Adherence studies have taken center stage due to the life-threatening risks associated with nonadherence to highly active antiretroviral therapy (HAART) regimens for people with HIV/AIDS. This study examines adherence through self-report of individuals on HAART regimens in a manner to account for demand characteristic bias, while still attempting to get an accurate portrayal of individuals' adherence to their medication. Results indicate that individuals who are more certain about their course of treatment, are more trusting of their medical provider, have family resources, and are more optimistic about life adhere better. The model predicts who will be adherent with 92% accuracy while accounting for 55% of the variance in adherence. Practice implications for increasing adherence and accurately assessing adherence are discussed. (Contains 3 tables.)