Most students do not struggle with clicking the right menu in SPSS. They struggle with two harder questions: which test does my data actually need, and what does this output table mean? This guide covers both, with reporting examples you can adapt.
Set your data up correctly first
Most SPSS problems trace back to data setup, not analysis. In Variable View, get three things right:
- Measure: Nominal (categories with no order), Ordinal (ordered categories), Scale (continuous). SPSS uses this to offer appropriate tests
- Values: label your codes, e.g. 1 = Male, 2 = Female, so output is readable
- Missing: define missing value codes explicitly rather than leaving blanks ambiguous
Choosing the right test
The test follows from three questions: what are you comparing, how many groups, and what type of data?
| Your question | Data type | Test |
|---|---|---|
| Difference between 2 independent groups | Scale outcome | Independent samples t-test |
| Difference between 2 related measures | Scale outcome | Paired samples t-test |
| Difference between 3+ groups | Scale outcome | One-way ANOVA |
| Relationship between 2 continuous variables | Scale | Pearson correlation |
| Relationship, non-normal or ordinal | Ordinal | Spearman correlation |
| Predicting an outcome from predictors | Scale outcome | Linear regression |
| Association between 2 categorical variables | Nominal | Chi-square test of independence |
If assumptions of a parametric test are violated, use the non-parametric equivalent: Mann-Whitney instead of independent t-test, Wilcoxon instead of paired t-test, Kruskal-Wallis instead of ANOVA.
Check assumptions before you trust results
- Normality: Shapiro-Wilk (p > .05 suggests normal) or inspect histograms and Q-Q plots
- Homogeneity of variance: Levene test in the t-test and ANOVA output
- Independence: determined by your design, not testable in SPSS
- Linearity: scatterplots for correlation and regression
- Multicollinearity: VIF values in regression, generally concerning above 10
With larger samples, mild normality violations matter less. Report which assumptions you checked and what you did when one failed.
Reading the output: what actually matters
Significance (p value)
p < .05 conventionally indicates a statistically significant result. Note that SPSS displays .000, which means p < .001, not zero. Never report p = .000.
Effect size
Significance tells you whether an effect likely exists; effect size tells you whether it matters. Report Cohen d for t-tests, eta squared or partial eta squared for ANOVA, r for correlations, and R squared for regression.
| Measure | Small | Medium | Large |
|---|---|---|---|
| Cohen d | 0.2 | 0.5 | 0.8 |
| Correlation r | 0.1 | 0.3 | 0.5 |
| Eta squared | 0.01 | 0.06 | 0.14 |
Reporting results in APA format
Report the test, degrees of freedom, test statistic, p value and effect size together.
- t-test: An independent samples t-test showed a significant difference in scores between groups, t(58) = 2.45, p = .017, d = 0.64.
- ANOVA: A one-way ANOVA revealed a significant effect of teaching method, F(2, 87) = 5.32, p = .007, eta squared = .11.
- Correlation: Study hours correlated positively with final grade, r(98) = .42, p < .001.
- Regression: The model explained 34% of variance, R squared = .34, F(3, 96) = 16.20, p < .001.
- Chi-square: There was a significant association between gender and choice, chi-square (1, N = 120) = 6.71, p = .010.
Italicise statistical symbols and report exact p values to three decimals unless below .001.
Common beginner mistakes
- Running a parametric test without checking assumptions
- Reporting p = .000 instead of p < .001
- Omitting effect sizes entirely
- Treating correlation as evidence of causation
- Using a t-test repeatedly for three or more groups instead of ANOVA, inflating error
- Reporting output tables verbatim instead of interpreting them
- Correct data setup in Variable View prevents most SPSS problems
- Test choice follows from comparison type, group count and data type
- Always check assumptions and report what you found
- Report effect size alongside significance, never alone
- SPSS .000 means p < .001 and must be reported that way
Frequently Asked Questions
Which test should I use for my data?
What does p = .000 mean in SPSS?
Do I always need to report effect size?
What if my data is not normally distributed?
Can I use Excel or R instead of SPSS?
Need help with your SPSS analysis?
Our statistics specialists guide you through test selection, assumption checking and interpreting output in plain English – so you can defend every number.
💬 Get Free ConsultationContact Us