5 SPSS Mistakes That Undermine Your Results
20 May 2026
SPSS makes it remarkably easy to produce output — a few clicks and you have a results table. What it doesn't do is tell you whether that table is actually answering your research question correctly. Here are five mistakes we see most often, and how to catch them before they make it into your final dissertation.
1. Skipping Assumption Checks
Tests like ANOVA and t-tests rely on assumptions — normality, homogeneity of variance — that SPSS will happily let you ignore. Running the test anyway doesn't produce an error message; it just produces results that may not be trustworthy. Always check relevant assumptions before running and reporting a test, and have a non-parametric alternative ready if they're not met.
2. Misreading the P-Value
A p-value above .05 doesn't mean "there's no relationship" — it means your data didn't provide strong enough evidence of one in this particular sample. This distinction matters in how you discuss non-significant findings, which should still be reported and discussed honestly, not treated as a failed test.
3. Incorrect Variable Coding
Entering categorical variables as plain numbers without properly labelling the values (in SPSS's Variable View) can silently distort certain analyses, or simply make your output unreadable later. Taking the extra few minutes to code and label variables properly at data entry saves considerably more time down the line.
4. Pasting Output Without Interpretation
An SPSS output table is not, on its own, a finding. Dropping a table into your results chapter without explaining what it means in plain language, tied back to your research question, is one of the most common and easily avoidable weaknesses examiners flag.
5. Ignoring Effect Size
Statistical significance tells you a result is unlikely to be due to chance; it doesn't tell you whether the effect is actually large or meaningful. Reporting effect size alongside significance gives a fuller, more honest picture of your findings — and is increasingly expected in strong dissertations.
The Common Thread
Notice that none of these are really about SPSS itself — they're about statistical understanding that SPSS doesn't enforce for you. The software will run almost anything you ask it to; making sure what you're asking for is actually correct is the part that takes genuine care.