Armstrong and Vogel (2009) proposed that the differences between self-efficacy and interests are a matter of measurement artifact rather than substance. In tests of this hypothesis, they conceived of self-efficacy and interest as observed indicators of larger RIASEC (Realistic, Investigative, Artistic, Social, Enterprising, and Conventional) types and as response method factors. We revisit the authors' theoretical assumptions, measurement procedures, analyses, and interpretation of findings. When viewing this study in the context of the larger literature, we find ample support for the construal of self-efficacy and interests as distinct but related constructs. In addition, we examine the authors' reanalysis of earlier longitudinal findings, reaching different conclusions than they did about the nature of the temporal relations among the social cognitive variables. Ultimately, whether one wishes to highlight or minimize the differences between interest and self-efficacy may largely depend on whether one's purpose is explanation (e.g., how do people make career-relevant choices?) or classification (e.g., which RIASEC type does a person most resemble?). (Contains 1 table.)