Influences of Parenting Practices on the Risk of Having a Chance to Try Cannabis
http://www.100md.com
《小儿科》
Department of Mental Health, Johns Hopkins University, Bloomberg School of Public Health, Baltimore, Maryland
Department of Epidemiology, Michigan State University, College of Human Medicine, East Lansing, Michigan
ABSTRACT
Background. Parenting practices have been linked with an array of adolescent health-compromising behaviors, but little is known about their possible long-lasting effects. In this study, we estimate the extent to which parental monitoring, parental involvement and reinforcement, and coercive parental discipline during primary school might exert a durable influence on the risk of transitioning into an early stage of youthful cannabis involvement, ie, the first chance to try cannabis.
Methods. Data were from a prospective study of first-graders who entered an urban public school system in the middle 1980s. Parenting was assessed in fourth grade, and cannabis experiences were evaluated during periodic assessments from middle childhood through young adulthood.
Results. The estimated risk of the first chance to try cannabis peaked around 16 to 18 years of age. Lower parental involvement and reinforcement and higher coercive parental discipline were associated modestly with a greater risk of cannabis exposure opportunity through the years of adolescence and into early adulthood (parental involvement and reinforcement: adjusted relative risk: 1.4; 95% confidence interval: 1.1–1.7; parental discipline: adjusted relative risk: 1.3; 95% confidence interval: 1.1–1.5); the estimated impact of parental monitoring was less durable.
Conclusions. Early-onset cannabis involvement can be hazardous. Certain parenting practices in the mid-primary school years may have a durable impact, perhaps helping to shield youths from having a chance to try cannabis throughout adolescence and into young adulthood.
Key Words: adolescent health early childhood illicit substance parental supervision parenting
Abbreviations: CI, confidence interval RR, relative risk aRR, adjusted relative risk estimate
With a longitudinal study design and an epidemiologic sample, the objective of this study was to estimate the extent to which 3 intercorrelated facets of parenting practices during the primary school years might exert a durable influence on the risk of transitioning into an early stage of youthful cannabis involvement, namely, the first chance to try this drug. Numerous studies have documented associations between parenting practices and an array of adolescent health-compromising behaviors, both cross-sectionally and prospectively, but few have had a span from primary school through early adulthood.1–4 Of particular pertinence to this study is a line of research about higher levels of parental monitoring and possibly reduced occurrence of tobacco use and illegal drug involvement.1,2,5,6 A mechanism posited to explain this observed inverse association is that highly monitored youths are less likely to associate with deviant or drug-using friends.7,8 Other dimensions of parenting practices also have been studied, with mixed findings in relation to adolescent problem behaviors.9–11 Parental discipline, for example, seemed to be less influential with respect to the onset of illegal drug use, but inept discipline may be associated with increased risk of conduct problems among children.10,11
This study is the first to investigate whether early parenting practices might help shield youths from early-onset cannabis involvement, which seems to be more hazardous than onset of use in adulthood, with a focus on a youth's first opportunity to try cannabis. Previous studies established the crucial role of drug exposure opportunity in the earliest stages of drug involvement and dependence.12–14 Notably, data from nationally representative samples in the United States showed that the majority of individuals with a cannabis exposure opportunity eventually transition to actual use; within 1 year of the first chance to try cannabis, an estimated two thirds start to use the drug.13 The chance to try cannabis also is associated with occurrence of other illegal drug involvement,15 apparently through mechanisms that include the first chance to try cocaine and other illegal drugs.16
Consistent with more general parenting models dominant in the 1980s, when this prospective study was designed, we measured cannabis exposure opportunity as well as measures of 3 interdependent dimensions of parenting practices, namely, parental monitoring, parental involvement/reinforcement, and coercive parental discipline.2,9,11,17 Of course, there is a possibility that a youth's early involvement with drugs (or behaviors associated with early cannabis use) might influence parenting.18 For this reason, our approach involved a focus on parenting in the middle years of primary school, at an age when involvement with illegal drugs among youths is rare, and we excluded from the analysis any children whose first chance to try cannabis preceded our study's assessment of parenting; we then estimated the subsequent risk of a cannabis exposure opportunity into young adulthood. We estimated the suspected influences of parental practices within the context of a more general conceptual model that included exogenous covariates (eg, early affiliation with drug-using) that were in place before cannabis involvement started.
METHODS
Study Population and Sample
The data were from a longitudinal prospective study completed within the context of a group-randomized prevention trial, with first-grade children being randomly assigned to 1 of 2 intervention classrooms (good behavior game or mastery learning) or to standard classroom conditions at the time of entry into primary school. Detailed information on the study design, sampling procedures, longitudinal assessment plan, and interventions has been presented elsewhere.1,2,19–24 In brief, during 2 successive school years in 1985–1986, a total of 2311 youths entered first grade in 19 primary schools in an urban public school system in a mid-Atlantic state of the United States and were recruited for the study at that time (1985: cohort I, n = 1196; 1986: cohort II, n = 1115). Selected characteristics of these 2311 children are presented in Table 1.
The 19 schools were located within 5 preidentified urban areas of the local school system. Families living in these areas ranged from very poor to middle class; there was a mix of race/ethnicity subgroups (mainly non-Hispanic black and white). Our study team conceptualized these 19 schools as an original social-ecologic niche for the 2311 participants and, during the childhood and adolescent years, we focused the periodic follow-up assessment protocols on the first-graders who remained within this same social-ecologic niche (ie, those who were growing up and going to school within the same relatively homogeneous urban context). Consistent with this perspective, we were able to negotiate regular access within the local school system, in a manner that allowed us to conduct standardized face-to-face interviews with the teachers of the children (at least once per year from first grade onward) and with the children themselves from middle primary school through 1994. This intensity of assessment was not possible when families moved to other parts of the country. However, we were able to secure resources that allowed us to attempt to track, monitor, obtain informed consent from, and reassess all 2311 children once they had reached their young adult years (mean of 21 years of age during the 2000–2002 calendar years).
For the present investigation of a possibly enduring influence of parenting, as practiced when the children were in their middle primary school years, we drew on information from all of the assessments just mentioned, including standardized teacher ratings of each child's behavior at the start of primary school, soon after entry into first grade, and our initial child interview assessments of parenting practices. Inclusion criteria that reduced the number of children in the analysis sample for the present investigation were (1) completion of the mid-primary school assessment of parenting practices (as part of our T0 baseline assessment for the present investigation), which took place near the end of the fourth school year after entry into primary school, at which time most of the children were 9 to 10 years of age, and (2) completion of at least 1 assessment of cannabis involvement after the T0 assessment of parenting practices. As noted above, to constrain a possible influence of early cannabis involvement on parenting practices, we excluded 120 youths who had experienced a first chance to try cannabis before the first assessment of parenting practices. After application of these inclusion and exclusion criteria, the analysis sample for the present investigation included 1222 participants, whose characteristics are presented in Table 1.
Follow-up assessments were available for the vast majority of the original 2311 youths. To complete the young adult interview assessments, between 2000 and 2002 we tracked, rerecruited, secured informed consent from, and interviewed almost 75% of the 2311 originally recruited first-graders (n = 1692). In quantitative terms, the main reasons we were unable to include all 2311 first-graders in the analysis sample for the present investigation were as follows: (1) children had moved to other school systems before the mid-primary school assessment of parenting (n = 933), (2) it was not possible to locate the child for follow-up assessment of cannabis involvement (n = 310), (3) the child was located but declined to be interviewed in young adulthood (n = 142), and (4) the child was located but was living overseas (eg, military posting) or too far away for completion of the face-to-face interview (n = 135). Through the National Death Index, we were able to confirm that 32 of the original 2311 participants had died by the time of the young adult follow-up assessments completed in calendar years 2000–2002.
Assessment and Measures
Trained interviewers conducted face-to-face standardized interviews to assess parenting and the other constructs under study, except as noted below. These interviews were conducted in a private setting (ie, within the school in 1989–1994 and at/near the place of young adult residence in 2000–2002) and on average required 40 to 60 minutes during childhood and 90 minutes in young adulthood.
The primary response construct in this investigation was the time elapsed from the T0 assessment of parenting practices to the event of each youth's first chance to try cannabis. As described above, every spring from the baseline T0 assessment to 1994, our research staff members interviewed the children remaining in the school system. As part of these interviews they asked whether and when this event (first chance to try cannabis) had occurred; a similar inquiry was included as part of the more recently completed young adult interviews. The assessment protocol to detect cannabis involvement drew a careful distinction between cannabis exposure opportunity and actual use and involved probing to disclose the child's familiarity with slang names for cannabis.25 To aid comprehension among the young children, the interviewers showed a cartoon sketch depicting youths being offered a chance to try cannabis and asked the following questions: "Have you ever been offered [cannabis]" and "How old were you the first time you were offered [cannabis] to smoke" For the young adulthood assessment, there was no cartoon sketch; lifetime experiences regarding cannabis opportunity were assessed through a standardized survey item that included a detailed definition, "By an ‘opportunity’ I mean someone either offered you alcohol or drugs, or you were present when others were using and you could have used if you wanted to. Thinking back your entire lifetime, how old were you the very first time you had an opportunity to use marijuana or hashish" A total of 690 participants reported the exact age at the first chance to try cannabis, either during the young adult assessment or during one of the spring interviews conducted between T0 and 1994. An additional 285 participants reported actual use of cannabis but information on the age at the first chance to try cannabis was missing (eg, the participant could not recall the age at the first chance to try cannabis). On the basis of national sample survey evidence, the median time from first opportunity to first use is <1 year.14 Therefore, we imputed values for the age of first cannabis exposure opportunity for these 285 participants as the age of first cannabis use minus 0.5 years.
The investigation's primary explanatory variables were measured during the T0 interview, which addressed 3 domains of parenting practices by means of a multi-item assessment adapted from prior scales.26 The parental monitoring construct was assessed with 10 items, such as, "How often, before you go out, do you tell your parents when you will be back" (Cronbach's = .6 at T0). The parental involvement/reinforcement construct was assessed with another set of 11 items, such as, "How many days in a week do you sit around and talk with your dad" ( = .7 at T0). Six items were adapted to measure coercive parental discipline, such as, "How often does your dad get angry when he punishes you" ( = .6 at T0). Standardized scores were generated to reflect the levels of these 3 interdependent facets of parenting practices. The scores then were divided into tertiles; a higher tertile reflects higher levels of parental monitoring, parental involvement/reinforcement, and parental discipline. The parental monitoring, parental involvement/reinforcement, and parental discipline items are listed in the Appendix.
An array of covariates thought to influence the occurrence of cannabis exposure opportunity also were taken into account, ie, (1) alcohol use, assessed as use of alcohol without permission before T0; (2) tobacco use, assessed as smoking >2 puffs of a cigarette before T0; (3) affiliation with cannabis-using peers at T0, assessed through a yes response to the question, "Do you have friends around your age who ever use marijuana/reefer"; (4) affiliation with peers involved in rule-breaking behaviors at T0, assessed with 5 items, such as being involved with peers who cheat on tests26; and (5) early childhood misbehavior. With respect to misbehavior, youths were classified into 3 groups according to the authority acceptance summary score of 10 standardized items on the Teacher Observation of Classroom Adaptation-Revised, which had been completed during an interview session with each first grade teacher 6 to 10 weeks after the child's entry into first grade.8,22 A central, computerized, school database provided data on birth dates, eligibility to receive subsidized lunch, gender, and race/ethnic minority status.
Statistical Analyses
Contingency table analyses and bivariate logistic regressions were conducted to estimate crude associations linking levels of parental practices with the first chance to try cannabis. Next, smoothed estimates of the crude hazard and survival functions were derived to plot the estimated instantaneous hazard rate of the first chance to try cannabis, year by year from T0 to 10 years later, stratified according to levels of parenting practices. The Wilcoxon-Breslow test was used to test departure of these data from the null model under which the occurrence of the first chance to try cannabis was assumed to be distributed equally across different levels of parenting practices. Our approach involved specifying the elapsed time from T0 until the first chance to try cannabis (on the basis of each youth's answers to our interview questions about whether and when this event had occurred). Within the analysis sample, there were 265 youths who had never had the chance to try cannabis up to the date of our last contact and interview. Consistent with standard survival analysis methods, these 265 youths with no cannabis involvement contributed person-years up to the time of that interview and then became "censored," in survival analysis terms.27
In the multiple regression context, we turned to a discrete time form of survival analysis proposed originally by Cox27 and others.28,29 The discrete time units were specified as years after T0 to the occurrence of the first chance to try cannabis or the age at the last interview, if no cannabis opportunity had occurred. Then, on the basis of these year-to-year data and a statistical model less restrictive than the semiparametric hazard model, adjusted relative risk estimates (aRRs) and 95% confidence intervals (CIs) were estimated from coefficients obtained through logistic regression, with other possible confounding effects taken into account through covariate adjustments. Here, youths were grouped within risk sets indexed in the measured units of time.29 All analyses were performed with the Stata statistical package (Stata Corp, College Station, TX).30
RESULTS
As noted above, Table 1 shows the distribution of selected sample characteristics for the original sample of 2311 first-graders and for the present investigation's analysis sample. As documented elsewhere1 and consistent with census and administrative data for the urban areas and public school system under study, a majority of the youths were members of a racial/ethnicity minority group (primarily black); more white than African heritage families moved out of the area or transferred children to other school systems with each passing year of the study. A majority of the children were eligible for free or subsidized lunch in first grade; this proportion increased over time, reflecting differential out-migration of families with socioeconomic advantages. Before the first assessment of parenting midway through primary school (T0), an estimated 5% of the youths had already smoked tobacco cigarettes at least once and 5% had tried alcohol without parental permission at least once; 5% had peers who had smoked cannabis. By young adulthood, the vast majority of youths had experienced a chance to try cannabis (n = 957 [78% of the analysis sample]) (Table 1).
Estimates of the smoothed instantaneous hazard rate for the first chance to try cannabis are shown in Figure 1, separately for each of the 3 dimensions of parenting practices. Youths with lower levels of parental monitoring at T0 had the steepest hazard rate. This indicates that children who were less well monitored by parents in mid-primary school were more likely to have earlier chances to try cannabis, compared with better-monitored children. Similar hazard curve patterns were found for parental involvement/reinforcement. Children who were at the lower levels of coercive parental discipline in mid-primary school were less likely to experience a chance to try cannabis, compared with those at the higher levels of coercive discipline. Results from the Wilcoxon-Breslow tests tended to support conclusions gained by visual inspection of the curves for parental monitoring (2 = 9.13; df = 2; P = .001) and parental discipline (2 = 11.74; df = 2; P = .003); support with respect to parental involvement/reinforcement was not as robust (2 = 5.13; df = 2; P = .08).
Overall, in the analyses linking the risk of a first chance to try cannabis with parental involvement/reinforcement or coercive parental discipline, the RR estimates remained quite stable as we made statistical adjustments for the first sets of covariates that might have had confounding effects on the associations under study. For example, as shown for model 4 in Table 2, we found that children with the lowest levels of parental involvement/reinforcement at T0 were an estimated 40% more likely to have had the chance to try cannabis than were children with higher parental involvement/reinforcement levels (aRRlow: 1.4; 95% CI: 1.1–1.7; P = .002). Independent of parental monitoring and parental involvement/reinforcement, higher levels of coercive discipline were associated modestly with a greater risk of having a chance to try cannabis (aRRhigh: 1.2; 95% CI: 1.0–1.5; P = .03). The RR estimates linking lower levels of parental monitoring and cannabis exposure opportunity were attenuated with covariate adjustment (aRRlow: 1.0; 95% CI: 0.8–1.2; P = .96).
DISCUSSION
In prior investigations of this prospectively assessed sample, our research group observed delayed and reduced occurrence of adolescent drug involvement (including tobacco smoking) among youthful participants who had had the benefit of higher levels of parental supervision and monitoring during the middle years of primary school.1,2 Extending this line of research into the young adult years, we held constant early use of tobacco and other drugs, levels of affiliation with cannabis-using and deviant peers, and other covariates. Although our measures of parenting practices were recorded 4 school years after the children entered first grade and had psychometric constraints because of the young ages of the respondents at that time (9–10 years), we were able to observe delayed and reduced occurrence of the first chance to try cannabis among the children who had had the highest levels of parental involvement/reinforcement, as well as those with lower levels of coercive parental discipline, through the years of adolescence and early adulthood; however, the evidence did not remain statistically robust for parental monitoring. The results showed that the estimated impact of coercive parental discipline was most tangible in the first 5 years after our T0 assessments and estimates were attenuated 6 to 7 years later; the estimated impact of parental reinforcement/involvement seemed to be more durable. Moreover, the magnitude of any durable risk-reducing impact of mid-primary school parenting practices seemed to be modest. Nevertheless, for both coercive parental discipline and parental reinforcement/involvement, the associated RR estimates were statistically robust over the span of a 10-year follow-up period, even with covariate adjustment. Some observers might find it surprising that there is any predictive and explanatory value of a parenting construct measured during the middle of primary school, with respect to illegal drug experiences measured across a subsequent time span of up to 10 years.
Some potential limitations of this study should be considered before more detailed discussion. One limitation was attrition. As noted in the Methods section, we had parental practice information for 60% of the originally recruited sample; in general, the other children had moved out of the area or to a different social-ecologic niche (eg, transfer to a private school). As we acknowledged elsewhere,31 some families who withdrew their children from the urban public school system or moved to another area might have migrated because they had concerns about risk exposures of the type under study here. If families with higher levels of parental monitoring and parental involvement were over-represented among those who moved away, then one result would be an attenuation of this study's estimates of the impact of higher levels of parental monitoring and involvement, which pertain to families who kept their children in the same urban public school system and continued to live in this urban area through 4 school years after school entry at first grade.
Our survival analyses and plots revealed that the risk of the first chance to try cannabis had a peak near 16 to 18 years of age in this urban sample, which were years of life after our spring assessment protocols ended in 1994. Consequently, although our assessment of cannabis involvement started quite early, in the middle of primary school, approximately two thirds of the data with respect to the age of first cannabis opportunity came from the young adult interviews, with recollection of experiences from 1994 (or earlier) to 2000–2002. Fortunately, with respect to this potential limitation of the study data, the recalled age at the first cannabis experience has been found to be quite reliable.32,33
We also must acknowledge that our assessments of parenting practices for the present investigation were based entirely on the children's reports, as elicited with the multi-item parental monitoring, parental involvement/reinforcement, and parental discipline scales at T0. In the best studies of parenting practices, there are multi-informant assessments from both parents and children.34 In later years of this study, we were able to secure parenting information from 1 caregiver (typically the mother) and, as reported elsewhere, the pattern of findings based on the caregivers' reports about parenting followed the pattern of findings based on the children's reports.1
We also note that a considerable number of children in this sample had had the chance to try cannabis before our T0 assessment of parenting practices. We excluded these children from the analysis sample because we were concerned that early cannabis involvement might have influenced parenting practices, rather than vice versa. For example, a child's precocious cannabis involvement might lead a parent toward greater monitoring of that child. However, in a post hoc exploratory analysis of parental monitoring levels among children who had experienced their first chance to try cannabis before fourth grade, we examined this possibility and found no predictive or explanatory relationship linking cannabis involvement with subsequent levels of monitoring by parents. If precocious cannabis involvement sometimes leads to greater parental monitoring, then this increase might be counter-balanced by shifts in the other direction, eg, parents might "give up on a bad apple" and neglect to monitor the child once early cannabis involvement has started. Clearly, if these possibilities are to be examined, then new longitudinal research is needed, with parenting and drug involvement measurements being recorded earlier than the fourth year of schooling.
We acknowledge that this investigation represents an initial examination of the possible long-term or carry-forward effects of parental practices in midchildhood as they might account for opportunities to try cannabis in late childhood, during adolescence, and within the first years of young adulthood. Therefore, we focused on our first assessment of parental practices at approximately age 9 to 10, without integrating possible effects of prior parenting, possible changes in parental practices over time, and possible effects of these changes on cannabis experiences. Future research incorporating multiple assessments of parental practices from the preschool years onward would allow us to investigate the constancy of and changes in multidimensional parenting practices, in relation to adolescent drug-related experiences, in greater detail.
Notwithstanding limitations such as these, a distinguishing feature of our study is that parental practices were assessed within 4 years after the children entered first grade, and the experiences with respect to the first chance to try cannabis were observed prospectively, through the first years of young adulthood. This approach establishes a solid temporal sequence for the observed associations and also reduces possible reciprocities arising as parenting practices are influenced by a child's cannabis-related experiences and behaviors. In addition, by taking a multidimensional approach to the study of parenting practices, we were able to identify potential variations in the links between parenting behaviors and early cannabis involvement that were ignored in previous studies that focused on only 1 dimension of parenting.1,2,35
Previous studies suggested that higher levels of parental monitoring were related to lower risks of alcohol, tobacco, and cannabis use and other health-compromising behaviors in adolescence.1,2,4,10,35,36 Extending the outcome to the earliest stage of cannabis involvement, our study suggests that higher levels of parental monitoring (eg, parents' knowledge of their child's whereabouts), measured midway through primary school, might have had little or no impact on delaying or reducing the risk of exposure to try cannabis after early adolescence. It is possible that higher levels of parental monitoring play a more important role in lowering youth risk to start actual dug use once exposed to or given a drug opportunity, rather than reducing the risk of a drug exposure opportunity.16 In addition, a durable impact of parental supervision might be seen in future studies of youths growing up where preadolescent cannabis use is less common than it was found to be in this urban sample of public school enrollees.
Previous findings on the associations linking parental involvement/reinforcement and coercive discipline with adolescent drug use behaviors were mixed.11 Our study suggested that higher levels of parental involvement/reinforcement might have helped shield the youths from cannabis exposure opportunity from midchildhood to early adulthood, indicating a possible nonspecific autarceologic mechanism of protection against early-onset drug involvement.37 In addition, higher levels of coercive discipline were found to be related to a modestly increased risk of cannabis involvement, independent of the levels of parental monitoring and involvement/reinforcement. In the few previous studies that probed the dimension of coercive parental discipline in relation to drug-using behaviors, the evidence suggested no association.10,11 The inconsistency of these results may be attributable, in part, to environmental conditions (ie, urban areas of this study), race/ethnicity differences, the age ranges of the study populations, or differences in measurement strategies.
CONCLUSIONS
The results of this study expand existing evidence regarding the suspected link between parenting practices in the middle years of childhood and later illegal drug experiences. Certain parenting practices in the middle years of childhood may have a durable impact, shielding youths from having a chance to try cannabis throughout adolescence and into young adulthood, the at-risk period during which most individuals start to engage in drug-related activities. Given the established role of cannabis opportunity in the transition to actual use of cannabis and other illegal drugs,13,16 our findings may be helpful as a guide to prevention efforts that target delayed or reduced risk of adolescent illegal drug involvement.7 A current focal point for mass media campaigns in the United States is parental monitoring and involvement, bolstered by the aforementioned research that indicates proximal influences of these parenting practices with respect to the onset of actual drug use. Continued attention to parental monitoring can be justified, even if this study's evidence of modest long-term influence, with respect to supervision in mid-primary school and the risk of later cannabis experience, is correct. In comparison, coercive parental discipline has been neglected in these mass media campaigns; this facet of parenting may deserve greater scrutiny and attention in drug prevention programs, if other research can confirm the evidence from this investigation.
APPENDIX: PARENTING PRACTICES
Parental Monitoring and Supervision
What time do your mom or dad or parents expect you to come home from school (clear, unclear)
What time do they expect you to come home on a weekend night (clear, unclear)
How often would your mom or dad or parents or a sitter know if you came home an hour late on weekends (always, most of the time, sometimes, hardly ever, never)
Are there kids your mom or dad or parents don't allow you to play with (always, most of the time, sometimes, hardly ever, never)
How often, before you go out, do you tell your parents when you will be back (always, most of the time, sometimes, hardly ever, never)
If your parents or a sitter are not at home, how often do you leave a note for them about where you are going (always, most of the time, sometimes, hardly ever, never)
How often do you check in with your parents or a sitter after school before going to play (always, most of the time, sometimes, hardly ever, never)
When you get home from school, how often is someone there within 1 hour (always, most of the time, sometimes, hardly ever, never)
If you are at home when your parents are not, how often do you know how to get in touch with them (always, most of the time, sometimes, hardly ever, never)
How often do you talk with your mom or dad or parents about your plans for the coming day For instance, talk about what will happen at school or with friends (always, most of the time, sometimes, hardly ever, never)
Parental Involvement/Reinforcement
Within the last year, has your mom (or step-mom) tried to teach you how to do something that she likes to do (yes, no)
Would you say it [what you did with your parent] was very much fun, mostly fun, somewhat fun, slightly fun, or not al all fun
Within the last year, has your dad (or step-dad) tried to teach you how to do something that he likes to do (yes, no)
Would you say it was very much fun, mostly fun, somewhat fun, slightly fun, or not al all fun
How many days in a week do you sit around and talk with your mom (every day, most days, 3–4 days, 1–2 days, 0 days)
How many days in a week do you sit around and talk with your dad (every day, most days, 3–4 days, 1–2 days, 0 days)
Not counting watching television or eating meals together, about how many hours each week do you spend with your mom doing something special that you enjoy (usually none, 1–2 hours per week, 2–3 hours per week, 3.5–7.0 hours per week, 7.0 or more hours per week)
Not counting watching television or eating meals together, about how many hours each week do you spend with your dad doing something special that you enjoy (usually none, 1–2 hours per week, 2–3 hours per week, 3.5–7.0 hours per week, 7.0 or more hours per week)
On a day-to-day basis, how often do your mom and dad or parents notice you are doing a good job and let you know it (always, most of the time, sometimes, hardly ever, never)
How often does your mom show you she likes it when you help around the house without being told (always, most of the time, sometimes, hardly ever, never)
How often does your dad show you he likes it when you help around the house without being told (always, most of the time, sometimes, hardly ever, never)
Coercive Parental Discipline
If your mom and dad or your parents say you will get punished if you don't stop doing something and you keep doing it, how often will they punish you (always, most of the time, sometimes, hardly ever, never)
When you break rules and your parents know about it, how often will you get away with no punishment (always, most of the time, sometimes, hardly ever, never)
How often does your mom get angry when she punishes you (always, most of the time, sometimes, hardly ever, never)
How often does your dad get angry when he punishes you (always, most of the time, sometimes, hardly ever, never)
How often do you know what kind of punishment to expect when you have done something wrong (always, most of the time, sometimes, hardly ever, never)
How often do you think that the punishment you get depends on how your mom or dad or parents feel at the time (always, most of the time, sometimes, hardly ever, never)
ACKNOWLEDGMENTS
This work was supported by grant awards from the National Institute on Drug Abuse (grants R01DA09897, R01DA04392, and K05DA015799) and the National Institute of Mental Health (grants R01MH042968 and POMH38725) and by a Development Award to Dr Chen from Johns Hopkins University Center for Adolescent Health Promotion and Disease Prevention.
Special thanks go to Dr Sheppard Kellam for initiating this line of research and to the participating youths, parents, school administration, teachers, and staff over the years of this long-span study.
FOOTNOTES
Accepted Jan 10, 2005.
James C. Anthony, PhD, Department of Epidemiology, College of Human Medicine, Michigan State University, B601, West Fee Hall, 6th Floor, East Lansing, MI 48824. E-mail address: janthony@msu.edu
No conflict of interest declared.
REFERENCES
Chilcoat HD, Anthony JC. Impact of parent monitoring on initiation of drug use through late childhood. J Am Acad Child Adolesc Psychiatry. 1996;35 :91 –100
Chilcoat HD, Dishion TJ, Anthony JC. Parental monitoring the incidence of drug sampling in urban elementary school children. Am J Epidemiol. 1995;141 :25 –31
Reboussin BA, Anthony JC. Latent class marginal regression models for longitudinal data: modeling youthful drug involvement and its suspected influences. Stat Med. 2001;20 :623 –639
Dishion TJ, McMahon RJ. Parental monitoring and the prevention of child and adolescent problem behavior: a conceptual and empirical formulation. Clin Child Fam Psychol Rev. 1998;1 :61 –75
DiClemente RJ, Wingood GM, Crosby R, et al. Parental monitoring: association with adolescents' risk behaviors. Pediatrics. 2001;107 :1363 –1368
Chassin L, Pillow DR, Curran PJ, et al. Relation of parental alcoholism to early adolescent substance use: a test of three mediating mechanisms. J Abnorm Psychol. 1993;102 :3 –19
Dishion TJ, Andrews DW, Crosby L. Antisocial boys and their friends in early adolescence: relationship characteristics, quality, and interactional process. Child Dev. 1995;66 :139 –151
Lloyd JJ, Anthony JC. Hanging out with the wrong crowds: how much difference can parents make in an urban environment J Urban Health. 2003;80 :383 –399
Kilgore K, Snyder J, Lentz C. The contribution of parental discipline, parental monitoring, and school risk to early-onset conduct problems in African American boys and girls. Dev Psychol. 2000;36 :835 –845
Guo J, Hill KG, Hawkins JD, et al. A developmental analysis of sociodemographic, family, and peer effects on adolescent illicit drug initiation. J Am Acad Child Adolesc Psychiatry. 2002;41 :838 –845
Kaplow JB, Curran PJ, Dodge KA, Conduct Problems Prevention Research Group. Child, parent, and peer predictors of early-onset substance use: a multisite longitudinal study. J Abnorm Child Psychol. 2002;20 :199 –216
Stenbacka M, Allebeck P, Romelsjo A. Initiation into drug abuse: the pathway from being offered drug to trying cannabis and progression to intravenous drug abuse. Scand J Soc Med. 1993;21 :31 –39
Van Etten ML, Neumark YD, Anthony JC. Initial opportunity to use marijuana and the transition to first use: the United States, 1979–1994. Drug Alcohol Depend. 1997;49 :1 –7
Van Etten ML, Anthony JC. Comparative epidemiology of initial drug opportunities and transitions to first use: marijuana, cocaine, hallucinogen and heroin. Drug Alcohol Depend. 1999;54 :117 –125
Lynskey MT, Heath AC, Bucholz KK, et al. Escalation of drug use in early-onset cannabis users vs co-twin controls. JAMA. 2003;289 :427 –433
Wagner FA, Anthony JC. Into the world of illegal drug use: exposure opportunity and other mechanisms linking the use of alcohol, tobacco, marijuana and cocaine. Am J Epidemiol. 2002;155 :918 –925
Darling N, Cumsille P. Theory, measurement, and methods in the study of family influences on adolescent smoking. Addiction. 2003;98 :21 –36
Neiderhiser JM, Reiss D, Hetherington EM, et al. Relationships between parenting and adolescent adjustment over time: genetic and environmental contributions. Dev Psychol. 1999;35 :680 –692
Kellam SG, Werthamer-Larsson L, Dolan LJ, et al. Developmental epidemiologically based preventive trials: baseline modeling of early target behaviors and depressive symptoms. Am J Community Psychol. 1991;19 :563 –584
Kellam SG, Anthony JC. Targeting antecedents to prevent tobacco smoking: findings from an epidemiologically based randomized field trial. Am J Public Health. 1998;99 :1490 –1495
Block J, Block JH, Keyes S. Longitudinally foretelling drug usage in adolescence: early childhood personality and environmental precursors. Child Dev. 1988;59 :336 –355
Werthamer-Larrson L, Kellam S, Wheller L. Effect of first-grade classroom environment on shy behavior, aggressive behavior, and concentration problems. Am J Community Psychol. 1991;19 :585 –602
Barrish HH, Saunders M, Wolfe MD. Good behavior game: effects of individual contingencies for group consequences and disruptive behavior in a classroom. J Appl Behav Anal. 1969;2 :119 –124
Dolan LJ, Kellam SG, Brown CH, et al. The short-term impact of two classroom-based preventive interventions on aggressive and shy behaviors and poor achievement. J Appl Dev Psychol. 1993;14 :317 –345
Wilcox HC, Benoit MZ, Storr CL, Anthony JC. Ethnicity as manifest in drug-specific vocabulary and subsequent risk to start cannabis use in early adolescents. Ethn Dis. 2005;15 :In press
Patterson GR, Capaldi DM. Psychometric Properties of Fourteen Latent Constructs From the Oregon Youth Study. New York, NY: Springer-Verlag; 1998
Cox DR. Regression models and life tables. J R Stat Soc. 1972;34 :187 –220
Singer JD, Willett JB. It's about time: using discrete-time survival analysis to study duration and the timing of events. J Educ Stat. 1993;15 :155 –195
Singer JD, Willett JB. Applied Longitudinal Data Analysis: Modeling Change and Event Occurrence. New York, NY: Oxford University Press; 2003
Stata Corp. Stata Statistical Software: Release 7.0. College Station, TX: Stata Corp; 2001
Scharfstein DO, Manski CF, Anthony JC. On the construction of bunds in prospective studies with missing ordinal outcomes: application to the good behavior game. Biometrics. 2004;60 :154 –164
Shillington AM, Clapp JD. Self-report stability of adolescent substance use: are there differences for gender, ethnicity and age Drug Alcohol Depend. 2000;60 :19 –27
Shillington AM, Cottler LB, Mager DE, et al. Self-report stability for substance use over 10 years: data from the St. Louis Epidemiologic Catchment Study. Drug Alcohol Depend. 1995;40 :103 –109
McMahon RJ, Metzler CW. Selecting parenting measures for assessing family-based prevention interventions. In: Ashery RS, Robertson EB, Kumpfer KL, eds. Drug Abuse Prevention Through Family Intervention. NIDA Research Monograph 177. Rockville, MD: National Institute on Drug Abuse; 1998:94–323
Gosebruch G, Sanchez M, Delva J, et al. Family attention and tobacco smoking among adolescents in Central America, Panama, and the Dominican Republic. Subst Use Misuse. 2003;38 :1037 –1062
Borawski EA, Ievers-Landis CE, Lovegreen LD, et al. Parental monitoring, negotiated unsupervised time, and parental trust: the role of perceived parenting practices in adolescent health risk behaviors. J Adolesc Health. 2003;33 :60 –70
Chen CY, Dormitzer CM, Bejarano J, et al. Religiosity and the earliest stages of adolescent drug involvement in seven countries of Latin America. Am J Epidemiol. 2004;159 :1180 –1189(Chuan-Yu Chen, PhD, Carla)
Department of Epidemiology, Michigan State University, College of Human Medicine, East Lansing, Michigan
ABSTRACT
Background. Parenting practices have been linked with an array of adolescent health-compromising behaviors, but little is known about their possible long-lasting effects. In this study, we estimate the extent to which parental monitoring, parental involvement and reinforcement, and coercive parental discipline during primary school might exert a durable influence on the risk of transitioning into an early stage of youthful cannabis involvement, ie, the first chance to try cannabis.
Methods. Data were from a prospective study of first-graders who entered an urban public school system in the middle 1980s. Parenting was assessed in fourth grade, and cannabis experiences were evaluated during periodic assessments from middle childhood through young adulthood.
Results. The estimated risk of the first chance to try cannabis peaked around 16 to 18 years of age. Lower parental involvement and reinforcement and higher coercive parental discipline were associated modestly with a greater risk of cannabis exposure opportunity through the years of adolescence and into early adulthood (parental involvement and reinforcement: adjusted relative risk: 1.4; 95% confidence interval: 1.1–1.7; parental discipline: adjusted relative risk: 1.3; 95% confidence interval: 1.1–1.5); the estimated impact of parental monitoring was less durable.
Conclusions. Early-onset cannabis involvement can be hazardous. Certain parenting practices in the mid-primary school years may have a durable impact, perhaps helping to shield youths from having a chance to try cannabis throughout adolescence and into young adulthood.
Key Words: adolescent health early childhood illicit substance parental supervision parenting
Abbreviations: CI, confidence interval RR, relative risk aRR, adjusted relative risk estimate
With a longitudinal study design and an epidemiologic sample, the objective of this study was to estimate the extent to which 3 intercorrelated facets of parenting practices during the primary school years might exert a durable influence on the risk of transitioning into an early stage of youthful cannabis involvement, namely, the first chance to try this drug. Numerous studies have documented associations between parenting practices and an array of adolescent health-compromising behaviors, both cross-sectionally and prospectively, but few have had a span from primary school through early adulthood.1–4 Of particular pertinence to this study is a line of research about higher levels of parental monitoring and possibly reduced occurrence of tobacco use and illegal drug involvement.1,2,5,6 A mechanism posited to explain this observed inverse association is that highly monitored youths are less likely to associate with deviant or drug-using friends.7,8 Other dimensions of parenting practices also have been studied, with mixed findings in relation to adolescent problem behaviors.9–11 Parental discipline, for example, seemed to be less influential with respect to the onset of illegal drug use, but inept discipline may be associated with increased risk of conduct problems among children.10,11
This study is the first to investigate whether early parenting practices might help shield youths from early-onset cannabis involvement, which seems to be more hazardous than onset of use in adulthood, with a focus on a youth's first opportunity to try cannabis. Previous studies established the crucial role of drug exposure opportunity in the earliest stages of drug involvement and dependence.12–14 Notably, data from nationally representative samples in the United States showed that the majority of individuals with a cannabis exposure opportunity eventually transition to actual use; within 1 year of the first chance to try cannabis, an estimated two thirds start to use the drug.13 The chance to try cannabis also is associated with occurrence of other illegal drug involvement,15 apparently through mechanisms that include the first chance to try cocaine and other illegal drugs.16
Consistent with more general parenting models dominant in the 1980s, when this prospective study was designed, we measured cannabis exposure opportunity as well as measures of 3 interdependent dimensions of parenting practices, namely, parental monitoring, parental involvement/reinforcement, and coercive parental discipline.2,9,11,17 Of course, there is a possibility that a youth's early involvement with drugs (or behaviors associated with early cannabis use) might influence parenting.18 For this reason, our approach involved a focus on parenting in the middle years of primary school, at an age when involvement with illegal drugs among youths is rare, and we excluded from the analysis any children whose first chance to try cannabis preceded our study's assessment of parenting; we then estimated the subsequent risk of a cannabis exposure opportunity into young adulthood. We estimated the suspected influences of parental practices within the context of a more general conceptual model that included exogenous covariates (eg, early affiliation with drug-using) that were in place before cannabis involvement started.
METHODS
Study Population and Sample
The data were from a longitudinal prospective study completed within the context of a group-randomized prevention trial, with first-grade children being randomly assigned to 1 of 2 intervention classrooms (good behavior game or mastery learning) or to standard classroom conditions at the time of entry into primary school. Detailed information on the study design, sampling procedures, longitudinal assessment plan, and interventions has been presented elsewhere.1,2,19–24 In brief, during 2 successive school years in 1985–1986, a total of 2311 youths entered first grade in 19 primary schools in an urban public school system in a mid-Atlantic state of the United States and were recruited for the study at that time (1985: cohort I, n = 1196; 1986: cohort II, n = 1115). Selected characteristics of these 2311 children are presented in Table 1.
The 19 schools were located within 5 preidentified urban areas of the local school system. Families living in these areas ranged from very poor to middle class; there was a mix of race/ethnicity subgroups (mainly non-Hispanic black and white). Our study team conceptualized these 19 schools as an original social-ecologic niche for the 2311 participants and, during the childhood and adolescent years, we focused the periodic follow-up assessment protocols on the first-graders who remained within this same social-ecologic niche (ie, those who were growing up and going to school within the same relatively homogeneous urban context). Consistent with this perspective, we were able to negotiate regular access within the local school system, in a manner that allowed us to conduct standardized face-to-face interviews with the teachers of the children (at least once per year from first grade onward) and with the children themselves from middle primary school through 1994. This intensity of assessment was not possible when families moved to other parts of the country. However, we were able to secure resources that allowed us to attempt to track, monitor, obtain informed consent from, and reassess all 2311 children once they had reached their young adult years (mean of 21 years of age during the 2000–2002 calendar years).
For the present investigation of a possibly enduring influence of parenting, as practiced when the children were in their middle primary school years, we drew on information from all of the assessments just mentioned, including standardized teacher ratings of each child's behavior at the start of primary school, soon after entry into first grade, and our initial child interview assessments of parenting practices. Inclusion criteria that reduced the number of children in the analysis sample for the present investigation were (1) completion of the mid-primary school assessment of parenting practices (as part of our T0 baseline assessment for the present investigation), which took place near the end of the fourth school year after entry into primary school, at which time most of the children were 9 to 10 years of age, and (2) completion of at least 1 assessment of cannabis involvement after the T0 assessment of parenting practices. As noted above, to constrain a possible influence of early cannabis involvement on parenting practices, we excluded 120 youths who had experienced a first chance to try cannabis before the first assessment of parenting practices. After application of these inclusion and exclusion criteria, the analysis sample for the present investigation included 1222 participants, whose characteristics are presented in Table 1.
Follow-up assessments were available for the vast majority of the original 2311 youths. To complete the young adult interview assessments, between 2000 and 2002 we tracked, rerecruited, secured informed consent from, and interviewed almost 75% of the 2311 originally recruited first-graders (n = 1692). In quantitative terms, the main reasons we were unable to include all 2311 first-graders in the analysis sample for the present investigation were as follows: (1) children had moved to other school systems before the mid-primary school assessment of parenting (n = 933), (2) it was not possible to locate the child for follow-up assessment of cannabis involvement (n = 310), (3) the child was located but declined to be interviewed in young adulthood (n = 142), and (4) the child was located but was living overseas (eg, military posting) or too far away for completion of the face-to-face interview (n = 135). Through the National Death Index, we were able to confirm that 32 of the original 2311 participants had died by the time of the young adult follow-up assessments completed in calendar years 2000–2002.
Assessment and Measures
Trained interviewers conducted face-to-face standardized interviews to assess parenting and the other constructs under study, except as noted below. These interviews were conducted in a private setting (ie, within the school in 1989–1994 and at/near the place of young adult residence in 2000–2002) and on average required 40 to 60 minutes during childhood and 90 minutes in young adulthood.
The primary response construct in this investigation was the time elapsed from the T0 assessment of parenting practices to the event of each youth's first chance to try cannabis. As described above, every spring from the baseline T0 assessment to 1994, our research staff members interviewed the children remaining in the school system. As part of these interviews they asked whether and when this event (first chance to try cannabis) had occurred; a similar inquiry was included as part of the more recently completed young adult interviews. The assessment protocol to detect cannabis involvement drew a careful distinction between cannabis exposure opportunity and actual use and involved probing to disclose the child's familiarity with slang names for cannabis.25 To aid comprehension among the young children, the interviewers showed a cartoon sketch depicting youths being offered a chance to try cannabis and asked the following questions: "Have you ever been offered [cannabis]" and "How old were you the first time you were offered [cannabis] to smoke" For the young adulthood assessment, there was no cartoon sketch; lifetime experiences regarding cannabis opportunity were assessed through a standardized survey item that included a detailed definition, "By an ‘opportunity’ I mean someone either offered you alcohol or drugs, or you were present when others were using and you could have used if you wanted to. Thinking back your entire lifetime, how old were you the very first time you had an opportunity to use marijuana or hashish" A total of 690 participants reported the exact age at the first chance to try cannabis, either during the young adult assessment or during one of the spring interviews conducted between T0 and 1994. An additional 285 participants reported actual use of cannabis but information on the age at the first chance to try cannabis was missing (eg, the participant could not recall the age at the first chance to try cannabis). On the basis of national sample survey evidence, the median time from first opportunity to first use is <1 year.14 Therefore, we imputed values for the age of first cannabis exposure opportunity for these 285 participants as the age of first cannabis use minus 0.5 years.
The investigation's primary explanatory variables were measured during the T0 interview, which addressed 3 domains of parenting practices by means of a multi-item assessment adapted from prior scales.26 The parental monitoring construct was assessed with 10 items, such as, "How often, before you go out, do you tell your parents when you will be back" (Cronbach's = .6 at T0). The parental involvement/reinforcement construct was assessed with another set of 11 items, such as, "How many days in a week do you sit around and talk with your dad" ( = .7 at T0). Six items were adapted to measure coercive parental discipline, such as, "How often does your dad get angry when he punishes you" ( = .6 at T0). Standardized scores were generated to reflect the levels of these 3 interdependent facets of parenting practices. The scores then were divided into tertiles; a higher tertile reflects higher levels of parental monitoring, parental involvement/reinforcement, and parental discipline. The parental monitoring, parental involvement/reinforcement, and parental discipline items are listed in the Appendix.
An array of covariates thought to influence the occurrence of cannabis exposure opportunity also were taken into account, ie, (1) alcohol use, assessed as use of alcohol without permission before T0; (2) tobacco use, assessed as smoking >2 puffs of a cigarette before T0; (3) affiliation with cannabis-using peers at T0, assessed through a yes response to the question, "Do you have friends around your age who ever use marijuana/reefer"; (4) affiliation with peers involved in rule-breaking behaviors at T0, assessed with 5 items, such as being involved with peers who cheat on tests26; and (5) early childhood misbehavior. With respect to misbehavior, youths were classified into 3 groups according to the authority acceptance summary score of 10 standardized items on the Teacher Observation of Classroom Adaptation-Revised, which had been completed during an interview session with each first grade teacher 6 to 10 weeks after the child's entry into first grade.8,22 A central, computerized, school database provided data on birth dates, eligibility to receive subsidized lunch, gender, and race/ethnic minority status.
Statistical Analyses
Contingency table analyses and bivariate logistic regressions were conducted to estimate crude associations linking levels of parental practices with the first chance to try cannabis. Next, smoothed estimates of the crude hazard and survival functions were derived to plot the estimated instantaneous hazard rate of the first chance to try cannabis, year by year from T0 to 10 years later, stratified according to levels of parenting practices. The Wilcoxon-Breslow test was used to test departure of these data from the null model under which the occurrence of the first chance to try cannabis was assumed to be distributed equally across different levels of parenting practices. Our approach involved specifying the elapsed time from T0 until the first chance to try cannabis (on the basis of each youth's answers to our interview questions about whether and when this event had occurred). Within the analysis sample, there were 265 youths who had never had the chance to try cannabis up to the date of our last contact and interview. Consistent with standard survival analysis methods, these 265 youths with no cannabis involvement contributed person-years up to the time of that interview and then became "censored," in survival analysis terms.27
In the multiple regression context, we turned to a discrete time form of survival analysis proposed originally by Cox27 and others.28,29 The discrete time units were specified as years after T0 to the occurrence of the first chance to try cannabis or the age at the last interview, if no cannabis opportunity had occurred. Then, on the basis of these year-to-year data and a statistical model less restrictive than the semiparametric hazard model, adjusted relative risk estimates (aRRs) and 95% confidence intervals (CIs) were estimated from coefficients obtained through logistic regression, with other possible confounding effects taken into account through covariate adjustments. Here, youths were grouped within risk sets indexed in the measured units of time.29 All analyses were performed with the Stata statistical package (Stata Corp, College Station, TX).30
RESULTS
As noted above, Table 1 shows the distribution of selected sample characteristics for the original sample of 2311 first-graders and for the present investigation's analysis sample. As documented elsewhere1 and consistent with census and administrative data for the urban areas and public school system under study, a majority of the youths were members of a racial/ethnicity minority group (primarily black); more white than African heritage families moved out of the area or transferred children to other school systems with each passing year of the study. A majority of the children were eligible for free or subsidized lunch in first grade; this proportion increased over time, reflecting differential out-migration of families with socioeconomic advantages. Before the first assessment of parenting midway through primary school (T0), an estimated 5% of the youths had already smoked tobacco cigarettes at least once and 5% had tried alcohol without parental permission at least once; 5% had peers who had smoked cannabis. By young adulthood, the vast majority of youths had experienced a chance to try cannabis (n = 957 [78% of the analysis sample]) (Table 1).
Estimates of the smoothed instantaneous hazard rate for the first chance to try cannabis are shown in Figure 1, separately for each of the 3 dimensions of parenting practices. Youths with lower levels of parental monitoring at T0 had the steepest hazard rate. This indicates that children who were less well monitored by parents in mid-primary school were more likely to have earlier chances to try cannabis, compared with better-monitored children. Similar hazard curve patterns were found for parental involvement/reinforcement. Children who were at the lower levels of coercive parental discipline in mid-primary school were less likely to experience a chance to try cannabis, compared with those at the higher levels of coercive discipline. Results from the Wilcoxon-Breslow tests tended to support conclusions gained by visual inspection of the curves for parental monitoring (2 = 9.13; df = 2; P = .001) and parental discipline (2 = 11.74; df = 2; P = .003); support with respect to parental involvement/reinforcement was not as robust (2 = 5.13; df = 2; P = .08).
Overall, in the analyses linking the risk of a first chance to try cannabis with parental involvement/reinforcement or coercive parental discipline, the RR estimates remained quite stable as we made statistical adjustments for the first sets of covariates that might have had confounding effects on the associations under study. For example, as shown for model 4 in Table 2, we found that children with the lowest levels of parental involvement/reinforcement at T0 were an estimated 40% more likely to have had the chance to try cannabis than were children with higher parental involvement/reinforcement levels (aRRlow: 1.4; 95% CI: 1.1–1.7; P = .002). Independent of parental monitoring and parental involvement/reinforcement, higher levels of coercive discipline were associated modestly with a greater risk of having a chance to try cannabis (aRRhigh: 1.2; 95% CI: 1.0–1.5; P = .03). The RR estimates linking lower levels of parental monitoring and cannabis exposure opportunity were attenuated with covariate adjustment (aRRlow: 1.0; 95% CI: 0.8–1.2; P = .96).
DISCUSSION
In prior investigations of this prospectively assessed sample, our research group observed delayed and reduced occurrence of adolescent drug involvement (including tobacco smoking) among youthful participants who had had the benefit of higher levels of parental supervision and monitoring during the middle years of primary school.1,2 Extending this line of research into the young adult years, we held constant early use of tobacco and other drugs, levels of affiliation with cannabis-using and deviant peers, and other covariates. Although our measures of parenting practices were recorded 4 school years after the children entered first grade and had psychometric constraints because of the young ages of the respondents at that time (9–10 years), we were able to observe delayed and reduced occurrence of the first chance to try cannabis among the children who had had the highest levels of parental involvement/reinforcement, as well as those with lower levels of coercive parental discipline, through the years of adolescence and early adulthood; however, the evidence did not remain statistically robust for parental monitoring. The results showed that the estimated impact of coercive parental discipline was most tangible in the first 5 years after our T0 assessments and estimates were attenuated 6 to 7 years later; the estimated impact of parental reinforcement/involvement seemed to be more durable. Moreover, the magnitude of any durable risk-reducing impact of mid-primary school parenting practices seemed to be modest. Nevertheless, for both coercive parental discipline and parental reinforcement/involvement, the associated RR estimates were statistically robust over the span of a 10-year follow-up period, even with covariate adjustment. Some observers might find it surprising that there is any predictive and explanatory value of a parenting construct measured during the middle of primary school, with respect to illegal drug experiences measured across a subsequent time span of up to 10 years.
Some potential limitations of this study should be considered before more detailed discussion. One limitation was attrition. As noted in the Methods section, we had parental practice information for 60% of the originally recruited sample; in general, the other children had moved out of the area or to a different social-ecologic niche (eg, transfer to a private school). As we acknowledged elsewhere,31 some families who withdrew their children from the urban public school system or moved to another area might have migrated because they had concerns about risk exposures of the type under study here. If families with higher levels of parental monitoring and parental involvement were over-represented among those who moved away, then one result would be an attenuation of this study's estimates of the impact of higher levels of parental monitoring and involvement, which pertain to families who kept their children in the same urban public school system and continued to live in this urban area through 4 school years after school entry at first grade.
Our survival analyses and plots revealed that the risk of the first chance to try cannabis had a peak near 16 to 18 years of age in this urban sample, which were years of life after our spring assessment protocols ended in 1994. Consequently, although our assessment of cannabis involvement started quite early, in the middle of primary school, approximately two thirds of the data with respect to the age of first cannabis opportunity came from the young adult interviews, with recollection of experiences from 1994 (or earlier) to 2000–2002. Fortunately, with respect to this potential limitation of the study data, the recalled age at the first cannabis experience has been found to be quite reliable.32,33
We also must acknowledge that our assessments of parenting practices for the present investigation were based entirely on the children's reports, as elicited with the multi-item parental monitoring, parental involvement/reinforcement, and parental discipline scales at T0. In the best studies of parenting practices, there are multi-informant assessments from both parents and children.34 In later years of this study, we were able to secure parenting information from 1 caregiver (typically the mother) and, as reported elsewhere, the pattern of findings based on the caregivers' reports about parenting followed the pattern of findings based on the children's reports.1
We also note that a considerable number of children in this sample had had the chance to try cannabis before our T0 assessment of parenting practices. We excluded these children from the analysis sample because we were concerned that early cannabis involvement might have influenced parenting practices, rather than vice versa. For example, a child's precocious cannabis involvement might lead a parent toward greater monitoring of that child. However, in a post hoc exploratory analysis of parental monitoring levels among children who had experienced their first chance to try cannabis before fourth grade, we examined this possibility and found no predictive or explanatory relationship linking cannabis involvement with subsequent levels of monitoring by parents. If precocious cannabis involvement sometimes leads to greater parental monitoring, then this increase might be counter-balanced by shifts in the other direction, eg, parents might "give up on a bad apple" and neglect to monitor the child once early cannabis involvement has started. Clearly, if these possibilities are to be examined, then new longitudinal research is needed, with parenting and drug involvement measurements being recorded earlier than the fourth year of schooling.
We acknowledge that this investigation represents an initial examination of the possible long-term or carry-forward effects of parental practices in midchildhood as they might account for opportunities to try cannabis in late childhood, during adolescence, and within the first years of young adulthood. Therefore, we focused on our first assessment of parental practices at approximately age 9 to 10, without integrating possible effects of prior parenting, possible changes in parental practices over time, and possible effects of these changes on cannabis experiences. Future research incorporating multiple assessments of parental practices from the preschool years onward would allow us to investigate the constancy of and changes in multidimensional parenting practices, in relation to adolescent drug-related experiences, in greater detail.
Notwithstanding limitations such as these, a distinguishing feature of our study is that parental practices were assessed within 4 years after the children entered first grade, and the experiences with respect to the first chance to try cannabis were observed prospectively, through the first years of young adulthood. This approach establishes a solid temporal sequence for the observed associations and also reduces possible reciprocities arising as parenting practices are influenced by a child's cannabis-related experiences and behaviors. In addition, by taking a multidimensional approach to the study of parenting practices, we were able to identify potential variations in the links between parenting behaviors and early cannabis involvement that were ignored in previous studies that focused on only 1 dimension of parenting.1,2,35
Previous studies suggested that higher levels of parental monitoring were related to lower risks of alcohol, tobacco, and cannabis use and other health-compromising behaviors in adolescence.1,2,4,10,35,36 Extending the outcome to the earliest stage of cannabis involvement, our study suggests that higher levels of parental monitoring (eg, parents' knowledge of their child's whereabouts), measured midway through primary school, might have had little or no impact on delaying or reducing the risk of exposure to try cannabis after early adolescence. It is possible that higher levels of parental monitoring play a more important role in lowering youth risk to start actual dug use once exposed to or given a drug opportunity, rather than reducing the risk of a drug exposure opportunity.16 In addition, a durable impact of parental supervision might be seen in future studies of youths growing up where preadolescent cannabis use is less common than it was found to be in this urban sample of public school enrollees.
Previous findings on the associations linking parental involvement/reinforcement and coercive discipline with adolescent drug use behaviors were mixed.11 Our study suggested that higher levels of parental involvement/reinforcement might have helped shield the youths from cannabis exposure opportunity from midchildhood to early adulthood, indicating a possible nonspecific autarceologic mechanism of protection against early-onset drug involvement.37 In addition, higher levels of coercive discipline were found to be related to a modestly increased risk of cannabis involvement, independent of the levels of parental monitoring and involvement/reinforcement. In the few previous studies that probed the dimension of coercive parental discipline in relation to drug-using behaviors, the evidence suggested no association.10,11 The inconsistency of these results may be attributable, in part, to environmental conditions (ie, urban areas of this study), race/ethnicity differences, the age ranges of the study populations, or differences in measurement strategies.
CONCLUSIONS
The results of this study expand existing evidence regarding the suspected link between parenting practices in the middle years of childhood and later illegal drug experiences. Certain parenting practices in the middle years of childhood may have a durable impact, shielding youths from having a chance to try cannabis throughout adolescence and into young adulthood, the at-risk period during which most individuals start to engage in drug-related activities. Given the established role of cannabis opportunity in the transition to actual use of cannabis and other illegal drugs,13,16 our findings may be helpful as a guide to prevention efforts that target delayed or reduced risk of adolescent illegal drug involvement.7 A current focal point for mass media campaigns in the United States is parental monitoring and involvement, bolstered by the aforementioned research that indicates proximal influences of these parenting practices with respect to the onset of actual drug use. Continued attention to parental monitoring can be justified, even if this study's evidence of modest long-term influence, with respect to supervision in mid-primary school and the risk of later cannabis experience, is correct. In comparison, coercive parental discipline has been neglected in these mass media campaigns; this facet of parenting may deserve greater scrutiny and attention in drug prevention programs, if other research can confirm the evidence from this investigation.
APPENDIX: PARENTING PRACTICES
Parental Monitoring and Supervision
What time do your mom or dad or parents expect you to come home from school (clear, unclear)
What time do they expect you to come home on a weekend night (clear, unclear)
How often would your mom or dad or parents or a sitter know if you came home an hour late on weekends (always, most of the time, sometimes, hardly ever, never)
Are there kids your mom or dad or parents don't allow you to play with (always, most of the time, sometimes, hardly ever, never)
How often, before you go out, do you tell your parents when you will be back (always, most of the time, sometimes, hardly ever, never)
If your parents or a sitter are not at home, how often do you leave a note for them about where you are going (always, most of the time, sometimes, hardly ever, never)
How often do you check in with your parents or a sitter after school before going to play (always, most of the time, sometimes, hardly ever, never)
When you get home from school, how often is someone there within 1 hour (always, most of the time, sometimes, hardly ever, never)
If you are at home when your parents are not, how often do you know how to get in touch with them (always, most of the time, sometimes, hardly ever, never)
How often do you talk with your mom or dad or parents about your plans for the coming day For instance, talk about what will happen at school or with friends (always, most of the time, sometimes, hardly ever, never)
Parental Involvement/Reinforcement
Within the last year, has your mom (or step-mom) tried to teach you how to do something that she likes to do (yes, no)
Would you say it [what you did with your parent] was very much fun, mostly fun, somewhat fun, slightly fun, or not al all fun
Within the last year, has your dad (or step-dad) tried to teach you how to do something that he likes to do (yes, no)
Would you say it was very much fun, mostly fun, somewhat fun, slightly fun, or not al all fun
How many days in a week do you sit around and talk with your mom (every day, most days, 3–4 days, 1–2 days, 0 days)
How many days in a week do you sit around and talk with your dad (every day, most days, 3–4 days, 1–2 days, 0 days)
Not counting watching television or eating meals together, about how many hours each week do you spend with your mom doing something special that you enjoy (usually none, 1–2 hours per week, 2–3 hours per week, 3.5–7.0 hours per week, 7.0 or more hours per week)
Not counting watching television or eating meals together, about how many hours each week do you spend with your dad doing something special that you enjoy (usually none, 1–2 hours per week, 2–3 hours per week, 3.5–7.0 hours per week, 7.0 or more hours per week)
On a day-to-day basis, how often do your mom and dad or parents notice you are doing a good job and let you know it (always, most of the time, sometimes, hardly ever, never)
How often does your mom show you she likes it when you help around the house without being told (always, most of the time, sometimes, hardly ever, never)
How often does your dad show you he likes it when you help around the house without being told (always, most of the time, sometimes, hardly ever, never)
Coercive Parental Discipline
If your mom and dad or your parents say you will get punished if you don't stop doing something and you keep doing it, how often will they punish you (always, most of the time, sometimes, hardly ever, never)
When you break rules and your parents know about it, how often will you get away with no punishment (always, most of the time, sometimes, hardly ever, never)
How often does your mom get angry when she punishes you (always, most of the time, sometimes, hardly ever, never)
How often does your dad get angry when he punishes you (always, most of the time, sometimes, hardly ever, never)
How often do you know what kind of punishment to expect when you have done something wrong (always, most of the time, sometimes, hardly ever, never)
How often do you think that the punishment you get depends on how your mom or dad or parents feel at the time (always, most of the time, sometimes, hardly ever, never)
ACKNOWLEDGMENTS
This work was supported by grant awards from the National Institute on Drug Abuse (grants R01DA09897, R01DA04392, and K05DA015799) and the National Institute of Mental Health (grants R01MH042968 and POMH38725) and by a Development Award to Dr Chen from Johns Hopkins University Center for Adolescent Health Promotion and Disease Prevention.
Special thanks go to Dr Sheppard Kellam for initiating this line of research and to the participating youths, parents, school administration, teachers, and staff over the years of this long-span study.
FOOTNOTES
Accepted Jan 10, 2005.
James C. Anthony, PhD, Department of Epidemiology, College of Human Medicine, Michigan State University, B601, West Fee Hall, 6th Floor, East Lansing, MI 48824. E-mail address: janthony@msu.edu
No conflict of interest declared.
REFERENCES
Chilcoat HD, Anthony JC. Impact of parent monitoring on initiation of drug use through late childhood. J Am Acad Child Adolesc Psychiatry. 1996;35 :91 –100
Chilcoat HD, Dishion TJ, Anthony JC. Parental monitoring the incidence of drug sampling in urban elementary school children. Am J Epidemiol. 1995;141 :25 –31
Reboussin BA, Anthony JC. Latent class marginal regression models for longitudinal data: modeling youthful drug involvement and its suspected influences. Stat Med. 2001;20 :623 –639
Dishion TJ, McMahon RJ. Parental monitoring and the prevention of child and adolescent problem behavior: a conceptual and empirical formulation. Clin Child Fam Psychol Rev. 1998;1 :61 –75
DiClemente RJ, Wingood GM, Crosby R, et al. Parental monitoring: association with adolescents' risk behaviors. Pediatrics. 2001;107 :1363 –1368
Chassin L, Pillow DR, Curran PJ, et al. Relation of parental alcoholism to early adolescent substance use: a test of three mediating mechanisms. J Abnorm Psychol. 1993;102 :3 –19
Dishion TJ, Andrews DW, Crosby L. Antisocial boys and their friends in early adolescence: relationship characteristics, quality, and interactional process. Child Dev. 1995;66 :139 –151
Lloyd JJ, Anthony JC. Hanging out with the wrong crowds: how much difference can parents make in an urban environment J Urban Health. 2003;80 :383 –399
Kilgore K, Snyder J, Lentz C. The contribution of parental discipline, parental monitoring, and school risk to early-onset conduct problems in African American boys and girls. Dev Psychol. 2000;36 :835 –845
Guo J, Hill KG, Hawkins JD, et al. A developmental analysis of sociodemographic, family, and peer effects on adolescent illicit drug initiation. J Am Acad Child Adolesc Psychiatry. 2002;41 :838 –845
Kaplow JB, Curran PJ, Dodge KA, Conduct Problems Prevention Research Group. Child, parent, and peer predictors of early-onset substance use: a multisite longitudinal study. J Abnorm Child Psychol. 2002;20 :199 –216
Stenbacka M, Allebeck P, Romelsjo A. Initiation into drug abuse: the pathway from being offered drug to trying cannabis and progression to intravenous drug abuse. Scand J Soc Med. 1993;21 :31 –39
Van Etten ML, Neumark YD, Anthony JC. Initial opportunity to use marijuana and the transition to first use: the United States, 1979–1994. Drug Alcohol Depend. 1997;49 :1 –7
Van Etten ML, Anthony JC. Comparative epidemiology of initial drug opportunities and transitions to first use: marijuana, cocaine, hallucinogen and heroin. Drug Alcohol Depend. 1999;54 :117 –125
Lynskey MT, Heath AC, Bucholz KK, et al. Escalation of drug use in early-onset cannabis users vs co-twin controls. JAMA. 2003;289 :427 –433
Wagner FA, Anthony JC. Into the world of illegal drug use: exposure opportunity and other mechanisms linking the use of alcohol, tobacco, marijuana and cocaine. Am J Epidemiol. 2002;155 :918 –925
Darling N, Cumsille P. Theory, measurement, and methods in the study of family influences on adolescent smoking. Addiction. 2003;98 :21 –36
Neiderhiser JM, Reiss D, Hetherington EM, et al. Relationships between parenting and adolescent adjustment over time: genetic and environmental contributions. Dev Psychol. 1999;35 :680 –692
Kellam SG, Werthamer-Larsson L, Dolan LJ, et al. Developmental epidemiologically based preventive trials: baseline modeling of early target behaviors and depressive symptoms. Am J Community Psychol. 1991;19 :563 –584
Kellam SG, Anthony JC. Targeting antecedents to prevent tobacco smoking: findings from an epidemiologically based randomized field trial. Am J Public Health. 1998;99 :1490 –1495
Block J, Block JH, Keyes S. Longitudinally foretelling drug usage in adolescence: early childhood personality and environmental precursors. Child Dev. 1988;59 :336 –355
Werthamer-Larrson L, Kellam S, Wheller L. Effect of first-grade classroom environment on shy behavior, aggressive behavior, and concentration problems. Am J Community Psychol. 1991;19 :585 –602
Barrish HH, Saunders M, Wolfe MD. Good behavior game: effects of individual contingencies for group consequences and disruptive behavior in a classroom. J Appl Behav Anal. 1969;2 :119 –124
Dolan LJ, Kellam SG, Brown CH, et al. The short-term impact of two classroom-based preventive interventions on aggressive and shy behaviors and poor achievement. J Appl Dev Psychol. 1993;14 :317 –345
Wilcox HC, Benoit MZ, Storr CL, Anthony JC. Ethnicity as manifest in drug-specific vocabulary and subsequent risk to start cannabis use in early adolescents. Ethn Dis. 2005;15 :In press
Patterson GR, Capaldi DM. Psychometric Properties of Fourteen Latent Constructs From the Oregon Youth Study. New York, NY: Springer-Verlag; 1998
Cox DR. Regression models and life tables. J R Stat Soc. 1972;34 :187 –220
Singer JD, Willett JB. It's about time: using discrete-time survival analysis to study duration and the timing of events. J Educ Stat. 1993;15 :155 –195
Singer JD, Willett JB. Applied Longitudinal Data Analysis: Modeling Change and Event Occurrence. New York, NY: Oxford University Press; 2003
Stata Corp. Stata Statistical Software: Release 7.0. College Station, TX: Stata Corp; 2001
Scharfstein DO, Manski CF, Anthony JC. On the construction of bunds in prospective studies with missing ordinal outcomes: application to the good behavior game. Biometrics. 2004;60 :154 –164
Shillington AM, Clapp JD. Self-report stability of adolescent substance use: are there differences for gender, ethnicity and age Drug Alcohol Depend. 2000;60 :19 –27
Shillington AM, Cottler LB, Mager DE, et al. Self-report stability for substance use over 10 years: data from the St. Louis Epidemiologic Catchment Study. Drug Alcohol Depend. 1995;40 :103 –109
McMahon RJ, Metzler CW. Selecting parenting measures for assessing family-based prevention interventions. In: Ashery RS, Robertson EB, Kumpfer KL, eds. Drug Abuse Prevention Through Family Intervention. NIDA Research Monograph 177. Rockville, MD: National Institute on Drug Abuse; 1998:94–323
Gosebruch G, Sanchez M, Delva J, et al. Family attention and tobacco smoking among adolescents in Central America, Panama, and the Dominican Republic. Subst Use Misuse. 2003;38 :1037 –1062
Borawski EA, Ievers-Landis CE, Lovegreen LD, et al. Parental monitoring, negotiated unsupervised time, and parental trust: the role of perceived parenting practices in adolescent health risk behaviors. J Adolesc Health. 2003;33 :60 –70
Chen CY, Dormitzer CM, Bejarano J, et al. Religiosity and the earliest stages of adolescent drug involvement in seven countries of Latin America. Am J Epidemiol. 2004;159 :1180 –1189(Chuan-Yu Chen, PhD, Carla)