The Impact of Classroom Peer Groups on Pupil GCSE Results (original) (raw)
Related papers
2010
The magnitude and characteristics of the effect of a child's peers on their outcomes has long interested researchers and policy makers. In this paper, I take advantage of the correlation between the average outcomes a child's peer group attains with the distribution of ages within the cohort to construct an instrument for the ability of the peer group in order to estimate the peers effects on children's outcomes at age 11. IV results suggest there is a significant positive effect of a more able peer group. Furthermore, the results suggest that there is more benefit for children who are close to the ability of the peer group than those whose ability is not close.
National Bureau of Economic Research, 2009
In this paper, we study ability peer effects in secondary schools in England and identify which segments of the peer ability distribution drive the impact of peer quality on students achievements. To do so, we use census data for four cohorts of pupils taking their age-14 national tests, and measure students ability by their prior achievements at age-11. We employ a new identification strategy based on within-pupil regressions that exploit variation in achievements across the three compulsory subjects (English, Mathematics and Science) tested both at age-14 and age-11. We find significant and sizeable negative peer effects arising from bad peers at the very bottom of the ability distribution, but little evidence that average peer quality and very good peers significantly affect pupils academic achievements. However, these results mask some significant heterogeneity along the gender dimension, with girls significantly benefiting from the presence of very academically bright peers, and boys marginally losing out.
2009
In this paper, we study ability peer effects in secondary schools in England and identify which segments of the peer ability distribution drive the impact of peer quality on students achievements. To do so, we use census data for four cohorts of pupils taking their age-14 national tests, and measure students ability by their prior achievements at age-11. We employ a new identification strategy based on within-pupil regressions that exploit variation in achievements across the three compulsory subjects (English, Mathematics and Science) tested both at age-14 and age-11. We find significant and sizeable negative peer effects arising from bad peers at the very bottom of the ability distribution, but little evidence that average peer quality and very good peers significantly affect pupils academic achievements. However, these results mask some significant heterogeneity along the gender dimension, with girls significantly benefiting from the presence of very academically bright peers, and boys marginally losing out.
Peer Heterogeneity, School Tracking and Students’ Performances: Evidence from PISA 2006
2011
The empirical literature using large international students' assessments tends to neglect the role of school composition variables in order not to incur in a misidentification of peer effects. However, this could lead to an error of higher logical type since the learning environment crucially depends on peer variables. In this paper, using PISA 2006, we show how peer heterogeneity is a key determinant of students' attainments. Interestingly, the effect of peer variables differs depending on the country tracking policy: peer heterogeneity reduces efficiency in comprehensive systems whereas it has a non-linear impact in early-tracking ones. In turn, linear peer effects are larger in early-tracking systems. Results remain robust in both student-and school-level regressions and when we add school-level dummies and several controls correlated with the school choice to alleviate the selectivity bias. JEL: I21, I28, J24
Economics of Education Review, 2012
Investigation of peer effects on achievement with sample survey data on schools may mean that only a random sample of the population of peers is observed for each individual. This generates measurement error in peer variables similar in form to the textbook case of errors-in-variables, resulting in the estimated peer group effects in an OLS regression model being biased towards zero. We investigate the problem using survey data for England from the Programme for International Student Assessment (PISA) linked to administrative microdata recording information for each PISA sample member's entire year cohort. We calculate a peer group measure based on these complete data and compare its use with a variable based on peers in just the PISA sample. We also use a Monte Carlo experiment to show how the extent of the attenuation bias rises as peer sample size falls. On average, the estimated peer effect is biased downwards by about one third when drawing a sample of peers of the size implied by the PISA survey design.
Peer heterogeneity, school tracking and students'performances: evidence from Pisa 2006
2011
This paper analyses the interaction between school tracking policies and peer effects in OECD countries. Using the PISA 2006 dataset, we show that the linear peer effects are stronger and more concave-shaped in the early-tracking educational system than in the comprehensive one. Second, and more interestingly, the effect of peer heterogeneity goes in opposite directions in the two systems. In both student-and school-level estimates, peer heterogeneity reduces students' achievements in the comprehensive system while it has a positive impact in the early-tracking one. For late tracking countries, this result appears driven by pupils attending vocationally-oriented programs. Finally, peer effects are stronger for low ability students in both groups of countries. JEL Codes: I21, I28, J24.
Attainment in secondary school
Oxford Economic Papers, 1999
This paper studies attainment in secondary schools. We estimate an education production function in which attainment depends upon parental inputs, peer group inputs and schooling inputs. We find that the most powerful parental input is parental interest in children, as assessed by teachers. We find a strong peer group effect. The school pupil-teacher ratio does not enter significantly. The only strongly endogenous variable is initial attainment. We argue that this is due to measurement error. There is some evidence that parental interest is endogenous but we do not find peer group variables to be so.
Measuring Peer Effects in the Brazilian School System
2011
This paper investigates the existence and magnitude of peer e¤ects among pupils in Brazilian schools using a dataset on achievement of …fth-graders (around eleven years old) in Math, which is accompanied by detailed questionnaires completed by students, teachers and principals. We address the self-selection of group formation through a selection on observables hypothesis and use teacher's wage as a proxy for unobserved teacher's ability. A robustness check using as instrumental variable a question answered by the principal on how classrooms are formed at school was performed and similar results were found. Our results suggest that both contextual (exogenous) and endogenous peer e¤ects are important determinants of students' achievement in Brazil. Furthermore, by being able to separate out contextual e¤ects from endogenous peer e¤ects, we show that even though boys perform slightly better than girls in Math, children in classes with larger proportions of boys tend to have worse results in Math.
2008
Administrative datasets frequently lack information on the protocol used to assign subjects to treatment conditions. In this paper, we show that under certain conditions, consistent causal estimates can be recovered from data where the actual assignment protocol is uncertain, but the probability of random assignment is known. Predictably, estimators are less consistent when the probability of random assignment is unknown and must itself be inferred. We introduce a method of inferring the probability of random assignment ex post based on a scalar measure of the degree of balance in observed characteristics across treatment conditions. We then apply this method to the study of peer effects in education production. Results indicate that a considerable portion of OLS peer effects estimates reflect selection – intuitively, the results tend to be stronger in schools that depart obviously from random assignment. One relatively robust result, however, is that average achievement is higher i...
Peer effects in private and public schools across countries
Journal of Policy Analysis and Management, 2000
Many argue that the composition of a school or classroom-that is, the characteristics of the students themselves-affect the educational attainment of an individual student. This influence of the students in a classroom is often referred to as a peer effect. There have been few systematic studies that empirically examine the peer effect in the educational process. In this research, we examine the peer effect with a unique data set that includes individual student achievement scores and comprehensive characteristics of the students' families, teachers, other school characteristics, and peers for five countries. The data allow an examination of peer effects in both private and public schools in all countries. Our analysis indicates that peer effects are a significant determinant of educational achievement; the effects of peers appear to be greater for low-ability students than for high-ability students. The finding is robust across countries but not robust across school type.