Panel Paper: The Effect of Relative Age On Education and Crime

Thursday, November 8, 2012 : 3:40 PM
Preston (Sheraton Baltimore City Center Hotel)

*Names in bold indicate Presenter

Songman Kang, Philip Cook and Clara Muschkin, Duke University

Researchers repeatedly find that older students score higher on standardized tests than their younger peers, especially during the early years of schooling.  This potential benefit of relative maturity on academic achievement has motivated many parents to delay sending their children to school for a year.  However, it remains unclear whether the academic advantages older children have in the early years of formal schooling persist over time, and influence other long-run life outcomes.  In this paper, we examine how relative maturity influences academic achievement and educational attainment over the K-12 years and criminal outcomes during adolescence.

Our empirical analyses are based on administrative data on the cohort of third graders enrolled in North Carolina public school system in 1999-2000.  Our datasets contain information on their birth certificates, test scores from standardized reading and math exams, high school transcript, juvenile criminal complaint and adult crime conviction records.  We exploit age variation among students enrolled in North Carolina public schools, caused by the state school entry policy.  A child in North Carolina is eligible to enter kindergarten only if he reaches the age of 5 on or before October 16 in the relevant year.  As a result, children born in November 1990 are likely to be assigned to the same grade with children born in October 1991, while those born in October 1990 are likely to be placed in a full academic grade above. We use this policy-induced age variation among academic cohorts to estimate the effects of relative age on educational and criminal outcomes.  In order to account for academic redshirting and early grade retention, we instrument students' age with their birth month and estimate the relationship between a series of academic and criminal outcomes and predicted age.

The IV estimation results show that relative age has strong and persistent effects on academic outcomes throughout primary and secondary education.  Older students tend to score higher on standardized tests and are less likely to repeat grades between grades 3 and 8.  This academic advantage appears to be persistent; older students are more likely to take honors/AP courses in the last year of high school.  At the same time, we find that older students are less likely to graduate from high school; a one-year difference in age in third grade is associated with a 6.5% difference in graduation rate.  These advantages and disadvantages are in turn reflected in criminal outcomes as juveniles and adults.  Consistent with the in-school academic advantages older students have, the rate of juvenile criminal complaints is lower among older students.  The adverse effect on high school graduation seems to negate the advantage, however, as older students are more likely to receive criminal convictions as adults.