Because of the multifactorial nature of neuropsychological tests, attention remains poorly defined from a neuropsychological perspective, and conclusions made regarding attention across studies may be limited due to the different nature of the measures used. Thus, a more definitive schema for this neurocognitive domain is needed. We assessed the applicability of Mirsky and Duncan's (2001) neuropsychological model of attention to a cohort of 104 HIV+ adults. Our analysis resulted in a five-factor structure similar to that of previous studies, which explained 74.5% of the variance. However, based on the psychometric characteristics of the measures comprising each factor, we offer an alternative interpretation of the factors. Findings also indicate that one factor, which is generally not assessed in clinical neuropsychology settings, may be more predictive of real-world behaviors (such as medication adherence) than those composed of traditional measures. Suggestions for further research in this important area are discussed.