Family instability during early and middle adolescence

Abstract

Two studies investigated associations between family instability (changes in parents' intimate partners, work hours, residence, children's schools) and adolescent adjustment. In Study 1 (N = 141, M age = 15.23 years), instability was associated with increased caregiver-reported externalizing and internalizing behaviors (including youth-reported cigarette use), reduced teacher-reported frustration tolerance, social skills, task orientation, and lower academic grades. Logistic regression results for instability exposure showed an increased risk for school suspensions, Person in Need of Supervision status, binge drinking, and marijuana use. In Study 2 (N = 225, M age = 13.37 years), instability was linked to adolescent-reported externalizing and internalizing behaviors, teacher-reported disruptions, and lower English and math grades. Key sociodemographic factors and negative life events were statistically controlled in regression analyses. Results indicate that a more theoretically coherent measure of family instability, which is distinct from negative life events, may prove valuable in understanding the potentially adverse influence of instability on youth.

Introduction

Across America, instability in the home and school lives of children has become the norm. Nearly half of all American children moved to a different home at least once between 1995 and 2000 (Franklin, 2003). During the early 1990s, one in six U.S. third graders attended three or more different schools since beginning first grade (U.S. General Accounting Office, 1994). Not including normative transitions between elementary, middle, and high schools, 40% of 12- to 17-year-olds in a national sample had changed schools at least once (Lugaila, 2003). Despite a static divorce rate over the past two decades, children increasingly experience discontinuity in family structure as primary caregivers shift in and out of marital-like relationships (Bumpass & Lu, 2000). Working mothers who are young, low-income, single, and non-White are most likely to have unpredictable work hours that change weekly (Golden, 2001, March, Han, 2005). To understand the cumulative effects of these changes on children, we examined the effects of instability within home, school, and neighborhood contexts on adolescent outcomes in two longitudinal samples of rural and semi-rural youth in upstate New York. Instability was defined as changes in (a) children's residences, (b) children's schools, (c) parents' intimate partners, and (d) parents' work hours.

Bronfenbrenner (1979) described proximal processes, exchanges of energy between the developing person and the environment, as the engines of human development. To be effective, such processes should occur on a regular basis and over extended periods of time. During childhood and early adolescence, proximal processes occur within three key ecological contexts: home, school, and neighborhood (Bronfenbrenner, 1979, Bronfenbrenner and Evans, 2000).

Regular, predictable, and supportive exchanges with adults and peers in these settings promote competent development. Alternatively, disruptions to the continuity and predictability of these key contexts threaten competent development (Bronfenbrenner, 1979, Bronfenbrenner and Evans, 2000, Bronfenbrenner and Morris, 1998). As such, family instability is distinguished from a broader class of life events that do not fundamentally threaten the coherence of key developmental contexts (although it is acknowledged that some of these life events are potentially stressful and require readjustments on the part of the child). Likewise, although instability may be correlated with poverty and other measures of socioeconomic status (e.g., parental education), it is not a proxy for these risk factors. Indeed, instability might mediate the effects of poverty on child outcomes (Evans, 2004).

To date, only a handful of studies have examined the influence of instability (Ackerman et al., 2003, Ackerman et al., 1999, Forman and Davies, 2003, Herrenkohl et al., 2003, Milan and Pinderhughes, 2006, Moore et al., 2000, Simmons et al., 1987). As definitions of instability vary somewhat across these studies, it is difficult to achieve clarity regarding the unique influence of instability on child adjustment. One of our objectives in the present paper is to provide greater specificity to the instability measure relative to other indicators of adversity such as negative life events, illness, or poverty. To do so, we focus on construct of development — What is instability? How does it differ from other constructs? — and tests of its unique effects. While work on explanatory factors is also needed, such questions are premature until greater clarity in the content of the construct is achieved. The present study helps to clarify instability because the heterogeneous economic backgrounds of the sample children in Study 2 enabled us to disentangle correlates of instability from those of income — a contribution that only two other studies have made (Milan and Pinderhughes, 2006, Moore et al., 2000) and only one in the context of adolescence (Moore et al., 2000). We also provide a multi-informant assessment of adolescent outcomes across two different samples.

Ackerman et al. (Ackerman et al., 1999, Ackerman et al., 2003) defined instability in terms of the number of residences and intimate adult relationships involving a child's primary caregiver, the number of families with whom the child lived, the number of serious illnesses the child experienced during the past five years and, finally, the number of negative life events, such as changes in parental employment or deaths of relatives, the child experienced over the most recent six months. Findings from their 1999 study revealed a significant positive relationship between family instability and externalizing behaviors among racially diverse, urban five-year-olds, and between family instability and internalizing behaviors among seven-year-olds. Moreover, relations between instability and maladjustment were stronger for temperamentally difficult children and among children who had pre-existing problem behaviors.

Milan et al. (2006) examined family instability on child externalizing and internalizing problems over a 6-year elementary school period as a part of the Fast Track investigation. In their study, family instability was defined as a count of seven types of events during the past year: residential moves, family member deaths, divorces or separations from cohabitating partners, remarriage/reconciliation/entries into homes of new partners, temporary parent–child separations, sibling births or new children's entries into the home, and parental job losses or changes in employment status. Cumulative family instability predicted children's externalizing and internalizing behavior trajectories during elementary school. After controls for early maladjustment and demographic factors were included in the model, timing effects were observed for levels of externalizing behaviors among third graders, such that stronger relations were detected for early experiences of family instability during kindergarten compared with cumulative family instability experiences in first through third grades.

Forman and Davies (2003) defined instability as the number of times disruptive events occurred in five family domains over the most recent five years: (a) changes in residence, (b) changes in the primary or secondary caregiver, (c) transitions in romantic relationships of the primary caregiver, (d) job and income loss, and (e) death or serious illness of a close family member. Their research with a predominantly White, suburban sample of early adolescents showed that instability contributed to adolescent maladjustment, and that this association was mediated by parenting difficulties and adolescents' perceptions of family insecurity.

Moore et al. (2000) also developed a measure of instability, which they called turbulence. In their study of 22,729 6- to 17-year-olds from the 1997 National Survey of America's Families, 6% of children nationwide experienced turbulence, defined as two or more of six possible changes in residence, school, parental employment, or health within the past year. For children below the poverty line, this number doubled to 13%.

Simmons et al. (1987) examined the effects of recent cumulative change across multiple contexts for 447 urban, White sixth and seventh graders. Cumulative change was defined by changes in schooling, residence, or family structure brought about by parental death, divorce, or remarriage, or pubertal development and early dating behavior. Overall, neither boys' nor girls' self-esteem was influenced by individual factors. In contrast, as cumulative transitions increased for boys, GPA and extracurricular participation in seventh grade decreased significantly even after sixth-grade levels of these outcomes were statistically controlled. For girls, the effects of transitions on GPA were non-linear, such that after one transition each additional transition brought lower GPAs for seventh-grade girls. These findings underscore that it is not simply the addition of discrete stressors that contributes to declines in academic performance, but rather the difficulty children experience when coping with transitions in several areas of their lives. Herrenkohl et al. (2003) also found that caregiver and residential instability predicted adolescents' problem behaviors over and above socioeconomic background and childhood maltreatment experiences.

We have defined instability more broadly, viewing it as comprising more than simply residential moves and partner changes, in order to capture structural changes within the micro-contexts of childhood and adolescence. We offer a more concise definition than some prior studies, however, by excluding life events (e.g., child's illness, death of a grandparent) that do not typically alter the coherence and predictability of immediate ecological contexts where most proximal processes are developed and sustained. In addition to these conceptual changes, we extend prior work on instability to a different sample of youth. As indicated above, most prior work on chaos and children and youth development has focused on urban, predominantly African American samples. Herein we examine instability among predominantly White youth living in small towns and rural areas.

The present studies had two aims. The first aim was to refine the instability construct. Although our instability construct dovetails with those of previous research, it is focused more precisely on structural changes in children's microsystems. We estimated the effects of instability net of socioeconomic influences, negative life events, and other forms of childhood adversity. To our knowledge this is the first paper on family instability to disaggregate structural changes in home, school, and neighborhood contexts from a more general class of negative life events.

The second aim was to replicate and extend previous research on the effects of family instability on child outcomes. We present two studies with similar measures of instability and similar outcomes. Prior research has tended to focus almost exclusively on younger children from urban low-income or ethnic minority samples or conversely early adolescents from more affluent, suburban families. The present studies contribute to the literature by examining the effects of instability beyond the developmental periods of childhood and early adolescence with families that span a wider range of income levels and who reside in rural America. Our review of the literature indicates that this is only the second paper on instability to exhibit such socioeconomic variability within an adolescent sample. Moreover, we included a broader range of outcomes than those associated with externalizing and internalizing behaviors and relied upon multiple methods to index adolescent adjustment.