This chapter reviews theoretical developments and empirical studies related to causal inference on social networks from both experimental and observational studies. Discussion is given to the effect of experimental interventions on outcomes and behaviors and how these effects relate to the presence of social ties, the position of individuals within the network, and the underlying structure and properties of the network. The effects of such experimental interventions on changing the network structure itself and potential feedback between behaviors and network changes are also discussed. With observational data, correlations in behavior or outcomes between individuals with network ties may be due to social influence, homophily, or environmental confounding. With cross-sectional data these three sources of correlation cannot be distinguished. Methods employing longitudinal observational data that can help distinguish between social influence, homophily, and environmental confounding are described, along with their limitations. Proposals are made regarding future research directions and methodological developments that would help put causal inference on social networks on a firmer theoretical footing.

Social Networks and Causal Inference

It is difficult to find an area within Western sociology that has been as thoroughly debated and examined as social stratification, and within the stratification sphere no single set of issues has commanded the attention of theoreticians, researchers, and social policy makers more than that which surrounds social mobility and its antecedents.' Although the conclusions that are drawn about rates of social mobility and their governing processes vary somewhat from country to country, two main features seem universal: First, the link between occupational prestige and socioeconomic status (or social class, depending on which side of the debate you fall) is both direct and strong. Second, the major precondition for acceptance into most of the occupations which rank at or near the top of the prestige hierarchy is an extensive amount of formal education. Since neither of these points comes as a surprise to the average layman, elaborate documentation of these kinds of facts is often viewed as a trivial exercise.2

https://www.jstor.org/stable/info/1084192

Status, Achievement, and Motivation in the American High School

or even unspecified psychological dimensions of the self. This expectation is consistent with historical emphasis of symbolic interactionists on the importance of Cooley's (1902) looking-glass self, and a belief that the self is seen to arise through interaction with others (Fine, 1993). More recently, this emphasis has been extended to include specific attention to role identities, role salience, and role performance (Hoelter, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1983, 1984, 1985, 1987; McCall & Simmons, 1966; Stryker, 1980, 1982), but the present measure is unique in that is has been customized to fit an adolescent population. Self-esteem and self-image are important variables to include in a model of parent-teen socialization, the first representing a construct well-researched in the literature, and the second representing a theoretical innovation which is appropriate to a role-based multi-level model of socialization. One of the key effects of the nature and quality of parent-teen interaction and relationships may be a significant influence on adolescent self-image or self-esteem (Barber, 1992; Clark, 1994; Demo, 1992; Hoffman, Ushpiz & Levy-Shiff, 1988). For both male and female adolescents, parental appraisals of worth and performance are very important, and for male adolescents, perhaps the most important determinant of self-evaluation (Hoelter, 1984). Self-appraisal or evaluation can be viewed as the first and foremost result of the socialization process, the culmination and of all macro-level and micro-level social influences, and the internalization of all significant social relationships. To the extent that it defines the evaluation of one's assumed or adopted orientation to the external social world, it can also serve as a hinge upon which many other significant behavioral and societal outcomes rest. As such, self-appraisal serves as an appropriate ending point in a comprehensive model of parent-teen socialization. A Brief Review of Chapter II In this chapter, I have suggested that parental socialization of adolescents may be appropriately and productively studied with multi-level modeling. Characteristics for effective multi-level models were enumerated, including the need for greater theoretical integration of model elements. A general model of social influence (Figure 2.1) was developed which locates variables of social influence within four levels of the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. social order: the macro system, exo system, meso system, and micro system (Bronfenbrenner, 1979). Social influence was described as any process by which the actual, imagined, or implied presence of persons, groups, or culture changes the essence or behavior of the same. This concept was differentiated from the more specific type of social influence we referred to as parental socialization o f adolescents, which specifies the actors involved in the social influence chain, the direction of influence being analyzed (from the social system and parents, as agents of the system, to teens), and the fact that the broad patterns of interactive influence on general teen outcomes were to be analyzed, not specific behaviors. An applied model of parent-teen socialization (Figure 2.2) was derived from this more general model of social influence, and variables included in the model were justified from the research literature and by theoretical derivation. To structural variables of social class, family type, and employment were added interactional variables such as parent-teen contact time, parental subroles, activity categories, and quality of parent-teen relationship. Teen outcome variables self-esteem, self-image, and school grades were presented as appropriate dependent variables in the socialization model. Research on all of the key model variables was reviewed, and the need for including all of the specified variables within the same multi-level model was stressed as a way to measure differential and interactive effects of model elements, and the combined effect of the overall model, on the dependent variables. The next chapter describes the research design used to test the model and to provide the empirical foundation for all three research papers. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Toward a Comprehensive Model of Parent-Teen Socialization: The Antecedents and Consequents of Parent-Teen Interaction

/pdf/student-achievement-as-related-to-the-teacher-role-student-210i02wcod.pdf

Student Achievement as Related to the Teacher Role, Student Role and Their Interface

The critical limitation of the interpretation of schools as independent normative systems is the failure to demonstrate that observed student orientations are acquired in schools rather than in other settings such as the home or community. The problem is approached here as a question of whether students' orientations vary with lengths of membership in the same school organization. Data were examined for 7,954 students in 19 schools which contained grade-lengths 7-9, 7-12, 9-12, and 10-12. The analyses indicated that the distinctiveness of school normative climates is slightly but positively associated with lengths of school membership when school size is controlled. While schools can be viewed as relatively distinct systems, they should also be seen as highly permeable in the midst of their community and home environments.

The Issue of Schools as Social Systems: Socialization Effects as Inferred from Lengths of Membership*

A study of a school system was undertaken to replicate, partially, Coleman's research and critically evaluate the subcultural model he used. Using Parsons' distinction between instrumental and integrative roles, students chose desired characteristics of leaders for each type of school role. The characteristics were divided into instrumental and integrative; the hypothesis being that each type of characteristic would be selected for the appropriate type of role. Findings indicate that, contrary to Coleman, students selected instrumental characteristics for both instrumental and integrative roles. These findings raise doubts about the utility of the subcultural model for understanding adolescent roles and social organization.

Stanley S. Robin

Papers

A Social System Analysis of Preferred Leadership Role Characteristics in High School.