How do you run a Pearson correlation in SPSS?

Open Analyze, then Correlate, then Bivariate. Move your two continuous variables into the Variables box, tick Pearson, keep the two-tailed test, and click OK. Read the Pearson Correlation value for the strength and direction, and the significance value against an alpha of .05.

When should you use Spearman instead of Pearson in SPSS?

Use Spearman's rho when the relationship is monotonic but not straight, when your data are ordinal or ranked, or when outliers distort the pattern. Pearson assumes a linear relationship between two continuous, roughly normal variables, so when those assumptions fail, the rank-based Spearman coefficient is the safer choice.

What does a partial correlation do in SPSS?

A partial correlation measures the association between two variables after removing the influence of a third control variable from both. Run it through Analyze, Correlate, Partial, and place the control in the Controlling for box. If a strong correlation shrinks toward zero once you control for a confound, the original link was partly spurious.

How do you report a correlation in APA style?

Report the coefficient with its degrees of freedom, which is the number of pairs minus two, then the exact p-value, in the form r(118) = .42, p < .001. You can add the coefficient of determination to show the shared variance. Always describe the result as an association, not as one variable causing the other.

How to run a correlation in SPSS

To run a correlation in SPSS, you open Analyze > Correlate > Bivariate, move your two continuous variables in, choose Pearson or a rank-based coefficient, and read the correlation coefficient and its significance from the matrix. The analysis that holds up depends on checking linearity with a scatterplot first, choosing Pearson against Spearman correctly, and reporting the coefficient of determination rather than implying cause. This guide covers the procedure and the judgement around it.

What a Pearson correlation measures

Pearson's r measures the strength and direction of the linear relationship between two continuous variables, on a scale from minus one to plus one. A value near plus one means that as one variable rises the other rises in step; near minus one means one falls as the other rises; near zero means no linear relationship. The word linear is the catch: Pearson's r can miss a strong curved relationship entirely, which is why you always look at a scatterplot before trusting the number. And whatever the coefficient, it never proves causation, the point laboured in our guide to correlation and causation.

The assumptions, and Pearson against Spearman

Pearson's correlation assumes both variables are continuous, that their relationship is linear, that the data are free of strong outliers, and that the variables are roughly bivariate normal. Inspect the scatterplot from Graphs > Chart Builder for a straight-line pattern and stray points before you proceed. When the relationship is monotonic but not straight, when you have ranked data, or when outliers distort the pattern, use Spearman's rho instead, the rank-based coefficient that the same dialog offers. The wider decision is set out in parametric versus nonparametric tests.

Running the correlation and the syntax to keep

Open Analyze > Correlate > Bivariate and move both variables into the Variables box. Tick Pearson (and Spearman if you want the robust comparison), keep the two-tailed test of significance unless you have a directional hypothesis, and tick Flag significant correlations. The reproducible syntax is:

CORRELATIONS
  /VARIABLES=motivation achievement
  /PRINT=TWOTAIL NOSIG
  /STATISTICS DESCRIPTIVES
  /MISSING=PAIRWISE.

Choose two-tailed when you are testing for any association and one-tailed only when theory commits you to a direction in advance, because deciding the tail after seeing the result is not legitimate.

Reading the correlation matrix

SPSS returns a Correlations matrix. Each cell gives three numbers: Pearson Correlation (the r value), Sig. (2-tailed) (the p-value), and N (the pairs analysed). The diagonal is always 1 because a variable correlates perfectly with itself, and the matrix is symmetric, so you read one triangle. Interpret the magnitude with a rough guide of about .10 for a small, .30 for a medium, and .50 for a large relationship, while judging it against your field. Square the coefficient to get the coefficient of determination, r squared, the proportion of shared variance, so an r of .40 means sixteen percent of the variance is shared. Reading these tables sits within making sense of SPSS output.

Controlling for a third variable with partial correlation

A plain correlation can be inflated or masked by a lurking third variable. A partial correlation removes the influence of a control variable from both measures and reports the association that remains. Run it through Analyze > Correlate > Partial, moving the control into the Controlling for box. If a strong zero-order correlation shrinks toward zero once you control for an obvious confound, that is direct evidence the original link was partly spurious, and it is one of the cleanest ways to show you took the causation problem seriously.

Reporting a correlation in APA style

Report the coefficient with its degrees of freedom (the number of pairs minus two), the exact p-value, and ideally the shared variance. A clean sentence reads: motivation and achievement were positively correlated, r(118) = .42, p < .001, with motivation accounting for about eighteen percent of the variance in achievement. For several variables, present the full correlation matrix as a table and report the coefficients in text only where they bear on a hypothesis. Word the finding as an association, never as one variable causing the other. The conventions are in the APA format for this result.

From correlation to prediction

Correlation quantifies the strength of a link, but it does not let you predict one variable from another or weigh several predictors at once. When your question becomes predictive, or you need to control for several variables formally rather than one, the next step is multiple regression in SPSS, which builds directly on the correlations you have just examined. Seen this way, a careful correlation analysis is both a finding in its own right and the groundwork for the modelling that follows.