Poster #125 - Skill Information on Job Postings and Student Sorting in the Co-op Market

• In Event: Friday Poster Reception

Abstract

Job postings are an important part of how employers and jobseekers find one another. In the labor market literature, research has shown that job postings influence who self-selects into submitting a job application and before formal screening begins. This project extends such work to co-ops, work-based learning experiential opportunities of college students where they take one semester off to work with an employer. We examine how the skill profile--their rarity and specificity--of co-op postings structure early career matching and inequality. Our research questions are 1) how do the skill profile requirements in co-op postings relate to the number of applications they receive? And 2) how are students sociodemographic characteristics associated with the skill profile of the co-op jobs they apply to? Using data from a private university (job postings=39,210 and applications=1,510,099, students=26,603), we combine students’ applications with their sociodemographic characteristics and the description of the job postings from 2020 to 2025. The platform made visible all co-op jobs to students. We extract the skills using ESCO and ESCOX, a transformer- and cosine-similarity-based library. ESCO is especially useful here because it is open source, transparent, expert-validated, and has substantial overlap with O*NET. After normalizing our extracted skills, I construct two measures: a rarity index that captures how rare a posting’s skill bundle is relative to the co-op market each year, and percent specificity, the share of skills classified as occupation-or sector-specific. Common skills in 2025 are Python, communication, and Microsoft office while operating specialized equipment is among the rarest. Preliminary results show that job postings with higher rarity of skills are strongly associated with lower application volume. Postings with higher skills specificity also attract less diverse applicants pool in terms of race and gender. Job postings where the applicants are largely from tech or business-related fields of study receive more applications on average compared to other fields, but the pool is also less racially diverse. Student-level analysis show that while field of study is an important predictor of which skills profiles students apply for, it does not entirely eliminate racial differences. For example, controlling for field of study, Asian students tend to apply to co-op job postings that have a higher percentage of specific skills than white students, while Hispanic and international students apply less to these types of postings: this is consistent with the idea that underrepresented minorities have unequal access to labor market information or professional networks. These results suggest that how skills are incorporated in job postings relates who applies where and may reproduce unequal outcomes in a setting where all co-op job opportunities are practically visible to all. Implications of this work suggest that colleges can relocate efforts from late-stage placement (e.g. resume review, interview preparation) to early-stage application formation (e.g. pre-search advising and identifying when students decide which co-op jobs are for them). Reducing informal uncertainty around skill requirements, exposing students from underrepresented groups and encouraging them to apply to jobs with high specificity co-op jobs can also be another area of intervention.

Author

• Enrique Eduardo Valencia Lopez, University of California Irvine
  Presenting Author