Two columns I wrote to explain p-values to clinicians have been posted as open access. They went through many drafts being tested on non-statisticians until they seem to explain the issue adequately. Read them in order:
The goal of this column is to help working clinicians understand the statistical calculations involved in interpreting the results of diagnostic questionnaires. Using the Patient Health Questionnaire-9 as an example, the author explains how to determine the most appropriate cutoffs to choose, depending on the population involved, the probabilities of error ( and ), and the expected losses associated with each kind of error in the context in which the test is being administered. (Journal of Psychiatric Practice. 2011;17:57–60).
Hypothesis tests and p-values
The goal of this column is to help working clinicians understand p-values and the machinery of hypothesis testing underlying them, using a fictional study of an anti-insomnia drug. The author discusses what p-values really mean and common misconceptions concerning them. (Journal of Psychiatric Practice 2011;17:288–291)