This paper uses a detailed example to illustrate how to detect individual observations that disproportionately influence the results of hypothesis testing with relative risk regression models.