What is the null hypothesis in an outbreak analytic study, and when would you reject it?

• A. Null: there is no association between exposure and disease; reject if p < alpha or CI excludes the null value.
• B. Null: exposure causes disease.
• C. Null: data are perfect.
• D. Null: there is an association.

Answer: (C)

In outbreak analytic studies, you’re testing whether exposure and disease are related. The null hypothesis is that there is no association between exposure and disease. You would reject this null when the data show a statistically significant result, typically if the p-value is below the chosen alpha level or if the confidence interval for the measure of association excludes the null value (for a risk ratio or odds ratio, that null value is 1). If you don’t have statistical significance, you don’t reject the null. The idea that data are perfect isn’t the hypothesis being tested; hypotheses concern population-level relationships, not data quality.