Spearman Rank Correlation in SPSS: Run & Interpret

When your data do not behave the way Pearson’s correlation expects, you do not have to abandon the question. If you want to know whether two variables move together — as one goes up, does the other tend to go up (or down)? — but your variables are ranked, skewed, or measured on an ordinal scale, Spearman’s rank correlation is often the right tool. It is one of the most useful nonparametric tests in SPSS, and it is much easier to run than many students expect.

Table of Contents

In this tutorial, you will learn when to use Spearman’s rho, how to check its assumptions, how to run it in SPSS click by click, how to read the output, and how to report the result in APA style (7th edition). No advanced statistics background is needed.

What Is Spearman’s Rank Correlation?

Spearman’s rank correlation, written as rho (ρ) or rs, measures the strength and direction of a monotonic relationship between two variables. A monotonic relationship means that as one variable increases, the other tends to increase (a positive relationship) or decrease (a negative relationship) — but not necessarily at a constant rate, and not necessarily in a straight line.

That last point is the key difference from Pearson’s correlation. Pearson’s r measures linear relationships: how well the points fit a straight line. Spearman’s rho first converts each variable’s values to ranks, then looks at the relationship between those ranks. Because it works with ranks rather than raw scores, it is less sensitive to outliers and to skewed distributions, and it can be used when the relationship is curved but still consistently increasing or decreasing.

Spearman’s rho ranges from -1 to +1. A value near +1 means a strong positive monotonic relationship, a value near -1 means a strong negative monotonic relationship, and a value near 0 means little or no monotonic relationship. As a rough guide, many textbooks describe values around 0.10 as small, around 0.30 as medium, and around 0.50 or above as large — but always interpret the size in the context of your field, and never rely on the number alone without looking at your data.

When Should You Use Spearman Instead of Pearson?

Choose Spearman’s rho when one or more of these apply:

Your variables are ordinal. Examples include Likert-scale ratings, satisfaction scores, ranked preferences, or class positions. Pearson’s correlation assumes interval or ratio data, so it is not the natural choice for ranked data.
Your data are skewed or contain outliers. Because Spearman works with ranks, a single extreme score cannot distort the result the way it can with Pearson’s r.
The relationship looks monotonic but not linear. If a scatterplot shows scores rising together but levelling off or curving, Pearson may understate the relationship, while Spearman can still capture it.
You have paired observations. Each case contributes one score on each variable — for example, each student has both a stress rating and a sleep-quality rating. The two variables must be measured on the same cases.

Do not use Spearman if your variables are nominal categories with no order (such as gender or subject area), or if each case does not have a score on both variables. In those situations, a different test is needed.

Checking the Assumptions

Spearman’s rho has fewer assumptions than Pearson’s, but you should still check these before you interpret the result:

Both variables are ordinal, interval, or ratio. Ranked or continuous scores are fine; unordered categories are not.
The two variables are paired observations. The scores must come from the same cases.
The relationship is monotonic. This is the one assumption you should always check visually. In SPSS, go to Graphs, then Legacy Dialogs, then Scatter/Dot, and plot the two variables. You are looking for a pattern that generally rises or generally falls, even if it curves. If the points form a U-shape or show no pattern at all, Spearman’s rho will not describe the relationship well, and you should reconsider your analysis rather than reporting a misleading number.

You do not need to check normality for Spearman’s rho, and that is one of the reasons it is so popular with real-world thesis data, which are rarely perfectly normal.

Running Spearman’s Rho in SPSS: Click by Click

Suppose you have two variables in your SPSS data file: StressRating (a self-rated stress score) and SleepQuality (a sleep-quality rating). Each row is one participant. Here is the full procedure:

Open your data file in SPSS and check the Data View. Make sure both variables are numeric and that each row has a score on both variables. Missing values are handled pairwise by default, which means a case is used as long as it has both scores.
From the top menu, click Analyze, then Correlate, then Bivariate.
In the Bivariate Correlations dialog, move both variables from the left-hand list into the Variables box on the right. You can select them one at a time and click the arrow, or select both together.
Under Correlation Coefficients, uncheck Pearson if it is selected, and check Spearman. This is the step students most often forget — if Pearson is left checked, SPSS will give you Pearson’s r instead.
Under Test of Significance, leave Two-tailed selected unless you have a clear, pre-registered directional hypothesis. Two-tailed is the safe default for most theses and assignments.
Check Flag significant correlations if you want SPSS to mark significant results with asterisks in the output. This is helpful but optional.
Click OK. SPSS will produce a Correlations table in the Output Viewer.

That is the whole procedure. If you need to repeat the analysis with different variables, you can return to the same dialog, or use syntax, which is faster and leaves a record of exactly what you ran.

Paste-Ready SPSS Syntax

If you prefer syntax, or if your supervisor asks for a reproducible record of your analysis, paste the following into a syntax window and run it. Replace StressRating and SleepQuality with your own variable names:

NONPAR CORR
 /VARIABLES = StressRating SleepQuality
 /PRINT = SPEARMAN TWOTAIL NOSIG
 /MISSING = PAIRWISE.

The NONPAR CORR command tells SPSS to run a nonparametric correlation. The VARIABLES subcommand lists the two variables, PRINT = SPEARMAN requests Spearman’s rho, TWOTAIL requests a two-tailed significance test, and MISSING = PAIRWISE means each pair of variables uses all cases that have scores on both. Running the analysis from syntax gives exactly the same output as the menus, with the advantage that you can save the syntax file and re-run it later without clicking through the dialogs again.

How to Interpret the SPSS Output

The SPSS output for Spearman’s rho is a single Correlations table. It is short, but students often misread it, so work through it row by row.

First, find the cell where the row for one variable meets the column for the other. The table is symmetrical, so you will see the same value twice. The value labelled Correlation Coefficient is Spearman’s rho. Suppose your output shows rho = .62. That is a strong positive monotonic relationship: as stress ratings increase, sleep-quality ratings tend to move in the same direction in this example — always check which direction your own variables are coded before describing the finding, because a higher score does not always mean better sleep.

Next, read Sig. (2-tailed). This is the p-value. If it is less than .05, the correlation is statistically significant at the conventional level, meaning the relationship is unlikely to be due to chance alone in a sample of your size. SPSS may display the value as .000. That does not mean the probability is exactly zero; it means the value is smaller than .001 after rounding. Report it as p < .001, never as p = .000.

Finally, read N. This is the number of cases used in the analysis. Report it alongside rho, because a correlation based on 20 cases and one based on 200 cases do not carry the same weight, even if the rho values look identical.

A complete reading of an example output might therefore sound like this: there was a strong, positive, statistically significant monotonic relationship between the two ratings, rho = .62, p < .001, N = 48. Notice that this sentence reports direction, strength, significance, the statistic, the p-value, and the sample size — everything a reader needs, and nothing the output did not show.

Also report what you checked. A single sentence such as “A scatterplot confirmed a monotonic relationship” tells your reader that you did not simply run the test blindly. If the scatterplot showed a non-monotonic pattern, say so and explain what you did instead, rather than reporting a rho value that does not fit the data.

Reporting in APA Style (7th ed.)

APA style (7th edition) reports correlations in the text, usually without a separate table unless you have many correlations to present. Use the Greek letter rho (ρ) or rs for Spearman’s correlation, give the value to two decimal places, and include the sample size and the exact p-value where possible.

A template you can adapt is:

A Spearman rank-order correlation showed a strong positive monotonic relationship between stress ratings and sleep-quality ratings, rs(46) = .62, p < .001.

Note the details: rs is followed by the degrees of freedom in parentheses, which for a correlation is N minus 2 (here, 48 minus 2 equals 46). The correlation value has no leading zero before the decimal point, because a correlation cannot exceed 1. The p-value is given exactly when SPSS provides it, or as p < .001 when SPSS rounds it to .000.

If the result is not significant, report it just as clearly. For example: “A Spearman rank-order correlation showed no significant monotonic relationship between the two ratings, rs(46) = .11, p = .45.” Non-significant results are still results, and hiding them because they did not reach .05 is poor practice.

If you report several correlations, present them in a table with the variable names, rs values, p-values, and sample sizes, and refer to the table in the text. Keep the table simple: one row per pair of variables is usually enough.

Common Mistakes to Avoid

Reporting Pearson’s r by accident. This happens when the Pearson box is left checked in the dialog. Always glance at the output heading: it should say Spearman’s rho, not Pearson Correlation.
Describing correlation as causation. A significant Spearman correlation means two variables move together monotonically. It does not prove that one causes the other. Avoid causal language such as “stress causes poor sleep” on the basis of a correlation alone.
Ignoring the scatterplot. Spearman’s rho assumes a monotonic relationship. If you skip the scatterplot, you may report a number for a pattern it does not describe.
Reporting p = .000. SPSS rounds very small p-values to .000. Always write p < .001 instead.
Confusing strength with significance. A small rho can be statistically significant in a very large sample, and a large rho can be non-significant in a very small one. Report both the size of rho and the p-value, and interpret them together.

Key Takeaways

Spearman’s rho measures the strength and direction of a monotonic relationship between two ranked or continuous variables.
Use it for ordinal data, skewed data, data with outliers, or relationships that are monotonic but not linear.
Always check a scatterplot for monotonicity before interpreting the result.
In SPSS: Analyze > Correlate > Bivariate, uncheck Pearson, check Spearman, and click OK.
Report rs, degrees of freedom (N – 2), the p-value, and the sample size, following APA style (7th ed.).
Correlation is not causation — describe what moves together, not what causes what.

Frequently Asked Questions

What is the difference between Spearman and Pearson correlation?

Pearson’s r measures linear relationships between interval or ratio variables and assumes roughly normal data without extreme outliers. Spearman’s rho measures monotonic relationships using ranks, so it works for ordinal data, skewed data, and relationships that curve but still consistently rise or fall.

Can I use Spearman’s rho with Likert-scale data?

Yes. Likert-scale responses are ordinal, which makes Spearman’s rho a natural choice. Pearson’s correlation is sometimes used with Likert sums, but when you are correlating individual ordinal items or ratings, Spearman is the safer and more defensible option.

What does a negative Spearman correlation mean?

A negative rho means that as one variable increases, the other tends to decrease. For example, a rho of -.58 would indicate a strong negative monotonic relationship. The interpretation of strength is the same as for a positive value; only the direction differs.

Do I need to check normality before running Spearman’s rho?

No. Spearman’s rho does not assume normality, because it works with ranks. You should still check that the relationship is monotonic by inspecting a scatterplot, and you should check that your variables are paired observations on the same cases.

How do I report Spearman’s rho in APA style?

Write the statistic as rs with degrees of freedom (N – 2) in parentheses, followed by the rho value to two decimal places and the p-value. For example: rs(46) = .62, p < .001. Remember that SPSS displays very small p-values as .000, which you should report as p < .001.

Need this done on your data? Get a free quote → /services/