Increasing STEM undergraduate participation in innovative activities: Field experimental evidence.

Workers trained in STEM are generally viewed as essential for innovation-led economic growth. Yet, recent statistics suggest that a majority of STEM undergraduates do not go on to pursue innovation-focused careers in their fields of study. We investigate whether STEM students who do not self-select into innovative tasks are doing so because they are less capable than their peers who do. We find that monetary inducement among STEM students increases aggregate innovative output, but that low-GPA students who were induced significantly underperform relative to low-GPA students who self-selected; however, induced and self-selected high-GPA students perform statistically the same. In contrast, words of encouragement appears to benefit those students with the lowest GPAs. Our results highlight the value of efforts to increase the pool of STEM students who pursue innovative careers and underscore the importance of interventions targeted at specific student subgroups to maximize the returns on those efforts.