Peer effects on the educational outcomes of immigrant youth: heterogeneity by generation and school context

ABSTRACT

Previous research suggests that peer social networks affect immigrant youths’ academics. However, little is known about whether peer parental education spillovers on educational outcomes differ by generational status and whether students are exposed to more same-ethnic peers in school. This study tests: (1) whether attending classes with peers with educated parents is associated with academic outcomes, (2) whether this relationship varies by generation, and (3) whether school contexts with more peers of the same race/ethnicity influence peer parental spillover effects. To overcome the fact that students select into schools, we use the National Longitudinal Study of Adolescent to Adult Health and leverage quasi-exogenous variation in peer exposure. We find that peer parental education spillover effects vary by immigrant generation, but the extent of the difference varies by outcome. For academic effort, first-generation immigrant youth are less affected by peer parental education effects compared with others. For GPA, the peer spillover effect is smaller for second-generation youth compared with others. In schools with more same race/ethnicity peers, children of immigrants see lower peer spillover effects on educational outcomes compared with their counterparts in more diverse contexts. Findings suggest that network based social processes influencing academic outcomes are distinct for immigrant youth.

KEYWORDS:

• Peers
• immigrant generation
• school context
• assimilation
• education

Acknowledgements

This research uses data from Add Health, a programme project directed by Kathleen Mullan Harris and designed by J. Richard Udry, Peter S. Bearman, and Kathleen Mullan Harris at the University of North Carolina at Chapel Hill, and funded by grant P01-HD31921 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development, with cooperative funding from 23 other federal agencies and foundations. Special acknowledgment is due Ronald R. Rindfuss and Barbara Entwisle for assistance in the original design. Information on how to obtain the Add Health data files is available on the Add Health website (http://www.cpc.unc.edu/addhealth). No direct support was received from grant P01-HD31921 for this analysis. Analysis scripts are available at GitHub (https://github.com/socpub/peer_effects). The data that support the findings of this study are available from the Carolina Population Center. Restrictions apply to the availability of these data, which were used under a license for this study. Data are available at https://data.cpc.unc.edu/ with the permission of the Carolina Population Center.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 This method is known to work effectively especially in randomized trials where the treatment is not endogenous because randomization implies that baseline covariates are balanced between treatment and control groups (Groenwold et al. 2012). In supplementary analyses, we found that, despite a large reduction in the sample size, complete case analysis (listwise deletion) yields substantially similar results as the ones reported in this paper (Table S7 of the Online Supplementary Data file).

2 Scores from principal component analysis produced the same pattern of results in multivariate analyses (Table S8 of the Online Supplementary Data file).

3 Parental education was measured by the highest years of schooling completed by either parent. The number of years shown in parentheses was assigned to each response: eighth grade or less (8); more than eighth grade, but did not graduate from high school (10); high school graduate (12); completed a GED (12); went to a business, trade, or vocational school after high school (13); went to college but did not graduate (14); graduated from college or a university (16); professional training beyond a four-year college (16); She went to school, but I don't know what level (9); and She never went to school (0).

4 In addition to its methodological merit (i.e. addressing school selection and unobserved school-level confounding), there are additional reasons to define classmates as a peer group of interest in this study. First, unlike close friendships, classmates provide a stable and consistent normative environment for adolescents to make sense of their own and peers’ attitudes and behaviors (Legewie and DiPrete 2012; Choi et al. 2008). Second, determining the effect of classmates may be policy relevant because a wider group composition can be better targeted and subject to direct policy influence compared with close friendship dyads (Billings, Deming, and Rockoff 2014; Cook and Ludwig 2006).

5 Previous sociological studies on peer effects have employed a similar methodological approach as that used in the current study (Hermansen and Birkelund 2015; Lee and Lee 2020; Legewie and DiPrete 2012).

6 See Fletcher, Ross, and Zhang (2020) for a discussion of the potential bias induced by this mechanical correlation. Despite this possibility, however, the results obtained without using the Guryan et al. correction were virtually identical to those reported in this paper.

7 For those who reported multiple racial/ethnic backgrounds, we assigned percentage of school peers with a more obvious racial/ethnic background (Black, Hispanic, Asian, and others in order).