Peer Quality or Input Quality?: Evidence from Trinidad and Tobago

Using exogenous secondary school assignments to remove self-selection bias to schools and peers, I obtain credible estimates of (1) the effect of attending schools with higher-achieving peers, and (2) the direct effect of peer quality improvements within schools, on the same population. While students at schools with higherachieving peers have better academic achievement, within-school increases in peer achievement improve outcomes only at high-achievement schools. Peer quality can account for about one tenth of school value-added on average, but over one-third among the top quartile of schools. The results reveal large and important differences by gender. There is mounting evidence indicating that attending higher-achieving schools improves students' test scores (Hastings and Weinstein 2007, Pop-Eleches and Urquiola 2008, Jackson 2009) and broader outcomes such as course taking, secondary-school graduation, disciplinary incidents, and arrest rates (Culllen, Jacob and Levitt 2006, Clark 2007, Deming 2010).In most nations, high-achieving schools have student bodies with better incoming achievement than the average school. This may be due to residential sorting by socioeconomic status in conjunction with local funding for public schools, explicit across-school ability-grouping, or selection into schools. Since students may benefit directly from exposure to higher-achieving peers (C. Hoxby 2000, Hoxby and Weingarth 2006, Sacerdote 2001, Zimmerman 2003), part of the benefit to attending a high-achievement school may be attributed to the direct benefits of having higherachieving peers. Supporting this notion, researchers have found that parents may chose schools based on the potential peer group rather than a productive advantage (Willms and Echols 1992, Black 1999, Rothstein 2006) ─ suggesting that an important part of the benefits to attending a school with higher-achieving peers has to do with the peers themselves. From a policy perspective understanding the relative contribution of direct peer effects is important, as it speaks to the ability to replicate success across schools. For example, if having high concentrations of Catholic students (whose parents place a higher value than average on education) engenders an environment particularly conducive to learning, it would imply that 1 This is closely related to studies showing positive effects of attending private schools (Angrist, Bettinger, et al. 2002, Rouse 1998) and the effect of attending Catholic schools (Evans and Schwab 1995, Neal 1997). 2 Also, in the long run, the quality of other school inputs such as teachers may be endogenous to peer quality (Hanushek, Kain and Rivkin 2004, Jackson 2009, Boyd, et al. 2008). As such, while students may benefit from attending schools with higher achieving peers, it is unclear how much of the benefits to attending a better school are through the direct effect of being exposed to higher achieving peers.