The Impacts of Hard and Soft Skills on STEM Careers for Low-Income Graduates

India has more engineers than almost any country, yet 43% of graduates are unemployed. Disadvantaged students are more likely to attend lower-quality engineering colleges, face discrimination, and have limited exposure to professional norms, soft skills, and networks. As a result, many women, low-caste, and low-SES students struggle to secure careers in STEM. DeepTech tackles these barriers by providing low-SES students with (1) hard skill training (e.g., programming languages), (2) soft skills mentorship (e.g., decision-making, awareness, and motivation), and (3) job applications guidance and networking support. Our questions are: Does DeepTech increase employment and earnings? Does adding soft skills mentorship (i) improve the previous outcomes and (ii) increase the retention of hard skills (through, e.g., improved self management in learning)? (iii) Is providing soft skills mentorship cost effective? (iv) For whom does soft skill mentorship matter the most?

We will randomly assign 70% of 5000 recent engineering graduates to Deep Tech’s hard skills component, and randomly offer 450-600 of these students the accompanying soft-skills mentorship training. We will measure employment outcomes (e.g., job placement, job type, time in employment, earnings) up to 2.5 years following course completion. We will use machine learning to study heterogeneous returns to understand for whom soft skills matter most. The evaluation will guide our partner’s scaling decisions and inform other programs on integrating soft skills. DeepTech addresses inequities in STEM (a high-earning sector) by redistributing opportunities to underrepresented groups, bridging skill gaps, and boosting productivity in the sector by broadening the talent pool.