How to Interpret Spearman Correlation Output in SPSS (and Report It in APA 7)

SPSS reports Spearman correlations as a matrix that looks identical to the Pearson matrix, but uses ranks instead of raw scores, so it works for ordinal data and non-linear monotonic relationships. The table layout is the same: every coefficient appears twice, every variable is correlated with itself, and each cell stacks three numbers. This guide walks you through it row by row with a three-variable worked example, explains what changes when you choose Spearman over Pearson, and gives the APA 7 write-up. The reporting format alone is in the Spearman correlation APA guide.

What SPSS gives you

Analyze > Correlate > Bivariate, tick Spearman under Correlation Coefficients (and untick Pearson and Kendall's tau to avoid confusion), and click OK. SPSS produces:

1. Correlations: the matrix, with rows for Correlation Coefficient, Sig. (2-tailed), and N.

For descriptive statistics (means, SDs, medians, ranges), use Analyze > Descriptive Statistics > Frequencies or Descriptives.

The worked example

Forty team members rated their job satisfaction and engagement (both Likert 1–5 scales, ordinal) and reported their experience in months. All three variables went into one Bivariate Correlations run with Spearman ticked. Satisfaction: M = 3.35, SD = 1.189; Engagement: M = 3.90, SD = .900; Experience: M = 5.40 months, SD = 3.357.

Table 1: Correlations

How the grid is built. Each variable has a row block and a column. The row block shows three rows: Correlation Coefficient, Sig. (2-tailed), and N. The cell where a row meets a column holds the value for that pair.

• Correlation Coefficient: rs, from −1 to 1. Satisfaction and engagement: rs = .436, a moderate positive correlation. Satisfaction and experience: rs = .314, a smaller positive correlation (moderate by Cohen's benchmarks of .10, .30 and .50).
• Sig. (2-tailed): the two-sided p value for that rs. Satisfaction and engagement: .005, which is significant at the p < .01 level.
• N: the number of people with values on both variables.

The diagonal is always 1.000. Every variable correlates perfectly with itself. Its Sig. cell is blank. Ignore it.

The matrix is symmetric. The cells above and below the diagonal are the same numbers: Satisfaction × Engagement (.436) and Engagement × Satisfaction (.436) are one correlation. Three variables give three correlations, not six.

The stars. With Flag significant correlations ticked (the default), SPSS adds ** to rs when p < .01 and * when p < .05, and explains them in footnotes: "Correlation is significant at the 0.01 level (2-tailed)" and "Correlation is significant at the 0.05 level (2-tailed)". In APA text you report the exact p instead; stars belong only in a correlation table, with a note explaining them.

N when data are missing. By default SPSS uses pairwise deletion: each correlation uses everyone with both values, so N can differ from cell to cell. Report the N (or df) of each correlation, or switch to listwise deletion under Options so every cell uses the same people.

Reading the three pairs:

• Satisfaction and engagement: rs = .436, Sig. = .005. Moderate positive, highly significant.
• Satisfaction and experience: rs = .314, Sig. = .048. Moderate positive (by Cohen's benchmarks), significant at the .05 level.
• Engagement and experience: rs = .094, Sig. = .566. Not significant.

Misreadings that cost marks

• Writing p = .000. It means p < .001.
• Using Pearson's r symbol. With Spearman, write rs (r subscript s), not r, so the reader knows you used ranks.
• Reporting the diagonal 1s, or each correlation twice. Read the lower (or upper) triangle only.
• Using N as the degrees of freedom. A correlation's df is N − 2: rs(38), not rs(40).
• Leading zeros. APA drops the zero for values that cannot exceed 1: rs = .44, p = .005.
• Calling the relationship linear. Spearman measures a monotonic relationship (consistently increasing or decreasing), which need not be a straight line.
• Reading the stars as the p value. ** means p < .01, not p = .01. Report the exact Sig.
• Calling a correlation an effect. "rs = .44 means engagement causes satisfaction" is not what the correlation shows. Describe an association.
• Ignoring the scatterplot. Plot the data first. Spearman ranks measure monotonic association. Ties in the original data (common in Likert scales) are handled using average ranks.

When to use Spearman instead of Pearson

Use Spearman's correlation when:

• One or both variables are ordinal (Likert-type ratings, ranks, or other rankings), or
• The relationship is monotonic but not linear (consistently increasing or decreasing, but curved), or
• You have outliers that would distort Pearson's r, since ranks are robust to extreme values.

With ordinal data and a linear relationship, Spearman and Pearson often agree; with non-linear monotonic relationships, Spearman (or Kendall's tau-b) is appropriate.

The APA 7 format template

There was a [weak/moderate/strong] [positive/negative] monotonic correlation between [variable 1] and [variable 2], rs(df) = .XX, p = .XXX.

The write-up for this example

Spearman correlations were computed between satisfaction, engagement and experience (N = 40). Satisfaction and engagement showed a moderate positive monotonic correlation, rs(38) = .44, p = .005. Satisfaction and experience showed a smaller, moderate positive monotonic correlation, rs(38) = .31, p = .048. Engagement and experience were not significantly correlated, rs(38) = .09, p = .566.

With three or more variables, a correlation table with means and SDs is often clearer than prose; the APA correlation table guide shows the layout.

Let KyroStat do the write-up for you

Already have the data? Upload your spreadsheet to KyroStat, which runs the Spearman correlations on your data, draws the scatterplots, and writes the APA sentences with the correct rs notation and the code behind them.

Also useful for the write-up (affiliate links): Proofreading for your thesis or paper (20% off with code PAP20): Paperpal

Frequently asked questions

Why does every correlation appear twice in the SPSS table? The matrix shows every variable against every other, so the cells above and below the diagonal mirror each other. Report each pair once.

What does "Correlation is significant at the 0.01 level (2-tailed)" mean? It is the footnote for the ** flag: that correlation's two-sided p is below .01. Report the exact p from the Sig. (2-tailed) row instead of the stars.

What symbol do I use for Spearman? rs, an italic r with a subscript s. Some journals also accept the Greek letter rho (ρ).

What are the degrees of freedom for a Spearman correlation? N − 2, the same as Pearson. With N = 40, report rs(38).

When the relationship is not linear, does Spearman always work? Spearman works for any monotonic relationship (consistently increasing or decreasing), linear or curved. If the relationship reverses direction (rises then falls, or falls then rises), a single correlation coefficient does not describe it well. In this case, plot the data to show the pattern, or model it with a polynomial term (for example, a quadratic term if the relationship is inverted U-shaped).

Can I use Spearman with continuous data? Yes, as long as you rank-transform the data first (or let SPSS do it). Spearman is valid for continuous data if the relationship is non-linear or if outliers are a concern.

Which related guides might I need? Often these: Pearson correlation when both variables are continuous and linear, Kendall's tau as an alternative rank-based correlation, and partial correlation to control for a third variable.