Key points at a glance
- Identify the model’s actual assumption.
- Read plots alongside tests.
- Judge deviations in analytical context.
Inspect the relevant distribution
Not every statistical method requires normally distributed raw variables. In linear regression the assumption commonly concerns errors or residuals under the model conditions. First identify your method and the quantity to which its assumption applies.
Look beyond one threshold
Example: A residual Q-Q plot mostly follows the line but diverges sharply at the upper tail. Check outliers, data entry and skew. A Shapiro-Wilk p-value alone cannot tell you how consequential the departure is for your question.
Consider sample size and robustness
With small samples, tests may miss departures; with large ones, tiny unimportant deviations may be detected. Consider histogram, Q-Q plot and design together. Do not switch methods mechanically after one test.
Report a reasoned decision
State what you checked, which plot/test you used, the pattern found and what you did about it. Justify transformations or alternative methods substantively. Assess other assumptions such as independence and variance structure separately.
Final submission context
Formal details can feel like a separate writing task, but they matter just as much for a printed submission. Anything missing, misplaced or inconsistently formatted in the document will also appear in the bound copy.
Use this guide together with your cover page, table of contents, page numbers, source notes and appendices. Prepare the final PDF for printing and binding only after the complete file has been checked.
Practical check before PDF export
- Do headings, chapter structure and page numbers match?
- Are sources, figures and tables included completely?
- Are required elements such as declarations, appendices or the cover page included where required?
- Did you open and check the final PDF after exporting it?
Checklist
- Model assumption identified.
- Q-Q plot examined.
- Sample size considered.
- Decision justified.
Common mistakes
- Testing every raw variable indiscriminately.
- Using a p-value as the sole rule.
- Forgetting other assumptions.
Prepare the finished thesis for submission
Once the content and university requirements are checked, review your PDF and the available printing and binding options.
Review printing optionsRelated content
Explore related guidance on thesis structure, sources, formatting and final submission.
Frequently asked questions
Must all data be normal?
No. It depends on the chosen model’s assumption.
Is Shapiro-Wilk enough?
No. Also inspect plots and context.
What if the assumption is doubtful?
Check data quality, design, robustness and suitable alternatives.