Navigation

Adolescent Drug Use

INTRODUCTION

Children of drug dependent parents are at an elevated risk for drug use, delinquency, and the development of other problem behaviors. [1-3] In studies comparing them to demographically matched controls, children of substance abusers exhibit more behavioral and emotional problems, less socially adaptive behavior, higher rates of psychiatric disorder, and greater use of illicit drugs. [4-8] Research on children of substance abusers is still limited, but evidence indicates that youth in drug-affected families are exposed to a multitude of conditions identified as risk factors for the development of problems in childhood and adolescence. [9,10]

Families headed by drug dependent parents typically are characterized by poor family management practices, high family conflict, low family cohesion, and high stress. [2,4,11-13] Children in these families are subject to the consequences of their parents' many difficulties, including poverty, employment problems, poor living conditions, family violence, physical and mental illness, trouble with the law, and low educational attainment. [2,14-17] Given both high levels of environmental adversity and high rates of behavioral problems among children of substance abusers, researchers have begun examining the heterogeneity in this population to identify particular risk factors associated with problematic child outcomes. [18-21]

The present study focuses on effects of the family instability and disruptions in household composition which are so often experienced by children of substance abusing parents. [12,15,22-24] Employing a prospective design which controls for baseline child behaviors, parental risks, and family conflict, the study investigates the influence of parent figure transitions on the likelihood of delinquency and drug use among a sample of early adolescents with parents receiving treatment for opiate addiction.

Family Disruption and Youth Adjustment Problems

Almost all conceptual and empirical work relating family composition to the development of problem behaviors in youth, such as drug use and delinquency, has focused on the marital status of parents. [25-28] This literature indicates that children and adolescents who experience family disturbances due to divorce and remarriage typically demonstrate higher levels of aggressive, defiant, and delinquent behavior. [26,27,29-31] One explanation is that marital breakup produces conditions and consequences that have an adverse effect on children. Mechanisms proposed to explain the link between marital disruption and youth behavior problems include parental absence, economic hardship, parental distress, family conflict, damaged parent-child relationships, and problems in family functioning resulting from divorce. [25,28,32,33] Remarriage can reverse financial hardship and provide new parenting resources, but it also can introduce additional strain as family members adapt to new roles and relationships and as children struggle with issues of loyalty to the new parent figure. [28,34]

An alternative explanation for the link between marital disruption and youth adjustment problems is that both are manifestations of a single underlying cause. Certain parental factors, such as psychological distress, antisocial behavior, or a conflictive interpersonal style, may make multiple marital transitions more likely and also contribute to parenting practices that place children at risk for poor outcomes regardless of family disruptions. [32,35,36] This possibility certainly deserves attention with substance abusing parents, given the prevalence of mental health problems, antisocial behavior, and family conflict which likely precede family disruptions. Furthermore, one of the most consistent findings regarding children of substance abusers is that co-occurring parental psychiatric disturbance greatly increases the risk of psychopathology among the children. [8,18,21]

The explanations presented above are not mutually exclusive, and several mechanisms may operate simultaneously in a complex, interactive fashion to affect children's adjustment. [26,32] In fact, the combination of several risk factors noted above may be highly relevant in explaining the consequences of family transitions. [25] Evidence suggests that the accumulation of multiple risk factors has the greatest negative impact on child adjustment and development. [37-39] Adolescents who encounter a high number of stressful life events are more likely to engage in drug use, aggression, and delinquency. [40,41] Furthermore, the accumulation of risk can have a temporal component, suggesting that chronic or frequently repeated stresses would be related to the initiation and escalation of drug use, delinquency, and other behavior problems. [25,42]

Following this line of reasoning, a series of negative life events associated with a history of multiple family disruptions would be particularly detrimental. [43] Although marital breakup is a stressful event that can precipitate emotional distress and behavioral disorders, many middle class children successfully adapt to their changed life circumstances within the first 2 years following a break-up. [44] However, regaining equilibrium is less likely if a child encounters further disturbances, such as those caused by remarriage or other subsequent changes in family structure. [28] As Brody et al. [43] observe, "when family destabilizations happen repeatedly and frequently, adjustment to one transition is immediately followed by another, perhaps before adjustment to the previous one is complete." In such cases, effects from an unresolved initial divorce may be amplified by each succeeding disruption. Chronic family instability would presumably be harder to overcome for children of substance abusers who also confront multiple other risks.

Research on Multiple Marital Disruptions

Studies specifically testing the cumulative effect of multiple transitions due to divorce and remarriage have found the degree of antisocial behavior tends to vary with the number of disruptions experienced. In most of these studies, the number of parenting transitions was deduced retrospectively from the parent's marital status at a given point in time. For example, Furstenberg and Seltzer [45] counted children living with both parents as having no parenting transitions; children living with a single parent had one parenting transition due to divorce; divorce and remarriage represented two transitions; and additional divorces or remarriages signified multiple parenting transitions. These researchers reported successively higher levels of fighting and school discipline problems as the number of parenting transitions increased. [45]

Other studies have applied the same approach for classifying parenting transitions. Using three independent cross-sectional samples of early adolescents, Kurdek et al. [46] found support for an inverse relationship between the number of transitions and measures of adjustment. Capaldi and Patterson [36] also detected the hypothesized linear relationship between the number of parenting transitions and scores on a composite measure of problem behaviors in multiple domains (antisocial behavior, peer relations, academic problems, self-esteem, deviant peers, drug exploration, and delinquency) in a cross-sectional sample of boys in grade 4. Further analyses reported in this study revealed that parents who exhibited antisocial behaviors were more likely to experience relationship difficulties and to demonstrate poor parenting practices, suggesting that the boys experiencing transitions might have been exposed to multiple risk factors prior to family disruptions. [36]

Several studies on family structure have reported the same pattern of increasing problem behaviors when comparing adolescents from intact, divorced, and remarried families. [47,48] Demo and Acock [49] observed higher scores on socio-emotional adjustment, academic performance, and global well-being for adolescents in intact two parent families when compared with adolescents in divorced or remarried families. However, adolescents living with a continuously single mother were not significantly different from their peers in intact two parent families. These researchers speculated that adolescents whose mothers had never married benefited from their "intact, nondisrupted family history". [49]

Noting that "simply coding the home as either intact or broken obscures data concerning the stability of the family," Mednick et al. [50] constructed a variable to indicate family instability by summing the total number of changes in the adult configuration of the home during the course of their longitudinal study on a birth cohort of Danish males. Although a simple dichotomous measure of whether a child was from a broken home did not predict crime in young adulthood, the family instability measure reflecting the total number of changes in family constellation during the first 18 years was a highly significant predictor of later criminal behavior, even after controlling for paternal crime and socioeconomic status. This study also suggested that the timing of the instability was influential, with the degree of instability during adolescence having a strong linear relationship with criminality. [50]

Methodological Considerations for Investigating Multiple Transitions

With the exception of the study by Mednick et al., [50] researchers have generally defined family configurations and family disruptions with reference only to parents' marital status. This approach potentially distorts information regarding the presence or absence of other adult figures in the household, since the prevalence of cohabitation with an unmarried partner has grown significantly in recent decades. [51-53] In addition, many other household arrangements involve adults who, to a greater or lesser extent, assume the social and economic roles of an absent parent. [54] Approximately one-third of female single parents have lived with their own parents at some time, [51] and increasing numbers of grandparents are taking full responsibility for parenting their grandchildren. [55] In other cases, children are referred into state custody and placed into homes through the child welfare system, and these children frequently experience multiple foster family placements. [56,57] Children of substance using parents are especially likely to find themselves in the care of relatives or in the child welfare system. [12,15,22-24]

The majority of studies investigating how change in family structure affects children's behavior have, in fact, been cross-sectional. Although adjustment to change in family structure is a developmental process, most cross-sectional research fails to account for the child's developmental status at the time of the disruption, age at the time of outcome assessment, or the length of time elapsed since the disruption. [26,32,58,59] Furthermore, since cross-sectional indicators of family structure do not reflect the variation in parenting configurations that children may experience over time, longitudinal examination of the consequences for youth exposed to a series of family transitions and reorganizations is more appropriate. [27,28]

Cross-sectional research is also unable to account for parent, child, or family functioning prior to disruption. Only a handful of prospective studies have compared behavioral outcomes for children experiencing a divorce to those from intact families while controlling for pre-divorce levels of the behaviors. Some of these prospective studies have found diminished or negligible effects of divorce. [60-62]

Finally, the question of whether vulnerability to disruptions might vary by gender has not been adequately addressed. Two key studies investigating multiple transitions have employed samples composed only of males. [36,50] Furstenberg and Seltzer [45] did not account for gender in their examination of parent transitions, but a subsequent study using the same data showed greater negative effects of marital dissolution for girls. [63] Kurdek et al. [46] found a gender difference in the association between parent transitions and drug abstinence scores. Among adolescents with multiple parent transitions, boys reported low levels of drug abstinence while girls reported high levels of abstinence. In general, evidence for gender differences in the response to divorce is not as strong as once believed, but studies indicate that boys adapt to remarriages (usually involving a stepfather) more favorably than girls do. [26,32]

Present Study

The prospective study reported here examines whether family disruptions predict early adolescent drug use and delinquency in a sample of early adolescents with parents enrolled in treatment for opiate addiction. Children of substance abusers have not been represented in the previous research on parenting transitions despite being at risk for experiencing repeated family disruptions. Early adolescence deserves attention because it is a developmental period when youth may be more susceptible to family disruption and more likely to initiate behavior problems. [28] The youth in the study were assessed for baseline levels of drug use and delinquency and three follow-up interviews were conducted over a 2.5 year period. Parent figure transitions during the study period were examined for associations with youth behaviors at the final assessment. Identification of parent figures was based on the youth's own designation of who was fulfilling either the mother or father role, so transitions reflected changes in meaningful parenting figures rather than indicating only changes in marital status. The longitudinal design of the study permitted us to: (1) control for levels of youth drug use and delinquency preceding the transitions, (2) focus on exposure to disruptions in the early adolescent developmental period; and (3) limit the time elapsed between disruption and the outcome to the duration of the study period.

The first objective of the study was to investigate associations between parent figure transitions and children's subsequent drug usage and delinquency after controlling for corresponding baseline levels of these behaviors. Based on previous research regarding the impact of multiple parenting disruptions, experiencing a greater number of parent figure transitions was expected to increase the likelihood of subsequent drug use and delinquent behavior. The second objective was to determine whether any associations identified between parenting transitions and youth outcomes would remain after controlling for child characteristics (age, gender), pre-existing parent difficulties (depressive symptoms, criminal history), and initial levels of family conflict. The third objective was to examine potential interactive effects between parenting transitions and child gender. Finally, to highlight distinctions between family structure and stability, the analyses also contrasted stable two parent families with stable single parent families and families experiencing different patterns of parent figure instability.

METHOD

Sample and Study Overview

The sample of 67 youths is derived from a family-based intervention study entitled Focus on Families. The primary purpose was to evaluate the preventive effect of parenting education and case management services on the levels of drug use and delinquency in the children of opiate-addicted parents. [64,65] At the time of enrollment in the study, the parents were receiving methadone treatment and had children between the ages of 3 and 14 living with them at least 50% of the time. Recruitment occurred at two methadone clinics. A total of 144 parents in 130 families, 78% of those eligible for the study, voluntarily agreed to participate. Details regarding the intervention and results of the experimental study have been reported elsewhere. [65,66]

The 130 families in the project had 178 children between the ages of 3 and 14 at the time of enrollment. Children age 6 and older were interviewed at baseline and at 6, 12, and 24 months after the four month intervention period. Follow-up interviews with the children were pursued even when the parent in methadone treatment no longer had custody. Three different developmentally appropriate interviews were used for the children according to age at the time of interview (6- 8, 9-10, 11 years and older). Each more advanced interview included previous items and added more extensive and developmentally appropriate measures of family functioning, peer interaction, and child behaviors. Interviews with the parenting adults enrolled in the study were also obtained. The structured, fixed response interviews were conducted in-person by highly experienced data collectors. Parents were paid $20 for completion of each interview, and youth respondents received a toy or a gift certificate for their cooperation.

To maintain measurement consistency and to focus on youth in the age range more likely to engage in the outcome behaviors, the sample selected for the present analysis was restricted to youths who were aged 9 years or older at the time of their baseline interview and 11 years or older at the final follow-up interview. Inclusion was dependent upon completion of the baseline and final follow-up interviews and at least one of the two interviews in between these two interviews. Several of the 89 child subjects meeting these criteria were members of sibling groups. To assure the independence of observations, one child from each of these families was randomly selected to be retained in the sample, yielding a sample of 69 youths eligible for the analysis. Comparison of the randomly excluded vs. included siblings revealed no statistically significant differences on any variables used in the analysis. Missing data on parent criminal history resulted in the deletion of two additional cases, yielding a final sample of 67.

The mean age for the sample was 11.4 years (SD = 1.7) at baseline and 13.8 years (SD = 1.6) at the time of the final interview. The sample was 45% female. The racial composition was 63% Caucasian, 25% African American, and 12% were of other racial/ethnic heritage. The subjects had a mother figure (82%), father figure (12%), or both parents (6%) in methadone treatment at baseline. The portion of the sample in the experimental condition was 55%. Almost all of the children (97%) lived with their biological mothers at baseline; the remainder lived with a stepmother (1.5%) or another mother figure (1.5%). At the final interview, however, only 81% of the children lived with their biological mothers, and there were corresponding increases in other arrangements such as stepmother (3%), foster mother (3%), other mother figure (7%), or no mother figure (6%). At the baseline interview more children were in a household with no father figure present (33%) than were living with a biological father (29%), but father figures also included stepfathers (18%), mothers' boyfriends (18%), and others (2%). By the final interview children were even more likely to live without a father figure (40%) than with a biological father present (25%). Other father figures reported at baseline included stepfathers (8%), adoptive fathers (2%), mothers' boyfriends (16%), or other father figures (9%).

At the time of enrollment in the study, the mean age of the parents was 36.6 years (SD = 3.9). The parents indicated the following levels of educational attainment: no high school completion (15%), high school diploma or GED (42%), some post-secondary education (39%), college degree (4%). In terms of initial marital status, fewer than one-quarter of the parents (23%) were married, 21% were divorced, 20% were separated, 6% were widowed, and 30% had never been married. These figures changed somewhat over the course of the study, with final interview percentages being 29% married, 31% divorced, 12% separated, 5% widowed, and 23% never married. The sample was predominantly low income, with the majority of parents (63%) receiving public assistance at the time of the final interview.

Measures

Parent Figure Transitions

Measures of parent figure transitions were derived from child responses to two questions at each interview. Interviewers prefaced the questions by stating that they referred to "your parents (or the people you live with who act as parents to you)" and the exact wording was "Who do you live with most of the time?" Interviewers then gave the following seven response options for the mother figure: natural mother, stepmother, adoptive mother, foster mother, girlfriend of your father, do not live with a person acting as mother, or other. The same format was used for a question pertaining to the father figure. Based on the pattern of responses over the four interviews, the total number of mother figure entries and exits experienced by each child was determined. For example, if the child lived with the natural mother at the first interview but no person acting as a mother at the second interview, one maternal transition (an exit) would be registered. If the same child indicated living with a foster mother at the third interview, another transition (an entry) would be added for a count of two. If the child was living with the natural mother again at the final interview, two more transitions (foster mother exit, natural mother entry) would combine for a total of four maternal transitions during the study period. Entries and exits were given equal weight because each represents a break in continuity or stability from the child's perspective. The number of father figure transitions was derived in the same fashion. Finally, the number of mother figure changes was combined with the father figure changes to arrive at a measure of the total number of parent figure transitions experienced.

In addition to the continuous variable indicating the quantity of parental transitions, a categorical variable was constructed to represent four types of transition patterns: (a) two parent figures--no transitions; (b) single parent--no transitions; (c) partial stability--one consistent parent figure, other changes; (d) no stability--no consistent parent figures.

Youth Delinquent Behavior

The outcome measure for delinquent behavior, taken from the final follow-up interview, was based on child self-report on questions asking about involvement in 10 types of antisocial behaviors during the preceding 6 months (graffiti, stealing, throwing rocks or bottles at cars/people, breaking into buildings, destroying property, picking fights, hitting someone hard enough to injure, mugging, buying/selling stolen property, and selling drugs). Participants indicating they had engaged in a particular activity were also asked to report the frequency of the behavior, but examination of the frequency distributions revealed the typical skew of low base rate behaviors and suggested the key distinction was any involvement vs. no involvement. The percentages of youth involved in each type of activity are presented in Table 1. Because each of the delinquent behaviors could be classified as moderate to serious in severity, [20] a dichotomous measure was created to indicate participation in any of the behaviors.

A similar dichotomous variable was created for delinquent behavior at the baseline interview. However, since several youths were administered the age 9-10 interview at baseline, the baseline measure for all subjects was derived from the limited set of six behaviors asked of the younger children (graffiti, stealing, throwing rocks/bottles, breaking in, destroying property, and fighting).

Youth Drug Use

The self-report outcome measure for drug use was based on questions asking about the use of nine types of drugs within the 6 months preceding the interview (alcohol, marijuana, glue, cocaine/crack, amphetamines, tranquilizers, sedatives, psychedelics, and narcotics). Youth indicating the use of a particular drug were asked to report the frequency of consumption. Examination of the fre-quencies revealed highly skewed distributions, again suggesting the key distinction was any use vs. no use. The percentages of youth reporting use of each type of substance are presented in Table 1. In this case, a high overlap of drug types was observed. Most youth who used marijuana also used alcohol (80%), and among those using drugs other than marijuana all but one also used alcohol (97%). Consequently, a dichotomous variable representing use of any substances during the previous 6 months was constructed. The baseline measure of drug use for all subjects was derived from the much more limited list of drug types asked of younger children on the age 9-10 interview at baseline, i.e., only alcohol and marijuana.

Parent Depressive Symptoms

At the baseline interview, the parent receiving methadone treatment was asked about experiencing symptoms of depression corresponding to diagnostic criteria during the year prior to the interview. If respondents indicated feeling depressed, sad, and hopeless or disinterested in enjoyable activities for a period of more than 2 weeks, they were asked about the presence of nine depressive symptoms (e.g., poor appetite, trouble sleeping, fatigue, guilt, trouble concentrating or making decisions, suicidal thoughts, etc.). An index reflecting the severity of mood disturbance was created by counting the number of depressive symptoms reported. The mean number of depressive symptoms was 5.2 (SD = 2.3).

Parent Criminal History

At baseline, the parent receiving methadone treatment was asked "Have you ever been convicted of an offense or crime?" A dichotomous indicator of criminal history was based on the response (negative = 0, affirmative = 1). Two out of three parents (66%) had been convicted of a crime.

Family Conflict

At baseline, the parent receiving methadone treatment rated how often people in the family had argued, been critical of each other, and become upset with each other (1 = never, 5 = always). This three item scale had adequate internal consistency (alpha = 0.75). The mean score for this scale was 3.4 (SD = 0.6).

Statistical Analyses

Because participants in the intervention differed from their counterparts in the control condition on child baseline drug use (t = 2.09, p = 0.04) and number of parent figure transitions (t = 1.88, p = 0.07), experimental group membership was included as a control variable in all analyses. Analyses for both delinquency and drug use revealed that experimental condition did not interact with the number of transitions. To focus on the effects of parental transitions during the study period, the baseline levels of youth drug use or delinquency were also used as control variables in all analyses. Logistic regression was used to test for associations between parenting transitions and the drug use and delinquency outcomes. Additional analyses were performed to determine whether parenting transitions contributed to the models after controlling for child demographic variables or parent functioning at baseline. Separate analyses were performed using first the number of parental transitions and then the type of parental transition pattern as the predictor variable. The estimated logistic regression parameters were exponentiated to present results in terms of odds ratios, i.e., odds of the outcome in test group divided by odds in reference group.

RESULTS

Descriptive Information on Parent Transitions, Delinquency, and Drug Use

The majority of youths in this high-risk sample (70%) experienced some type of parent figure transition during the study period, and 28% reported three or more parent figure entries/exits. The mean number of transitions for the entire sample was 2.2 (SD = 2.5), ranging from 0 to 10. When the number of transitions was tested for associations with the baseline demographic characteristics of children (age, gender, and race) and parents (age, education, public assistance, and initial marital status), only child age was significantly correlated with transitions (r = 0.26, p = 0.03).

Children who did not experience a transition were evenly divided between the "two parent-stable" and "single parent-stable" categories (15% each). The most common transition pattern was "partial stability" (45%), and the mean number of transitions for this group was 1.8 (SD = 0.7). Not surprisingly, the youths in the "no stability" group (25%) reported the most disruptions, with an average of 5.5 parent figure transitions (SD = 2.6). The only association found between the transition pattern and child and parent demographic variables was the expected correspondence with baseline marital status ([chi square] = 27.84, df = 12, p = 0.006). Most in the two parent-stable category had married parents (70%); many in the single parent-stable category had a never married parent (40%); partial stability was twice as common with a never married parent (41%) than with married or divorced parents (21% each); and no stability was most common with a divorced (41%) or separated parent (29%).

At the final interview, delinquent behavior was reported by over half the sample, with 60% engaging in at least one of the listed behaviors. Among those indicating any delinquency, the number of behaviors reported ranged from 1 to 8, with a mean of 2.8 (SD = 2.0). The most common forms of delinquency were picking fights (63%), stealing (53%), and graffiti (33%). A fair percentage also reported dealing in stolen goods (28%), destroying property (25%), hitting people (23%), and dealing drugs (19%). There was a statistically significant association with baseline delinquency ([chi square] = 4.6, df = 1, p = 0.03); the majority exhibiting delinquent behavior (80%) had also reported similar activity at the baseline interview.

At the final interview, drug usage was reported by just over half the youth (51%). Among those who indicated any use, 50% used only one substance, 27% used two different substances, and 23% used three or more different substances. The most commonly used substances were alcohol (85%), marijuana (59%), psychedelics (14%), and narcotics (12%). The relationship with baseline drug use was not statistically significant ([chi square] = 0.9, df = 1, p = 0.34); fewer than half (38%) had also reported use at the baseline interview.

There was an association between delinquency and drug use at the final interview ([chi square] = 8.1, df = 1, p = 0.005). With reference to the entire sample, 39% were engaged in both delinquency and drug use, 21% reported delinquency only, 12% reported drug use only, and 28% were not involved in either type of activity.

Number of Parent Figure Transitions and Delinquency

A statistically significant association between delinquent behavior and the number of parent figure transitions was found (Table 2, Model 1). Controlling for baseline delinquency and experimental condition, the likelihood of involvement in delinquent behavior increased with each additional parent figure transition (odds ratio = 1.57, p = 0.01). With a condition of no parenting transitions as a reference, having two parent figure entries or exits raised the odds of delinquency by 2.5 times, and the odds of delinquency were 6.0 times greater with four transitions. Adjusting for baseline delinquency and experimental condition, the estimated probability of delinquency for the reference group with no transitions was 0.39, compared to a probability of 0.51 with one transition, 0.62 with two transitions, and 0.80 with four transitions.

When controlling for child characteristics, baseline parental difficulties, and baseline family conflict, the association between the number of parent figure transitions and delinquency was only slightly attenuated, and it remained statistically significant (odds ratio = 1.47, p = 0.03) (Table 2, Model 2). Of the additional variables, only parent criminal history had a statistically significant association with child delinquency. Children of parents with criminal convictions were over five times more likely to engage in delinquent behavior (odds ratio = 5.12, p = 0.02). Child age and gender had no direct effect on delinquency. No interactive effect between gender and the number of transitions was found (Table 2, Model 3).

Number of Parent Figure Transitions and Drug Use

The number of parent figure transitions had a statistically significant association with drug usage after controlling for experimental condition and baseline levels of drug use (odds ratio = 1.29, p = 0.04) (Table 3, Model 1). The adjusted probability of drug use with no transitions was 0.38, compared to a probability of 0.44 with one transition, 0.50 with two transitions, and 0.62 with four transitions. However, this association became nonsignificant after accounting for child characteristics, parent difficulties, and family relations (Table 3, Model 2). The only statistically significant predictor in the expanded equation was child age, with each additional year more than doubling the odds of drug use (odds ratio = 2.32, p < 0.001).

In the case of drug use, however, a significant interaction between gender and parent transitions substantially improved the fit of the model (Table 3, Model 3). To determine the nature of the interaction, the probability of drug use was calculated for both males and females while holding constant all other variables at their respective mean values. For females, the likelihood of drug use increased dramatically with more parent figure transitions. In contrast, the probability of drug use declined somewhat for males experiencing more parent figure transitions. To further explore gender differences in the prediction of drug use, the full regression model was analyzed separately for males and females (Table 4). The number of parent figure transitions was the only variable significantly associated with the probability of drug use for females (odds ratio = 1.92, p = 0.04). Age had a strong positive association with the likelihood of drug use for males (odds ratio = 4.91, p = 0.004), but parent depressive symptoms reduced the odds (odds ratio = 0.56, p = 0.05).

Pattern of Parent Figure Transitions and Delinquency

The pattern of parental transitions was related to youth delinquency over and above the effects of experimental condition and baseline delinquency (Table 5, Model 1). With a stable two parent family as the reference category, the odds of delinquent activity were 7.1 times greater if the youth experienced partial stability, and the odds increased by a factor of approximately 36 for children with no consistent parent figure. Although the likelihood of delinquency was not statistically different for children in the two stable family types, more children in stable single parent households reported delinquency than children in stable two parent families. Odds ratios for the different patterns of family stability were diminished by the addition of variables for child characteristics, parent difficulties, and family relations into the model (Table 5, Model 2). Nevertheless, children who had no consistent parent figure were still significantly more likely to be involved in delinquent behavior (odds ratio = 19.56, p = 0.02). As before, parent criminal history was the only additional variable that significantly increased the risk of delinquency (odds ratio = 4.54, p = 0.04). Figure 1 presents the predicted probability of delinquent behavior associated with each transition pattern with all other variables in the full model held constant at their mean values. Gender interactions were not tested due to small numbers within categorical groupings.

[FIGURE 1 OMITTED]

Pattern of Parent Figure Transitions and Drug Use

Controlling for experimental condition and baseline drug use, the categorical variable for transition pattern was tested with stable two parent families as the reference group (Table 5, Model 1). There was no statistically significant difference in drug use between children in stable two parent families and their counterparts in stable single parent families, but it is interesting to note that children in the single parent families were actually less likely to report drug use. Children experiencing partial parent instability were slightly more likely to use drugs than the reference group, but this difference was not statistically significant. A marginally significant contrast was found for children with no consistent parent figure, who were over five times more likely to use drugs (odds ratio = 5.40, p = 0.06). In the final model for drug use, which included child characteristics, parent difficulties, and family relations, only child age was found to have a significant effect on the outcome. Figure 1 presents the predicted probability of drug use associated with each transition pattern with all other variables in the final model held constant at their mean values. Again, small categories precluded testing interaction effects.

DISCUSSION

As children of drug addicted parents, all the youths in the study could be considered at heightened risk for both experiencing family disruptions and engaging in problem behaviors. [2,12,13] In fact, 70% of the youths experienced at least one parent figure transition, and nearly one-quarter did not have a single consistent parent figure during the course of the study. Likewise, at the conclusion of the study period, approximately half the youths reported involvement in drug use, and an even higher percentage participated in delinquent activities. The percentages of youth engaging in the specific problem behaviors listed in Table 1 were elevated but not extreme in comparison to national epidemiological data reported in the CDC's Youth Risk Behavior Surveillance for 9th graders during the same time period, [67] For alcohol, marijuana, and cocaine, the 6-month prevalence for drug use in this study is higher than the 1-month prevalence reported by the CDC (alcohol: 43% this study vs. 40.5% nationally; marijuana: 30% vs. 13.2%; cocaine: 3% vs. 1.6%). Furthermore, CDC Surveillance data for the metropolitan region from which the sample was drawn shows even less discrepancy (alcohol: 43% this study vs. 46.1% Seattle; marijuana: 30% vs. 22%; cocaine: 3% vs. 2.8%). The comparison to CDC data, presented simply for reference, needs to be considered in light of the small study sample size and the lack of direct equivalence in measurement and age range.

The primary purpose of the study was to determine whether parent figure transitions, defined as entries and exits of adults having parental roles, were associated with the delinquency and drug use observed within this sample of high-risk young adolescents. Even after controlling for baseline levels of delinquency, parent figure transitions were associated with an increased likelihood of later delinquent activity for both boys and girls. Furthermore, parent figure disruptions remained a key predictor of youth delinquency when other potentially confounding factors were considered. The baseline assessments of parent depression and family conflict were not associated with family disruption or youth outcomes. The lack of an effect for parent depression on youth behavior is at odds with previous reports implicating comorbidity of psychiatric and substance use disorders in the development of child psychopathology, [8,18] but study design and measurement issues provide plausible explanations for the discrepancy. For example, this study focused on youth delinquency and drug use in contrast to child mental health problems. Also, previous studies measured parental depression and child behaviors concurrently.

In contrast, parent criminality appears to be a risk factor for both parent figure transitions and youth delinquency. As reported in previous studies, [50] youth with a parent previously convicted of a crime had a greater likelihood of engaging in delinquent activity. Follow-up analyses showed that families with a convicted parent averaged one more parent figure transition than families with no parental criminal record, but this difference was not statistically significant (t = 1.36, p = 0.177). When looking at the categorical variable for parent transitions, there was an association with parent criminal history ([chi square] = 9.22 df = 3, p = 0.03). Stable two-parent households were least likely to have a convicted parent (30% convicted), stable single-parent households had a higher likelihood (50% convicted), and families with partial or no stability were most likely to have a convicted parent (77% convicted in both cases). In the logistic regression analyses, the estimated coefficients and the significance levels of the parent figure transition variables were slightly attenuated after adding parent criminal history and other variables to the model. Although there may be some shared variance between criminal history and parent figure transitions, each factor made a significant independent contribution to the prediction of youth delinquency.

The results of this study with regard to delinquent behavior are consistent with literature indicating that multiple, repeated family disruptions have negative effects on children. [36,45,50] The chronic accumulation of stress arising from a series of parent transitions following closely upon each other may contribute substantially to involvement in delinquent behavior. [25] Stressors that might place children at increased risk for adjustment problems as a consequence of "serial" parenting disruptions include chronic or recurrent exposure to inter-parental conflict, parental emotional distress, diminished parenting capacity, and income fluctuations. [32,33,43] Furthermore, these children often experience other potentially distressing circumstances associated with family disruption, such as changes in residence, school, and social network. [33,43] Young adolescents, who often have not developed strategies for coping with adversity, may be particularly vulnerable. [68] They are likely to react to stressful events with depression, frustration, and anger, [69] and experiencing a persistent or increasing number of stressful events over time has been shown to escalate the level of delinquency among early adolescents. [42]

Drug use also was associated with parent figure transitions, but only for females. Analyses exploring the moderating effect of gender suggested different mechanisms of influence for males and females. Counter to expectations, parent figure transitions and parental depression seemed to decrease the probability of drug use for males. Instead, age was the predominant factor affecting the likelihood of drug use for males, suggesting the importance of developmental processes that operate independently of family disruption. An increasing prevalence of drug use with age is well documented in the literature and has been attributed to greater influence by deviant peers, greater access to illicit substances, and adolescent rebellion and experimentation. [19,70,71] However, recent research by Hoffmann and Su [72] on gender differentiated predictors of adolescent substance use suggests that age and peer influences play a greater role for males than females. [72] These authors also report that stressful life events and low levels of parental support have a greater effect on substance use for females than for males. Likewise, Dornfeld and Kruttschnitt [73] found that recent parental divorce, marital discord, and prolonged single parenthood placed adolescent females at greater risk than males for substance use, although this type of gender interaction was also observed with delinquency as an outcome. In general, finding that parent figure transitions have differing consequences depending upon the gender of the child and the outcome evaluated is consistent with notions that males and females may have different vulnerabilities to a particular stressor and may express distress in different forms of behaviors and symptoms. [28,74,75]

Analyses based on the categorization of families according to household composition and transition pattern revealed that the relative amount of parental instability was a key factor in predicting involvement in delinquency and resulted in the identification of a subgroup of children at extreme risk for delinquency. To emphasize the importance of dynamic measures of family instability, it should be noted that parallel analyses using the initial marital status of the parent revealed no significant associations with either delinquency or drug use.

The results indicated no difference in the likelihood of delinquency among children in the two stable family groups, i.e., stable single parent vs. stable two parent families. Although this nonsignificant finding is likely due to low statistical power, it is consistent with previous research indicating that children with continuously single parents were not at elevated risk compared to intact two parent families. [49] Additionally, children in the stable single parent families actually had the lowest risk of drug use. When children had one consistent parent figure but experienced at least one change among other parent figures, they had greater odds of delinquency. However, the likelihood of delinquent behavior was dramatically increased for the group of children who experienced the greatest number of transitions and did not have at least one consistent parent figure during the study period.

Youth in this "no parent figure stability" category fit the profile of "serial" family disruptions described by Brody et al. [43] They had the greatest number of family disruptions, over three times more than the partial stability group, and their risk for delinquency was much higher than that of children in other families. These highly volatile family constellations may resemble households previously identified by social service providers as "chronically reconstituting" families and "family fragments". [76,77] Unable to maintain any particular household as a stable unit over time, chronically reconstituting families are plagued by constant role confusion and ambiguity, boundary redefinition, and conflicting loyalties. Accordingly, individuals in these families struggle with issues such as grieving, problems with commitment and trust, and poorly established identity and self-concept, [76]

An additional explanation for the high rate of delinquency among youths with "no parent stability" centers on the defining characteristic of that category--the lack of a single consistent parenting figure. Studies suggest that children exposed to adversity are less likely to develop negative behaviors when they have the support of at least one caring and consistent adult. [78-80] The attachment relationship with a primary caregiver typically evolves as an early adolescent seeks greater autonomy, but reliance on a dependable parent figure still remains a foundation of healthy development and threats to the relationship posed by a separation still cause emotional distress and grief. [81,82] Children can be distressed by separation from a parent even when the parent is abusive or neglectful. [83] In addition, the inability to count on at least one stable figure may lead these children to consider close relationships as transitory and untrustworthy.

Limitations

Although the present study addressed several of the methodological shortcomings in previous research examining the impact of family disruption on children's problem behavior, it is not without its own limitations. For example, the measures of drug use and delinquency were derived from reports of behavior during the 6 months prior to the final interview. Consequently, there was some overlap with the time period during which a parent transition could have occurred, and the temporal ordering of transitions before the outcome is not guaranteed. Another drawback was the inability to arrive at a precise count of parent figure transitions. Since parental transitions were based on the parenting situation at four widely spaced points in time and because changes from mother's old boyfriend to a new boyfriend could not be detected, these measures were probably conservative estimates of the true instability within these families. On the other hand, an assumption that transitions occur independently of interview timing (i.e., randomly when conditioned on interview time) would make the measures good indicators of the relative instability between families. Measurement of drug use and delinquency with dichotomous variables also warrants caution in the interpretation of the results. There was variation in the extent and severity of the outcome behaviors, and some early adolescents were classified for drug use and delinquency on the basis of very limited involvement in these activities. Nevertheless, given the relative youth of the sample, all of these activities are considered socially inappropriate, have legal sanctions associated with them, and could represent initiation into a pattern of more serious offending. Finally, because the study was based on a unique sample of youth with parents in methadone treatment, the results should not be generalized to families with different circumstances.

CONCLUSION

In summary, this study found that within a sample of young adolescents at risk for problem behavior due to parental substance abuse the likelihood of engaging in delinquent activity increased with the number of parent figure transitions. A similar association between parenting disruptions and the likelihood of drug use was observed for females only. The effect of parent figure transitions on problem behaviors among these youth remained significant after controlling for child characteristics, prior levels of delinquency, family conflict, parental depressive symptoms, and parental criminal history. The findings illustrate the importance of examining the dynamic nature of family composition over time, including parent figure disruptions not associated with a change in the parent's marital status.

Identification of parent figure transitions as a risk factor for an increased likelihood of adolescent delinquency and drug use in this population suggests that prevention and intervention with children in drug-affected families should address issues of family stability. To guide the development of effective preventive interventions, future research should examine in greater detail the dynamics of drug-affected families to determine the mechanisms through which parent figure transitions might influence the development of child problem behaviors. Interventions may need to incorporate strategies for dealing with grief and loss as well as the renegotiation of relationship boundaries, family roles, and parental authority.

Table 1. Percentage of Youth Engaging in Each Problem
Behavior in 6 Months Prior to Final Interview (N = 67)

Delinquent behaviors
  Picking fights                             37%
  Stealing                                   31%
  Graffiti                                   19%
  Buying/selling stolen property             16%
  Destroying property                        15%
  Throwing rocks or bottles at cars/people   13%
  Hitting someone hard enough to injure      13%
  Selling drugs                              12%
  Breaking into buildings                     6%
  Robbing/mugging                             2%
Drug use
  Alcohol                                    43%
  Marijuana                                  30%
  Psychedelics                                8%
  Narcotics                                   6%
  Glue                                        5%
  Cocaine/crack                               3%
  Amphetamines                                3%
  Tranquilizers                               3%
  Sedatives                                   0%

Table 2. Association Between Number of Parent Figure Transitions
and Delinquency

                                 Model 1      Model 2      Model 3

Parent figure transitions
  Total number of transitions    1.57 *       1.47 *       1.56 (+)
Control variables
  Experimental condition         0.28 *       0.22 *       0.212 *
  Child baseline delinquency     5.71 *       8.42 **      8.27 **
Parental difficulties
  Parent baseline depressive                  0.87         0.86
    symptoms
  Parent baseline criminal                    5.12 *       5.43 *
    history
Family relations
  Baseline family conflict                    1.06         1.09
Child characteristics
  Male                                        2.03         2.54
  Age                                         1.29         1.29
Gender interaction
  Male x Transitions                                       0.88
Overall model fit
  Chi-square (d.f.)             15.67 (3)    26.00 (8)    26.15 (9)
  Sensitivity/specificity        0.75/0.63    0.90/0.67    0.90/0.70
  Overall correct                0.70         0.81         0.82

The estimates are odds ratios (exp(B)) derived from logistics
regression models. Odds ratios greater than 1.0 (less than 1.0)
imply that delinquent behavior is more likely (less likely)
than in the reference category. (+) p < 0.10, * p < 0.05,
** p < 0.01.

Table 3. Association Between Number of Parent Figure
Transitions and Drug Use

                                 Model 1     Model 2      Model 3

Parent figure transitions
  Total number of transitions   1.29 *       1.19         2.29 *
Control variables
  Experimental condition        0.92         1.20         1.19
  Child baseline drug use       1.50         1.47         1.58
Parental difficulties
  Parent baseline depressive                 0.81         0.74 *
    symptoms
  Parent baseline criminal                   2.36         4.68 (+)
    history
Family relations
  Baseline family conflict                   1.44         2.24
Child characteristics
  Male                                       0.42         3.13
  Age                                        2.32 **      2.78 **
Gender interaction
  Male x Transitions                                      0.37 *
Overall model fit
  Chi-square (d.f.)             6.18 (3)    28.92 (8)    37.53 (9)
  Sensitivity/specificity       0.53/0.79    0.82/0.67    0.85/0.85
  Overall correct               0.66         0.75         0.85

The estimates are odds ratios (exp(B)) derived from logistic
regression models. Odds ratios greater than 1.0 (less than 1.0)
imply that drug use is more likely (less likely) than in
the reference category. (+) p < 0.10, * p < 0.05, ** p < 0.01.

Table 4. Interaction Between Gender and Number of Parent Figure
Transitions on Drug Use

                                     Male (N = 37)   Female (N = 30)

Parent figure transitions
  Total number of transitions         0.74            1.92 *
Control variables
  Experimental condition              2.40            0.68
  Child baseline drug use             1.50            2.16
Parental difficulties
  Parent baseline depressive          0.56 *          0.95
    symptoms
  Parent baseline criminal history   15.93            1.54
Family relations
  Baseline family conflict            2.26            1.59
Child characteristics
  Age                                 4.91 **         1.50
Overall model fit
  Chi-square (d.f.)                  32.1 (7)        10.57 (7)
  Sensitivity/specificity             0.88/0.90       0.71/0.70
  Overall correct                     0.89            0.70

The estimates are odds ratios (exp(B)) derived from logistic
regression models. Odds ratios greater than 1.0 (less than 1.0)
imply that drug use is more likely (less likely) than in
the reference category. (+) p < 0.10, * p < 0.05, ** p < 0.01.

Table 5. Associations Between Pattern of Parent Figure Stability
and Delinquency and Drug Use

                                         Delinquency

                                      Model 1      Model 2

Parent figure stability
  Two parent-stable
    (reference)
  Single parent-stable                3.87         2.86
  Partial stability                   7.10 *       3.04
  No stability                       35.97 **     19.56 *
Control variables
  Experimental condition              0.22 *       0.20 *
  Child baseline behavior             5.57 *       7.62 *
Parental difficulties
  Parent baseline depressive                       0.87
    symptoms
  Parent baseline criminal history                 4.54 *
Family relations
  Baseline family conflict                         1.04
Child characteristics
  Male                                             2.02
  Age                                              1.27
Overall model fit
  Chi-square                         18.49 (5)    26.75 (10)
  Sensitivity/specificity             0.85/0.52    0.90/0.67
  Overall correct                     0.72         0.81

                                          Drug Use

                                      Model 1     Model 2

Parent figure stability
  Two parent-stable
    (reference)
  Single parent-stable               0.72         0.65
  Partial stability                  1.42         1.47
  No stability                       5.40 (+)     2.93
Control variables
  Experimental condition             0.88         1.13
  Child baseline behavior            1.67         1.54
Parental difficulties
  Parent baseline depressive                      0.84
    symptoms
  Parent baseline criminal history                2.05
Family relations
  Baseline family conflict                        1.44
Child characteristics
  Male                                            0.43
  Age                                             2.36 **
Overall model fit
  Chi-square                         7.97 (5)    29.59 (10)
  Sensitivity/specificity            0.56/0.76    0.85/0.67
  Overall correct                    0.66         0.76

The estimates are odds ratios (exp(B)) derived from logistic
regression models. Odds ratios greater than 1.0 (less than 1.0)
imply that delinquent behavior or drug use is more likely
(less likely) than in the reference category.
(+) p < 0.10, * p < 0.05, ** p < 0.01.

ACKNOWLEDGMENTS

The authors gratefully acknowledge the researchers and staff of the Focus on Families project and the families that participated in the study.

REFERENCES

[1.] Johnson, J.L. Forgotten no Longer: An Overview of Research on Children of Chemically Dependent Parents. In Children of Chemically Dependent Parents: Multiperspectives from the Cutting Edge; Rivinus, T.M., Ed.; Brunner/Mazel Publishers: New York, 1991; 29-54.

[2.] Kumpfer, K.L. Special Populations: Etiology and Prevention of Vulnerability to Chemical Dependency in Children of Substance Abusers. In Youth at High Risk for Substance Abuse; Brown, B.S., Mills, A.R., Eds.; U.S. Government Printing Office: Washington, DC, 1987.

[3.] Sowder, B.J.; Burt, M.R. Children of Heroin Addicts: An Assessment of Health, Learning, Behavioral and Adjustment Problems; Praeger: New York, 1980.

[4.] Bauman, P.; Levine, S.A. The Development of Children of Drug Addicts. Int. J. Addict. 1986, 21, 849-863.

[5.] Gross, J.; McCaul, M.E. A Comparison of Drug Use and Adjustment in Urban Adolescent Children of Substance Abusers. Int. J. Addict. 1990-91, 25 (4A), 495-511.

[6.] Stanger, C.; Higgins, S.T.; Bickel, W.K.; Elk, R.; Grabowski, J.; Schmitz, J.; Amass, L.; Kirby, K.C.; Seracini, A.M. Behavioral and Emotional Problems Among Children of Cocaine and Opiate Dependent Parents. J. Am. Acad. Child Adolesc. Psychiatry 1999, 38 (4), 421-428.

[7.] de Cubas, M.M.; Field, T. Children of Methadone-Dependent Women: Developmental Outcomes. Am. J. Orthopsychiatry 1993, 63 (2), 266-276.

[8.] Nunes, E.V.; Weissman, M.M.; Goldstein, R.B.; McAvay, G.; Seracini, A.M.; Verdeli, H.; Wickramaratne, P.J. Psychopathology in Children of Parents with Opiate Dependence and/or Major Depression. J. Am. Acad. Child Adolesc. Psychiatry 1998, 37 (11), 1142-1151.

[9.] Dryfoos, J.G. Adolescents at Risk: Prevalence and Prevention; Oxford University Press: New York, 1990.

[10.] Hawkins, J.D.; Catalano, R.F.; Miller, J.Y. Risk and Protective Factors for Alcohol and other Drug Problems in Adolescence and Early Adulthood: Implications for Substance Abuse Prevention. Psychol. Bull. 1992, 112 (1), 64-105.

[11.] Tarter, R.E.; Blackson, T.C.; Martin, C.S.; Loeber, R.; Moss, H.B. Characteristics and Correlates of Child Discipline Practices in Substance Abuse and Normal Families. Am. J. Addict. 1993, 2 (1), 18-25.

[12.] Deren, S. Children of Substance Abusers: A Review of the Literature. J. Subst. Abuse Treat. 1986, 3 (2), 77-94.

[13.] Mayes, L.C. Substance Abuse and Parenting. In Handbook of Parenting: Volume 4, Applied and Practical Parenting; Bornstein, M.H., Ed.; Lawrence Erlbaum Associates: Mahwah, NJ, 1995; 101-125.

[14.] Powis, B.; Gossop, M.; Bury, C.; Payne, K.; Griffiths, P. Drug-Using Mothers: Social, Psychological and Substance Use Problems of Women Opiate Users with Children. Drug Alcohol Rev. 2000, 19, 171-180.

[15.] Kolar, A.F.; Brown, B.S.; Haertzen, C.A.; Michaelson, M.A. Children of Substance Abusers: The Life Experiences of Children of Opiate Addicts in Methadone Maintenance. Am. J. Drug Alcohol Abuse 1994, 202, 159-171.

[16.] Kumpfer, K.; DeMarsh, J.P. Family Environmental and Genetic Influences on Children's Future Chemical Dependency. In Childhood and Chemical Abuse: Prevention and Intervention; Ezekoye, S., Kumpfer, K., Bukoski, W., Eds.; Haworth: New York, 1986; 49-91.

[17.] Suffet, F.; Bryce-Buchanan, C.; Brotman, R. Pregnant Addicts in a Comprehensive Care Program: Results of a Follow-up Survey. Am. J. Orthopsychiatry 1981, 51, 297-306.

[18.] Luthar, S.S.; Cushing, G.; Merikangas, K.R.; Rounsaville, B.J. Multiple Jeopardy: Risk and Protective Factors Among Addicted Mothers' Offspring. Dev. Psychopathol. 1998, 10, 117-136.

[19.] Gainey, R.R.; Catalano, R.F.; Haggerty, K.P.; Hoppe, M.J. Deviance Among the Children of Heroin Addicts in Treatment: Impact of Parents and Peers. Deviant Behav. 1997, 18 (2), 143-159.

[20.] Nurco, D.N.; Blatchley, R.J.; Hanlon, T.E.; O'Grady, K.E. Early Deviance and Related Risk Factors in the Children of Narcotic Addicts. Am. J. Drug Alcohol Abuse 1999, 25 (1), 25-45.

[21.] Weissman, M.M.; McAvay, G.; Goldstein, R.B.; Nunes, E.V.; Verdeli, H.; Wickramaratne, P.J. Risk/Protective Factors Among Addicted Mothers' Offspring: A Replication Study. Am. J. Drug Alcohol Abuse 1999, 25 (4), 661-679.

[22.] Lawson, M.S.; Wilson, G.S. Parenting Among Women Addicted to Narcotics. Child Welfare 1980, 59 (2), 67-79.

[23.] Marcenko, M.-O.; Spence, M. Psychosocial Correlates of Child Out-of-Home Living Arrangements Among at-Risk Pregnant Women. Fam. Soc. 1995, 76 (6), 369-375.

[24.] Wilson, G.S. Clinical Studies of Infants and Children Exposed Prenatally to Heroin. Ann. NY Acad. Sci. 1989, 562, 183-194.

[25.] Amato, P.R. Children's Adjustment to Divorce: Theories, Hypotheses, and Empirical Support. J. Marriage Fam. 1993, 55, 23-38.

[26.] Amato, P.R.; Keith, B. Parental Divorce and the Well-being of Children: A Meta-Analysis. Psychol. Bull. 1991, 110 (10), 26-46.

[27.] Demo, D.H.; Acock, A.C. The Impact of Divorce on Children. J. Marriage Fam. 1988, 50, 619-648.

[28.] Hetherington, E.M.; Stanley-Hagan; Anderson, E.R. Marital Transitions: A Child's Perspective. Am. Psychol. 1989, 44 (2), 303-312.

[29.] Ganong, L.H.; Coleman, M.A. Comparison of Clinical and Empirical Literature on Children in Stepfamilies. J. Marriage Fam. 1986, 48, 309-318.

[30.] Lipsey, M.W.; Derzon, J.H. Predictors of Violent or Serious Delinquency in Adolescence and Early Adulthood. In Serious & Violent Juvenile Offenders: Risk Factors and Successful Interventions; Loeber, R., Farrington, D.P., Eds.; Sage Publications: Thousand Oaks, CA, 1998; 86-105.

[31.] Wells, E.L.; Rankin, J.H. Families and Delinquency: A Meta-Analysis of the Impact of Broken Homes. Soc. Probl. 1991, 38 (1), 71-93.

[32.] Hetherington, E.M.; Bridges, M.; Insabella, G.M. What Matters? What Does Not? Five Perspectives on the Association Between Marital Transitions and Children's Adjustment. Am. Psychol. 1998, 53 (2), 167-184.

[33.] Emery, R.E. Marriage, Divorce, and Children's Adjustment, 2nd Ed. Sage Publications: Thousand Oaks, CA, 1999.

[34.] Cherlin, A.J. Marriage, Divorce, Remarriage; Harvard University Press: Cambridge, MA, 1992.

[35.] Lahey, B.B.; Hartdagen, S.E.; Frick, P.J.; McBurnett, K.; Connor, R.; Hynd, G.W. Conduct Disorder: Parsing the Confounded Relation to Parental Divorce and Antisocial Personality. J. Abnorm. Psychol. 1988, 97 (3), 334-337.

[36.] Capaldi, D.M.; Patterson, G.R. Relation of Parental Transitions to Boys' Adjustment Problems: I. A Linear Hypothesis. II. Mothers at Risk for Transitions and Unskilled Parenting. Dev. Psychol. 1991, 27 (3), 489-504.

[37.] Forehand, R.; Biggar, H.; Kotchick, B.A. Cumulative Risk Across Family Stressors: Short- and Long-Term Effects for Adolescents. J. Abnorm. Child Psychol. 1998, 26 (2), 119-128.

[38.] Newcomb, M.D.; Maddahian, E.; Bentler, P.M. Risk Factors for Drug Use Among Adolescents: Concurrent and Longitudinal Analyses. Am. J. Public Health 1986, 76, 525-530.

[39.] Sameroff, A.J.; Seifer, R.; Bartko, W.T. Environmental Perspectives on Adaptation During Childhood and Adolescence. In Developmental Psychopathology: Perspectives on Adjustment, Risk, and Disorder; Luthar, S.S., Burack, J.A., Eds.; Cambridge University Press: New York, 1997; 507-526.

[40.] Tolan, P. Socioeconomic, Family, and Social Stress Correlates of Adolescent Antisocial and Delinquent Behavior. J. Abnorm. Child Psychol. 1988, 16 (3), 317-331.

[41.] Vaux, A.; Ruggiero, M. Stressful Life Change and Delinquent Behavior. Am. J. Community Psychol. 1983, 11 (2), 169-183.

[42.] Hoffmann, J.P.; Cerbone, F.G. Stressful Life Events and Delinquency Escalation in Early Adolescence. Criminology 1999, 37 (2), 343-373.

[43.] Brody, G.H.; Neubaum, E.; Forehand, R. Serial Marriage: A Heuristic Analysis of an Emerging Family Form. Psychol. Bull. 1988, 103 (2), 211-222.

[44.] Hetherington, E.M.; Cox, M.; Cox, R. Long-Term Effects of Divorce and Remarriage on the Adjustment of Children. J. Am. Acad. Child Psychiatry 1985, 24 (5), 518-530.

[45.] Furstenberg, F.F., Jr.; Seltzer, J.A. Divorce and Child Development. Sociological Studies of Child Development; JAI Press: New York, 1986; Vol. 1, 137-160.

[46.] Kurdek, L.A.; Fine, M.A.; Sinclair, R.J. The Relation Between Parenting Transitions and Adjustment in Young Adolescents: A Multisample Investigation. J. Early Adolesc. 1994, 14 (4), 412-432.

[47.] Hoffmann, J.P.; Johnson, R.A. A National Portrait of Family Structure and Adolescent Drug Use. J. Marriage Fam. 1998, 60, 633-645.

[48.] Flewelling, R.L.; Bauman, K.E. Family Structure as a Predictor of Initial Substance Use and Sexual Intercourse in Early Adolescence. J. Marriage Fam. 1990, 52, 171-181.

[49.] Demo, D.H.; Acock, A.C. Family Structure, Family Process, and Adolescent Well-being. J. Res. Adolesc. 1996, 6 (4), 457-488.

[50.] Mednick, B.R.; Baker, R.L.; Carothers, L.E. Patterns of Family Instability and Crime: The Association of Timing of the Family's Disruption with Subsequent Adolescent and Young Adult Criminality. J. Youth Adolesc. 1990, 19 (3), 201-220.

[51.] Bumpass, L.L.; Raley, R.K. Redifining Single Parent Families: Cohabitation and Changing Family Reality. Demography 1995, 32 (1), 97-109.

[52.] Bumpass, L.L.; Sweet, J.A. Children's Experience in Single-Parent Families: Implications of Cohabitation and Marital Transitions. Fam. Plann. Perspect. 1989, 21 (6), 256-260.

[53.] Cherlin, A.J. The Weakening Link Between Marriage and the Care of Children. Fam. Plann. Perspect. 1988, 20 (6), 302-306.

[54.] Hawkins, A.J.; Eggebeen, D.J. Are Fathers Fungible? Patterns of Coresident Adult Men in Maritally Disrupted Families and Children's Well-being. J. Marriage Fam. 1991, 53, 958-972.

[55.] Jendrek, M.-P. Grandparents Who Parent Their Grandchildren: Circumstances and Decisions. Gerontologist 1994, 34 (2), 206-216.

[56.] Fein, E.; Maluccio, A.N.; Kluger, M. No More Partings: An Examination of Long-Term Foster Family Care; Child Welfare League of America: Washington, DC, 1990.

[57.] Goerge, R.M.; Wulczyn, F.H.; Harden, A. Foster Care Dynamics, 1983-1992: A Report of the Multistate Foster Care Data Archive; Chapin Hall Center for Children: Chicago, 1994.

[58.] Brody, G.H.; Forehand, R. Prospective Associations Among Family Form, Family Processes, and Adolescents' Alcohol and Drug Use. Behav. Res. Ther. 1993, 31 (6), 587-593.

[59.] Emery, R.E.; Kitzman, K.M. The Child in the Family: Disruptions in Family Functions. In Developmental Psychopathology, Vol. 2: Risk, Disorder, and Adaptation; Cicchetti, D., Cohen, D.J., Eds.; John Wiley and Sons: New York, 1995; 3-31.

[60.] Baydar, N. Effects of Parental Separation and Reentry into Union on the Emotional Well-being of Children. J. Marriage Fam. 1988, 50, 967-981.

[61.] Block, J.H.; Block, J.; Gjerde, P.F. The Personality of Children Prior to Divorce: A Prospective Study. Child Dev. 1986, 57, 827-840.

[62.] Cherlin, A.J.; Furstenberg, F.F., Jr.; Chase-Lansdale, P.L.; Kiernan, K.E.; Robins, P.K.; Morrison, D.R.; Teitler, J.O. Longitudinal Studies of Effects of Divorce on Children in Great Britain and the United States. Science 1991, 252, 1386-1389.

[63.] Allison, P.D.; Furstenberg, F.F., Jr. How Marital Dissolution Affects Children: Variations by Age and Sex. Dev. Psychol. 1989, 25, 540-549.

[64.] Catalano, R.F.; Haggerty, K.P.; Gainey, R.R. Prevention Approaches in Methadone Treatment Settings: Children of Drug Abuse Treatment Clients. In Handbook of Drug Abuse Theory, Science and Practice; Bukoski, W.J., Sloboda, Z., Eds.; Plenum: in press.

[65.] Catalano, R.F.; Gainey, R.R.; Fleming, C.B.; Haggerty, K.P.; Johnson, N.O. An Experimental Intervention with Families of Substance Abusers: One-Year Follow-up of the Focus on Families Project. Addiction 1999, 94 (2), 241-254.

[66.] Catalano, R.F.; Haggerty, K.P.; Gainey, R.R.; Hoppe, M.J. Reducing Parental Risk Factors for Children's Substance Misuse: Preliminary Outcomes with Opiate-Addicted Parents. Subst. Use Misuse 1997, 32 (6), 699-721.

[67.] Centers for Disease Control and Prevention. C.D.C. Surveillance Summaries, March, 24, 1995. M.M.W.R. 1995:44 (No. Ss-1).

[68.] Compas, B.E. Coping with Stress During Childhood and Adolescence. Psychol. Bull. 1987, 101 (3), 393-403.

[69.] Compas, B.E.; Malcame, V.L.; Fondacaro, K.M. Coping with Stressful Events in Older Children and Adolescents. J. Consult. Clin. Psychol. 1988, 56, 405-411.

[70.] Johnston, L.; O'Malley, P.M.; Bachman, J.G. The Monitoring the Future National Results on Adolescent Drug Use: Overview of Key Findings, 1999; National Institute on Drug Abuse: Bethesda, MD, 2000.

[71.] Elliott, D.S.; Huizinga, D.; Ageton, S.S. Explaining Delinquency and Drug Use; Sage: Beverly Hills, CA, 1985.

[72.] Hoffmann, J.P.; Su, S.S. Stressful Life Events and Adolescent Substance Use and Depression: Conditional and Gender Differentiated Effects. Subst. Use Misuse 1998, 33 (11), 2219-2262.

[73.] Dornfeld, M.; Kruttschnitt, C. Do the Stereotypes Fit? Mapping Gender-Specific Outcomes and Risk Factors. Criminology 1992, 30 (3), 397-419.

[74.] Aneshensel, C.S.; Rutter, C.M.; Lachenbruch, P.A. Social Structure, Stress, and Mental Health: Competing Conceptual and Analytic Models. Am. Sociol. Rev. 1991, 56, 166-178.

[75.] Leadbeater, B.J.; Blatt, S.J.; Quinlan, D.M. Gender-Linked Vulnerabilities to Depressive Symptoms, Stress, and Problem Behaviors in Adolescents. J. Res. Adolesc. 1995, 5 (1), 1-29.

[76.] Hunter, J.E.; Schuman, N. Chronic Reconstitution as a Family Style. Soc. Work 1980, 25 (6), 446-451.

[77.] Polansky, N.; Gaudin, J.M.J.; Kilpatrick, A.C. Family Radicals. Child. Youth Serv. Rev. 1992, 14, 19-26.

[78.] Masten, A.S.; Best, K.M.; Garmezy, N. Resilience and Development: Contributions from the Study of Children Who Overcome Adversity. Dev. Psychopathol. 1990, 2, 425-444.

[79.] Rutter, M. Resilience in the Face of Adversity: Protective Factors and Resistance to Psychiatric Disorder. Br. J. Psychiatry 1985, 147, 598-611.

[80.] Werner, E.E. High-Risk Children in Young Adulthood: A Longitudinal Study from Birth to 32 Years. Am. J. Orthopsychiatry 1989, 59 (1), 72-81.

[81.] Allen, J.P.; Land, D. Attachment in Adolescence. In Handbook of Attachment: Theory, Research, and Clinical Applications; Cassidy, J., Shaver, P.R., Eds.; Guilford Press: New York, 1999; 319-335.

[82.] Ainsworth, M.D.S. Attachments Beyond Infancy. Am. Psychol. 1989, 44 (4), 709-716.

[83.] Crittenden, P.M.; Ainsworth, M.D.S. Child Maltreatment and Attachment Theory. In Child Maltreatment: Theory and Research on the Causes and Consequences of Child Abuse and Neglect; Cicchetti, D., Carlson, V., Eds.; Cambridge University Press: New York, 1989; 432-463.

Thomas E. Keller, (1), ([dagger]) Richard F. Catalano, (2) Kevin P. Haggerty, (2) and Charles B. Fleming (2)

(1) School of Social Service Administration, University of Chicago, 969 East 60th St., Chicago, IL 60637

(2) Social Development Research Group, University of Washington, 9725 3rd Avenue NE, Seattle, WA 98115

* The research reported here was supported by a grant from the National Institute on Drug Abuse (R01 DA05824-02). Preparation of this paper was supported in part by a National Institute of Mental Health National Research Service Award (MH20010).

([dagger]) Corresponding author. E-mail: tkeller@uchicago.edu

COPYRIGHT 2002 Marcel Dekker, Inc.
COPYRIGHT 2003 Gale Group