Choosing a statistical test feels complicated, but it usually comes down to two questions: what are you trying to find out, and what type of data do you have? Try our free interactive test chooser too.
Step 1: Know your data types
- Categorical (nominal) — groups with no order: gender, department, yes/no.
- Ordinal — ordered categories: education level, single Likert items.
- Continuous (scale) — numbers with meaningful distances: age, marks, income, and the average score of a multi-item Likert scale (commonly treated as continuous).
Step 2: What is your question?
| Your question | Data | Test (parametric) | Alternative (non-parametric) |
|---|---|---|---|
| Compare 2 independent groups (e.g. male vs female) | Group + continuous | Independent samples t-test | Mann–Whitney U |
| Compare the same people twice (before / after) | Continuous ×2 | Paired samples t-test | Wilcoxon signed-rank |
| Compare 3+ groups (e.g. three departments) | Group + continuous | One-way ANOVA (+ post-hoc) | Kruskal–Wallis |
| Is there a relationship between two variables? | Continuous ×2 | Pearson correlation | Spearman correlation |
| Does X predict / affect Y? | Continuous outcome | Simple / multiple linear regression | — |
| Predict a yes/no outcome | Binary outcome | Binary logistic regression | — |
| Are two categorical variables associated? | Categorical ×2 | Chi-square test of independence | Fisher’s exact (small samples) |
Step 3: Check assumptions
Parametric tests assume roughly normal data (or a reasonably large sample), similar variances between groups and independent observations. Check normality with skewness/kurtosis or the Shapiro–Wilk test, and Levene’s test for equal variances. If assumptions fail badly, use the non-parametric alternative.
Examples
- “Is there a difference in job satisfaction between public and private school teachers?” → independent t-test.
- “Does social support predict stress among nurses?” → simple linear regression.
- “Is smoking status associated with gender?” → chi-square.
Mediation and moderation
If your framework has a mediator (X → M → Y) or moderator, regression-based approaches such as Hayes’ PROCESS macro in SPSS, or structural equation modelling (SEM), are commonly used.
Decide your tests while writing your objectives — not after collecting data. Then your questionnaire will collect exactly the type of data those tests need.
🤝 Need help with SPSS & Data Analysis?
Tell us your topic, level and deadline. We reply with a clear plan and price — nothing is charged before you agree.
📲 Found this useful? Share it with a classmate on WhatsApp.
