Independent Samples t-Test in SPSS: Run, Interpret and Report

The independent samples t-test compares the means of two separate groups, for example women and men, or buyers and non-buyers. In SPSS, go to Analyze > Compare Means > Independent-Samples T Test, define the two groups, and read the row that matches Levene's test. Always add an effect size: a significant difference is not always a meaningful one.

When to use it

• One continuous dependent variable (for example a purchase intention score).
• One grouping variable with exactly two independent groups.
• Each respondent belongs to only one group. For the same people measured twice, use the paired samples t-test instead.

With three groups or more, move to one-way ANOVA.

Assumptions

1. Independence: the two groups contain different respondents.
2. Approximate normality of the dependent variable in each group. With 30 or more per group, the test is robust to moderate departures. See how to check normality in SPSS.
3. Homogeneity of variances: SPSS tests it with Levene's test and gives a corrected result when it is violated.

Run the test in SPSS

1. Go to Analyze > Compare Means > Independent-Samples T Test.
2. Move purchase intention (PI) to Test Variable(s).
3. Move gender to Grouping Variable, click Define Groups and type the two codes (for example 1 = women, 2 = men).
4. In recent versions (SPSS 27 and later), tick Estimate effect sizes.
5. Click OK.

Read the output

Group Statistics

Independent Samples Test

This table has two rows. Choose the right one with Levene's test:

• Levene's p > .05: variances are similar, read the row Equal variances assumed.
• Levene's p ≤ .05: read Equal variances not assumed (Welch's t-test).

Levene's test is not significant (p = .21), so we read the first row: women report higher purchase intention than men, t(298) = 2.90, p = .004.

Many methodologists now recommend reading the Welch row by default, since it performs well whether or not variances are equal (Delacre et al., 2017). Both rows lead to the same conclusion here.

Effect size

SPSS gives Cohen's d in the Independent Samples Effect Sizes table. Here d = 0.34.

The difference is significant but small to moderate. Say so in the discussion rather than presenting it as a strong gender effect.

How to report the result

An independent samples t-test showed that women (M = 3.78, SD = 0.82) reported higher purchase intention than men (M = 3.49, SD = 0.91), t(298) = 2.90, p = .004, d = 0.34, 95% CI of the difference [0.09, 0.49]. Levene's test indicated equal variances (F = 1.58, p = .21).

If the assumptions are not met

When the data are strongly skewed and groups are small, use the non-parametric Mann-Whitney U test: Analyze > Nonparametric Tests > Legacy Dialogs > 2 Independent Samples. Report the medians and the U statistic.

FAQ

What does "one-sided p" mean in recent SPSS versions?

Since SPSS 27, the output shows one-sided and two-sided p-values. Report the two-sided value unless you stated a directional hypothesis in advance.

Can I compare more than two groups with several t-tests?

No. Multiple t-tests inflate the risk of a false positive. Use ANOVA with post hoc tests.

My groups are very unequal in size. Is that a problem?

Unequal sizes are fine, but combined with unequal variances they bias the standard t-test. Read the Welch row in that case.

Should I report Cohen's d or Hedges' g?

For small samples (under about 20 per group), Hedges' g corrects a slight bias. SPSS gives both.

References

• Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Lawrence Erlbaum.
• Delacre, M., Lakens, D., & Leys, C. (2017). Why psychologists should by default use Welch's t-test instead of Student's t-test. International Review of Social Psychology, 30(1), 92–101. https://doi.org/10.5334/irsp.82
• Lakens, D. (2013). Calculating and reporting effect sizes to facilitate cumulative science: A practical primer for t-tests and ANOVAs. Frontiers in Psychology, 4, 863. https://doi.org/10.3389/fpsyg.2013.00863