Using Data to Change Public Policy
This summer, I worked with Professor Judd Kessler from the Business Economics and Public Policy department. Of the 2 research projects that I was involved with, both dealt with the importance of data in changing public policy. The first project I focused on revolved around Transitional Grants and Loans for Teach For America (TFA). TFA is a prestigious teacher placement program that allows outstanding teachers to be placed at a low-income school district for at least 2 years. The research that I focused on was an extension of past research conducted in 2015, 2016, and 2017. Past research revolved around what amount and forms of grants led to the longest tenures amongst applicants. Now, we need to focus on following up with the TFA applicants, including those who successfully completed the program and those who rejected the program. Instead of introducing unnecessary bias through survey data, we utilized an automated academic Linkedin web scraper to search for TFA applicant profiles. Since the web scraper was not 100% accurate, it was my job to data scrape each applicant's profile thoroughly, ensuring to note their past and present professions, academic institutions attended, class years, and other identifying personal information. In total, I data scraped over 4,300 applicant profiles. This information will be used to analyze and understand how liquidity impacts career choices and how this might mitigate the impacts of the teacher shortage within the U.S.
The second project that I focused on was Incentivized Resume Rating (IRR), a methodology that aims to elicit employer preferences without bias, allowing both the recruiter and candidates seeking a job to acquire satisfaction without deception. IRR specifically involves employers evaluating hypothetical resumes. Each resume is created with randomized characteristics. Primary characteristics include leadership experience, internship experience, name, race, and gender. Of the hypothetical resumes that employers find suitable for their job openings, real resumes will be matched to satisfy employer preferences. This creates an incentive for both candidates and employers which is what motivates the deception-less method. Part of IRR involved understanding its applications in broader spheres. As a result, I conducted literature reviews on IRR applications in discrimination in the venture capital industry, benefits of obesity in low-income countries, gender differences in higher education, organ transplantation, and more. It was equally important to create a new tool for IRR to better understand whether employers can identify candidate characteristics based upon leadership and activities without seeing the candidate's name. To do this, I data scraped 674 resumes and analyzed the initial data collected to identify leading demographics indicators.