If all students in a particular course receive high grades, it is an example of

Inhaltsverzeichnis Show

JEL classification
Data availability
Cited by (0)
Introduction
Section snippets
Empirical methodology
Data and descriptive statistics
The effect of course repeaters on first-time course-takers
Acknowledgements
Recommended articles (6)
What are two types of affectivity?
Which of the following is an example of an extrinsic work value?
Which of the following refers to the perceived fairness of the outcome received?
Which term refers to the degree to which people accept as normal an unequal distribution of power?

PDFView PDF

Under a Creative Commons license

Open access

Highlights

•

I examine the relationship between university students’ appearance and grades.

•

When education is in-person, attractive students receive higher grades.

•

The effect is only present in courses with significant teacher–student interaction.

•

Grades of attractive females declined when teaching was conducted remotely.

•

For males, there was a beauty premium even after the switch to online teaching.

Abstract

This paper examines the role of student facial attractiveness on academic outcomes under various forms of instruction, using data from engineering students in Sweden. When education is in-person, attractive students receive higher grades in non-quantitative subjects, in which teachers tend to interact more with students compared to quantitative courses. This finding holds both for males and females. When instruction moved online during the COVID-19 pandemic, the grades of attractive female students deteriorated in non-quantitative subjects. However, the beauty premium persisted for males, suggesting that discrimination is a salient factor in explaining the grade beauty premium for females only.

JEL classification

D91

I23

J16

Z13

Keywords

Attractiveness

Beauty

COVID-19

Discrimination

Data availability

Data will be made available on request.

Cited by (0)

Introduction

The questions of whether low-achieving students should be retained in a grade or required to repeat a failed course are answered by the extent to which grade or course repetition affects the retained or repeating individual and the extent to which grade or course repeaters affect their classmates. An extensive literature has investigated the effect of grade retention on the individual, but there is less evidence on the potential effects of grade or course repeaters on their classmates.1 This paper provides new evidence that course repeaters negatively affect their course-mates in US high schools. Using unique longitudinal transcript data in which individual and school-specific course fixed effects control for ability and course selection, an increase in the share of repeaters in a given high school mathematics course causes an increase in the probability of course failure for first-time course-takers.

Many US students fail high school mathematics courses. Seven percent of students in the data used in this paper repeat a failed mathematics course in high school, and this increases to fifteen percent for students taking Algebra I. Furthermore, there has been a widespread push to raise standards in American high school education.2 In the past decade, many US states have both increased the number of mathematics credits required for high school graduation and specified particular mathematics courses that need to be passed (Reys, Dingman, Nevels, & Teuscher, 2007). Tables 8 and 9 in the appendix summarize the mathematics requirements for high school graduation by state. The median state requires students to obtain three credits of high school mathematics and pass Algebra I in order to graduate. Media reports indicate that these requirements may have increased the likelihood of course repetition for students who fail high school mathematics courses, and, in particular, reveal diverse public opinion on whether Algebra I should be required for graduation (Hacker, 2012, Helfand, 2006). The effects of repetition in high school mathematics courses are clearly important to understand. The negative externalities exerted by repeaters on their classmates found in this paper suggest a cost to course repetition ignored by much of the previous analyses, and, to the extent that policies such as requiring Algebra I to graduate increase course repetition, a cost to minimum mathematics requirements that may have been overlooked by policy-makers.3

Understanding the externalities imposed by repeaters in high school mathematics courses may also inform the grade retention debate. This is because both grade retention and course repetition result in students being exposed to a set of low-achieving classmates who are likely to share similar characteristics.4 To the extent that repeating and retained students exert similar externalities on their classmates, this paper suggests grade retention analyses should include effects exerted on classmates of the retained individual.

The paper uses a fixed effects strategy on longitudinal transcript data for multiple cohorts of US high school students to estimate the causal effect of course repeaters on their course-mates, students taking the same course but not necessarily in the same class. Essentially, the study compares the achievement of first-time course-takers in the same mathematics course (such as Algebra I) in the same high school in different years using year-to-year variation in the share of repeaters in the course to identify the effect. It is assumed that unobserved year-specific shocks to education production in the previous year provide variation in the share of course repeaters in the current year. These may be classroom-specific shocks (such as increased teacher absenteeism in a specific course causing a higher course failure rate that year) or be caused by natural variation in cohort course aptitude.

The academic achievement of first-time course-takers is shown to be negatively correlated with the share of repeaters in the course that year. First-time course-takers are more likely to fail when exposed to greater shares of course repeaters, and this is particularly evident when the share of repeaters reaches a threshold of five to ten percent of the course.

Grading to a curve cannot be ruled out as a cause of the effect. It is possible that repeaters push a subset of first-time course-takers down a potentially fixed grade distribution, although it is argued that other features of the data are inconsistent with grading to a curve. The use of course failure rather than GPA as the outcome of interest is partly motivated by the consideration that even if effects are generated by a “nominal”’ rather than “real” phenomenon, course failure may have serious long-run consequences. Rose and Betts (2004) find that advanced high school mathematics courses have greater effects on students’ earnings a decade after graduation than less advanced courses, and Joensen and Nielsen (2009) establish a causal relationship between advanced high school mathematics and earnings, so to the extent that course failure determines the decision to enroll in advanced mathematics courses, the estimated course spillover may affect labor market outcomes. Course repetition externalities are therefore important irrespective of their cause and should be considered when thinking about high school course progression and failure policies.

Course repeaters may exert externalities on their course-mates in a variety of ways. These course composition effects can be grouped into two categories: general effects arising from repeaters being low-achievers and specific repeater effects not exerted by other low-achievers. Low-achieving students are likely to disproportionately extract teacher inputs or redirect teacher inputs away from first-time course-takers. They may need more time to understand concepts, slowing the pace of the class, and may also be more likely to misbehave in the classroom given that disruptive behavior is generally correlated with classroom ability, requiring teacher intervention. Low-achieving classmates may also be more likely to directly distract their classmates, lowering education production even without affecting teacher inputs.

In addition to these low-achiever effects, course repeaters may exert additional externalities specifically related to failing and retaking a course. They may be bored and inattentive when encountering course material for the second time, increasing the likelihood of disruptive behavior. Repeaters may also have a poor attitude or be uncooperative because they failed the course the previous year, and this may negatively affect both their classmates and the teacher.5

Course repeaters may also exert externalities through course size, and, for courses with more than one class, class assignment. Course size effects are captured by using the share rather than number of repeaters in a course. Class assignment may matter if repeaters are assigned to classes non-randomly. For example, repeaters may be assigned to the best teacher for a particular course if failing for a second time is particularly costly (either from the perspective of the school or the student). This may increase the likelihood of first-time course-takers being assigned to another class with a worse teacher, leading to poorer performance for first-time course-takers.6

The primary focus of this study is an analysis of the combined low achiever and repeater effects that course repeaters exert on their course-mates. This is the appropriate level of analysis for an overall evaluation of course repetition effects. Further results attempt to separate the general low-achiever and specific repeater externalities. This has important policy implications. The negative externalities course repeaters exert may persist irrespective of subsequent course assignment if the effects are driven by repeaters being low achievers; they will retain this property in most courses, although it could be minimized in less mathematically-challenging courses. However, if the effect is specifically related to course repetition, the negative externality could be eliminated by failed students not repeating the course.

The results in this paper are best compared with those obtained by Lavy, Paserman, et al. (2012). Defining low-ability students as students who are old for their grade (most likely having repeated kindergarten or first grade), they find that the proportion of low-ability peers is negatively correlated with the academic achievement of regular students. Variation in the composition of seven adjacent cohorts of 10th grade students in Israeli high schools (from 1994 to 2000) is used to identify the effect. It is argued that the majority of students had little experience with their peers prior to entering high school, so results are not driven by common cohort-specific shocks.

This paper has three key distinctions from Lavy, Paserman, et al. (2012). First, we observe course enrollment and achievement for all students in a set of high schools for multiple years allowing the inclusion of both individual and school-specific course fixed effects. This approach deals with potentially confounding individual effects (such as cohort-specific shocks and ability differences) and course effects (such as repeaters being more likely to repeat difficult courses) that cannot be dealt with using repeated cross-sectional data. Second, it isolates the effects of course-mates rather than grade-mates. Students in the same grade may have little interaction and may not take many of the same courses, which would attenuate effects for analyses performed at the grade level. And, third, it focuses on high school mathematics courses in the US, which is particularly relevant given policies stipulating minimum mathematics requirements for graduation in US high schools increasing the likelihood of mathematics course repetition.

Repeaters are low-achieving peers for first-time course-takers. Results can therefore be compared with the literature investigating ability peer effects in high school. These papers exploit a variety of identification strategies and typically find moderately sized, negative achievement effects for individuals exposed to low-ability peers.7 In particular, Lavy, Silva, and Weinhardt (2012) find evidence of negative peer effects at the bottom of the ability distribution for English secondary school students, which is consistent with the finding that course repeater effects operate at the low and middle parts of the ability distribution. Burke and Sass (2013) investigate peer effects using Florida public schools data, finding that ability peer effects are stronger at the class level than at the grade level.

This paper also contributes to the broader literature on spillovers in the classroom. For example, Carrell and Hoekstra (2010) find that children exposed to domestic violence at home significantly decrease the test scores of their peers and increase misbehavior in the classroom, and Aizer (2008) uses exogenous timing in ADD diagnosis and treatment to show that improvements in peer behavior increase student achievement. Externalities exerted by course repeaters are related, but worthy of particular attention as course failure and repetition is under the control of the school, therefore offering a more direct opportunity for policy intervention.

Finally, the externalities exerted by course repeaters may also be placed in the context of the related literature investigating the effects of grade retention. To the extent that the effect of grade retention on the retained is similar to the effect of course repetition on the repeating, we can draw on results from this literature to infer whether negative course repetition externalities are likely to be mitigated or aggravated by effects of course repetition on the individual. The literature investigating the causal effect of retention on the retained has exploited a variety of policies to overcome selection into retention. It provides evidence of positive effects of retention for students up to the third grade (Dong, 2010, Eide and Showalter, 2001, Greene and Winters, 2009, Jacob and Lefgren, 2004),8 but flat or negative effects for older students (Jacob and Lefgren, 2004, Jacob and Lefgren, 2009, Manacorda, 2012). Fruehwirth, Navarro, and Takahashi (2011) recognize that retention effects are likely to differ by the grade at which the student is retained and the unobservable behavioral and cognitive abilities of the student. They allow for heterogeneous effects in their econometric model and obtain generally negative effects from retention, suggesting grade retention is not an effective policy for raising the performance of low-ability students. If grade retention and course repetition effects on the individual are in the same direction, combining the results from the grade retention literature with the findings in this paper, course repetition may be costly both to the repeating student and their course-mates.

This remainder of this paper is organized in the usual way: methodology, data, results and then interpretation.

Section snippets

Empirical methodology

This paper considers course-specific rather than class-specific externalities. A course-level analysis has two favorable features. First, as discussed in the introduction, externalities may extend beyond the classroom if repeating students are systematically assigned to better (or worse) teachers. A class-level analysis would not capture these effects. And, second, we do not need to be concerned about sorting into classes within courses. For example, students may be assigned to classes

Data and descriptive statistics

This paper uses data from the National Longitudinal Study of Adolescent Health (Add Health). The Add Health is a school-based longitudinal study of a nationally representative sample of US adolescents who were in grades 7–12 during the 1994–1995 school year. A core sample was selected to participate in a series of detailed surveys. Complete high school transcript data (grades 9–12) are available for individuals selected for the core sample.11

The effect of course repeaters on first-time course-takers

Table 5 reports the primary set of results. The first three columns report results for Eq. (2). Controls are added sequentially across columns. The coefficient on the share of students failed and repeating of 0.26 in the first column is estimated without school-course and individual fixed effects, showing a positive correlation between the share of repeaters in a course and course failure rates. The specification in the third column includes school-course fixed effects to control for course

Conclusion

Mathematics is difficult for many students, and course repetition in high school mathematics courses is common. This repetition is partly promoted by policies in several US states that stipulate a minimum level of mathematics for high school graduation. As shown by Rose and Betts (2004) and Joensen and Nielsen (2009), mathematics is also important for future job market success, acting as further encouragement for students to repeat failed mathematics courses. This paper takes a new step by

Acknowledgements

I wish to thank Nicole Fortin for providing many hours of advice and supervision, Thomas Lemieux and Craig Riddell for very helpful suggestions, and David Green, Kevin Milligan, Marit Rehavi, and seminar participants at the University of British Columbia and University of South Carolina for useful feedback. Errors remain my own.

This research uses data from Add Health, a program project designed by J. Richard Udry, Peter S. Bearman, and Kathleen Mullan Harris, and funded by grant P01-HD31921