Immigration of different ethnicity entering the United States is a common phenomenon over the history, which causes the variety of the composition in U.S. society. This paper categorizes four types of immigration status and five types of ethnicity for adolescents in the U.S. between 1994 - 1995. Individuals with similar background are likely to make friends because of homophily. This clustering behavior has impact on adolescents' outcomes. We use the National Longitudinal Study of Adolescent Health (Add Health) data to study how the peer effects, when considering different immigrant status and ethnicity, affect the education outcomes. For the friendship information, we match the friends using the Add Health friendship nomination data. The purpose of this paper is to study when taking the immigrant status and ethnicity into consideration, how the peer effects affect the education outcomes. Using a modified linear-in-mean model, we also try to figure out whether the endogenous and exogenous peer effects can be identified separately.

In this paper, immigration status is defined as four different categories: Native, First generation immigrant, Second generation immigrant and Other immigration status. Native in this paper follows the definition that an individual whose parents were both U.S. born and the individual was born U.S. citizen regardless of his birthplace. For First generation immigrants, both their parents and they were foreign born and they came to U.S. themselves. Some of the non-U.S. born migrants settled in U.S. and started to reproduce in their second hometown. Their offspring is second generation immigrants. They were born in a family with the feature of the parents' origin in the context of U.S. society. Similarly, we construct ethnicity variable with five different possibilities: White, Black, Hispanic, Asian and Other Ethnicity.

This paper will study how an individual's academic achievement is affected by peer effects from his reference group. Manski (1993) defines endogenous and exogenous peer effects. Applying his definition, endogenous peer effect is how the individual academic grades tends to vary with the average achievement of the friends in his network when controlling all else. Exogenous peer effect is how the individual achievement tends to vary with the socioeconomic composition of the reference group.

This paper takes into consideration the immigration status and ethnicity of the peer groups. We want to study the peer effect from the same-type friends and different-type friends. For example, compared with friends with different immigration status, friends with same immigration status might have different peer effects on the subject’s outcome. One of the reasons is that individuals tend to cluster because of their similar features. This is explained in sociology as homophily, the tendency of individuals to associate and bond with similar others. The connections which make people feel comfortable being with similarities include age, gender, class, organizational role, beliefs, values and education, etc. This paper is studying the social network with or without the similarity in the ethnicity and immigration status. We want to identify the endogenous and exogenous peer effects from same-type peer group and different-type peer group.