Conventional wisdom suggests that assessment length is positively related to the likelihood that applicants will opt out of the assessment phase. However, restricting assessment length can negatively impact the utility of a selection system by reducing the reliability of its construct scores and constraining coverage of the relevant criterion domain. Given the costly nature of these tradeoffs, is it better for managers to prioritize (a) shortening assessments to reduce applicant attrition rates or (b) ensuring optimal reliability and validity of their assessment scores? In the present study, we use data from 222,772 job-seekers nested within 69 selection systems to challenge the popular notion that selection system length predicts applicant attrition behavior. Specifically, we argue that the majority of applicant attrition occurs very early in the assessment phase and that attrition risk decreases, not increases, as a function of time spent in assessment. Our findings supported these predictions, revealing that the majority of applicants that quit assessments, did so within the first 20 minutes of the assessment phase. Consequently, selection system length did not predict rates of applicant attrition. In fact, when controlling for observed system length and various job characteristics, we found that systems providing more conservative (i.e., longer) estimates of assessment length produced lower overall attrition rates. Collectively, these findings suggest that efforts to curtail applicant attrition by shortening assessment length may be misguided.