GENDER, FRIENDSHIP NETWORKS, AND DELINQUENCY: A DYNAMIC NETWORK APPROACH

• DOI: http://doi.org/10.1111/1745-9125.12052
• Date: 01 November 2014
• Author: BRIAN SOLLER,NATHAN J. DOOGAN,DANA L. HAYNIE
• Published date: 01 November 2014

GENDER, FRIENDSHIP NETWORKS,

AND DELINQUENCY: A DYNAMIC NETWORK

APPROACH∗

DANA L. HAYNIE,1NATHAN J. DOOGAN,2and BRIAN SOLLER3

1Department of Sociology, The Ohio State University

2College of Public Health, The Ohio State University

3Department of Sociology and Criminology, University of New Mexico

KEYWORDS: peer influence, social networks, gender

Researchers have examined selection and influence processes in shaping delinquency

similarity among friends, but little is known about the role of gender in moderating

these relationships. Our objective is to examine differences between adolescent boys

and girls regarding delinquency-based selection and influence processes. Using lon-

gitudinal network data from adolescents attending two large schools in AddHealth

(N =1,857) and stochastic actor-oriented models, we evaluate whether girls are in-

fluenced to a greater degree by friends’ violence or delinquency than boys (influence

hypothesis) and whether girls are more likely to select friends based on violent or delin-

quent behavior than boys (selection hypothesis). The results indicate that girls are more

likely than boys to be influenced by their friends’ involvement in violence. Although

a similar pattern emerges for nonviolent delinquency, the gender differences are not

significant. Some evidence shows that boys are influenced toward increasing their vi-

olence or delinquency when exposed to more delinquent or violent friends but are

immune to reducing their violence or delinquency when associating with less violent or

delinquent friends. In terms of selection dynamics, although both boys and girls have

a tendency to select friends based on friends’ behavior, girls have a stronger tendency

to do so, suggesting that among girls, friends’ involvement in violence or delinquency

is an especially decisive factor for determining friendship ties.

Criminologists generally agree that delinquency and crime are committed dispro-

portionately by males and that the gender gap in offending becomes even larger when

the focus turns toward violent offenses (Steffensmeier et al., 2005). Two explanations

have been offered for the gender gap in offending. “Differential exposure” expla-

nations argue that boys and girls are differentially exposed to risk factors that are

conducive to crime or delinquency. “Differential reaction” explanations argue that

∗Data for this article were drawn from the National Longitudinal Study of Adolescent Health (Add

Health), a program project designed by J. Richard Udry and Peter Bearman, and funded by a grant

from the National Institute of Child Health and Human Development (HD31921). This research

was supported in part by R24-HD058484 from the Eunice Kennedy Shriver National Institute of

Child Health & Human Development awarded to the Institute for Population Research at The

Ohio State University. Direct correspondence to Dana L. Haynie, Department of Sociology, The

Ohio State University, 238 Townshend Hall, 1885 Neil Ave Mall, Columbus, OH 43210 (e-mail:

haynie.7@osu.edu).

C2014 American Society of Criminology doi: 10.1111/1745-9125.12052

CRIMINOLOGY Volume 52 Number 4 688–722 2014 688

GENDER, DELINQUENCY, AND NETWORKS 689

although boys and girls may be exposed to similar risk factors, they will be affected

differently by them. Although these explanations are not mutually exclusive, criminol-

ogists often draw on one or the other especially when considering gender differences

in delinquency. One factor often used to evaluate these perspectives is the role of

delinquent peer exposure. Indeed, research has suggested that peer delinquency can

account for some of the gender gap in delinquency (Mears, Ploeger, and Warr, 1998;

Piquero et al., 2005). That is, girls are exposed to lower levels of peer delinquency

than boys (differential exposure), and when exposed, girls are influenced differentially

by delinquent peers compared with boys (differential reaction). In addition, there is

increasing awareness of the role of friendship selection in shaping peer-delinquency

similarity and the need to account for selection processes when examining the role

of delinquent peer exposure for youths’ involvement in delinquency. The goal of the

current study is to use longitudinal network models to test whether girls are differentially

influenced by their friends’ delinquency (differential reaction) and whether girls are

more likely than boys to make friends based on their friends’ delinquent behaviors

(selection).

The number of studies examining the effects of peers on delinquency has grown

substantially over the last two decades (Brechwald and Prinstein, 2011; Haynie, 2001;

Matsueda and Anderson, 1998; Weerman and Bijleveld, 2007; Weerman and Hoeve, 2012;

Zimmerman and Messner, 2010). Much of this research has suggested that the association

between an adolescent’s delinquent behavior and his or her friends is stronger than that of

other risk factors considered (Birkbeck and LaFree, 1993; Warr, 2002). Recently, scholars

have argued that prior findings are compromised by overlooking the network structures in

which adolescents are embedded (Haynie, 2001; Haynie and Osgood, 2005; Weerman and

Hoeve, 2012). In response, research has begun to apply longitudinal network methods to

determine the role of selection and influence processes in shaping delinquency similarity

among adolescents who are friends (Dijkstra et al., 2010; Weerman, 2011; Weerman and

Bijleveld, 2007).

Absent from the longitudinal network studies on delinquent peers has been a focus on

gender dynamics and the role they play in shaping delinquency similarity among friends.

This absence is surprising as many studies have indicated that girls’ friendships differ in

several important ways from boys’ friendships (Erwin, 1998; Rose and Rudolph, 2006;

Weerman and Hoeve, 2012). Additionally, recent research has indicated gender variation

in the association between peer delinquency and individual delinquency (Zimmerman

and Messner, 2010). Such differences may imply that peers influence girls differently than

boys, or that girls are more likely than boys to select friends based on shared delinquency

profiles.

The current study adds to the understanding of gender dynamics in delinquency by

applying dynamic longitudinal network methods to determine whether gender moder-

ates the effect of influence and selection on the tendency for adolescents to be similar to

their friends. We focus on two outcomes: involvement in violence or nonviolent delin-

quency. Using longitudinal friendship network data from adolescents attending two large

schools participating in the National Longitudinal Study of Adolescent Health (hereafter,

AddHealth), we test whether girls are more likely to be influenced by friends’ behavior

than boys and whether girls are more likely to select friends based on violent or delin-

quent behavior than boys.

690 HAYNIE, DOOGAN, & SOLLER

BACKGROUND

The role of friend and peer influence is central to explanations of crime, delinquency,

and other problem behaviors. Compared with children and adults, adolescents attribute

greater importance to friends, spend more time socializing with friends, and are more

strongly influenced by friends’ behaviors and attitudes (Giordano, Cernkovich, and

Holland, 2003). Not surprisingly, then, one of the most consistent findings in the crimi-

nological literature is that individuals with delinquent friends are likely to be delinquent

themselves. Robust associations between peer and individual delinquency have led some

to argue that peer processes are among the most important in explaining delinquent

outcomes, regardless of the type of delinquency considered (Akers, 1973; Haynie, 2001;

Warr, 2002).

Even though all agree that there is similarity in delinquency among friends, prominent

criminological theories present different mechanisms by which similarity emerges. In par-

ticular, control theories (Gottfredson and Hirschi, 1990; Hirschi, 1969) and influence the-

ories (Akers, 1973; Sutherland, 1947) are prevailing perspectives that discuss the role of

peers for adolescent delinquency. However, these theories offer opposing explanations

regarding the ways that peers shape individual delinquency. Accordingly, criminologists

have focused on different features of peer contexts to assess the plausibility of the differ-

ent proposed theoretical mechanisms.

Gottfredson and Hirschi’s (1990) general theory of crime describes the process through

which individuals self-select into peer groups. In their 1990 book, they argued that

peers have little to no influence on individual offending; rather, individual variation in

self-control (i.e., the ability to regulate impulsive behavior), which is relatively stable

by early adolescence, shapes how adolescents cluster together in peer settings. Delin-

quent adolescents with low self-control are likely to end up with other delinquents as

friends, as a result of their similar levels of self-control. Apart from determining the

types of friends one makes, low self-control is also a primary cause of delinquent be-

havior. Thus, delinquency and associations with delinquent others are both directly

caused by low self-control.1Therefore, according to control theorists, the association be-

tween delinquent peers and individual offending is spurious and the peer–delinquency

association is a result of selection processes and not because friends influence individual

delinquency.

In contrast to traditional control theories, Sutherland’s (1947) differential association

theory suggests that delinquency is learned from intimate social relationships with oth-

ers, such as friendships, through the transference of attitudes and definitions that en-

courage criminal behavior. Akers’s (1973) social learning theory builds on this by em-

phasizing behavioral modeling and operant conditioning’s role in the learning process.

This model assumes that the adoption of delinquent behavior occurs through the imi-

tation of peers’ behavior and the observation of its positive or negative consequences.

Both Sutherland’s and Akers’s theories represent prototypical influence theories in

their argument that delinquency, like any behavior, is learned in intimate relationships,

1. Hirschi’s (1969) social control theory is also based on the premise that “birds of a feather flock to-

gether,” such that although delinquent youth select other delinquents as friends, the friends them-

selves do not influence delinquency.