Some 40,000 new HIV infections occur each year in the United States, and 1 in 7 of those living with HIV are unaware that they have the virus. We analyzed data from two U.S. health surveys, the Behavioral Risk Factor Surveillance System (which conducts interviews by phone) and the National Health and Nutrition Examination Survey (which includes in-person interviews and physical examinations) separately to look at the prevalence of ever testing for HIV and how testing changed over the period 2011-2016. We employed t tests and linear contrasts to determine whether prevalence remained stable, increased, or decreased. We compared results from parametric regression models (linear and logistic regression) to confirm monotonic trends and to estimate annualized rates of change. We used quadratic contrasts to examine non-monotonic trends during the period. Results for the two surveys were similar, although the vastly different sample sizes for the two surveys affected the magnitude of changes we could detect. In this presentation, we comment on how we interpreted analytical results and communicated methods for both internal reports and publications.