Adolescent behavior and achievement, social capital, and the timing of geographic mobility

Advances in Life Course Research 18 (2013) 223–233

Contents lists available at SciVerse ScienceDirect

Advances in Life Course Research journal homepage: www.elsevier.com/locate/alcr

Adolescent behavior and achievement, social capital, and the timing of geographic mobility Brian Joseph Gillespie Department of Sociology, University of California, Irvine , 3151 Social Science Plaza, Irvine, CA 92697-5100, USA

A R T I C L E I N F O

A B S T R A C T

Article history: Received 12 December 2012 Received in revised form 17 May 2013 Accepted 1 July 2013

This paper examines the relationship between geographic mobility and adolescent academic achievement and behavior problems. Specifically, it addresses how the effects of moving differ by age and how social capital moderates the impact of moving on children (aged 6 to 15). Children’s behavior problems and academic achievement test scores were compared across four survey waves of the National Longitudinal Survey of Youth (2000, 2002, 2004, and 2006) and matched to data from their mothers’ reports from the National Longitudinal Survey of Youth 1979. The findings indicate that the negative behavioral effects of geographic mobility on adolescents are most pronounced for individuals relocating to a new city, county, or state as opposed to those moving locally (i.e., within the same city). Furthermore, as suggested by a life-course perspective, the negative effects of moving on behavior problems decrease as children get older. The results also show that several social capital factors moderate the effects of moving on behavior but not achievement. ß 2013 Elsevier Ltd. All rights reserved.

Keywords: Academic achievement Behavior problems Geographic mobility Life course Residential mobility Social capital

Past research has shown that moving is detrimental for children (Hagan, MacMillan, & Wheaton, 1996; Rumberger & Larson, 1998; Scanlon & Devine, 2001). Research focuses on two specific outcomes of mobility on children: academic achievement and behavior problems. Scholars have consistently found that compared with non-mobile children, mobile children experience significantly more behavior problems (Haynie, South, & Bose, 2006; Simpson & Fowler, 1994) as well as negative academic outcomes, such as dropping out (Hofferth, Boisjoly, & Duncan, 1998; South, Haynie, & Bose, 2005), decreased academic performance (Ingersoll, Scamman, & Eckerling, 1989; Tucker, Marx, & Long, 1998; Wood, Halfon, Scarlata, Newacheck, & Nessim, 1993), and grade retention (Simpson & Fowler, 1994). One possible reason given for the differences in outcomes between mobile and non-mobile children is the loss of social capital experienced by both the child and the parents in the move (Coleman, 1988; Pettit & McLanahan,

E-mail address: [email protected]. 1040-2608/$ – see front matter ß 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.alcr.2013.07.001

2003; Pribesh & Downey, 1999; Stack, 1994; Gillespie, in press). Household characteristics that predict selection into migration can complicate the picture, as moves can be instigated by family disruptions, such as divorce (Astone & McLanahan, 1994; Norford & Medway, 2002; Tucker et al., 1998), and employment changes (Brett, 1982) that negatively affect child outcomes. Research has shown that the impact of important life events, such as parental divorce (Perry-Jenkins, Repetti, & Crouter, 2000) and parent entrance into the work force (Cooksey, Menaghan, & Jekielek, 1997), depends largely on when they occur in the child’s life. Moving, like other important child events, might also have distinct effects depending on how old the child is at the time. This paper makes several contributions. Unlike prior studies of the effects of geographic mobility on child outcomes, it explicitly considers differences by age of the child. Age interactions are included to determine whether moving has more pronounced negative effects on the behavior and achievement levels of older or younger children. Further, this study clarifies the academic and behavioral effects of geographic mobility on children by

224

B.J. Gillespie / Advances in Life Course Research 18 (2013) 223–233

including measures for social capital before and after a move. Specifically, mother and child data from four waves of the National Longitudinal Survey of Youth are matched to examine the effects of mobility on child outcomes and to test the extent to which they can be explained by changes in social capital and in individual as well as household characteristics. 1. Theoretical background Research on the effects of geographic mobility on child outcomes has, for the most part, shown that moving is harmful for children (Humke & Schaefer, 1995; South & Haynie, 2004; Gillespie and Bostean, in press). For instance, Hendershott (1989) and Norford and Medway (2002) found that moving increases behavior problems in children. Others have found a significant relationship between geographic mobility and dropping out of school (Coleman, 1988; Hagan et al., 1996). Ingersoll et al. (1989) as well as Pribesh and Downey (1999) found significant effects of geographic mobility on poor academic performance and other researchers (McLanahan & Sandefur, 1994; Astone & McLanahan, 1994) have found that moving is also correlated with school drop-out and low educational attainment net of selection into moving. Outside of individual and household predictors, such as marital disruption (Madigan & Hogan, 1991), the major debates on geographic mobility and child outcomes have centered largely on social capital, the quality and quantity of one’s interpersonal relations. Coleman’s (1988) work on social capital has inspired scholars to view where a person lives as promoting the formation and maintenance of social ties that are paramount to a child’s ability to excel in educational settings. Coleman’s key point is that interactions within and outside of the household (e.g., among children, parents, teachers, schools, and community) are resources that provide children with assets that increase their abilities, achievement-levels, and general welfare. Social interactions among parents (parent–community closure), between parent and child (intergenerational closure), and among children (child–community closure) provide pathways to socialize, facilitate control, share information and resources, and establish and reinforce norms and expectations. These important social and community ties are broken when a family relocates, resulting in a loss of social capital. Negative effects may be even worse when families relocate repeatedly: ‘‘. . .for families that have moved often, the social relations that constitute social capital are broken at each move’’ (Coleman, 1988, p. 113). With the notable exception of Gasper, DeLuca, and Estacion (2010), the literature has generally supported Coleman’s (1988) claim. Gasper et al. (2010) found that the association between behavior problems and geographic mobility across city, county, or state lines is mostly due to selection into moving based on preexisting characteristics of mobile individuals and families. They found that geographically mobile children are usually poorer and have lower academic achievement than non-mobile children (which also puts them at risk for behavior problems). South and colleagues (South & Haynie, 2004;

South et al., 2005) found that social capital has limited predictive effects on the educational attainment of high schoolers who change residences or schools versus those who do not. Children who move are more likely than children who do not move to make friends who negatively affect their educational performance and aspirations. In light of this mixed evidence, comparative research needs to explore how geographic mobility is associated with different adolescent outcome domains using consistent measures. It is well established that parent community and school involvement has a positive influence on adolescent academic (Epstein & Sanders, 2002; Grolnick & Ryan, 1989) and behavioral (Domina, 2005) outcomes. Parental involvement leads to richer social networks, but also greater information passing and greater knowledge of children’s academic and behavioral well-being (Spera, 2005). Social capital is not only reflected in the parent– school relationship but also in the parent’s knowledge of their child’s social network. Muller (1998) found that children scored higher on achievement tests if their parents were acquainted with their friends. These children also receive better grades in school (Crouter, MacDermid, McHale, & Perry-Jenkins, 1990). Geographic mobility, however, runs the risk of damaging these beneficial relationships. In fact, Pettit and McLanahan (2003) found that geographic mobility is associated with a reduced likelihood of parents talking with the parents of their children’s friends. Research on the effects of the parent–child relationship (intergenerational closure) on adolescent outcomes has shown that high quality intergenerational relationships are beneficial for children. Aseltine, Gore, and Colten (1998) found a significant relationship between parent– child closeness and decreased depression and externalizing behavior problems in children. Others (Conger, Ge, Elder, Lorenz, & Simons, 1994) found that positive parent– child relationships buffer the negative emotional and behavioral effects of a disruptive event, divorce, on children. Of course, the stressful act of moving might also undermine the parent–child relationship in many ways. Moving results in decreased parent engagement (Pettit & McLanahan, 2003), changes in activities and routines (Brett, 1982), a breach in the strength of social ties (Coleman, 1988). As such, the parent–child relationship may be jeopardized by a move. However, the parent–child relationship (arguably the ‘‘closure’’ left most intact after a move) may also buffer the negative effects of moving on children. Lastly, research on the importance of social capital in adolescence (child–community closure) has focused on the many positive social and emotional benefits of having friends during childhood (Ahn, 2012). Several reasons have been advanced for why friendships are more important in late adolescence than in earlier childhood. Douvan and Adelson (1966) argue that the value of friendship is to minimize the tumult that accompanies the onset of puberty. Berndt (1982) disagrees and hypothesizes that cognitive growth facilitates better understanding of sharing and reciprocity in friendships. Both agree that friendship is an important social qualifier for adolescents,

B.J. Gillespie / Advances in Life Course Research 18 (2013) 223–233

because they are establishing their independent identities separate from their parents (and other authority figures) and are thus focusing more on the acceptance, opinions, and ideas of their peers. Indeed, geographic mobility severs these important social ties. Mobile adolescents have fewer friends than non-mobile adolescents (Humke & Schaefer, 1995), which contributes to the negative effects of mobility (Pribesh & Downey, 1999). Since loners engage in fewer delinquent activities than non-loners (but also spend less time on school work and receive worse grades) (Demuth, 2004), we might expect the effects of geographic mobility to affect academic outcomes for some and behavioral outcomes for others. Taken together, it is expected that social capital (child–community, parent–community, and intergenerational closure) will moderate the effect of geographic mobility on adolescent behavioral and academic outcomes. Further, it is expected that this moderation will be stronger for behavioral than academic outcomes. Nevertheless, the intensity of a breach in social capital also depends largely on the type of move made. The type of relocation that occurs – often categorized as local or long distance – is closely linked to the reasons driving mobility (Fischer, 2002; Long, Tucker, & Urton, 1988). Local residential mobility is defined as a move that occurs within the same county or neighborhood and is often a result of personal preference, family changes, and other housing considerations (Frey, 2003; Long, 1988; Rossi, 1980). A long distance move (i.e., geographic mobility) is usually defined as relocation across city, county, or state lines (U.S. Bureau of the Census, 2012). This type of move is typically made for employment or healthrelated reasons and macro-economic conditions, such as job transfers or shifts in labor market demand (Kilborn, 2009). Given that community ties are more likely to be broken as a result of distance mobility, it is expected that a short-distance move – while stressful and disruptive to school continuity – will not be as disruptive of social capital and community ties as a long distance move. Moving is a decision made almost completely by parents, albeit sometimes with children’s interests as a major concern. A life course perspective suggests that the timing of such an important event may have distinct effects depending upon how old the child is at the time (Elder, 1987, 1998). Important studies on adolescents have found support for Elder’s ‘‘timing in lives’’ claim with children’s age at their parents’ divorce (Perry-Jenkins et al., 2000) and the timing of parental employment (Cooksey et al., 1997). On one hand, it is expected that older children will have developed more friendships and social ties than younger children (Brown, 1990); as a result, moving will have greater negative effects for them. However, older children may be less affected by a move, because they are more resilient, are better able to seek out and establish social capital in their new location, or have a greater developmental need to do so. Moving is difficult. In fact, geographic relocation is widely accepted as one of life’s most stressful events (Holmes & Rahe, 1967). It is also seen as a major household event that has the potential to be a turning point in the life of an adolescent (Hagan et al., 1996). Changing residences

225

often results in significant changes in roles, habits, and identities (Brett, 1982) that induces stress and requires adaptive behavior in the new residence. Together, these arguments suggest that the stress induced by the interruption of daily life and routine, the emotional upheaval brought on by a move, and the time consumed by the act of moving will have a harsher effect on adolescent behavior and achievement. Thus, geographic mobility is expected to affect adolescents above and beyond individual and family characteristics and social capital measures. 1.1. Hypotheses Following from the research, six main hypotheses are considered. A general hypothesis links mobility to poor outcomes: H1A. Geographic mobility will be negatively associated with academic achievement. H1B. Geographic mobility will be positively associated with behavior problems. Because mobility will disrupt social ties and the exchange of information between parents and children, parents and community, and children and community, adolescents who lose social capital are expected to experience more negative outcomes from moving: H2. A distance move will lead to more pronounced negative adolescent outcomes than a local move. H3. Social capital will moderate the effects of geographic mobility on adolescents. Since geographic mobility has weaker effects for academic than behavioral outcomes and social capital has greater effects on behavioral than cognitive outcomes: H4. Social capital moderation of mobility and child outcomes is expected to be stronger for behavior problems than academic achievement. There are competing expectations regarding which ages will be most affected by a move. Older children will have developed qualitatively and quantitatively richer friendships and social ties than younger children and, as a result, moving may disadvantage them more. Alternately, older children may be less affected by a move, because they are more resilient or are able to seek out and establish social capital on their own. H5A. Moving will be associated with more disadvantageous outcomes for older than younger children (mobility– disruption hypothesis). H5B. Moving will be associated with more disadvantageous outcomes for younger than older children (mobility– resilience hypothesis). Lastly, it is hypothesized that some of the negative effects of geographic relocation on adolescents result from the stresses induced as a result of the general interruption and disruption of daily life and routine. As such:

226

B.J. Gillespie / Advances in Life Course Research 18 (2013) 223–233

H6. Geographic mobility will negatively affect adolescents above and beyond individual and family characteristics and social capital measures.

2. Data and methods 2.1. National Longitudinal Survey of Youth (NLSY) Information and Sampling The National Longitudinal Survey of Youth (NLSY79) is a longitudinal study of a representative sample of American men and women aged 14 to 21 in 1979. The children of the female NLSY79 respondents are also surveyed biennially starting in 1986 and these NLSY79 Child and Young Adult data files can be linked with the original NLSY79 to assess intergenerational phenomena and outcomes. Thus, the children in this sample are representative of all children born to the nationally representative cohort of women sampled in 1979. The present study utilizes data only from the 2000, 2002, 2004, and 2006 survey waves, because respondents’ geographic mobility was not assessed in the NLSY until 2000. The ages for the children under study range from 6 to 15. Multiple children are surveyed in each household, giving the NLSY a useful hierarchical design. High response rates (between 85 and 90%) also contribute to the validity of the analysis. This rate has been maintained because data are primarily collected in the respondent’s home through face-to-face interviews. An important limitation of these data is that in 1986, when the children of these 21–28-year-old mothers were first assessed, the oldest children had been born to very young women. Thus, the sample may exclude some children born to younger women, because they had already left the sample before the 2000 wave. In order to have complete trajectories (including the baseline score) in the analysis, ages are based on respondents’ test-ready age two years prior to the first survey wave. Further, to avoid the confounding effects of marital disruption on relocation, households experiencing divorce and/or separation between survey waves (less than 5%) were removed from the analysis. Owing to these restrictions and attrition across waves, this study utilizes the records of 2835 adolescent respondents (a sample of 11,340 child-years). The average household income for families in the 2000 wave was approximately $ 55,000 per year. The majority of the children (52%) were boys, comparable to the entire NLSY population sample (51%). The mean age of children at the 2000 wave was 10.8 (SD = 2.5). Half of the sample mothers identified as non-black, non-hispanic (50%), 29% identified as black, and 21% identified as hispanic. In 2000, about three-fourths of the sample (77%) did not move while 15% moved locally and 8% relocated across city, county, or state lines. In 2002, 752 people (25%) moved; in 2004, 734 (25%) and in 2006, 788 (25%) respondents were residentially mobile. Less than five percent of households had multiple moves at any wave. The sample statistics on mobility are reasonably consistent with recent reported rates of geographic mobility in

Table 1 Descriptive statistics in 2000 (Wave 1). Mean (%) Dependent variables BPI2000 PIAT2000 BPI baseline1998 PIAT baseline1998 Individual/household Age2000 Male Black Income Father in household Birth order Urban Mother’s education Mother’s age at child’s birth Children in household Mother never married Social capital Church attendance Closeness to mother Friends by name Catholic Loneliness Extracurricular activities Geographic mobility Did not move Local move Distance move

Standard deviation

53.88 57.95 56.31 56.27

28.61 25.11 28.4 24.94

10.81 52% 29% 54,668 65% 2.23 72% 13.03 27.88 2.56 11%

2.45

2.07 3.58 2.9 0.06 1.4 0.63

52,763 1.18 2.48 3.12 1.22

1.1 0.69 1.5 0.24 0.6 0.48

77% 14.5% 8.43%

N = 2835.

the United States (U.S. Bureau of the Census, 2012). Table 1 presents the descriptive statistics for the restricted sample population in 2000. 3. Variables and measures 3.1. Child outcomes Academic achievement was measured using the NLSY79 Child and Youth respondent’s 2000, 2002, 2004, and 2006 Peabody Individual Achievement Test (PIAT) scores. The PIAT is a widely used measure of academic achievement for children (Dunn & Markwardt, 1970). Since 1986, the children in this study have been assessed biennially between ages five and 15. Each assessment begins with five age-appropriate questions and progresses to more advanced concepts. The reading recognition test measures word recognition and pronunciation ability, and the math test measures basic math skills and concepts. The score is the mean of the child’s age-standardized percentile scores on subsets in mathematics, reading comprehension, and reading recognition. The behavioral problems measure is based on Peterson and Zill’s (1986) Behavior Problems Index. This index consists of 28 questions, drawn primarily from the widely used Child Behavior Checklist (Achenbach & Edelbrock, 1981) along with other widely used child behavior scales. The respondent’s mother indicates whether a statement about the child’s behavior is ‘‘often true,’’ ‘‘sometimes

B.J. Gillespie / Advances in Life Course Research 18 (2013) 223–233

true,’’ or ‘‘never true.’’ The composite score is a measure with higher numbers indicating more behavior problems. Because the purpose of this analysis is to assess the effects of moving and several other theoretical predictor variables on changes in educational achievement and behavior problems, it is important to use longitudinal data in order to include measures of the predictors and outcomes in a person-year format. This allows for consideration of social capital for each survey wave in the analysis. Adequately controlling for past behaviors before a move occurs is crucial; otherwise, associated changes in child outcomes after moving cannot be determined confidently. The sample consists of children who completed the PIAT and BPI for the 1998 (baseline), 2000, 2002, 2004, and 2006 survey rounds. The PIAT is administered starting at age five, and the behavior problems assessment begins at age four; neither examination is recorded after age 15. 3.2. Independent variables 3.2.1. Individual and household characteristics Individual and household characteristics include time variant variables, such as annual household income (logged) and age. A dummy variable for parent marital status indicated whether or not a respondent’s mother was never married at each survey wave. Time-invariant variables include the child’s sex, birth order, mother’s age at child’s birth, mother’s highest year of education completed (measured once in 2000), family structure (father in household or not), and the number of children in the respondent’s household. Children were assigned to racial groups based on the primary racial identification of their mothers as black; hispanic; or non-black, nonhispanic. All other variables in the analysis vary across survey waves. Urban residence was measured as whether or not the respondent lived in an urban area (=1, else = 0). 3.3. Social capital Interaction between parents and community institutions was measured using two variables: how many of the friends the child’s parent knows by sight and name, coded as none of them (0), only a few (1), about half (2), most of them (3), or all of them (4) (Teachman, Paasch, & Carver, 1996) and, following Coleman (1988), a dichotomous variable marking whether or not a child attends Catholic school. Interaction between child and community institutions was measured by whether or not the child participates in extracurricular activities (White & Gager, 2007); how often he or she attends religious services coded as (0) never, (1) a few times a year, (2) about once a month, and (3) about once a week (Parcel & Dufur, 2001); and how often the child feels lonely or wishes he or she has more friends as measured as never or hardly ever (1), sometimes (2), or often (3). Intergenerational interaction was measured by the level of closeness the respondent feels to his or her mother, reported as being not very close (1), fairly close (2), quite close (3), or extremely close (4) (Pryor, 1999).

227

3.4. Geographic mobility The potentially disruptive act of moving is captured by two dummy variables indicating whether a respondent (1) relocated within the same city or (2) relocated to a new city, county, or state between each survey wave. The omitted category for comparison is not moving. 4. Analytic strategy Linear mixed modeling (LMM), an extension of hierarchical linear modeling, was used to examine the effects of social capital on the relationship between geographic mobility and child educational achievement and behavior problems. Models were run separately for each of these two child outcomes. LMM is a flexible and powerful method for the analysis of longitudinal data. In LMM, independent observations are not assumed, meaning that betweensubject and within-subject effects are both considered. This modeling structure is also flexible in its use of missing information. Other models use listwise deletion of cases if a complete trajectory is not available for an individual. LMM, on the other hand, accounts for all respondents in the data set and is, therefore, arguably a better model for unbalanced panel data sets like the NLSY where not every respondent is observed in every year. Lastly, LMM allows for the analysis of hierarchically organized data. In this study, four models (A through D) were tested on three levels using a linear mixed modeling structure. The first of these three levels consisted of households, the second was the individual child nested in each household and the last level, survey wave or ‘‘time,’’ was measured by interview round and nested within each child. This study applied an upward two-step preliminary modeling procedure employed by Singer and Willett (2003): (a) an unconditional means model, and (b) an unconditional growth model. First, the unconditional means model is the preliminary verification for whether this is an appropriate analysis by partitioning the total variation in the outcome variable (BPI and PIAT). The intraclass correlation coefficient (ICC) measures the proportion of variance in the outcome variable that is due to betweenchildren differences rather than differences within children over time. The unconditional growth model was run to (a) assess the effects of aging on child achievement in academics and behavior problems and (b) detect whether there was significant variance to be explained from household-level characteristics. The baseline models and hypotheses presented above combine to suggest the following linear mixed model. 4.1. Final model Level 1 (time): PIAT/BPIti = b0i + b1i (AGEti) + b2i rti + uti. Level 2 (child): b0i = b0 + b1 (household characteristics)i + b2 (social capital main effects)i + b3 (social capital interactive effects)i + b4 (baseline PIAT/BPI score)i + v0i, where b1i = b5 + w1i and b2i = b6.

228

B.J. Gillespie / Advances in Life Course Research 18 (2013) 223–233

Here, PIAT/BPIti denotes the time t score on the PIAT/ BPI of individual i. b0 is the overall average PIAT (or BPI) score for each individual over time, and the covariates listed are status controls for household characteristics, social capital, and baseline score. v0i is the child-level error term with the assumption that it is normally distributed with variance t. The Hausman specification test validated these models. LMM assumes that the dependent variable be conditionally normal. Shapiro–Wilkins, Skewness, and Kurtosis testing indicate that both dependent variables were distributed reasonably normally. Stata 12 estimated the fixed and random effects as well as the reliability and correlation coefficients. Variance inflation factors were checked in order to assess any severe multicollinearity in the model (average VIF: 1.31).

5. Results In 2000, mobile children were significantly more likely to come from an urban or suburban area (76%) than a rural area (24%) [X2(1) = 10.2, p < 001]. Children whose mothers were never married in 2000 were more likely to move (36%) than children whose mothers had been wed (21%) [X2(1) = 37.3, p < 000]. Mobile children were also more likely to be black or hispanic (29%) than non-black, nonhispanic (17%) [X2(1) = 69.89, p < 000]. 5.1. Academic achievement Table 2 presents a summary of the series of theoretically important variable cluster models fitted to the data on educational achievement. For each dependent variable,

Table 2 Linear mixed model of geographic mobility on academic achievement. Peabody individual achievement test (PIAT)

Geographic mobility Local mobility Local mobility  age Distance mobility Distance mobility  age Individual/household Age Male Black Income (logged) Father in household Urban mother’s education Birth order Mother’s age at child’s birth Children in household Mother never Married Social capital Church attendance Closeness to mother Friends by name Catholic Loneliness Extracurricular activities Interaction terms Distance move  church Attendance distance move  closeness to mom distance move  friends by name 0.54 Distance move  catholic Distance move  loneliness Distance move  extracurricular PIAT baseline Constant Variance components Level 2 Level 3 (initial status) Rate of change Covariance Residual Note: Robust standard errors. a p < 10. * p < 05. ** p < 01. *** p < .001.

Model A, b

Model B, b

0.08

26.02** 2.1** 10.65 0.94

1.15

Model C, b 13.94 0.96 6.74 0.60

Model D, b 14.27 0.99 4.7 0.24

0.73*** 1.63* 6.45*** 3.16* 0.01 0.45 0.91*** 0.62 0.34* 0.19 1.85

0.73*** 1.68* 6.43*** 3.22* 0.04 0.46 0.9*** 0.62 0.34* 0.18 1.78

0.16 0.15 0.12 2.15 0.01 1.73**

0.20 0.29 0.17 2.36 0.003 1.59* 0.58 2.54 11.18 0.2 1.29

58.25*** 300.32* 1271.4* 6.06* 79.66 250.62*

0.77*** 13.68*** 86.6* 630.37* 3.61* 41.0 253.8*

0.68*** 6.73 93.46* 815.1* 4.42* 54.25 246.15*

.68*** 6.61 93.36* 811.67* 4.41* 54.1 246.15*

B.J. Gillespie / Advances in Life Course Research 18 (2013) 223–233

four models were tested on three levels using a linear mixed modeling structure. The average PIAT academic achievement score for the respondents in the study is 57.95. More importantly, the estimated within-person, within-household, and between-person residual variance differs significantly from 0, indicating statistically significant individual differences in average academic achievement. Since the variance components are significantly different from 0, additional time-variant and time-invariant predictors may improve model fit. Model A in Table 2 suggests that there is no association between geographic mobility and academic achievement. In Model B, geographic mobility effects on academic achievement are being measured with the inclusion of interaction terms between local mobility and age, as well as distance mobility and age. Both local mobility and the age interaction with local mobility are significant in this model (p < 01). However, the insignificance of the main effect of local mobility (in Model A) implies a null effect of geographic mobility on academic achievement. Thus, in conjunction with Model A, these results suggest that the main effect is meaningless in the presence of the interaction term between local mobility and age. It is important to turn to the next models (Models C and D) in order to more fully understand the relationship between local mobility and academic achievement. Model C was the first theoretical model in which social capital and household and individual-level characteristics were added to the best fitting unconditional growth model. In Model C, the added variables for individual and household include age, sex, black or non-black status, logged household income, father in household, number of children in the household, birth order, mother’s age at birth of child, mother’s education, urbanicity, and a dummy variable for mother never married. The added variables are for social capital are parental knowledge of the child’s friends and the respondent child’s Catholic school attendance (parent–child community connectivity); the child’s participation in extracurricular activities, religious service attendance, and child’s self-reported level of loneliness (child–community connectivity); and the child’s report of closeness to his or her mother (parent–child connectivity). Analyses not shown here suggest that we cannot distinguish statistically between the academic achievement trajectories of children with non-black, non-hispanic and hispanic mothers in this model. For this reason, these models include just one race predictor (black). For social capital, only child participation in extracurricular activities after moving is significantly and positively associated with academic achievement (p < .01). As expected, the baseline PIAT score is a highly significant and positively associated predictor of later academic achievement. The linear age term is also significant; indicating that academic achievement scores decrease over time. Once again, geographic mobility does not have a significant effect on academic achievement in this model. Overall, Models A through C do not support the hypothesis (H1A) that geographic mobility will be negatively associated with academic achievement. The full and final model, Model D, adds social capital interactive effects with mobility to Model C. Participation

229

in extracurricular activities is still significant (p < .05). Interestingly, however, social capital does not have any moderating effects on geographic mobility and academic achievement. Further, geographic mobility was not significant in the final achievement models, nor was the interaction between geographic mobility and age. However, several individual and household characteristics are significant predictors of achievement (age, male, black, income, and mother’s education). Taken together, these results suggest that neither geographic mobility nor the loss of social capital associated with moving are associated with changes in academic achievement at any age through adolescence. 5.2. Behavior problems Similar to academic achievement, using the BPI for behavioral problems as a dependent variable, analysis was run in four models on three hierarchical levels (time within children and children within households). Table 3 shows the estimation of behavior problems in these four steps: Model A shows the total effect of geographic mobility, Model B adds interactions to test for age effects. Model C adds adjustment for social capital and individual and household characteristics, and Model D adds the series of interactions for social capital and mobility. With regard to the nature of the relationship between geographic mobility and behavior problems, Model A in Table 3 shows significant differences between nonmovers, distance movers, and local movers for the respondent child’s BPI score. Model B includes interaction terms for age and both types of geographic mobility. The coefficients for distance mobility and the interaction term between age and mobility are significantly associated with behavior problems. Moreover, the significant effect of distance mobility on behavior problems in Model A suggests that the main effect – and the interaction term in Model B – are indeed independently statistically and meaningfully significant. This indicates that while distance mobility is significant for increasing adolescent behavior problems, the effect is mitigated as the respondent ages. Model C in Table 3 adds relevant household and individual-level predictors of child behavior problems. Unlike achievement, there are several significant social capital predictors of behavior problems. The coefficients for closeness to mother (p < 01), child’s friends known by name (p < 001), and loneliness (p < 001) are all highly significant in this model. This provides preliminary support for Hypothesis 4—that social capital is more strongly associated with behavior problems than achievement. Importantly, distance mobility is significant (p < 01) and the age interaction with distance mobility is also significant (p < 01). Thus, Model C provides support for H1B; geographic mobility is positively associated with adolescent behavior problems. Model D tests whether or not the social capital significantly moderates the effects of moving on behavior problems net of individual and household-level characteristics. As this final model indicates, several social capital

230

B.J. Gillespie / Advances in Life Course Research 18 (2013) 223–233

Table 3 Linear mixed model of geographic mobility on behavior problems. Behavior problems index (BPI) Model A, b Geographic mobility Local mobility Local mobility  age Distance mobility Distance mobility  age Individual/household Age Male Black Income (logged) Father in household Urban Mother’s education Birth order Mother’s age at child’s birth Children in household Never married Social capital Church attendance Closeness to mother Friends by name Catholic Loneliness Extracurricular activities Interaction terms Distance move  church attendance Distance move  closeness to mom Distance move  friends by name Distance move  catholic Distance move  loneliness Distance move  extracurricular BPI baseline Constant Variance components Level 2 Level 3 (initial status) Rate of change Covariance Residual

2.97a 3.74*

Model B, b

Model C, b

13.68 0.95 29.75* 2.19*

16.22 1.18 37.17** 2.8**

Model D, b 10.62 0.71 45.2* 3.1**

0.11 2.61** 0.09 0.03 0.43 1.1 0.1 0.06 0.13 0.09 1.13

0.13 2.65** 0.1 0.08 0.36 1.1 0.12 0.11 0.17 0.12 1.2

0.56 1.64** 1.88*** 1.79 3.36*** 0.34

0.51 1.67** 1.68*** 1.63 3.44*** 0.76 0.64 0.45 2.39a 10.89** 1.03 5.61a

0.57*** 24.22***

55.14*** 331.66* 1130.8* 5.3* 70.1 90.3*

0.55*** 26.3***

61.17* 1000.8* 5.49* 69.17* 87.19*

.55*** 28.99**

39.97* 767.03* 4.1* 51.3 87.37*

41.28* 775.2* 4.1* 51.67 87.3*

Note: robust standard errors. a p < 10. * p < 05. ** p < 01. *** p < .001.

than the individual and household-level models. Level-1 and level-2 random effects remain significant in each model, meaning that additional level 1 and 2 predictors may improve model fit. 50

Behavior Problems Index

moderators have at least marginally significant explanatory value for behavior. There are significant interactive effects for parent–community closure (knowing child’s friend’s parents and Catholic school attendance) and child– community closure (loneliness) on behavior problems. Model D also shows support for H5B (the mobility– resilience hypothesis). There is a significant negative effect for the interaction between age and geographic mobility. Therefore, moving significantly contributes to the behavior problems of children, and these effects lessen as children grow older (see Fig. 1). This final model also shows partial support for H6—that the negative effects of moving exist over and above individual, household, and social capital characteristics. These effects do persist; however, they only do so for behavior problems (not achievement). The considerable decrease in information criterion fit statistics (not shown) clearly indicates that the behavior and achievement final models are a significantly better fit

40 Local 30

Distance

20 10 0 Low Age

High Age

-10

Fig. 1. Graph of geographic mobility and age interaction on behavior problems.

B.J. Gillespie / Advances in Life Course Research 18 (2013) 223–233

6. Discussion and conclusion Moving is difficult. It is arguably among life’s most stressful commonly-occurring events. The stress induced by the interruption of daily life and routine, the emotional upheaval brought on by a move, and the time consumption by the act of moving may be enough to have a significant impact on children’s behavior without employing alternate explanations, such as the loss of social capital. The current study tested the effects of geographic mobility on adolescent outcomes while also considering the effects of individual and household demographics and changes in social capital. To test these hypotheses with the greatest rigor, a longitudinal analysis was conducted using four waves from the linked mother–child data from the National Longitudinal Survey of Youth 1979. The results of this research both confirm and extend the findings of previous researchers. In support of the first hypothesis (H1B), and consistent with previous research (Haynie et al., 2006; Jelleyman & Spencer, 2008), behavior problems are positively associated with moving. However, this relationship does not hold for academic achievement (H1A) where geographic mobility is insignificant even in the baseline model. This latter finding contradicts recent findings that geographic mobility is associated with lower standardized test scores (Ersing, Sutphen, & Loeffler, 2009; Lyle, 2006). Why might geographic mobility matter for children’s behavior problems but not academic achievement? One possible explanation for this is that school, as opposed to residential, mobility may be associated with negative academic outcomes. This qualitative difference in mobility should be addressed in future research. Another reason may be that most other studies on the effects of geographic mobility on achievement outcomes have used small-scale non-representative data (Lyle, 2006), cross sectional analysis (Ersing et al., 2009), or short-term longitudinal studies (South et al., 2005). Of course, as with most empirical studies of correlational data, there is a possibility that unobserved parent and household characteristics account for the geographic mobility. This is consistent with findings that moving negatively influences educational outcomes because of preexisting factors that select individuals into moving, such as family disruption, parent’s income, or education (Pettit & McLanahan, 2003; Gasper et al., 2010). These analyses provide support for this explanation. Individual and household characteristics of the mobile population are the strongest predictors of academic achievement. It may be that ascribed characteristics that select individuals into geographic mobility also affect adolescent outcomes. It was hypothesized that a loss of social capital associated with a distance move would be more detrimental to adolescent outcomes than a local move (H2). Additional hypotheses posit that social capital would moderate the effects of geographic mobility on behavior problems (H3) and less so on adolescent academic outcomes (H4). These hypotheses were supported by the analysis. The social capital main effects model (Tables 2 and 3, Model C) shows that, at least generally, social capital is

231

strongly associated with behavior problems and not achievement. Model D in Tables 2 and 3 examines the effect of social capital on child outcomes as it relates to geographic mobility. Several social capital measures were found to significantly moderate the negative effects of moving on behavior problems. However, the only significant social capital predictor for academic achievement is whether or not the child participates in extracurricular activities. The findings extend previous research that contends that social capital has greater effects on behavioral than cognitive outcomes (McNeal, 1999), even when considering the loss of social capital associated with geographic relocation. Previous research shows that social capital and friendship during adolescence are markedly more important than at any other age across the life course (South & Haynie, 2004). It was expected that ‘‘timing in life’’ would play an important part in the detrimental effects of moving on children. Neither geographic mobility nor the movingby-age interaction have significant effects on adolescent academic achievement, net of the other theoretically important variables in the analysis. However, in support of Hypothesis 4B (the mobility–resilience hypothesis), moving is associated with behavior problems but the effects lessen as the child approaches young adulthood. This may be an indication that older children are better able to seek out social capital and support, such as school counselors, etc., to offset the negative behavioral effects of a move. This is an important finding, because research that is limited to older children may reach conclusions that do not hold for younger ones. Finally, the negative effects of geographic mobility were expected to exist above and beyond individual and family characteristics and social capital (H5). For academic achievement, moving does not have significant effects in any model. However, mobility was found to affect adolescent behavior problems above and beyond social capital characteristics. These analyses are subject to several caveats. Measuring child outcomes across only four waves of this longitudinal survey does not allow for analysis of behavior and achievement effects that take longer than two, four, or six years to develop. In fact, some evidence exists that geographic mobility in childhood is related to subjective well-being even into adulthood (Oishi & Schimmack, 2010). Reverse causation may also be present in the models above. For instance, problem behavior may cause children to have distant relationships with their parents, or families may be more or less likely to move because of their child’s preexisting behavior problems or school achievement. The sample has its own limitations. As mentioned above, there may be a selection issue related to the young age of some mothers at the time of the original interview (and young maternal age has been linked to impaired child development; Geronimus, Korenman, & Hillemeier, 1994). Further, because only the children of NLSY female respondents are surveyed, father–child and father–community interaction (other than what is reported by the mother) cannot be assessed as a component of social capital. Also, because of the NLSY design, children raised in

232

B.J. Gillespie / Advances in Life Course Research 18 (2013) 223–233

single-father homes are not included. Lastly, the fact that the BPI is based on the report of the parent might pose additional problems. However, studies have found parents’ perceptions of their children’s behavior to be both valid and reliable (Bird, Gould, & Staghezza, 1992). Connecting longitudinal data on children’s behavior problems and school achievement to parents’ involvement and household demographic characteristics lends greater understanding to how, why, and under what circumstances geographically mobile children develop behavioral or academic problems. The findings of the current study clarify four important issues. First, different predictors matter for behavior problems than for academic achievement in adolescence. Prior researchers had posited similarly negative mobility effects for a range of child outcomes (Coleman, 1988; Hagan et al., 1996; Hendershott, 1989; Norford & Medway, 2002). It appears that educational achievement is less affected by mobility than it is by individual and household-based characteristics, such as the child’s sex and parental education. Behavior problems, on the other hand, are affected by the act of moving. Second, these findings highlight the importance of analyzing effects as they vary by child’s age. The negative behavioral effects of geographic mobility significantly decrease for children as they age. This is an important finding, because research that is limited to older children may reach conclusions that do not hold for younger ones. Third, different types of mobility affect children differently. Distance mobility, which is associated with employment relocation, family changes, and other housing considerations, often implies a greater loss of community ties and social capital. The findings of this study substantiate that distance mobility (and the attendant loss of social capital) has more prominent negative behavioral effects on children than local mobility. Nevertheless, measuring only distance mobility may be downplaying the stress that accompanies moving at all. Thus, fourth, in respect to behavioral problems, the negative effects of geographic mobility exist above and beyond individual and family characteristics and social capital measures. At least for behavior problems, widely accepted mechanisms may not be the only, best, or even simplest explanations for the negative effects of moving on children. Acknowledgments The author would like to thank Judith Treas, George Farkas, and the two anonymous reviewers for their advice and comments on earlier drafts of this manuscript. References Achenbach, T. M., & Edelbrock, C. S. (1981). Behavioral problems and competencies reported by parents of normal and disturbed children aged four through sixteen. Monograph of the Society for Research in Child Development, 46(1), 1–82. Ahn, J. (2012). Teenagers’ experiences with social network sites: relationships to bridging and bonding social capital. The Information Society, 28, 99–109. Aseltine, R. H., Jr., Gore, S., & Colten, M. E. (1998). The co-occurrence of depression and substance abuse in late adolescence. Development and Psychopathology, 10, 549–570.

Astone, N. M., & McLanahan, S. S. (1994). Family structure, residential mobility, and school dropout: a research note. Demography, 31(4), 575–584. Berndt, T. J. (1982). The features and effects of friendship in early adolescence. Child Development, 53(6), 1447–1460. Bird, H. R., Gould, M. S., & Staghezza, B. (1992). Aggregating data from multiple informants in child psychiatry epidemiological research. Journal of the American Academy of Child and Adolescent Psychiatry, 31(1), 78– 85. Brett, J. M. (1982). Job transfer and well-being. Journal of Applied Psychology, 67(4), 450–463. Brown, B. B. (1990). Peer groups and peer cultures. In S. S. Feldman & G. R. Elliott (Eds.), At the threshold: the developing adolescent (pp. 171–196). Cambridge, MA: Harvard University Press. Coleman, J. S. (1988). Social capital in the creation of human capital. American Journal of Sociology, 94, 95–120, Supplement: Organizations and Institutions: Sociological and Economic Approaches to the Analysis of Social Structure. Conger, R. D., Ge, X., Elder, G. H., Jr., Lorenz, F. O., & Simons, R. L. (1994). Economic stress, coercive family process, and developmental problems of adolescents. Child Development, 65(2), 541–561. Cooksey, E. C., Menaghan, E. G., & Jekielek, S. M. (1997). Life-course effects of work and family circumstances on children. Social Forces, 76(2), 637– 667. Crouter, A. C., MacDermid, S. M., McHale, S. M., & Perry-Jenkins, M. (1990). Parental monitoring and perceptions of children’s school performance and conduct in dual- and single-earner families. Developmental Psychology, 26(4), 649–657. Demuth, S. (2004). Understanding the delinquency and social relationships of loners. Youth & Society, 35(3), 366–392. Domina, T. (2005). Leveling the home advantage: assessing the effectiveness of parental involvement in elementary school. Sociology of Education, 78(3), 233–249. Douvan, E., & Adelson, J. (1966). The adolescent experience. New York: Wiley. Dunn, L. M., & Markwardt, F. C. (1970). Examiner’s manual: peabody individual achievement test. Circle Pines, MN: American Guidance Service. Elder, G. H., Jr. (1987). War mobilization and the life course: a cohort of World War II veterans. Sociological Forum, 2(3), 449–472. Elder, G. H., Jr. (1998). The life course as developmental theory. Child Development, 69(1), 1–12. Epstein, J. L., & Sanders, M. G. (2002). Family, school, and community partnerships. In M. H. Bornstein (Ed.), Handbook of parenting, vol. 5: practical issues in parenting (pp. 407–437). Mahwah, NJ: Erlbaum. Ersing, R. L., Sutphen, R. L., & Loeffler, D. N. (2009). Exploring the impact and implications of residential mobility: from the neighborhood to the school. Advances in Social Work, 10(1), 1–18. Fischer, C. S. (2002). Ever-more rooted Americans. City and Community, 1(2), 177–198. Frey, W.H., (2003). Who moves where: A 2000 Census survey. Washington, DC: Population Reference Bureau. Gasper, J., DeLuca, S. A., & Estacion, A. (2010). Coming and going: explaining the effects of residential and school mobility on adolescent delinquency. Social Science Research, 39, 459–476. Geronimus, A., Korenman, S., & Hillemeier, M. M. (1994). Does young maternal age adversely affect child development? Evidence from cousin comparisons in the United States. Population and Development Review, 20(3), 585–609. Gillespie, B. J., in press. Does intergenerational solidarity buffer the negative effects of residential mobility? Evidence for U.S. children. Field Action Science Reports. Gillespie, B. J., & Bostean, G. in press. Socioeconomic status, residential mobility, and health. In K. Fitzpatrick (Ed.), Poverty in America: health and well-being among the vulnerable, volume II: Place, race and health. Santa Barbara: Praeger. (chapter 2). Grolnick, W. S., & Ryan, R. M. (1989). Parent styles associated with children’s self-regulation and competence in school. Journal of Educational Psychology, 81(2), 143–154. Hagan, J., MacMillan, R., & Wheaton, B. (1996). New kid in town: social capital and the life course effects of family migration on children. American Sociological Review, 61(3), 368–385. Haynie, D. L., South, S. J., & Bose, S. (2006). The company you keep: adolescent mobility and peer behavior. Sociological Inquiry, 76(3), 397–426. Hendershott, A. B. (1989). Residential mobility, social support and adolescent self-concept. Adolescence, 24(93), 217–232. Hofferth, S. L., Boisjoly, J., & Duncan, G. J. (1998). Parents’ extrafamilial resources and children’s school adjustment. Sociology of Education, 71(3), 246–268. Holmes, T. H., & Rahe, R. H. (1967). The social readjustment rating scale. Journal of Psychosomatic Research, 11, 213–218.

B.J. Gillespie / Advances in Life Course Research 18 (2013) 223–233 Humke, C., & Schaefer, C. (1995). Relocation: a review of the effects of residential mobility on children and adolescents. Psychology, 32, 18–24. Ingersoll, G. M., Scamman, J. P., & Eckerling, W. D. (1989). Geographic mobility and student achievement in an urban setting. Educational Evaluation and Policy Analysis, 11(2), 143–149. Jelleyman, T., & Spencer, N. (2008). Residential mobility in childhood and health outcomes: a systematic review. Journal of Epidemiology and Community Health, 62, 584–592. Kilborn, P. T. (2009). Next stop, Reloville: life inside America’s new rootless professional class. New York: Henry Holt and Company. Long, L. H. (1988). Migration and residential mobility in the United States. New York: Russell Sage Foundation. Long, L., Tucker, C. J., & Urton, W. L. (1988). Migration distances: an international comparison. Demography, 25(4), 633–640. Lyle, D. S. (2006). Using military deployments and job assignments to estimate the effect of parental absences and household relocations on children’s academic achievement. Journal of Labor Economics, 24(2), 319–350. Madigan, T. J., & Hogan, D. P. (1991). Kin access and residential mobility among young mothers. Social Science Quarterly, 72(3), 615–622. McLanahan, S., & Sandefur, G. (1994). Growing up with a single parent: what hurts, what helps. Cambridge. MA Harvard University Press. McNeal, R. B., Jr. (1999). Parental involvement as social capital: differential effectiveness on science achievement, truancy, and dropping out. Social Forces, 78(1), 117–144. Muller, C. (1998). Gender differences in parent involvement and adolescents’ mathematics achievement. Sociology of Education, 71, 336–356. Norford, B. C., & Medway, F. J. (2002). Adolescents’ mobility histories and present social adjustment. Psychology in the Schools, 39(1), 51–62. Oishi, S., & Schimmack, U. (2010). Residential mobility, well-being, and mortality. Journal of Personality and Social Psychology, 98(6), 980–994. Parcel, T. L., & Dufur, M. J. (2001). Capital at home and at school: effects on child social adjustment. Journal of Marriage and the Family, 63(1), 32–47. Perry-Jenkins, M., Repetti, R. L., & Crouter, A. C. (2000). Work and family in the 1990. Journal of Marriage and the Family, 62(4), 981–998. Peterson, J. L., & Zill, N. (1986). Marital disruption, parent–child relationships, and behavior problems in children. Journal of Marriage and the Family, 48(2), 295–307. Pettit, B., & McLanahan, S. (2003). Residential mobility and children’s social capital: evidence from an experiment. Social Science Quarterly, 84(3), 632–649.

233

Pribesh, S., & Downey, D. B. (1999). Why are residential and school moves associated with poor school performance? Demography, 36(4), 521–534. Pryor, J. (1999). Waiting until they leave home: the experiences of young adults whose parents separate. Journal of Divorce and Remarriage, 32(1– 2), 47–61. Rossi, P. H. (1980). Why families move. Beverly Hills, CA: Sage Publications. Rumberger, R. W., & Larson, K. W. (1998). Student mobility and the increased risk of high school dropout. American Journal of Education, 107(1), 1–35. Scanlon, J., & Devine, K. (2001). Residential mobility and youth well-being: research, policy, and practice issues. Journal of Sociology and Social Welfare, 28(1), 119–138. Simpson, G. A., & Fowler, M. G. (1994). Geographic mobility and children’s emotional/behavioral adjustment and school functioning. Pediatrics, 93(2), 303–309. Singer, J. D., & Willett, J. B. (2003). Applied longitudinal data analysis: modeling change and event occurrence. New York, NY: Oxford University Press. South, S. J., & Haynie, D. L. (2004). Friendship networks of mobile adolescents. Social Forces, 83(1), 315–350. South, S. J., Haynie, D. L., & Bose, S. (2005). Student mobility and school dropout. Social Science Research, 36, 68–94. Spera, C. (2005). A review of the relationship among parenting practices, parenting styles, and adolescent school achievement. Educational Psychology Review, 17(2), 125–146. Stack, S. (1994). The effect of geographic mobility on premarital sex. Journal of Marriage and the Family, 56(1), 204–208. Teachman, J. D., Paasch, K., & Carver, K. (1996). Social capital and dropping out of school early. Journal of Marriage and the Family, 58(3), 773–783. Tucker, C. J., Marx, J., & Long, L. (1998). ‘‘Moving on’’: residential mobility and children’s school lives. Sociology of Education, 71(2), 111–129. U.S. Bureau of the Census. 2012. Geographic mobility: March 2010 to March 2011. Current population reports, Washington DC: U.S. Government Printing Office. White, A. M., & Gager, C. T. (2007). Idle hands and empty pockets? Youth involvement in extracurricular activities, social capital, and economic status. Youth and Society, 39(1), 75–111. Wood, D., Halfon, N., Scarlata, D., Newacheck, P., & Nessim, S. (1993). Impact of family relocation on children’s growth, development, school function, and behavior. Journal of the American Medical Association, 270(11), 1334–1338.