QuickStats Healthan Ayom Healthcare venture
Synopsis and Dissertation

How to Interpret Statistical Results in a Dissertation

Turn statistical output into clear findings for your discussion and viva: effect size, confidence intervals, p-values, clinical importance and bias.

By the QuickStats Health editorial teamUpdated 10 November 20264 min readLevel: Intermediate

Software gives you tables of numbers. Your discussion and viva need you to explain what those numbers mean. This guide gives a simple sequence for interpreting any result, with example wording you can adapt.

Step 1: State the direction and size of the effect

What changed, and by how much? A mean difference, odds ratio or relative risk tells you the size. Example: patients receiving the new protocol had a mean pain score 1.8 points lower on a 10-point scale.

Step 2: Look at the confidence interval

The 95% CI shows the range of effects compatible with your data. Is the whole interval clinically important, or does it include trivial or harmful values? A wide interval means your estimate is imprecise, often because the sample is small.

Step 3: Then read the p-value

The p-value tells you how compatible the data are with no effect. Report it exactly, but do not let it replace steps 1 and 2. Our article on p-values and confidence intervals explains the details.

Step 4: Ask whether it matters clinically

A statistically significant result may be too small to matter, and a non-significant result may still be compatible with an important effect. Compare the effect with what is considered clinically meaningful in your field (for example, a minimal clinically important difference, where one is established).

Step 5: Consider other explanations

Before concluding that an exposure causes an outcome, consider:

  • Chance: multiple comparisons, small samples.
  • Bias: how participants were selected and how outcomes were measured.
  • Confounding: other factors linked to both exposure and outcome.
  • Reverse causation: especially in cross-sectional studies.

Step 6: Compare with other studies

Are your results similar to previous studies? If not, why might they differ: population, setting, definitions, design or sample size?

Example wording

Situation Suggested wording
Clear, important effect "Anaemia was associated with low dietary iron intake (OR 2.6, 95% CI 1.5 to 4.5), consistent with earlier studies in similar settings."
Significant but small "Although statistically significant, the difference of 1.2 mmHg is unlikely to be clinically important."
Not significant, wide CI "We did not find clear evidence of a difference; however, the confidence interval was wide and does not rule out a clinically important effect."
Cross-sectional association "Because exposure and outcome were measured at the same time, we cannot determine which came first."

Preparing for the viva

Be ready to explain in your own words: why you chose your sample size, why each test was appropriate, what your main effect sizes and confidence intervals mean, and what the main limitations are. Understanding your own tables matters more than memorising definitions.

Common mistakes

  • Interpreting p-values alone.
  • Writing "no association" from a non-significant result.
  • Claiming causation from observational data.
  • Comparing with studies without explaining differences in population or design.

Key takeaways

  • Interpret in order: size, confidence interval, p-value, clinical importance.
  • Consider chance, bias, confounding and reverse causation.
  • Use careful, honest language for associations and non-significant results.
  • Understand your own analysis well enough to explain it at the viva.

References

  1. Greenland S, Senn SJ, Rothman KJ, et al. Statistical tests, P values, confidence intervals, and power: a guide to misinterpretations. Eur J Epidemiol. 2016;31(4):337–350. doi:10.1007/s10654-016-0149-3
  2. Altman DG, Bland JM. Absence of evidence is not evidence of absence. BMJ. 1995;311(7003):485.
  3. Kirkwood BR, Sterne JAC. Essential Medical Statistics. 2nd ed. Oxford: Blackwell Science; 2003.

This article is for education. Always follow the requirements of your university, guide, ethics committee and target journal.