Appraise Evidence Beyond the P Value

Read estimate, uncertainty, and importance together

Start with the effect estimate in its units. A mean difference of 2 points, a risk ratio of 0.70, and a risk difference of 8 percentage points describe different quantities. Then examine the confidence interval for values compatible with the data and model. For a difference, the null value is zero; for a ratio, it is one. A narrow interval suggests greater precision than a wide interval, but precision does not correct bias. Finally compare the entire interval with a clinically meaningful threshold. A tiny effect can be statistically detectable in a very large sample, while a potentially important effect can remain uncertain in a small sample. Practical interpretation requires magnitude, direction, precision, outcome importance, time horizon, harms, and applicability, not a p-value label alone.