Statistical Test Selector
Answer 3 quick questions and find the right statistical test.
Ethical, compliant guidance: We provide academic support, mentoring, analysis, editing and structuring — not authorship. Your work stays compliant with university policies.
About this tool
‘Which statistical test should I use?’ derails more analysis chapters than any other question. Choose wrong and everything downstream — results, interpretation, viva answers — is built on sand.
This selector walks the same decision path a statistician does: what you're testing (difference, relationship, prediction), how many groups, whether measurements are independent or repeated, and what type of data you have. It then recommends the standard test, its assumptions, the SPSS menu path, the effect size to report, and the non-parametric alternative if assumptions fail.
It covers the tests that account for the vast majority of master's and doctoral analyses: t-tests, ANOVA, chi-square, correlation, regression and their non-parametric counterparts.
How to use it
- 1Choose your goal — comparing groups, testing a relationship, predicting an outcome, or comparing against a known value.
- 2Answer the follow-up questions about your groups, design and data type.
- 3Read the recommendation: run the test via the SPSS path shown, check every listed assumption first, and report the stated effect size alongside the p-value.
- 4If an assumption fails (e.g. normality), switch to the listed non-parametric alternative.
Frequently asked
How do I know if my data are 'normal enough'?+
Check skewness/kurtosis (roughly within ±1), inspect the histogram and Q–Q plot, and run Shapiro-Wilk for smaller samples. With n ≥ 30 per group, t-tests and ANOVA tolerate mild non-normality; with small or clearly skewed samples, prefer the non-parametric alternative.
Are Likert-scale variables continuous or ordinal?+
A single Likert item is ordinal. A multi-item scale score (the mean or sum of several items) is conventionally treated as continuous once reliability is established — which is why t-tests, ANOVA and regression appear so often in survey-based theses.
What if my model has multiple constructs and paths?+
Then you're beyond single tests — you need structural equation modelling: AMOS (CB-SEM) for confirming established theory, SmartPLS (PLS-SEM) for prediction-oriented or complex models with modest samples.
Go deeper
Quantitative Research guidesSampling Methods in Research: Types, Selection & Justification
How to Choose the Right Statistical Test: A Decision Guide
Qualitative vs Quantitative vs Mixed Methods: Which Fits Your Study?
Reliability and Validity Explained (With Examples)
Not sure where to start?
Book a free 15-minute consult. We'll map your next three steps — no obligation.