Key points at a glance
- A statistical association alone does not establish a cause.
- Consider timing, reverse direction and other relevant variables.
- Match your wording to the design and its assumptions.
How do association and causation differ?
Correlation describes how characteristics vary together. A causal claim additionally proposes that changing one characteristic produces a change in another. That requires a defensible identification strategy and appropriate assumptions, not just a striking correlation coefficient.
The distinction matters particularly for one-time surveys. When two variables are measured together, even their temporal order may be unclear. OpenStax explains why an observed association alone does not support a causal conclusion.
An original example: study time and examination results
Suppose an invented survey shows that people reporting more study time obtain higher examination scores. Initially, this establishes an association in those data. It does not show that an additional study hour would cause the observed difference for every individual.
Prior knowledge could affect both study behaviour and scores. The type of studying may differ between groups. Asking about study time after the examination may also influence responses. These possibilities do not refute the association, but they limit its causal interpretation.
What controls and longitudinal data can contribute
Adding control variables to a regression does not automatically make it causal. You need to explain why those variables are included and which assumptions the interpretation requires. Unobserved factors may still matter, while inappropriate controls can introduce additional problems.
Longitudinal data can clarify temporal order without resolving every identification problem. Properly conducted randomised experiments provide a different basis for inference. Describe your actual design and justify its reach rather than assigning causality to a method by name.
Use language proportionate to the evidence
For an association analysis, “is associated with” or “coincides with higher values in this sample” is often more appropriate than “causes” or “leads to”. Report the direction, uncertainty and context as well.
A precise statement for the invented example is: “Higher reported study time was associated with higher examination scores in the sample. The observational design does not establish a causal effect.” Add the actual statistical estimates when reporting a real study.
Check more than the results chapter. Abstracts, discussions, headings and conclusions can unintentionally strengthen cautious findings. Claims throughout the thesis should remain consistent with the same evidence.
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
- Identify the study design clearly.
- Consider temporal order.
- Assess plausible alternative explanations.
- Justify control variables theoretically.
- Keep abstract and conclusion claims consistent with the analysis.
Common mistakes
- Treating statistical significance as proof of causation.
- Assuming controls remove every source of bias.
- Promising an individual effect from an observed group pattern.
Print and bind your finished thesis
Once you have checked the content and final PDF, configure the printed copies to match your submission requirements.
Configure printing and bindingRelated content
Explore related guidance on thesis structure, sources, formatting and final submission.
Frequently asked questions
Does a strong correlation prove causation?
No. Even a strong association can reflect shared causes, selection processes or other mechanisms. Its size does not replace a causal justification.
Can I describe a regression result as an influence?
Consider whether the wording implies causation. For an observational association analysis, “association” or a precise description of the model coefficient is often clearer.
Is a non-causal study still useful?
Yes. It can describe patterns, assess hypotheses or guide further investigation. The key is to ask an appropriate question and avoid overstating what the results establish.