How to explain statistical results at the defence

What exactly was compared?

Before showing a table, clarify group composition and observation timing. Distinguish independent groups from repeated measurements in the same participants. Keep sample sizes and measure labels on the slide. Explain any change in observation count.

Mean and variability

State what follows ±. Standard deviation describes variation in observations, while standard error concerns the precision of an estimated mean. Label the quartiles accompanying a median. ICMJE recommendations emphasise identifying statistical measures and providing clear table notes.

Interpretation

Describe the magnitude and direction of a difference before the statistical assessment. A p-value does not replace a substantive answer. Association alone does not establish causal direction. For a prediction model, explain validation and the data used to assess it.

Cross-checking materials

Compare numbers across the dissertation, summary and presentation. Check rounding, group names and figure captions. Keep a separate table linking objectives to analysis methods so the basis for a calculation can be found quickly.

The question before the number

Start with what was compared or assessed: the groups, measure and time point. Then give the number and explain its meaning within the research question. This lets the audience follow the argument without decoding the table at the same time.

Two group means alone do not tell the audience whether observations are independent or measurements were collected consistently. These are design details. Prepare them beforehand so the statistics discussion need not reconstruct basic analysis conditions.

Sample sizes and missing values

Check the observation count for each calculation. The full dataset may contain more records than a particular table because of missing values, different assessment stages or selection criteria. Reporting the actual n makes the result clearer.

Do not silently replace missing values with zeros. Missing-data handling belongs in the analysis plan and requires justification. Prepare a brief explanation of the available data, records included and reasons. Keep these details consistent in the table, talk and dissertation.

Explaining complex notation

Create a short glossary of your notation, covering means, medians, standard deviations, interval estimates and association measures. Explain each in relation to your study. Do not substitute one measure for another merely to shorten a label.

A p-value on a slide should accompany a specific hypothesis and comparison. Do not reduce the scientific conclusion to one number. Discuss direction, magnitude, precision and limitations. Wording must match the analysis performed rather than the expected effect.

Supporting calculations

Keep the final table, methods description, data-processing log and reproducible calculations in a format your team can use. Prepare a backup slide explaining the most complex result. Exclude participants’ personal data from public materials.

Rehearse questions about sample sizes, method choice and the meaning of the main result. If answering requires searching among random file versions, organise the materials first. Good preparation lets you explain where a number comes from and the limits of its interpretation.