scispace - formally typeset
Search or ask a question
Journal ArticleDOI

Peer effects and academic achievement: a regression discontinuity approach

01 Oct 2013-Economics of Education Review (Pergamon)-Vol. 36, pp 108-121
TL;DR: In this article, a fuzzy regression discontinuity approach was used to identify ability peer effects among teenagers, where assignment into high-ability classes constitutes the source of identifying information and significant and sizable effect on the academic achievement of students around the assignment threshold.
About: This article is published in Economics of Education Review.The article was published on 2013-10-01 and is currently open access. It has received 47 citations till now. The article focuses on the topics: Academic achievement & Regression discontinuity design.

Summary (2 min read)

1 Introduction

  • Despite this belief there is no general consensus on the direction of the effect peers have on one another.
  • The problem when it comes to estimating peer effect is that, as the saying goes, birds of feather flock together1, and the same applies to students.
  • Furthermore, once students learn wheather they are in a normal or HA class their outside options are rather limited if they decide to drop out.
  • Jumps in the relationship between assignment grade and grades by the end of the first year in the neighborhood of the HA class threshold can therefore be taken as evidence of a treatment effect.
  • I will follow the same strategy in this paper.

2 Previous Literature

  • When identifying the causal effect of peer ability on educational outcomes two issues are particularly challenging.
  • Random class assignment is not that common in higher education, so using this method to test for ability peer effects is seldom feasible and researchers must therefore resort to other methods to identify a causal effect of peers’ ability in observational studies.
  • Schneeweis and Winter-Ebner (2007) use the Austrian subsample of the PISA data and employ school type fixed effects and school fixed effects to estimate peer effects.
  • They find positive effects on the academic achievement of students who were randomly assigned to academically stronger peers in the non-tracking schools.
  • The indirect effect, stemming from the fact that teachers are able to teach at a level more appropriate to the average student, will very likely not be the same in developed countries where student heterogeneity is not as great.

3 The dataset

  • Compulsory education in Iceland is organized in a single structure system, i.e., primary and lower secondary education belong to the same school level, and generally take place in the same school.
  • Upper secondary education is not compulsory, but anyone who has completed compulsory education has the right to enter a course of studies in an upper secondary school.
  • At the end of their second year students receive the Commercial Diploma which corresponds roughly to A-levels in the United Kingdom and the High School Diploma in the United States.
  • In particular, the same teachers teach normal classes and HA classes.
  • Table (3) then shows the number of students in each class, where classes within each year are ranked according to their average assignment grade.

4 Empirical approach

  • This cutoff in the sorting of students into HA classes constitutes a valuable source of identifying information.
  • Intuitively, the continuity assumption requires that differential assignment into classes is the only source of discontinuity in outcomes around the assignment threshold, 0, so that unobservables vary smoothly as a function of assignment grade and, in particular, do not jump at the cutoff.
  • This system of streaming students into HA and normal classes has never been official and students were therefore, in most cases, not aware of the system until they had started their studies.
  • It seems implausible that this control was perfect, so the key identifying assumption is likely to hold here.

5 Results

  • The first crucial assumption for being able to apply the RD design is that there is an observable assignment variable on which assignment is based and that there is a discontinuity at some cutoff value of the assignment variable in the level of treatment.
  • The figures show that there is a discontinuity in normalized spring exam result and year grade around the assignment threshold and therefore present evidence that academic achieve- ment, as measured by spring exam results or year grade, is affected by being assigned into a HA class.
  • The same holds for year grades, the estimates obtained when using a second and third degree polynomial are again very similar, indicating a positive effect of 0.235 and 0.224 standard deviations, respectively and 0.221 when using a third degree polynomial and controls.
  • Internal validity of the RD approach is based on the local continuity assumption, i.e., that the conditional expectation of the outcome variable is continuous around the discontinuity point.
  • Sorting of individuals around the cutoff may lead to different average characteristics of those above and below the threshold so the internal validity of the results would break in this case.

6 Conclusions

  • I have estimated ability peer effects using data for five cohorts of age 16 in an Icelandic high-school where I measure peers’ ability by their academic ability as recorded by standardized test scores and test scores from their previous schools (elementary school).
  • From a methodological perspective, I view my main contribution to be the approach taken to measure peer effects, where student assignment into HA classes constitutes the source of identifying information.
  • As far as I know, this has never been done before.
  • In terms of findings, my results suggest that assigning students to classes with peers of higher academic ability increases their own academic performance.
  • My results therefore fall within this range but are close to the upper end.

Did you find this useful? Give us your feedback

Citations
More filters
Journal ArticleDOI
TL;DR: In this paper, the authors evaluate a tracking program in a large urban district where schools with at least one gifted fourth grader create a separate "gifted/high achiever" classroom.
Abstract: We evaluate a tracking program in a large urban district where schools with at least one gifted fourth grader create a separate “gifted/high achiever” classroom. Most seats are filled by non-gifted high achievers, ranked by previous-year test scores. We study the program's effects on the high achievers using (i) a rank-based regres sion discontinuity design, and (ii) a between-school/cohort analysis. We find significant effects that are concentrated among black and Hispanic participants. Minorities gain 0.5 standard deviation units in fourth-grade reading and math scores, with persistent gains through sixth grade. We find no evidence of negative or positive spillovers on nonparticipants. (JEL I21, J21, J24)

59 citations


Additional excerpts

  • ... Two recent non‐U.S. studies (Duflo, Dupas and Kremer 2011;  Vardardottir 2013) use more rigorous designs and find positive effects on upper‐tracked students....

    [...]

01 Jan 2016
TL;DR: In this paper, the authors evaluate a tracking program in a large urban district where schools with at least one gifted fourth grader create a separate "gifted/high achiever" classroom.
Abstract: We evaluate a tracking program in a large urban district where schools with at least one gifted fourth grader create a separate “gifted/high achiever” classroom. Most seats are filled by non-gifted high achievers, ranked by previous-year test scores. We study the program's effects on the high achievers using (i) a rank-based regres sion discontinuity design, and (ii) a between-school/cohort analysis. We find significant effects that are concentrated among black and Hispanic participants. Minorities gain 0.5 standard deviation units in fourth-grade reading and math scores, with persistent gains through sixth grade. We find no evidence of negative or positive spillovers on nonparticipants. (JEL I21, J21, J24)

57 citations

Journal ArticleDOI
TL;DR: Instrumenting PSQI scores by sleep quality during secondary education, it is found that increasing total sleep quality with one standard deviation leads to 4.85 percentage point higher course marks.

51 citations


Cites background from "Peer effects and academic achieveme..."

  • ...Seminal studies such as Black et al. (2005), Ermisch & Francesconi (2001), McNabb et al. (2002), Ortiz & Dehon (2008), Sanbonmatsu et al. (2006), Vardardottir (2013) and Leos-Urbel et al. (2013) have identified prior accumulated human capital, the costs and returns of higher education, social…...

    [...]

ReportDOI
TL;DR: This paper studied the impacts of assignment to separate gifted classrooms on three distinct groups of fourth grade students: non-disadvantaged students with IQ scores ≥130, subsidized lunch participants and English language learners with IQ score ≥116; and students who missed the IQ thresholds but scored highest among their school/grade cohort in statewide achievement tests in the previous year.
Abstract: Education policy makers have struggled for decades with the question of how to best serve high ability K‐12 students. As in the debate over selective college admissions, a key issue is targeting. Should gifted and talented programs be allocated on the basis of cognitive ability, or a broader combination of ability and achievement? Should there be a single admission threshold, or a lower bar for disadvantaged students? We use data from a large urban school district to study the impacts of assignment to separate gifted classrooms on three distinct groups of fourth grade students: non-disadvantaged students with IQ scores ≥130; subsidized lunch participants and English language learners with IQ scores ≥116; and students who miss the IQ thresholds but scored highest among their school/grade cohort in state-wide achievement tests in the previous year. Regression discontinuity estimates based on the IQ thresholds for the first two groups show no effects on reading or math achievement at the end of fourth grade. In contrast, estimates based on test score ranks for the third group show significant gains in reading and math, concentrated among lower-income and black and Hispanic students. The math gains persist to fifth grade and are also reflected in fifth grade science scores. Our findings suggest that a separate classroom environment is more effective for students selected on past achievement - particularly disadvantaged students who are often excluded from gifted and talented programs.

49 citations


Additional excerpts

  • ... Journal of Econometrics 142(2) (2008), pp. 698–714..    Pop‐Eleches , Cristian and Miguel Urquiola....

    [...]

  • ... A few recent studies have found benefits of tracking in non‐U.S.  settings—including Kenya (Duflo, Dupas and Kremer 2011), Trinidad and Tobago (Jackson 2010), Romania (Pop‐ Eleches and Urquiola 2013), and Iceland (Vardardottir 2013)....

    [...]

Dissertation
01 Jan 2016
TL;DR: In this paper, the authors present a framework for research context and design in the field of bioinformatics, and present a review of the results of their work. ___________________________________________
Abstract: .......................................................................................................................................................... i Declaration ..................................................................................................................................................... ii Acknowledgements ....................................................................................................................................... iii Publications ................................................................................................................................................... v Part 1 Research context and design .................................................................. 1 Chapter

25 citations


Cites background from "Peer effects and academic achieveme..."

  • ...This meant that they were not challenged in class and any increase that might have accrued to their own academic performance from studying alongside high achieving peers was foreclosed (Vardardottir 2013)....

    [...]

References
More filters
Book
01 Jan 1981
TL;DR: The Enlarged Edition as mentioned in this paper provides an overview of the evolution of the family and the state Bibliography Index. But it does not discuss the relationship between fertility and the division of labor in families.
Abstract: Preface to the Enlarged Edition Introduction 1. Single-Person Households 2. Division of Labor in Households and Families Supplement: Human Capital, Effort, and the Sexual Division of Labor 3. Polygamy and Monogamy in Marriage Markets 4. Assortative Mating in Marriage Markets 5. The Demand for Children Supplement: A Reformulation of the Economic Theory of Fertility 6. Family Background and the Opportunities of Children 7. Inequality and Intergenerational Mobility Supplement: Human Capital and the Rise and Fall of Families 8. Altruism in the Family 9. Families in Nonhuman Species 10. Imperfect Information, Marriage, and Divorce 11. The Evolution of the Family Supplement: The Family and the State Bibliography Index

9,096 citations

Posted Content
01 Jan 2009
TL;DR: The core methods in today's econometric toolkit are linear regression for statistical control, instrumental variables methods for the analysis of natural experiments, and differences-in-differences methods that exploit policy changes.
Abstract: The core methods in today's econometric toolkit are linear regression for statistical control, instrumental variables methods for the analysis of natural experiments, and differences-in-differences methods that exploit policy changes. In the modern experimentalist paradigm, these techniques address clear causal questions such as: Do smaller classes increase learning? Should wife batterers be arrested? How much does education raise wages? Mostly Harmless Econometrics shows how the basic tools of applied econometrics allow the data to speak. In addition to econometric essentials, Mostly Harmless Econometrics covers important new extensions--regression-discontinuity designs and quantile regression--as well as how to get standard errors right. Joshua Angrist and Jorn-Steffen Pischke explain why fancier econometric techniques are typically unnecessary and even dangerous. The applied econometric methods emphasized in this book are easy to use and relevant for many areas of contemporary social science. An irreverent review of econometric essentials A focus on tools that applied researchers use most Chapters on regression-discontinuity designs, quantile regression, and standard errors Many empirical examples A clear and concise resource with wide applications

7,192 citations

Journal ArticleDOI

5,474 citations

Journal ArticleDOI
TL;DR: In this article, the authors use a particular model for causal inference (Holland and Rubin 1983; Rubin 1974) to critique the discussions of other writers on causation and causal inference.
Abstract: Problems involving causal inference have dogged at the heels of statistics since its earliest days. Correlation does not imply causation, and yet causal conclusions drawn from a carefully designed experiment are often valid. What can a statistical model say about causation? This question is addressed by using a particular model for causal inference (Holland and Rubin 1983; Rubin 1974) to critique the discussions of other writers on causation and causal inference. These include selected philosophers, medical researchers, statisticians, econometricians, and proponents of causal modeling.

4,845 citations

Journal ArticleDOI
TL;DR: The authors examined the reflection problem that arises when a researcher observing the distribution of behaviour in a population tries to infer whether the average behaviour in some group influences the behaviour of the individuals that comprise the group.
Abstract: This paper examines the reflection problem that arises when a researcher observing the distribution of behaviour in a population tries to infer whether the average behaviour in some group influences the behaviour of the individuals that comprise the group. It is found that inference is not possible unless the researcher has prior information specifying the compisition of reference groups. If this information is available, the prospects for inference depend critically on the population relationship between the variables defining reference groups and those directly affecting outcomes. Inference is difficult to implossible if these variables are functionally dependent or are statistically independent. The prospects are better if the variables defining reference groups and those directly affecting outcomes are moderately related in the population.

4,495 citations

Frequently Asked Questions (1)
Q1. What are the contributions in "Peer effects and academic achievement regression discontinuity approach" ?

In this paper, I study ability peer effects in an Icelandic high school. An important feature of this system is that the same teachers teach high-ability and normal classes, both types of classes follow a common curriculum and all students take the same exams. Furthermore, the system is unofficial so students are in most cases not aware of it before they have started their studies. I find that sorting students into high-ability classes does have significant and sizable effect on the academic achievement of students around the assignment threshold, i. e., the results suggest that being assigned to a high-ability class increases academic achievement by 0. 23 standard deviations.