10 SPSS Mistakes Thesis Students Make (and How to Fix Them)
Your thesis deadline is close, your SPSS output is 40 pages long, and something feels wrong — but you cannot tell what. Most thesis students do not fail because statistics is too hard. They fail because of a handful of avoidable mistakes that creep in before the first test is even run.
This guide covers the ten SPSS mistakes I see most often in thesis data, in the order you are most likely to make them — from data setup to the final results chapter. Each one comes with the exact fix, so you can check your own file as you read along.
1. Starting the analysis before checking the data
The mistake: opening SPSS, running the analysis straight away, and trusting whatever comes out. A single typo — a 99 where a 9 should be, a missing value coded as 0 — can quietly wreck every result that follows.
The fix: run Frequencies and Descriptives on every variable before you test anything. Check the minimum and maximum of each variable, look for impossible values, and confirm how many cases are missing. Five minutes here saves hours of rework later.
2. Forgetting to reverse-score questionnaire items
The mistake: computing a scale total from questionnaire items without reversing the negatively worded ones. If “I feel calm” is scored 1 to 5 and “I feel nervous” is also scored 1 to 5, adding them together cancels the scale out — and your reliability and correlations will look far weaker than they really are.
The fix: check the scoring key for every scale you use, reverse-score the negatively worded items with Transform > Recode into Different Variables, and only then compute the total. Always run Cronbach’s alpha after recoding, not before.
3. Choosing the test before checking the data type
The mistake: running a t-test or Pearson correlation on ordinal Likert data, or a chi-square test on a continuous outcome, simply because the test is familiar. The test has to match the measurement level of your variables — nominal, ordinal, or scale — or the result is meaningless.
The fix: before choosing a test, write down each variable’s measurement level and the number of groups you are comparing. Two groups with a scale outcome usually means a t-test; three or more groups means ANOVA; two ordinal variables usually means Spearman correlation. When in doubt, check a decision guide before you click.
4. Ignoring the assumptions of the test
The mistake: reporting a t-test or ANOVA without ever checking normality, equal variances, or linearity. Supervisors and examiners routinely ask about assumptions, and a violated assumption is the most common reason a results section gets sent back.
The fix: check assumptions as part of the analysis, not as an afterthought. For a t-test or ANOVA, look at normality within each group and run Levene’s test for equal variances. For regression, check linearity, homoscedasticity, and multicollinearity. If an assumption fails, switch to the nonparametric alternative or transform the variable — and say so in your thesis.
5. Reading the wrong line of the output
The mistake: copying the first number you see. In an independent-samples t-test, that means reporting the “equal variances assumed” row even when Levene’s test is significant. In regression, it means reporting the unstandardized coefficient when you meant the standardized one, or reading R-squared from the wrong model.
The fix: slow down at the output stage. For a t-test, read Levene’s test first — if its p-value is below .05, use the “equal variances not assumed” row. For regression, report the model summary and the coefficients table together, and double-check which model number you are reading when you ran several blocks.
6. Writing p = .000
The mistake: copying SPSS output literally and writing “p = .000” in the thesis. A p-value is never exactly zero — SPSS simply rounds anything below .001 to .000. Examiners notice this immediately, and it marks the work as unedited output.
The fix: whenever SPSS shows .000, write p < .001. For other values, report the exact p-value to two or three decimal places, for example p = .032. Never write p = .000 anywhere in your thesis.
7. Reporting significance without effect size
The mistake: writing “the difference was significant, p = .021” and stopping there. With a large sample, even a trivial difference becomes significant — so a p-value alone says nothing about whether the finding matters.
The fix: always report an effect size alongside the p-value: Cohen’s d for t-tests, eta-squared for ANOVA, and R-squared for regression. Then interpret it in plain language — a small effect in a large sample may not be worth a paragraph in your discussion.
8. Pointing and clicking without saving syntax
The mistake: doing everything through menus and never saving the syntax. When your supervisor asks you to rerun the analysis with one variable changed, you have to rebuild every dialog box from memory — and you will not remember exactly which options you ticked.
The fix: click Paste instead of OK in every dialog box, so SPSS writes the syntax into a syntax window. Save that syntax file with your data file. Re-running the whole analysis then takes one click, and your method section can describe exactly what you did.
9. Leaving variables and values unlabeled
The mistake: working with variables named VAR00001 and values coded 1 and 2 with no labels. Your output tables become unreadable, and a month later even you will not remember what “Group 2” was.
The fix: give every variable a clear name and a descriptive label, and define value labels for every coded variable (for example, 1 = Male, 2 = Female). Do this in Variable View before you analyse anything — your output tables will label themselves.
10. Pasting raw SPSS tables into the thesis
The mistake: copying SPSS output tables straight into Chapter 4. They are wide, cluttered, full of redundant columns, and formatted for a screen, not a thesis page. They also signal to the examiner that the results were never really processed.
The fix: rebuild each table in APA style: keep only the columns the reader needs, round to two decimal places, add a clear title and notes, and put the full SPSS output in an appendix if your supervisor wants it. One clean table says more than five pages of raw output.
A quick syntax check you can run on any dataset
Before you run any test, paste this syntax into a syntax window, replace the variable names with your own, and run it. It flags impossible values and missing data in seconds:
FREQUENCIES VARIABLES=Age Gender Score1 Score2 Score3
/STATISTICS=MINIMUM MAXIMUM MEAN
/ORDER=ANALYSIS.
DESCRIPTIVES VARIABLES=Score1 Score2 Score3
/STATISTICS=MEAN STDDEV MIN MAX.Scan the output for three things: minimum and maximum values that fall outside the possible range, a mean that looks impossible for the scale, and a missing-case count that is higher than you expected. Fix the data first, then analyse.
How to interpret the output
When your results come back, read them in this order. First, check the sample size — is the N what you expected, or did listwise deletion quietly drop half your cases? Second, check the test statistic and its degrees of freedom, which tell you the test ran on the data you intended. Third, read the p-value and the effect size together: the p-value tells you whether the result is unlikely under the null hypothesis, and the effect size tells you whether it matters. Finally, check the direction — which group scored higher, and by how much. If any of these four steps surprises you, go back to the data before you write a word of the results section.
Reporting in APA style (7th ed.)
Avoid the most common reporting errors by following these rules. Report exact p-values except when they fall below .001, in which case write p < .001. Italicise statistical symbols such as p, t, F, and r, and put a space on both sides of the equals sign. Give degrees of freedom in parentheses, for example t(58) = 2.41, p = .019, d = 0.62. Report descriptive statistics (means and standard deviations) before the test result, so the reader can see the size of the difference before the significance test. And always state which test you ran, on which variables, and why that test was appropriate — a results section without a rationale invites the examiner to ask for one.
Key Takeaways
- Check your data before you analyse it — impossible values and missing cases are the most common source of wrong results.
- Reverse-score questionnaire items before computing scale totals, and run reliability after recoding.
- Match the test to your variables’ measurement level, and check the test’s assumptions every time.
- Read Levene’s test before choosing a t-test row, and never write p = .000 — it is p < .001.
- Report an effect size with every p-value, save your syntax, label your variables, and rebuild tables in APA style.
Frequently Asked Questions
What is the most common SPSS mistake thesis students make?
Analysing data that was never checked. Missing values coded as zeros, items that were never reverse-scored, and out-of-range typos are the most common errors, and they distort every test that follows. A five-minute Frequencies and Descriptives check catches most of them.
Why does SPSS show p = .000?
SPSS rounds p-values below .001 to .000. The true value is not zero — it is simply smaller than .001. In your thesis, always write p < .001 instead of copying the rounded output.
Do I need to report effect sizes in my thesis?
Yes. Most universities and examiners now expect an effect size (Cohen’s d, eta-squared, or R-squared) alongside every significance test, because a p-value alone does not show how large or important a difference is.
Should I use menus or syntax in SPSS?
Use the menus if you like, but click Paste instead of OK so SPSS saves the syntax. A saved syntax file makes your analysis reproducible, lets you rerun everything in one click, and gives you an exact record for your method section.
How do I know if I used the wrong statistical test?
Check the measurement level of your variables and the number of groups, then match them to the test’s requirements. If your outcome is ordinal or your data is heavily skewed, a nonparametric test is usually the safer choice. When you are unsure, ask your supervisor before you write up the results — not after.
Stuck on your thesis analysis? Get expert SPSS help
If any of the mistakes above sound familiar — or if your supervisor has sent your results chapter back for a rework — you do not have to fix it alone. Our team works with thesis students every week: cleaning messy datasets, choosing the right tests, running the analysis in SPSS, and writing up results in proper APA style.
We can take your data from raw file to finished results chapter, including syntax files you can keep, APA-formatted tables, and a clear explanation of every result so you can defend it in your viva. Many students come to us after one of these ten mistakes has already cost them weeks — the earlier you get help, the less rework there is.
See exactly what we offer on the services page, from one-off analysis support to full dissertation help. Every project starts with a free quote and a clear timeline, so you know the cost and the delivery date before you commit.
Need this done on your data? Get a free quote → /services/