<p>The gender achievement gap in higher STEM education is well documented as STEM course grades often show a bias favoring men, even when controlling for overall student achievements. The current work, which was preregistered prior to data collection, replicates and extends prior work to evaluate several possible contributors to the gender gap in advanced STEM courses. The contributing factors include students’ STEM-preparedness, learning assessment design, and students’ activity level in the course. We analyzed the data of a cohort of 315 undergraduate engineering students in a leading technological university in Israel. A mixed-effect model analysis revealed a gender gap in both high-stakes and low-stakes assessments of learning. We found that the gap first appears in lower-division undergraduate mathematics courses and remains steady through advanced STEM courses, with relatively little variance between different courses. We further found that students’ activity in the course’s learning management system is not associated with a gap reduction. The analysis suggests that the gender gap can be traced back, at least in part, to students’ STEM-preparedness when entering college. Notably, although the gap is traced prior to college, high-school grades likely show a ceiling effect and are thus not representative of the gap shown later in college. Therefore, it may be that other preparatory factors contribute to this gap, outside of or in addition to high-school performance.</p>