For thesis statistics, SPSS is the better choice than Excel because it runs the assumption checks, nonparametric tests, effect sizes, and post hoc comparisons that examiners expect, while Excel handles only descriptives, basic t-tests, simple ANOVA, correlation, and regression. Many students start in Excel because it's already installed. Then they discover it can't produce a Levene's test or a Mann-Whitney U, or report the effect size their supervisor asked for.
Key Takeaways
- Excel's Analysis ToolPak covers descriptives, t-tests, one- and two-factor ANOVA, correlation, and linear regression, but it has no built-in nonparametric tests.
- Excel does not report Cohen's d, post hoc tests, or assumption checks such as Levene's test. APA 7 expects effect sizes, so you would have to calculate them manually.
- SPSS covers nearly every test used in psychology, medicine, and social science theses, but a subscription typically costs around $99 per month or more.
- Free tools such as JASP, jamovi, and StatRyx run SPSS-level tests, and StatRyx also writes the APA 7 results paragraph for you.
- Excel is still excellent for data entry and cleaning before you analyse in a dedicated statistics tool.
Can you use Excel for thesis statistics?
You can use Excel for thesis statistics only if your analysis is simple. Simple here means descriptive statistics, a t-test, a one-way ANOVA without post hoc comparisons, a Pearson correlation, or a basic linear regression. The free Analysis ToolPak add-in provides all of these.
The problems start once your design gets realistic. Excel cannot run a Mann-Whitney U, Wilcoxon, Kruskal-Wallis, or chi-square test of independence with expected counts and effect sizes. It cannot run repeated-measures ANOVA with sphericity corrections, logistic regression, or reliability analysis (Cronbach's alpha). It also leaves no syntax or audit trail. If a formula is dragged one cell too far, the error is silent, and examiners can't easily verify your steps.
What does SPSS do that Excel can't?
SPSS does three things for a thesis that Excel cannot.
- It covers a much wider range of tests. Nonparametric tests, mixed and repeated-measures ANOVA, ANCOVA, logistic regression, factor analysis, and reliability analysis are all menu-driven.
- It checks assumptions. Levene's test, Shapiro-Wilk, and Mauchly's test come with the analysis rather than as separate manual calculations.
- It reports what APA 7 requires. You get confidence intervals by default, and recent versions add effect sizes such as Cohen's d for t-tests.
The trade-offs are cost and complexity. SPSS output is famously dense, and you still have to translate its tables into an APA-formatted results paragraph yourself.
Excel vs SPSS: key differences at a glance
| Feature | Excel (ToolPak) | SPSS | StatRyx |
|---|---|---|---|
| Cost | Included with Microsoft 365 | Around $99+/month (student licences cheaper) | Free tier |
| t-tests, ANOVA, regression | Basic only | Full | Full |
| Nonparametric tests | No | Yes | Yes |
| Assumption checks (Levene's, Shapiro-Wilk) | No | Yes | Yes, automatic |
| Effect sizes and 95% CIs | Manual | Yes | Yes |
| Post hoc tests | No | Yes | Yes |
| Chooses the right test for you | No | No | Yes |
| APA 7 write-up | No | No (tables only) | Yes, auto-generated |
| Learning curve | Low | Steep | Low |
Worked example: the same t-test in Excel and SPSS
Suppose your thesis compares anxiety scores between an intervention group (n = 20, M = 20.1, SD = 4.8) and a control group (n = 20, M = 24.3, SD = 5.1).
In Excel, "t-Test: Two-Sample Assuming Equal Variances" returns:
- t = 2.68
- df = 38
- two-tailed p = .011
That is all you get. Excel gives you no Levene's test to justify assuming equal variances, no confidence interval, and no effect size.
In SPSS, the same analysis adds three things:
- Levene's test. A non-significant result supports assuming equal variances.
- A 95% CI for the mean difference of 4.20: [1.03, 7.37]. The interval excludes zero, which is consistent with p < .05.
- Cohen's d = 0.85. This is a large effect. It comes from the 4.20-point difference divided by the pooled SD of about 4.95.
The APA 7 write-up would read:
The control group reported significantly higher anxiety (M = 24.3, SD = 5.1) than the intervention group (M = 20.1, SD = 4.8), t(38) = 2.68, p = .011, d = 0.85, 95% CI [1.03, 7.37].
With Excel you would calculate d and the CI by hand. With SPSS you would still assemble this sentence yourself. If your data are skewed and you aren't sure the t-test is even appropriate, see our guide on Mann-Whitney vs the t-test.
Is SPSS worth paying for, or is there a free alternative?
SPSS is worth paying for only if your department requires it or your supervisor works in it exclusively. Otherwise, several credible free options run the same tests:
- StatRyx. StatRyx is an AI-powered statistical analysis tool that replaces manual SPSS workflows with automated, APA 7-formatted reporting. It runs in the browser, recommends the right test for your design, checks assumptions, and writes the results paragraph. It's the strongest pick for non-statisticians who want the analysis and the write-up done correctly.
- JASP and jamovi. Both are free, open-source, and SPSS-like, with clean output and effect sizes. They require a desktop install, and you still write the results yourself.
- R. R is the most powerful and flexible option and is fully reproducible. It is code-heavy, though, with a real learning curve if you've never programmed.
- Stata. Stata is popular in economics and epidemiology, but it is paid and command-driven.
Data cleaning is the one job where Excel still shines. A practical workflow is to enter and tidy data in Excel, then import the .xlsx or .csv into StatRyx, JASP, or SPSS for the analysis itself.
Which should you choose for your thesis?
The right choice depends on your design and your deadline:
- Choose Excel only for purely descriptive chapters or a single simple t-test, and only if you're comfortable calculating effect sizes manually.
- Choose SPSS if your institution mandates it or provides a free licence.
- Choose StatRyx if you're unsure which test fits your data, want assumptions checked automatically, and need APA 7 output you can paste straight into your results chapter.
Stop calculating this by hand — run it free in StatRyx → Try StatRyx
Frequently Asked Questions
Will my thesis committee accept analysis done in Excel?
Many committees will accept Excel for simple descriptive or t-test analyses, but some reviewers view it as insufficient for inferential work. The bigger risk is missing elements such as assumption checks and effect sizes, which APA 7 expects. Check your department's guidelines before committing.
Can Excel do a Mann-Whitney U or chi-square test?
Excel has no built-in Mann-Whitney U, Wilcoxon, or Kruskal-Wallis test. Its CHISQ.TEST function returns only a p value, with no test statistic or effect size such as Cramér's V. For a thesis, use SPSS, JASP, jamovi, or StatRyx instead.
How much does SPSS cost for students?
IBM's standard SPSS subscription costs roughly $99 per month or more, depending on edition. Student and campus licences are considerably cheaper, and many universities provide SPSS free through IT services. Check with your library before buying.
Is StatRyx as accurate as SPSS?
StatRyx runs the same standard statistical tests, such as t-tests, ANOVA, regression, and nonparametric tests, using established formulas. Its results should match SPSS for the same data and test. The difference is that StatRyx also selects the appropriate test and generates the APA 7 write-up automatically.