Walking in (the University of) Memphis: Which College Campuses Opened in Fall 2020? PDF
In Fall 2020, due to the COVID-19 pandemic, colleges faced the unusual decision of whether or not to open their campuses to students, and if so, how extensively. Some colleges opened fully, or near so, while others shifted entirely online. Which attributes explain the level of reopening for US two- and four-year institutions? I use mobile phone location data to produce a continuous measure of the level of reopening at each college. Some college features, such as number of dorms, private status, student racial mix, and out-of-state prevalence predict the degree of reopening. However, external local cues are highly important, with local COVID prevalence, the behavior of nearby large colleges, local political environment, and spatial autocorrelation meaningfully explaining variation in reopening levels.
Keeping an Eye Out: Behavior in the Presence of Riskless Ambiguity. CSU Fullerton Department of Economics Working Paper 2018/001. PDF
Past research on choice under ambiguity - decisions made when the probability of each outcome is unknown - has typically focused on scenarios in which ambiguity is presented alongside risks with known probabilities. Understanding the response to ambiguity is then difficult to distinguish from the response to risk, confounding our understanding of how subjects behave in conditions of ambiguity. In this paper I take advantage of the fact that all decisions are necessarily ambiguous to some extent to develop a scenario in which ambiguity is present but is otherwise risk-free. Respondents participate in a task that is very similar to a multi-armed bandit but with no random variation in payouts. I show that respondents anticipate ambiguity in payouts without prompting, continue to do so for long periods of time, respond differently to gains and losses in this context, and that this tendency is independent of risk preference. I offer a basic, clean framework for creating a riskless ambiguous environment for subjects, and argue that these exploratory results make the case for further study of ambiguity in riskless environments.
Most Likely to Succeed: Personality and Long-Run Economic Outcomes” with Elaina Rose.
We estimate the effect of college seniors’ personality traits on long-run career and economic outcomes. The analysis uses a novel data set that merges information from the 1980 - 1984 West Point yearbooks with detailed alumni records. Personality traits were inferred from peer-written narratives and faculty-conferred distinctions. The narratives were evaluated by West Point alumni from non-study classes and the multiple reports were aggregated using a Latent Class Model (LCM). We find that cadets identified by faculty as leaders, and those characterized by peers as leaders, determined, or social prove to be more successful in terms of military and civilian career outcomes 31-35 years after graduation.
The Search: The Effect of the College Scorecard on Interest in Colleges (2016). PDF.
Abstract: The College Scorecard is a website launched by the U.S. Department of Education in September 2015 that provides information about different colleges. This paper studies the effects of the website on interest in colleges, calculating both intent-to-treat and a local average treatment effect estimates of the effect of the College Scorecard. Interest is measured using Google search activity. The Scorecard led to more searches for keywords associated with high-earnings, high-graduation rate, and low-tuition colleges. However, the size of the effect is very small.
Diverging Prices Among Novel Goods with Alasdair Young (2017). CSUF Department of Economics Working Paper 2017/002. PDF.
Abstract: It is well-established that many markets exhibit price dispersion in equilibrium. However, relatively little attention is paid to how dispersion might be expected to change over time. The presiding assumption is that if dispersion is not already in equilibrium, it should naturally decrease over time. We develop a model in which shifts in demand combined with menu costs lead dispersion to naturally rise for some novel goods. We test the predictions of the model in the online market for collectible cards and find that price dispersion increases over time for more than half of the goods in the market.
College Curricular Dispersion: More Well Rounded or Less Well Trained? with Dan Goldhaber, Mark Long, and James Cowan (CEDR Working Paper 2015-6). PDF.
Students are typically given a large amount of freedom to choose the level of “curricular dispersion”: the tight focus or lack thereof in the courses they elect to take while in college.There is little evidence about what predicts students' curricular dispersion, whether it affects later college or labor force outcomes, or, in fact, how to measure of curricular dispersion. In this paper we develop a measure of curricular dispersion and use data from Washington State to explore its predictors and associated outcomes. We find that prior dispersion predicts future dispersion but not subsequent changes in college major. We report mixed findings on the associations between curricular dispersion and overall college GPA, the probability of graduation, and early career wages.
Consumption Value and the Demand for College Education (2016) PDF.
Abstract: The author estimates student demand for college education as a function of expectations of future earnings and consumption value using a detailed measurement of the non-financial consumption value of college education. Consumption value drives the educational plans of high school students more than does future earnings. Student responses to direct experiential utility and the avoidance of psychic costs are especially strong, but background and social encouragement are also important. Investment-focused economic models of educational choice focus on a relatively small part of the decision.
Are College-Bound Students Aware of What's "Promised" to Them? Evidence from Washington State with Grant H. Blume (2013). PDF.
Abstract: Students making decisions about their education must learn about the costs of attending a given college and the financial aid available to them. However, previous research demonstrates that students and parents have a poor understanding of the actual amount of aid available. Institutional promise scholarships, which advertise and ensure the availability of full tuition funding to low-income students, are an example of a policy designed to affect behavior largely by changing the information that students have about financial aid. We look for evidence that these policies relate to student information about financial aid. Using a data set of aid expectations from high school students in Washington State we find that low-income students expect more funding than their middle- and high-income peers, but that low-income students' expectations of aid do not fully incorporate all available information and are not sensitive to the presence of an institutional promise scholarship.