Pearson Correlation in SPSS and How to Build a Correlation Table

Search description: Run Pearson correlations in SPSS, interpret r, and build the publication-ready correlation table with means, SDs and alphas that journals expect.

Pearson's correlation measures the strength and direction of a linear relationship between two continuous variables. In SPSS, go to Analyze > Correlate > Bivariate, select your variables and click OK. Most papers present the result as a correlation matrix with means, standard deviations and reliability coefficients. This guide shows how to run it, read it and build a table that reviewers expect.

What r tells you

The coefficient r ranges from -1 to +1. A positive value means both variables increase together, a negative value means one decreases when the other increases, and 0 means no linear relationship. Cohen (1988) suggests these benchmarks for the absolute value:

Correlation does not imply causation, and a weak r does not mean no relationship: the link may be non-linear.

Assumptions

• Two continuous variables (composite scores from Likert items are usually fine).
• A linear relationship. Check it with a scatter plot: Graphs > Legacy Dialogs > Scatter/Dot.
• No extreme outliers, which can inflate or hide a correlation.
• Approximate normality. If it fails, use Spearman's rho, available in the same dialog.

Run the correlation in SPSS

1. Go to Analyze > Correlate > Bivariate.
2. Move all study variables into Variables: ATT, PV, PS and PI.
3. Keep Pearson and Two-tailed, and tick Flag significant correlations.
4. Click Options and tick Means and standard deviations. Choose Exclude cases pairwise or listwise depending on your missing data strategy.
5. Click OK.

Read the output

SPSS shows a square table where each cell contains three lines: Pearson correlation, Sig. (2-tailed) and N. The table is symmetric, so you only need the lower half. Asterisks mark significant values (* p < .05, ** p < .01).

Look at three things:

1. Direction and size of the correlations that match your hypotheses.
2. Very high correlations between predictors (above .80), which warn of multicollinearity and possible overlap between constructs.
3. Unexpected signs, which often reveal a coding error.

Build a publication-ready correlation table

Journals expect a compact table with descriptive statistics, correlations and reliabilities on the diagonal:

Note. N = 300. Cronbach's alpha on the diagonal in parentheses. * p < .05. ** p < .01.

Tip: copy the SPSS table into Excel, delete the Sig. and N rows, round to two decimals and remove the leading zeros before pasting it into Word.

How to report correlations in the text

Pearson correlations showed that attitude was strongly and positively related to purchase intention, r(298) = .66, p < .001. Perceived value was also positively related to purchase intention, r(298) = .58, p < .001, while price sensitivity showed a weak negative correlation, r(298) = -.21, p < .001.

The degrees of freedom for a correlation are N - 2.

Correlation and discriminant validity

A correlation matrix is also a first check of discriminant validity: correlations between different constructs should be clearly lower than the reliability of each construct. For a formal test, use the Fornell-Larcker criterion or HTMT in SmartPLS.

FAQ

Pearson or Spearman?

Use Pearson for continuous, roughly normal data with a linear link. Use Spearman for ordinal data, strong skewness or outliers.

Can I correlate a dichotomous variable such as gender?

Yes. The Pearson correlation with a 0/1 variable is called point-biserial correlation, and SPSS computes it the same way.

What does r² mean?

r² is the share of variance the two variables have in common. With r = .66, attitude and purchase intention share about 44% of their variance.

Pairwise or listwise deletion?

Listwise keeps the same N for all correlations and is easier to report. Pairwise uses more data but gives different N values across cells.

References

• American Psychological Association. (2020). Publication manual of the American Psychological Association (7th ed.).
• Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Lawrence Erlbaum.
• Schober, P., Boer, C., & Schwarte, L. A. (2018). Correlation coefficients: Appropriate use and interpretation. Anesthesia & Analgesia, 126(5), 1763–1768. https://doi.org/10.1213/ANE.0000000000002864