|
This document was developed under the direction of Soledad Sambrano, Ph.D., of the Center for Substance Abuse Prevention (CSAP) through contract #277-95-5002 with EMT Associates, Inc., and ORC Macro (formerly Macro International, Inc.).
The principal authors of this document are Elizabeth Sale, Ph.D., of EMT Associates, Inc.; Soledad Sambrano, Ph.D., of the Center for Substance Abuse Prevention; J. Fred Springer, Ph.D.; Jack Hermann, Ph.D., of ORC Macro; and Rafa Kasim, Ph.D., of EMT Associates, Inc. Special thanks to Paul J. Brounstein, Ph.D., and Fred Seitz, Ph.D., of CSAP; Charles Turner, Ph.D.; Wei Pan, Ph.D.; David Cordray, Ph.D.; Will Shadish, Ph.D.; Chrystalla Ellina, Ph.D.; and Mary Nistler, M.P.P.A., for their important contributions.
Special thanks to the CSAP staff who managed the projects and the 48 grantees who participated in the study. Also appreciation to Juana Mora, Ph.D; Craig Love, Ph.D.; and Jane Maxwell, Ph.D. for their review of these documents.
All material appearing in this report is in the public domain and may be reproduced or copied without permission from the Substance Abuse and Mental Health Services Administration. However, this publication may not be reproduced or distributed for a fee without specific, written authorization of the Office of Communications, SAMHSA, U.S. Department of Health and Human Services. Citation of the source is appreciated. Suggested citation: Substance Abuse and Mental Health Services Administration. The National Cross-Site Evaluation of High-Risk Youth Programs. Center for Substance Abuse Prevention, DHHS Publication No. SMA-003375. Rockville, MD, 2002.
Copies may be obtained, free of charge, from the National Clearinghouse for Alcohol and Drug Information (NCADI). NCADI is a service of the Substance Abuse and Mental Health Services Administration (SAMHSA). For copies of publications, please write or call:
National Clearinghouse for Alcohol and Drug Information (NCADI)This publication can be accessed electronically through the Internet World Wide Web connection listed below.
World Wide Web Internet connection: http://www.samhsa.gov
Originating OfficeThe Center for Substance Abuse Prevention (CSAP) in the Substance Abuse and Mental Health Services Administration (SAMHSA) is the Nation’s lead agency for substance abuse prevention. The Center funds community-based organizations, universities, behavioral health providers, and public agencies to identify effective prevention programs and practices and disseminates findings, program models, and other prevention materials to practitioners and policymakers across the country. This document summarizes the findings of CSAP’s National Cross-Site Evaluation of High-Risk Youth Programs. This large multiple-site evaluation was designed to assess 48 prevention programs and to identify those program characteristics that are associated with strong substance abuse prevention outcomes.
The National High-Risk Youth Demonstration (funded from 1987 to 1995) has been one of the most ambitious and productive of CSAP’s funding initiatives. In its early years, the demonstration focused on identifying promising approaches to prevention. As the demonstration matured, individual site and cross-site research produced knowledge about risk and protective factors related to substance use and helped to identify model programs for effective prevention. Research and experience in the High-Risk Youth Demonstration also contributed to the awareness of the importance of culturally sensitive, age-appropriate and gender-specific programming. This progress in prevention theory and practice laid the foundation for the research reported in this document, the largest and most comprehensive of CSAP’s High-Risk Youth studies.
In addition to individual-level information on substance use, risk, and protection, the evaluation includes detailed information on the nature and amount of prevention services in which each child participated, as well as systematic process information on the study programs. This monograph highlights the study’s contributions to the growing evidence that “Prevention Works.” The monograph also highlights the study’s contributions to the understanding of how substance use develops in youth at high risk, and the risk and protective factors that contribute to or protect against substance use. More specifically, this document summarizes major findings concerning changes in substance use, risk, and protection as youth mature through adolescence, summarizes findings concerning the pathways between external and internal risk and protective factors and substance use during adolescence, and identifies implications for policies and programs designed to prevent substance use.
This document is part of a series of Points of Prevention publications that document the contribution of CSAP’s National Cross-Site Evaluation of High-Risk Youth Programs to prevention knowledge and provide science-based guidance for improved prevention policy and practice.
Charles G. Curie, M.A., A.C.S.W.Substance use is increasingly recognized as one of the Nation's most pervasive, costly, and challenging health and social problems. The use, and particularly the early use, of tobacco, alcohol, marijuana, and other illicit drugs is intricately entwined with serious personal and social problems, including school failure, crime, family violence and abuse, and a host of additional social and personal problems that constitute a continuing national tragedy. For over a decade, the Center for Substance Abuse Prevention (CSAP) within the Substance Abuse and Mental Health Services Administration (SAMHSA) has been the Federal agency charged with providing leadership in preventing the profound negative consequences of substance use. Important components of this leadership responsibility include design and funding of demonstration substance use prevention programs, followed by evaluation of those programs to identify prevention services that work in real community settings.
The prevention concept is simple. It postulates that changing the social and personal conditions that promote and support substance use will benefit society and individuals more than trying to treat the physical and psychological results of use or control its social consequences. Implementing the prevention concept, however, is complex. A growing body of research has documented a “web of influence” through which circumstances of community, family, school, and peer group condition youths’ risk for substance use (CSAP, 1999).
| The science-based knowledge emerging from this study is strengthening our understanding of conditions that can influence substance use patterns among high-risk youth. |
This monograph addresses a topic that is important to both prevention practitioners and researchers. While the risk and protection framework heavily influences the work of both professional groups, the interrelationships among risk and protective factors have not been well understood. Previous efforts to model how risk and protection factors relate to substance use have been based on data from the general youth population, often using relatively small samples. The data collected by the CSAP National Cross-Site Evaluation of High-Risk Youth Programs provide an excellent opportunity to investigate these interrelationships and, through modeling, to advance our understanding of how certain risk and protective factors influence the behaviors of at-risk youth. The Cross-Site sample reflects a diversity of youth from many population subgroups. A robust model of risk, protection, and substance use can show practitioners which factors to target and suggest appropriate intervention strategies. Researchers can use such a model to identify critical variables, guide measurement selection decisions, and focus data analysis plans.
More specifically, this monograph contributes to understanding how prevention can strengthen conditions that help youth at risk avoid substance use by presenting evidence, findings, and recommendations in the following areas:
Major findings presented in this monograph include the following:
The National Cross-Site Evaluation of High-Risk Youth (HRY) Programs is a 5-year study that CSAP began in 1995 (Sambrano, Springer, and Hermann, 1997). Forty-eight HRY demonstration programs across the Nation participated in the study.1 These grantees were funded by CSAP to implement and assess programs to prevent and reduce the use of alcohol and illicit drugs among at-risk youth. A rigorous research design incorporating lessons from earlier evaluations of prevention service implementation and effectiveness guided the study. The evaluation used a quasi-experimental comparison group design to study the more than 6,000 youth who were participating in the 48 demonstration programs, comparing them with more than 4,500 similar youth in the same communities who were not participating in the programs.
The study has several unique features, including:
The study design allowed CSAP to test the effectiveness of programs by measuring changes in participants’ risk, protection, and substance use over time and comparing the results to changes in similar youth who did not receive program services.2
| Effective program strategy must involve messages and activities that are meaningful to the particular circumstances of youth. |
Effective program strategy must involve messages and activities that are meaningful to the particular circumstances of youth. The maturation process, which is rapid in the earlier years of life, provides clear examples. Prevention objectives that are appropriate for teenagers at a stage when tobacco and alcohol experimentation is widespread are not appropriate for 9-year-old children. As youth take steps toward the independence that is necessary for autonomy in adulthood, the protective and supervisory role of family evolves as well. Young men’s and young women’s risk and protection influences differ, pointing to the need for differing gender-based strategies. What faces the prevention field at this stage in its development is a “sorting out” process—one of identifying from among the many prevention objectives and strategies those approaches that prove most effective for youth at risk.
Prevention professionals can use the information generated by CSAP’s study to improve their aim at this moving target of approaches. This diverse high-risk study sample provides an excellent proving ground for identifying connective paths among risk and protection factors and substance use. It also provides a basis from which to assess the appropriateness of specific prevention objectives and practices in a variety of settings.
1 CSAP funded 94 programs in 1994 and 1995. Programs were not included in the Cross-Site Evaluation if they served children primarily under the age of 9. Approximately half of the 48 programs that participated in the Cross-Site study were funded in 1994 for 5 years and the other half were funded in 1995 for 3 years. These program were located throughout 22 States, including Alaska and Hawaii.
2 Findings on program effectiveness are reported in other monographs published in the CSAP Points of Prevention Series.
| This diverse high-risk study sample provides an excellent proving ground for identifying connective paths among risk and protection factors and substance use. |
This report cites findings using data from both program participants and comparison group youth because its objective is to convey the importance of differences in youth risk and protection characteristics as they relate to substance use. Typically, baseline data were collected within a few weeks before or after participating youth began receiving services, and at the same time for comparison group youth. The information that follows on study sample diversity, patterns of risk and protection by age, and patterns of substance use by age and gender draws on these baseline data.
The age at which to target youth for preventive interventions has been a topic of discussion and debate among prevention practitioners and researchers. Prevention or early intervention programs have been developed for all full age ranges of youth. These encompass preschool programs aimed at enhancing children’s development and addressing negative behaviors, as well as programs aimed at college binge-drinking and at safely negotiating life’s major transitions. Youth selected for inclusion in CSAP’s National Cross-Site Evaluation were limited to those between the ages of 9 and 18. Within that broad age range, from preadolescence through the late teens, the target focus of each program determined the age of the youth it served. Figure 1 displays age and gender information for youth (both program participants and comparison group youth) in the study sample.
Figure 1
Distribution of Sample by Age and Gender
(N = 10,473)
| This concentration of program effort in the middle school years reflects a planned response to the perception that middle school youth are at a transition point that puts them at particular risk for starting to use substances. |
Because 19 (40%) of the programs included in the study targeted female adolescents, there are many more females (66%) than males (34%) in the total sample. The gender-specific age profiles within Figure 1 indicate that, in this sample, the females are somewhat older than the males (mean female age = 12.84; male = 12.76), with a higher percentage of females in the 14- to 17-year-old age groups.
Many of the HRY prevention programs were community-based, focusing on particular community populations. Figure 2 indicates that these programs served a diversity of racial/ethnic groups. More than 33 percent of the youth were African American and around 25 percent were Hispanic. Of the remaining youth, approximately 10 percent were Native American, 10 percent were Asian/Pacific Islander, and 10 percent were White/non-Hispanic. There is cultural and/or regional diversity within the racial/ethnic groups as well (e.g., programs targeting youth in recent-immigrant communities).
Some programs used recruitment procedures that targeted special populations, adding diversity to the youth sample. Most programs recruited youth from high-risk settings: schools, neighborhoods, housing developments, or youth organizations. As an alternative, several programs based participant selection on a common individual behavioral or personal attribute. Specifically, two programs served youth who had been placed in a secure facility by court order; two programs targeted youth with disabilities (physical and developmental/emotional); one program focused on young women with histories of sexual abuse; and one program focused on youth in the foster care system.
Figure 2
| The diversity of the youth in the sample strengthens the study’s ability to identify general characteristics of at-risk youth. |
Over the past decade, research and practice in prevention have produced a dominant approach to thinking about how to prevent substance use and associated problems among youth. A large body of research informs this approach. It proposes that substance use and other problems are part of a consistent pattern of circumstances that tend to occur together. Community and neighborhood environment, school conditions and performance, family environment, and particularly, peer attitudes and behaviors all contribute to substance use. This literature also established that early behaviors are highly predictive of the development of problem behaviors as youth mature.
Articulation of the role of risk factors in the initiation and growth of substance use among youth had important implications for prevention programming. Early prevention efforts were often based on the assumption that young people used substances largely because they were not informed of their health, legal, and social dangers. Programs based on this assumption were limited in content and disappointing in result. Knowledge about risk factors gave prevention practitioners another focus for their efforts. If the conditions associated with substance use could be improved, substance use itself might be stopped. Targeting this broad range of conditions identified by the research, prevention practitioners developed programs aimed at reducing risk in the community, in schools, in families, and among peers. Research has shown that when risks in a child’s life are reduced, the child is less vulnerable to substance use and related social and health problems as he or she matures (Hawkins, Lishner, Jenson, and Catalano, 1987).
Risk factor research substantially influenced the ways in which prevention funding organizations, program designers, and service deliverers thought about their program activities and objectives. Parenting programs, community-wide prevention coalitions, and programs aimed at transforming schools as communities (Battistich, Schaps, Watson, and Solomon, 1996) all were added to educational and informational efforts as strategies to prevent substance use. However, the “risk” literature is not always a comfortable fit with the orientation of prevention workers, who want to improve the lives of youth and strengthen their communities. Some prevention researchers have dubbed the risk paradigm a “damage model” (Wolin and Wolin, 1995) that focuses on negative influences on youth. Research on protective factors has emerged to focus, alternatively, on “what is positive and healthy in young people” (CSAP, 1999, p. 3).
| Protective factors represent the influences, orientations, and behaviors in youth’s lives that contribute to positive development and help prevent negative behaviors and outcomes such as substance use. |
Program activities designed to reduce risk or promote protection in the external environments of community, school, and family are very different from prevention strategies that work directly with youth to develop internal protective factors. Therefore, the programmatic implications of changing external risk and protection are very different from those for working on the development of internal factors in young people.
| The diverse Cross-Site Evaluation sample of high-risk youth represents an exceptional opportunity to assess the ways in which risk and protection factors change with maturation, interact, and relate to substance use. |
The importance of external and internal risk and protection for the development of effective prevention interventions depends on the interrelationships of these factors and on their association with substance use. The diverse Cross-Site Evaluation sample of high-risk youth presents an exceptional opportunity to assess the ways in which risk and protection factors change with maturation, interact, and relate to substance use. The Cross-Site Evaluation measured specific dimensions of external and of internal risk and protection. Although not exhaustive of the many factors that have been proposed in the literature, the Cross-Site measures include important factors frequently targeted in prevention strategies.
The study also used a set of substance use norm variables. These focus on the attitudes and behaviors—specifically related to the use of alcohol, tobacco, and marijuana—of significant reference groups in the lives of youth. Past research has shown strong relationships between self-reported substance use and these normative contexts. Ability to assess the relationship of these social norm contexts to general risk and protective factors with known correlations to many youth behaviors is another benefit of this study.
All of the measures presented in this report are from the CSAP National Youth Survey completed by both participant and comparison group youth themselves. Self-report measures are consistent with a model of influence in which the individual youth remains at the core. All external influences are “processed, interpreted, and responded to based upon those characteristics that the individual brings to the situation” (CSAP, 1999, p. 2).
To summarize and order the existing research findings and program assumptions about risk and protection, CSAP developed a framework representing the “web of influence” through which risk and protection shape substance use (CSAP, 1999). As Figure 3 shows, this framework groups external influences on youth into five areas (domains): family, peers, school, community, and society.
Figure 3
The “Web of Influence” on Substance Use
The external risk and protective factors and substance use norms identified earlier in this report (and measured by the self-reported National Youth Survey) can be placed within these five domains. The internal risk and protective factors identified earlier shape each individual’s response to the external factors. Protective internal factors may strengthen protective external influences, or they may help youth resist the external influences that increase risk for substance use. The reverse is also true: internal risk factors may heighten the effect of external influences that increase risk or may blunt even protective external factors. In the National Cross-Site measurement of risk and protection, substance use norms are individual risk and protection factors that represent the youth’s perception of the substance use expectations and behaviors of significant others in her or his life.
| The framework raises important questions about the ways in which external influences and internal orientations interact to prevent or promote substance use among young people. |
Risk and protection research indicates that the risk for substance use during adolescence is strongly related to identifiable characteristics of the environment in which adolescents live. The family management practices of parents or family care-givers, the opportunities and policies of the school environment, and the quality of life and opportunities for participation that characterize the community all influence the development of positive or problem behaviors in youth. The National Youth Survey designed for the Cross-Site Evaluation asked youth respondents about their perceptions of their family, school, and community and defined these measures as follows:
| The challenge for prevention professionals is to make these environments more protective and less characterized by risk, to help youth connect more effectively with protective influences, and to help youth resist the negative influences of environmental components characterized by risk. |
The risk and protection framework suggests that more protection in each of these elements of a youth’s environment will increase the opportunities for positive development and outcomes and reduce the chances of negative outcomes. The challenge for prevention professionals is to make these environments more protective and less characterized by risk, to help youth connect more effectively with protective influences, and to help youth resist the negative influences of environmental components characterized by risk.
Strategies for helping youth successfully negotiate environmental influences must be appropriate to the changing relationship between young people and their environments as they mature. Figure 4 displays the responses of sample youth, aged 9 to 18, to the external risk and protection measures. Responses are baseline measures taken when the youth entered the study.
Figure 4
Distribution of Youths’ Perceptions of External Risk and Protective Factors By Age at Program Entry
(N = 10,473)
As youth age, their perceptions of the level of risk or protection offered by external circumstances shift considerably. For each of the four external risk and protection measures, the perceived level of protection decreases with age, most markedly after the age of 12. The decline in perceived protection with age is greatest for school prevention environment and family supervision. This pattern is partly expected and normal. As youth mature and become more autonomous, the character of family supervision and involvement changes to allow youth more responsibility in their daily lives. The pattern for school indicates that the school environment has fewer opportunities for promoting positive choices and development as youth move through middle school into the high school years. The pattern of decline is less dramatic with respect to the community and neighborhood measures, indicating a more moderate change in youth involvement in positive community activities, or in their perception of social disorganization in their neighborhood, as they age. The relative stability of involvement in positive community activities indicates potential for community-based prevention activities to influence youth. As youth age and become more independent of home, they form associations in the community. The challenge to prevention is to help make these associations positive ones.
| Although researchers do not agree on what makes a youth resilient, they often cite belief in one’s ability to produce a meaningful and worthwhile future as a component. |
| Developing a sense of connectedness to meaningful segments of the environment that provide and support positive opportunities is an important aspect of the internal protective orientation of young people in high-risk environments. |
Some resiliency researchers emphasize the importance of attachments to positive social influences, such as a caring and trusted adult (Benard, 1991). The influence of these attachments, sometimes called “bonding” or “connectedness,” is an emerging focus in the prevention literature. Most often, this bonding is seen as an individual (internal) attribute, the young person’s belief that this external connection is important and meaningful to him or her. Bonded youth feel a stake in the external environment. They believe that interacting and accomplishing in that environment are worthwhile and contribute to a positive future. Thus, developing a sense of connectedness (Resnick et al., 1997) to meaningful segments of the environment that provide and support positive opportunities is an important aspect of the internal protective orientation of young people in high-risk environments.
CSAP’s National Youth Survey provided self-reported information on the following internal risk and protective factors:
| Like external risk and protection, internal orientations may change as youth mature. |
Like external risk and protection, internal orientations may change as youth mature. Figure 5 displays the change in average responses for each of the internal risk and protection factors with age in the baseline sample.
Figure 5
Internal Risk and Protective Factors by Age at Program Entry
(N = 10,473)
Similar to the results for external risk and protection, the responses of younger children are indicative of more protective orientations, including stronger bonding with family and school, stronger perceptions of self-efficacy and self-control, and more belief in self. For this sample, all but one of the protective factors are highest for the youngest age group. The exception is social confidence, which rises slightly from preadolescence to the midteen years. The rest of the orientations drop toward greater risk between the ages of 11 and 15, indicating a reduction in the internal protective orientations during the middle and early high school years.
The change in protective factors with age is particularly dramatic with respect to the indicators of bonding or connectedness. Family bonding changes the most, with a precipitous drop from a positive family orientation at the age of 9 to a less strong feeling that the family is a place for meaningful communication, contribution, or recreation during the midteen years. School bonding shows a similar change. Other orientations, with the exception of social confidence, follow a similar but less pronounced pattern.
| Research has established peer attitudes and behaviors as one of the strongest correlates of self-reported substance use. |
CSAP’s National Youth Survey measured several substance use norms in those social contexts that are closest to youth: parents and friends. Youth were asked to report their perceptions of their parents’ reactions to substance use, as well as their perceptions of their friends’ substance use behaviors and their friends attitudes toward use:
As with the other risk and protection domains, youths’ perceptions of substance use norms in these significant reference groups change rapidly with age. Figure 6 shows that, within the 5-year age range from 11 to 17, youth perceive substance use norms as moving dramatically from prohibitive to permissive. To underscore this change, 17 percent of the 10-year-olds compared with 73 percent of the 16-year-olds thought their best friend used alcohol sometimes.
| In sum, as youth age, they perceive a decline in the protective nature of external factors, internal factors, and substance use norms toward greater risk. |
Figure 6
Substance Use Norms by Age at Program Entry
(N = 10,473)
In sum, as youth age, they perceive a decline in the protective nature of external factors, internal factors, and substance use norms toward greater risk. The following section looks at patterns of substance use as youth mature.
The CSAP HRY demonstration programs have as their main objective preventing or reducing substance use and its associated problems. Therefore, it is essential to know how many and which youth were using substances when they entered the programs. The self-report survey given to youth at the start of the programs asked respondents about their use of a number of substances within the preceding 30 days 3 and over their lifetime. Table 1 shows prior 30-day substance use rates for the portion of the study sample ages 12 through 17, by age subgroups. It also compares those results with data from the 1998 National Household Survey on Drug Abuse (NHSDA) (SAMHSA/OAS, 1998), a randomly sampled general population survey of persons 12 years of age or older.
Table 1
Comparison of NHSDA and Cross-Site Substance Use for 12- to 17-Year-Old Respondents
NOTE. National Household Survey on Drug Abuse (NHSDA) sample size for 12- to 17-year-olds (n = 6,778); Cross-Site sample size for 12- to 17-year-olds (n = 7,245).
| The Cross-Site programs serve youth in a higher-risk population than young people in the general population. |
Looking specifically at the percentage of youth in the Cross-Site sample reporting use of any one substance during the previous 30 days, the rate of use is low (Figure 7). The most frequently reported substances used are alcohol (18%), cigarettes (18%), and marijuana (14%). Relatively few youth reported recent use of drugs such as cocaine or crack, speed, tranquilizers, PCP, and heroin. This pattern suggests that the program sites recruited youth who were usually not regularly involved in substance use and were, therefore, appropriate participants in prevention or early intervention programs.
Figure 7
Self-Reported 30-Day Substance Use at Baseline: Percentage of Study Sample
(N = 10,473)
3 Youth were asked: “How many days in the last 30 days did you smoke a cigarette/have a drink of alcohol/use marijuana?” Responses ranged from 0 to 5 (six response categories). Possible responses included (0) none, (1) 1 or 2, (2) 3 to 5, (3) 6 to 9, (4) 10 to 19, and (5) 20 to 31.
To profile the sample, a composite measure of “30-day substance use” was constructed. This composite measure includes the use of any one of three substances—tobacco, alcohol, or marijuana—within the past 30 days. Figure 8 displays the percentage of youth who reported substance use by age and gender using this composite measure of 30-day use.
Figure 8 Self-Reported 30-Day Substance Use at Baseline by Age and Gender (N = 10,473)
| Use rates across the age groups are consistently higher among males than females. |
The findings on risk, protection, and substance use and the age of youth reveal a consistent pattern. As youth move through the adolescent years, there is a steady movement from the protective to the risk conditions in most of the external and internal factors identified here. That movement is particularly great in family bonding, school bonding, and peer attitudes and use, those factors that refer to the social environments to which youth are building attachments as they mature. The following section explores these risk and protective factors further by explicitly assessing their relation to substance use.
All risk and protection frameworks assume that each youth has a personal “web of influence”: a set of risk and protective factors, external and internal, that are related to that youth’s use or nonuse of substances. Testing this assumption involves identifying the degree to which a relationship exists between self-reported substance use and the presence of protective influences or internal orientations. These relationships for the entire youth sample and for males and females are examined separately to determine whether there are differences between these groups in use levels, increases in use, or other aspects of substance use related to risk and protective influences.
The analysis of these relationships uses correlation coefficients (Pearson’s r) to compare relationship strengths, with coefficients of ±1.0 indicating complete association and 0.0 meaning no association. Because of the very large sample size, statistical significance is not a meaningful indicator for this analysis; in fact, all reported correlations are highly significant.4 The external, internal, and substance use norm domains are analyzed separately. Substance use is measured as combined 30-day substance use of cigarettes, alcohol, or marijuana.
Figure 9 shows the strength of association between each of four external risk and protection factors and self-reported substance use. The four factors were selected from three external domains: one each from the family (family supervision) and school (school prevention environment) domains, and two from the community domain (community protection environment, neighborhood risk). Strength of association is shown separately for males and for females. In the case of protective factors, the degree of the negative association indicates the strength of the factor’s protective influence against substance use. In the case of a risk factor, the degree of the negative association indicates the tendency of the factor to make substance use more likely. The larger the negative correlation, the greater the degree of protection against use or risk for use the influence provides.
Figure 9
Bivariate Correlations Between External Risk and Protection Factors and Substance Use,
by Gender (N = 10,473)
The figure indicates that for this sample of youth, the stronger the external protection, the less substance use youths will experience. For example, in risk-free neighborhoods, youth are less likely to use substances than youth in high-risk neighborhoods. On the other hand, youth with little family supervision are more likely to use substances.
Males and females demonstrate different patterns of association between external influences and substance use. Family supervision influences male and female behaviors similarly, but each of the other external influences is more strongly 5 associated with substance use for males than for females, dramatically so in the case of neighborhood risk. This suggests that conditions in the neighborhoods have a greater influence on substance use in males than in females.
| The data suggest that conditions in the neighborhood have a greater influence on substance use in males than in females. |
4The statistical significance of an association is determined by the strength of the relationship between the two variables and the number of cases used in the analysis. As the size of a sample increases, weaker relationships become statistically significant.
5 For this analysis, values of 0.0 to -0.19 are considered weak, -0.20 to -0.34 are considered moderate, -0.35 to -0.49 are strong, and -0.50 and above are very strong.| The association between peer use and personal substance use is the strongest among all the risk and protection factors investigated in this study. |
In the “web of influence” framework, the substance use norms including the youths’ perception of parental attitudes, peer attitudes, and peer use fall into the external domains of family and peers. 6
Figure 10 summarizes the patterns of association between these external substance use norm influences and substance use. The results confirm the strong association between substance use by peers and personal use that has consistently been found in research literature. In fact, the association between peer use and personal substance use is the strongest among all the risk and protection factors investigated in this study. Parental attitudes are also very strongly associated with personal use, and peer attitudes are strongly associated. In all cases, these associations are slightly stronger among males than among females, particularly for peer attitudes and peer use.
Figure 10
Bivariate Correlations Between Substance Use Norms and Substance Use: Males and Females
(N = 10,473)
The patterns of association between internal risk and protection factors and substance use are shown in Figure 11. Of the six internal factors investigated, school bonding and family bonding—two factors that indicate connectedness to social contexts- clearly show the strongest associations with substance use, although the associations are in the “moderate” range. Belief in self, self-control, and self-efficacy are less associated with substance use, and social confidence is very weakly related to substance use.7
Figure 11
Bivariate Correlations Between Internal Risk and Protection Factors and Substance Use: Males and Females
(N = 10,473)
6 All risk and protective factors are coded so that a higher value indicates stronger protection. Therefore, negative correlations of parental and peer attitudes and peer use with substance use indicates negative relationships between these factors and substance use.
7 All internal factors are coded so that higher numbers represent greater protection.
The pattern of gender differences for the internal factors is very different from the pattern for both of the external influence
categories. The relationships between all the internal risk and protection factors and substance use are substantially
stronger for females than for males. This is especially true for family bonding.
The differences in association with substance use among the internal risk and
protection factors investigated in this study have implications for understanding
the influences of these factors on substance use and for increasing the effectiveness
of prevention interventions. This study’s findings reinforce the body of
existing prevention research on internal factors in several areas:
The relationships between all
the internal risk and protection
factors and substance use are
substantially stronger for
females than for males.
The findings from the Cross-Site Evaluation confirm the premises of CSAP’s “web of influence” and support use of the risk and protection framework as a foundation for prevention programming. Risk and protection factors across the internal and external domains correlate with substance use by youth. Furthermore, the results are consistent with findings from less comprehensive studies. Youth perceptions of peer norms are very strongly associated with use, and parental norm setting is strongly associated with use. School bonding, school prevention environment, family connectedness, and family supervision are moderately related to substance use. The consistent pattern across domains is that male substance use behavior is shaped more by the external environment than female behavior, whereas female substance use behavior is more sensitive to internal influences. Most important, the findings concerning internal risk and protection highlight the importance of the school and family connectedness orientations. These findings suggest that internal risk and protection is not primarily an “inoculation” against external influences, but a catalyst and facilitator for building connectedness with positive external environments.
| Internal risk and protection is not primarily an “inoculation” against external influences, but a catalyst and facilitator for building connectedness with positive external environments. |
The findings in the previous section provide insights into the structure of the relationships among individual risk and protective factors within the domains that are part of the “web of influence.” However, the findings do not reflect the interrelationships among the domains themselves, which are depicted by the arrows and lines in Figure 12. To investigate more specifically the relationships suggested in the “web of influence,” a modeling technique 8 was used to describe what can be called “pathways of influence” among the risk and protective factors and substance use. This technique does not test cause-and-effect relationships, but it does indicate whether such relationships are plausible. Figure 12 shows the model, which is based on data from the full (both participant and comparison youth) study sample.9
Figure 12
Structural Equation Model for Risk and Protection Factors and Substance Use Among High-Risk Youth
(N = 10,473)
Summary Statistics CFI = 0.90 RMSEA = 0.03 NNFI = 0.90 NOTE. Negative correlations between parental attitudes and peer substance use with personal substance use result because all measures with the exception of substance use are coded so that higher numbers represent more positive behaviors. Widest lines indicate correlations of 0.50 or more, medium- width lines are correlations ranging from 0.20 to 0.49, thin lines are correlations of less than 0.20.
The model includes the array of risk and protection factors and substance use norms explained in the previous section, reflecting a comprehensive application of the CSAP framework. The number next to each arrow indicates the strength of the relationship the arrow represents (i.e., the relationship between one factor and another).10 The model fits the data well (CFI = 0.90, RMSEA = 0.03, NNFI = 0.90),11 meeting the high standard for good model fit in these kinds of analyses.
8Structural equation modeling was conducted using LISREL (Joreskog and Sorbom, 1999).
9 Multigroup analyses contrasting treatment and comparison youth demonstrated no substantial differences in model fit or path coefficients, justifying use of the combined sample of youth at high risk.
10 The width of each arrow shaft indicates the strength of the association between the individual protective factor and substance use; i.e., the degree to which higher protection on the factor relates to less use. Wider arrows shafts are indicative of stronger relationships.
11 The Comparative Fit Index (CFI) statistic measures the goodness of fit. Models with indices of 0.90 or more are considered to be strong-fitting models.
To produce a lean model and improve statistical fit, the number of factors included in the model was reduced analytically. The following decisions were made during model building:
The model is consistent with and elaborates on prior prevention research using a social ecology approach. Hawkins and Weis (1985) showed that peer influences did not entirely explain the phenomenon of substance use and that external factors played a vital role in shaping the resiliency of youth and their ability to resist substance use. Kumpfer and Turner (1991) later supported and expanded on this work in studies of youth living in Utah.
The model in Figure 12 shows one possible explanation for the way risk and protection factors from different domains in the “web of influence” interact in their connection with substance use.
Based on the model, it is possible to form broader conclusions about the dynamics of the “web of influence.” The model clearly supports the interactive nature of protective influences—internal, external, and normative—against substance use.
The model shown earlier (see Figure 12) depicts pathways of influence for all the youth in the Cross-Site study sample. Prevention practitioners may question whether the dynamics of risk and protection presented in the model apply equally to subgroups of youth at risk. To answer that question, the study sample was analyzed by gender, age, and ethnicity.
The model shown in Figure 13 is partitioned by gender. (Coefficients in parentheses are for females.) There is a strong fit for both males and females with minimal differences, and the overall statistical fit of the model is good (CFI = 0.90). All of the hypothesized paths hold for both gender groups.
Figure 13
Structural Equation Model for Gender Groups
(Males, N = 3,596; Females, N = 6,941)
Summary Statistics
CFI = 0.90 RMSEA = 0.03 NNFI = 0.90 NOTE. Coefficients in parentheses are for females. Widest lines indicate correlations of 0.50 of more, medium-width lines are correlations ranging from 0.20 to 0.49, thin lines are correlations of less than 0.20.
Some interesting implications emerge from the few gender differences shown in the model. School connectedness is a stronger predictor of reported school performance (grades and school attendance) for females than for males, suggesting that school provides a more relevant forum of connectedness for females than for males. Parental attitudes toward substance use are slightly more strongly related to personal use for females (-0.42) than for males (-0.37). Both paths from neighborhood risk to peer substance use to the subject’s own substance use are slightly stronger for males (0.31 and -0.56) than females (14 and -0.48). These differences are minor within the overall similarity of the paths to substance use for both genders.
| School connectedness is a stronger predictor of reported school performance for females than for males. |
Partitioning the model by age reveals some important differences, as shown in Figure 14. (Coefficients in parentheses are for the older youth.) The model is strong for both younger (9- to 11-year-old) and older (12- to 17-year-old) youth.
Figure 14
Summary Statistics
CFI = 0.90 RMSEA = 0.025 NNFI = 0.90 NOTE. Coefficients in parentheses are for the 12- to 17-year-old youth. Widest lines indicate correlations of 0.50 or more, medium-width lines are correlations ranging from 0.20 to 0.49, thin lines are correlations of less than 0.20.
Several major age-related differences emerge in the model. The first is the importance of family factors for older youth. Family supervision and parental attitudes are stronger predictors of substance use behavior in youth age 12 years and older than they are for preteens. Some of this difference may be attributed to reduced variation in the younger group, where positive orientation toward family is higher in general. Nevertheless, the data indicate that family continues to play a critical protective role as youth develop through adolescence—despite the fact that teens report a decline in family supervision and family connectedness as they grow older (see Figures 4 and 5). Second, peers influence older youth more than younger youth. Peer attitudes and perceptions of peer use are stronger predictors of substance use among older youth than they are for those younger than 12. This reflects the developing sensitivity to social perceptions that begins in early adolescence and continues into adulthood.
| School performance is a stronger predictor of substance use behavior among older youth than in younger youth. |
Fourth, school performance is a stronger predictor of substance use behavior in older youth than in younger youth. For students older than 12, the model suggests that poor grades are associated with substance-using peers and personal substance use.
| The model has a high degree of consistency across the major ethnic groups and cultures of America. |
Figure 15 shows the model partitioned by ethnicity. Multiple-group structural modeling indicates that all the paths hypothesized in the model hold for all ethnic groups. Differences between the groups for most paths are typically small and idiosyncratic. A few of these differences do suggest particular areas of concern about risk and protection for specific racial and ethnic groups. For example, school performance among African-American youth is more strongly associated with family supervision and less strongly associated with the youth’s own school connectedness than for youth in other racial/ethnic groups. For Hispanic youth, the opposite pattern is present. Overall, however, the model has a high degree of consistency across the major ethnic groups and cultures of America.
Figure 15 Structural Equation Model for Racial/Ethnic Groups (N = 10,473)
NOTE. Coefficients are listed from top to bottom in the following order: Native American (n = 1313) Asian/Pacific Islander (n = 1164) African American (n = 3722) Hispanic (n = 2742) Non-Hispanic/White (n =1212)
| Family continues to play a critical protective role as youth develop through adolescence. Parental attitudes, peer attitudes, and perceptions of peer use are stronger predictors of substance use among older youth than they are for those younger than 12. |
In summary, the model of risk and protection factors for substance use among high-risk youth is robust. It is based on a large sample of at-risk youth, and it applies to females and males, younger and older youth, and across ethnic groups. The model emphasizes the critical importance of family, peers, and individual protective factors for buffering youth from substance use. More importantly, it suggests the interdependence of these domains. The key to prevention is not to make youth insensitive to their social environment, but to ensure that they are strongly connected to positive, healthy environments. Given the stability of the model across these important population subgroups, practitioners can use it to address important questions about which risk and protective factors prevention programs should be targeting. By reducing to a manageable number the important predictors of substance use, the model should help practitioners focus their resources efficiently to improve program outcomes.
The National Cross-Site Evaluation of High-Risk Youth Programs has created an excellent opportunity to expand knowledge about risk and protective factors and substance use among youth at high risk. The diversity in the sample allowed analyses of patterns in risk and protective factors and substance use as youth mature through the adolescent years, across genders, and across racial and ethnic groupings. These analyses produced important findings with profound implications for prevention policy and program design.
Finally, structural modeling brings additional coherence and focus to the ways in which risk, protective, and normative factors may influence substance use among youth. The analyses produced a robust model of adolescent substance use and its correlates among high-risk youth populations. The relationships displayed in this model substantiate prior findings on the relationships among risk and protective factor measures and substance use. The model also expands on this research by showing the indirect and direct relationships among individual, family, peer, school, and community factors. The model is relatively stable across age, gender, and ethnic groups. There is some evidence that family influences are more important for females than males and that neighborhood risk influences are stronger for males.
By specifying plausible pathways within and between external, internal, and normative risk factors and substance use, the model adds detail to understanding of the “web of influence” on adolescent substance use.
| Successful and fulfilling interaction with a positive social environment is a key to positive development. The model places connectedness with the social settings of family and school at the core of prevention programming. |
Finally, by highlighting the importance of the family and the school, the model suggests that youth develop important personal competencies (such as cooperativeness, a positive view of the future, a belief in self, and a feeling of personal efficacy) through positive orientations to, and interactions with, the social settings that are central to our society. Successful and fulfilling interaction with a positive social environment is a key to positive development. Thus the model places “connectedness” with the social settings of family and school at the core of prevention programming.
All of this has direct implications for prevention policy and practice. The trends in risk, protection, and use confirm the critical status of the middle school years as a period of experimentation and initiation of substance use, as well as a critical transition in the social bonds that youth form. The data also indicate that programs serving youth at risk in middle school must reach some youth who have already initiated use. These are important years for prevention.
The overall findings also confirm the importance of comprehensive prevention that addresses the range of environmental factors as well as the individual orientations and behaviors of the youth themselves. Community, family, school, and peer circumstances clearly condition substance use. For schools, youth's connection and performance are the central issue with respect to preventing substance use. The school is the major forum for accomplishment and recognition in youth at these ages. Those who are connected with that environment use cigarettes, alcohol, and marijuana less.
The central implication for prevention is the need to build connections to positive and meaningful social environments for youth. Though important, just changing the environment, or just changing individual orientations, is not enough. Protection and positive development requires connection between the two. Building and supporting these connections is a central challenge to prevention and a positive promise to youth.
Akers, R. L. (1977). Deviant behavior: A social learning approach. Belmont, CA: Wadsworth.
Battistich, V., Schaps, E., Watson, M. and Solomon, D. (1996). Prevention effects on the Child Development Project: Early findings from an ongoing multi-site demonstration trial. Journal of Adolescent Research, 11, 6-11.
Benard, B. (1991). Fostering resiliency in kids: Protective factors in the family, school, and community. Northwest Regional Educational Laboratory/Western Regional Center for Drug-Free Schools and Communities.
Center for Substance Abuse Prevention. (1999). Understanding substance abuse prevention, toward the 21st century: A primer on effective programs. CSAP Monograph, DHHS Publication No. (SMA) 99-3301. Rockville, MD: Substance Abuse and Mental Health Services Administration, U.S. Department of Health and Human Services.
Elliott, D., Huizinga, D., and Ageton, S. S. (1982). Explaining delinquency and drug use. Report No. 21. Boulder, CO: Behavioral Research Institute.
Ford, D. S. (1983). Factors related to the anticipated use of drugs by urban junior high school students. Journal of Drug Education, 13(2), 187-197.
Fors, S. W. and Rojek, D. G. (1983). The social and demographic correlates of adolescent drug use patterns. Journal of Drug Education, 13, 205-222.
Foster, B. (1984). Upper middle class adolescents’ drug use: Patterns and factors. Advances in Alcohol and Substance Abuse. 4, 27-36.
Hansen, W. B. (1997). Prevention programs: Factors that individually focused programs must address. In Resource papers for the Secretary’s Youth Substance Abuse Prevention Initiative, (pp. 53-66). Rockville, MD: Center for Substance Abuse Prevention, Substance Abuse and Mental Health Services Administration.
Hawkins, J. D., Lishner, D., Jenson, J. and Catalano, R. (1987). Delinquents and drugs: What the evidence suggests about prevention and treatment programming. Youth at Risk for Substance Abuse: National Institute on Drug Abuse 1987, 81- 131.
Hawkins, J. D. and Weiss, J. G. (1985). The social development model: An integrated approach to delinquency prevention. Journal of Primary Prevention, 6, 73-97.
Joreskog, K. G. and Sorbom, D. (1999). Lisrel 8: New statistical features. Chicago: SSI.
Keith, J. G., and Perkins, D. F. (1995). 13,000 adolescents speak: A profile of Michigan youth. East Lansing, MI: Michigan State University, Institute for Children, Youth, and Families.
Kumpfer, K. L. and Turner, C. W. (1991). The social ecology model of adolescent substance use: Implications for prevention. The International Journal of Addictions, 25, 435-463.
Newcomb, M. D., Huba, G. J., and Bentler, P. M. (1986). Risk factors for drug use among adolescents: Concurrent and longitudinal analyses. American Journal of Public Health, 776, 525-531.
Newcomb, M. D., and Felix-Ortiz, M. (1992). Multiple protective and risk factors for drug use and abuse: Cross-sectional and prospective findings. Journal of Personality and Social Psychology, 63(2), 280-296.
Oetting, E. E., and Beauvais, F. (1986). Peer cluster theory: Drugs and the adolescent. Journal of Counseling and Development, 65, 17-22.
Resnick, M. D. (1999). The impact of caring and connectedness on adolescent health and well-being: Research summary, in getting results: Developing safe and healthy kids Update 1, 22. Sacramento, CA: California Department of Education.
Resnick, M. D., Bearman, P. S., Blum, R. W., Bauman, K. E., Harris, K. M., Jones, J., Tabor, J., Beunning, T., Sieving, R.,
Shew, M., Ireland, M., Bearinger, L., and Udry, J. R. (1997). Protecting adolescents from harm: Findings from the National Longitudinal Study on Adolescent Health. Journal of the American Medical Association, 278, 823-832.
Resnick, M. D., Harris, L. J., and Blum, R. W. (1993). The impact of caring and connectedness on adolescent health and well being. Journal of Pediatrics and Child Health, 29(1), 83-89.
Sambrano, S., Springer, J. F., and Hermann, J. (1997). Informing the next generation of prevention programs: CSAP’s Cross-Site Evaluation of the 1994-1995 High-Risk Youth Grantees. Journal of Community Psychology, 25, 375-395. SAMHSA/OAS. (1998). National Household Survey on Drug Abuse population estimates 1998. Rockville, MD: Office of Applied Studies.
Wolin, S. J. and Wolin, S. (1993). The resilient self: How survivors of troubled families rise above adversity. New York: Villard Books.
Wolin, S., and Wolin, S. (1995). Resilience among youth growing up in substance-abusing families. Pediatric Clinics of North America, 42, 415-429.
| |||||||||||||||||||||
| Initiatives | Funding | Home | ||||||||
|
Site Map
Help
Account
Cart
Home
