This dissertation consists of three essays in the fields of labor economics and education economics. The first chapter examines the effect of residential segregation on neighbor-based informal hiring. Existing works in the neighborhood effects literature have documented mixed evidence of community characteristics on employment and earnings. Yet most studies are silent on or unable to pinpoint the exact mechanisms that drive their results, making it hard to reconcile the conflicting findings. As a departure, in the first chapter, I start with a specific mechanism—job search via neighbor networks—and explore how segregation at the place of residence affects employment through this channel. In the remaining essays, I turn to the economics of education.
The second chapter homes in on a specific unintended consequence of standardized testing: cheating between students on exams. While outright cheating is a common tactic to cope with grade pressure, it has received little attention from economists. My second chapter thus contributes to the sparse literature on how inordinate emphasis on exams can distort student behavior even to the test day. Finally, the third chapter revisits the popular belief that education necessarily improves cognitive skills. Insofar as one of the primary goals of school is to develop student intellect, are all years of schooling created equal? Along these lines, I question the value of the first year of high school to Chinese students in the context of the Program for International Student Assessment (PISA), an educational initiative of the Organization for Economic Cooperation and Development (OECD).
To briefly recap, the first chapter studies how residential segregation by race and by education affects job search via neighbor networks. Using confidential data from the US Census Bureau, I measure segregation for each characteristic at both the individual level and the neighborhood level. Causality is achieved by interweaving a spatial framework with a mover-stayer design. The spatial component entails comparison between individuals across different hyperlocal neighborhoods (blocks) within the same reference area (block group). The focus of the comparison is existing residents on a block (stayers) with respect to newcomers (movers). Specifically, I ask: what is the likelihood that an incumbent resident—conditional on changing jobs—will join a firm that has employed a new neighbor on their block? How is this probability mediated by residential segregation? My answers to these questions are manifold. At the individual level, I find that future coworkership with new block neighbors is less likely among segregated stayers than among integrated stayers, irrespective of races and levels of schooling. The impacts are heterogeneous in magnitude, being most adverse for the most socioeconomically disadvantaged demographics: Blacks and those without a high school education. At the block level, however, higher segregation along either dimension raises the likelihood of “any” future coworkership with new block neighbors for all racial or educational “groups.” My hybrid identification strategy, capitalizing on data granularity, allows a causal interpretation of these results. Together, they point to the coexistence of homophily and in-group competition for job opportunities in linking residential segregation to neighbor-based informal hiring. My subtle findings have important implications for policy-making.
The second chapter is an investigation of student cheating on high-stakes exams, a relatively understudied topic in the economics of education. The setting is Vietnam, the relevant assessment is the country’s national high school exit exams, and the (mis)behavior of interest is cheating between non-elite students and elite students who happen to sit in the same test room on test day. To quantify the pervasiveness of this misconduct, I exploit the quasi-random assignment of students from schools of varying quality into test rooms. Using micro-data from a large Vietnamese province, I find that the fraction of elite students in the same room has significantly positive effects on non-elite students’ scores when the tests are non-competitive (2007-2013). The effects are concentrated in the multiple-choice/quantitative subject tests and absent in the essay/qualitative subject tests. The average gains due to same-room elite density vary across subjects and can be as large as one point on a 0-10 grading scale. However, these effects disappear after an exam redesign in 2015 raises the stakes of the assessment (2018-2019). Similar patterns emerge when instead of quantity, the quality of the elite students in the same room is the main explanatory variable. Backed by institutional details, these findings provide credible evidence that discreet interpersonal cheating is present pre-reform, but vanishes as the reform reshuffles student incentives.
The third and last chapter explores the implications for cognitive skills of increased absolute schooling at an important juncture in a student’s academic career: transition between junior high school (or middle school) and senior high school (or high school). In particular, I ask if the first year of senior secondary school (Grade 10) affects 15-year-old Chinese students’ performance on the PISA 2015. Using a fuzzy regression discontinuity design, I find that on average, this additional year of schooling has no discernible effects on Science, Math, and Reading test scores. However, there is evidence that in the PISA 2015, Chinese tenth graders have fewer hours of in-school class time in these subjects and enjoy Science and peer cooperation less than comparable Chinese ninth graders. These observations add to the disappointment left by the lack of effects on test scores, even when they are insufficient to explain it away.
| Identifer | oai:union.ndltd.org:columbia.edu/oai:academiccommons.columbia.edu:10.7916/rt1k-9057 |
| Date | January 2022 |
| Creators | Mai, Tam |
| Source Sets | Columbia University |
| Language | English |
| Detected Language | English |
| Type | Theses |
Page generated in 0.003 seconds