Working Papers

Using High-Stakes Grades to Incentivize Learning
Abstract: I investigate whether policymakers can increase human capital production by introducing merit-based enrollment through a natural experiment in Norwegian high schools. By exploiting variation across space and time I compare the performance of students taking the same exit exam in compulsory school, but where the test is high-stakes for only a subset of students. Using a staggered triple-difference framework, I find that exam grades increase in the high-stakes setting if students have a sufficient number of prospective schools within traveling distance. Results from low-stakes ability assessments suggest actual learning—and not test-taking strategy—could largely explain the effect.

R&R, Economics of Education Review

Reducing the Gender Gap in Early Learning: Evidence From a Field Experiment in Norwegian Preschools
(With Mari Rege, Ingeborg F. Solli, and Ingunn Størksen)

Abstract: Early-childhood programs attract considerable policy interest as a tool to prepare children for formal schooling. However, we have limited knowledge about whether conditions for early learning are similar for all children. Program attendance seems to have less effect on boys, but so far there is little evidence suggesting reasons for why this might be the case. In this field experiment, we investigate whether a more structured curriculum can reduce the gender gap in early learning. While girls have higher skills at baseline, we find that the intervention primarily benefits boys, with effects persisting into formal schooling.

Under review, European Economic Review

Alumni Satisfaction, Rankings, and College Recommendations
(With Eric Bettinger)

Abstract: While college-access organizations, state and federal governments invest millions of dollars in informational campaigns, rankings, and college websites, students often rely more on parents and trusted adults for information on college options. We use new data from a nationally representative survey of adults to identify the extent to which those who attended college are willing to recommend their own university experience. We demonstrate that alumni place more value on their personal experiences than prominent rankings in formulating recommendations. Interestingly, individuals with low economic outcomes also express high levels of satisfaction, and willingness to recommend their own college experience to others. We discuss the implications of such strong preferences and the potential of satisfaction as an additional metric for college evaluation.

Under review, Educational Evaluation & Policy Analysis

The Role of Incentives for Improving Student Motivation and Performance

Abstract: Even though the returns to education can be substantial, many students put too little effort into their schooling. To mitigate this underinvestment problem, policymakers are often eager to try to motivate students using extrinsic incentives, such as cash payments and merit scholarships, stricter grading standards, and more competitive admission processes. The design, scope, and implementation of such incentive policies have been a fruitful area of economic research over the last 30 years. However, the evidence on their potency for improving student performance is mixed at best. In particular, the use of extrinsic incentives often elicit strategic responses from students, resulting in behavior that might improve performance metrics, but are not productive in producing human capital. Many incentive policies therefore end up producing unintended consequences that goes contrary to the policy objective. This essay surveys some of the most commonly used incentive mechanisms, their effect on academic achievement, and discuss avenues for future research on student incentives.

Invited submission, Oxford Research Encyclopedia of Economics & Finance

Work in Progress

Exposure to High-Stakes Testing and Attitudes Towards Math

Abstract: With growing demand for STEM-skilled labor, policymakers are concerned with encouraging more students to pursue science and technology-related education. Emerging evidence indicate that early performance signals can have substantial effect on the choices students make with regards to course profiles or field of study. However, there is still much to be learned about how and why such signals shape student beliefs and affect future investment decisions. I revisit data collected for a randomized controlled trial aimed at promoting growth mindset among Norwegian high schoolers, and leverage unique features of the Norwegian exam system to identify the causal effect of being exposed to a high-stakes test on subsequent attitudes towards mathematics — a key prerequisite subject for most STEM degrees. I demonstrate that students randomly selected to sit for a math test, as opposed to a similar test in a language subject, report having less anxiety towards mathematics, having more confidence that they will do well in mathematics in the future and being more interested in math as a topic. Administrative records show that these students are also more likely to choose advanced math subjects in high school. Although average effects are sizable and robust, they are particularly strong among students for whom math is a relatively weak subject and those who overachieve on the test, relative to teacher assessed grades.

Sibling Spillovers in High-Stakes Testing

Abstract: Despite their ubiquity, the role of siblings in the education production function has long been overlooked in economic research. However, a small, but emerging literature is now turning its focus towards direct spillovers in academic outcomes between siblings. While this body of research so far has focused on demonstrating the existence of such spillovers, we know little to nothing about the potential channels through which these effects might flow. In this paper I argue for one potential hypothesis, which I refer to as the inside information channel. Specifically, the hypothesis posits that students with older siblings have an advantage stemming from the fact that older sibling having already completed the schooling that the younger sibling is currently facing, and therefore has useful information and experiences to share. To investigate the plausibility of this mechanism I exploit the particularity that all Norwegian students graduating from compulsory school are obligated to sit for a standardized, written national exam, with a randomly drawn subject. The paper explores to what extent being randomly drawn to the same subject as your older sibling is beneficial, compared to students where the older sibling had a different subject. My results indicate that being drawn to the same exam subject on average increases exam grades by 1 - 1.5\% of a standard deviation. However, this effect size masks considerable heterogeneity, as I estimate treatment effects for boys and students from low-income households that are 3 - 5 times larger. These results suggest that vulnerable student subgroups could disproportionately benefit from such inside information and mentoring.

Family Structure, Parental Investment Behavior, and Human Capital Development
(With Maximiliaan Thijssen)

For teaching experience and other work activities please refer to my CV