With New Grant, Researchers Aim to Provide Community College Graduates with Personalized Job Matches


Audrey Breen

A research team at UVA aims to improve the job search process so that investments in a college education translate into good jobs for graduates.

Millions of Americans have lost employment during the COVID-19 crisis. Returning to college to obtain additional training and credentials may be important for many adults who seek new employment as the country moves towards economic recovery.

With nearly $300,000 in new funding from Ascendium Education Group, a research team at the University of Virginia is focusing their attention on how to improve the job search process so that investments in a college education translate into good jobs for graduates from the Virginia Community College System and the College System of Tennessee.

The research team will develop a personalized job matching algorithm and an accompanying career advising and nudging model that, together, aims to support lower-income community college students in their efforts to find employment after graduation.

“With this grant, we will develop an algorithm that will provide graduates with personalized information about currently open jobs that are well-aligned with their program of study and academic performance,” said Brian Kim, a doctoral student in education policy at the UVA School of Education and Human Development and one of the lead researchers on the project.

“Ascendium is pleased to support this project which aims to smooth the transition between postsecondary education and workforce entry for community college students from low-income backgrounds by developing scalable tools that provide personalized information and support for these learners,” said Carolynn Lee, program officer at Ascendium. “The research team shares Ascendium’s interest in aligning postsecondary and workforce data systems. This project will help learners find success in the labor market as they are exposed to jobs they are well-qualified for and provide stable, family-sustaining employment and compensation.”

The algorithm will also consider the individual student’s academic performance, as well as jobs that similar students from prior cohorts have had success pursuing.

“Through this project we want to put this data into action to support better labor market outcomes, especially for lower-income community college graduates in Virginia and Tennessee,” said Ben Castleman, Newton and Rita Meyers Associate Professor in the Economics of Education at UVA and researcher working on the project.

According to the researchers, community college students can lack access to the high-quality information, career advising, or social and professional networks to find good jobs. Similar to having less information about how to apply to college or for financial aid, students may not have the same social network and guidance from which to draw potential job opportunities.

“Many of the best practices and norms of applying for jobs for college graduates, like how to write a compelling cover letter or resume, might be unclear to those inexperienced with the process,” Castleman said.

“To generate these recommendations, we will combine rich data from prior cohorts, which we have the opportunity to work with through our multi-year research-policy partnership with VCCS with historical job posting data,” Kim said.

“The College System of Tennessee is committed to the mission of student success and workforce development,” said Russ Deaton, executive vice chancellor for policy and strategy at TBR—The College System of Tennessee. “Central to this mission is colleges’ work to serve and improve the condition of individuals, families, and the communities across the state. With our partners at the University of Virginia, we are eager to develop data insights that help Tennessee’s community colleges provide students with timely and personalized information about employment opportunities.”

Using machine learning methods, the team will train the algorithm to identify well matched jobs for VCCS graduates. They will refine the algorithm through feedback from data science experts and institution leaders and then evaluate the performance of the algorithm to identify areas for improvement.

The researchers will also develop a career advising and nudging model that will encourage students to meet with a career advisor and connect the algorithm’s job matches to the student. While recent studies have shown the diminishing impact of large-scale generic nudge interventions, the team will test the impact of providing customized job-match information, delivered primarily through career advising, to improve the workforce outcomes for community college graduates.

With a proof-of-concept algorithm created and tested in collaboration with VCCS and data science experts, the team will integrate the advising and nudging intervention and begin identifying partners with whom they will pilot and study the program in 2022.

“Ultimately, we hope to improve the workforce outcomes of lower-income community college students by providing them with well-matched jobs,” Castleman said.

This project is one of several that are part of the Virginia Policy Partnership Collaboration (VPPC). The VPPC connects University of Virginia faculty and students with education policy makers to address pressing education problems in the Commonwealth through careful research.

###