"Old Peer Effects on Cognitive and Noncognitive Skills: Mixing up vs Keeping Classmates in Elementary School," with Juan Murguia, Iowa State University

Thu, Oct 18, 2012 3:40 pm - 5:00 pm @ 360 Heady Hall

Abstract: I study old peer effects on early education achievement by exploding theunique nature of the Tennessee Student Teacher Achievement Ratio (STAR).I benefit from the randomized mixing up policy of STAR program in theidenti cation of old peer effects and estimate value-added models with and without school fixed effects. Specifcally, I analyze the relationship between the chance of passing first grade, as well as noncognitive skills, and the proportion of kindergarten classmates kept into first grade. Results show that keeping all kindergarten classmates, vs. losing all of them, increases the probability of passing first grade by 12 percent. In addition, noncognitive skills are improved with more kindergarten classmates kept into first grade. Interestingly, females show less old peer effect on grade passing but more on self-confidence than males. This paper presents evidence supporting the idea that the duration of peer connections is relevant in the estimation of peer effects and that mixing up classes in early education might be detrimental to the development of cognitive and noncognitive skills.