices. The American Statistician, 56(4), 316–324.
  5. Holm, S. (1979). A simple sequentially rejective multiple test procedure. Scandinavian Journal of Statistics, 6(2), 65–70.