SPSS for Your Dissertation: Which Test to Use

Most dissertation students do not struggle with SPSS buttons. They struggle with choosing the right test, checking that it is allowed, and reporting the result in the way their marker expects. This guide takes you through that in six steps, in plain English.

Step 1: Start with your question and your data

Write your research question in one sentence. Then answer three things: what type of data is the outcome (numbers on a scale, or categories), how many groups are you comparing, and are the groups separate people or the same people measured twice?

Your question	Test (if assumptions hold)	If they do not hold	SPSS menu
Two separate groups, scale outcome	Independent-samples t-test	Mann-Whitney U	Analyze > Compare Means > Independent-Samples T Test
Same people, two time points	Paired-samples t-test	Wilcoxon signed-rank	Analyze > Compare Means > Paired-Samples T Test
Three or more separate groups	One-way ANOVA	Kruskal-Wallis	Analyze > Compare Means > One-Way ANOVA
Two categorical variables	Chi-square test	Fisher's exact test	Analyze > Descriptive Statistics > Crosstabs
Link between two scale variables	Pearson correlation	Spearman correlation	Analyze > Correlate > Bivariate
Predict an outcome from several variables	Multiple linear regression	Check with your supervisor	Analyze > Regression > Linear

Step 2: Clean and check your data

Look at the frequencies of every variable (Analyze > Descriptive Statistics > Frequencies). Typing mistakes such as an age of 250 show up fast.
Decide how to handle missing answers and write the decision down. Your method chapter must explain it.
Check that each variable is the right type in Variable View: scale, ordinal or nominal.
Reverse-code any negatively worded questionnaire items before you add scores together.

Step 3: Check the assumptions

Each test only works if its assumptions are reasonable. For the common tests, check these.

Normality of the scores in each group. In SPSS use Analyze > Descriptive Statistics > Explore, then Plots, then tick "Normality plots with tests". Look at the Shapiro-Wilk result and the histograms. With large samples, small departures often do not matter.
Equal variances between groups. The independent t-test output includes Levene's test and gives you the right row to read.
Independence. Each person should appear once and should not influence another person's score. This comes from how you collected the data.
For correlation and regression, check for a straight-line pattern in a scatterplot and for extreme outliers.

Step 4: Run the test

Use the menu path in the table. Put your outcome variable in the "Test Variable" box and your grouping variable in the "Grouping Variable" box, then define the groups. Ask for descriptive statistics and an effect size where SPSS offers them. Keep the output file, because your supervisor may ask to see it.

Step 5: Read the output

The p value (the "Sig." column) tells you how surprising the result would be if there were really no effect. A value below .05 is the usual cut-off, but it is not a measure of importance.
The effect size tells you how big the difference or link is. For two-group comparisons Cohen's d is common, with about 0.2 small, 0.5 medium and 0.8 large as rough guides.
The means and standard deviations tell the story in real numbers. Always report them.

Step 6: Report it in APA style

Here is an invented example of how a result is usually written. Replace it with your own figures.

An independent-samples t-test showed that satisfaction was higher for students who used the tutoring service (M = 73.0, SD = 8.0) than for those who did not (M = 68.0, SD = 9.0), t(58) = 2.27, p = .027, d = 0.59.

Then add one sentence that says what it means for your research question. A result with no interpretation earns few marks.

Common mistakes that cost marks

Running many tests and reporting only the ones that came out significant.
Reporting a p value with no effect size or descriptive statistics.
Ignoring assumptions, or never mentioning that you checked them.
Treating a five-point rating scale as if it were exact measurement without justifying it.
Pasting SPSS tables into the document unedited. Rebuild them in APA format.

For help planning the rest of the project, read our guide to UK dissertation structure. Our SPSS dissertation help and statistics help pages explain what support is available.

Where TutorSuite fits

We can explain which test fits your design, show you how to check assumptions and give feedback on how you have written up your own results. TutorSuite is an academic support service, so we help you understand and improve your own work. Always follow your university's rules on what help you may use, and submit only work that is your own. Read our academic integrity page for details.

Frequently asked questions

Which test do I use to compare two groups in SPSS?

If the groups are separate people and the outcome is a score on a scale, use an independent-samples t-test. If the same people are measured twice, use a paired-samples t-test. If the assumptions fail, use Mann-Whitney U or Wilcoxon signed-rank.

What does p < .05 mean?

It means the result would be unlikely if there were truly no effect. It does not tell you how large or important the effect is, so always report an effect size too.

What is an effect size?

It is a number that shows how big a difference or relationship is. Common examples are Cohen's d for group differences and r for correlations.

What if my data is not normally distributed?

Check how far from normal it is and how large your sample is. If the problem is serious, use the non-parametric test and say why in your method chapter.

Do I need to report the Shapiro-Wilk test?

Many markers like to see that you checked normality. A short sentence with the result, plus a histogram in an appendix, is usually enough. Follow your department's guidance.

Last checked October 2026. Menu names can differ slightly between SPSS versions, so check yours.

SPSS Data Analysis for a Dissertation: A Beginner's Guide to Choosing the Right Test