This chapter delineates two perspectives about the determinants of SJT performance: the contextualized perspective views SJTs as measures of job-specific knowledge, whereas the other perspective views SJTs as measures of general domain knowledge. Many current SJTs are situated somewhere between the two. Both perspectives are useful but have different SJT design implications. One perspective suggests further investing in more realistic stimulus and response formats. Conversely, the other perspective suggests streamlining SJTs. An important practical implication of the first perspective is the promise of improved predictive power involved in more realistic SJTs, while the second perspective posits that criterion- and construct-related validity would not suffer and indeed could benefit from designing more generic SJTs allowing broader predictions. This might especially hold for entry-level selection purposes because contextualization appears to be of higher importance for advanced-level selection. In the future, it seems beneficial that a clearer demarcation is used between these two perspectives. We also provide recommendations for practice and a research agenda for more comparative research between these two SJT perspectives in terms of key selection variables.