4
Modeling the Payoffs of Interventions to Reduce Adolescent Vulnerability

Martha R. Burt, Janine M. Zweig, and John Roman

INTRODUCTION

Public policy often has been blind to adolescents, except when it has focused on aspects of their behavior that trouble their elders. Too often, policy makers limit their attention to artificially narrow and isolated aspects of youth behavior. They consider only health, or only criminal, or only educational issues. In addition, the payoff of youth vulnerability and our failure to ameliorate it are rarely addressed. The few existing treatments of the cost of adolescent risk behaviors have likewise focused on single behaviors (e.g., teen childbearing—Burt, 1985, 1986; Burt and Levy, 1987) or narrowly defined patterns (e.g., being a career criminal—Cohen, 1998). A just-released report identifying important future research issues related to youth (Millstein et al., 2000) does not even mention cost, either as the cost of outcomes to society or the cost of interventions or approaches to produce better outcomes. The absence of cost concerns is even more striking as Millstein and her colleagues review and summarize a decade of published documents that in their turn summarize and integrate research on adolescence and make recommendations for future research.

Compared to very young children and the elderly, adolescents suffer

Although the authors are affiliated with the Urban Institute, the views expressed in this chapter are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 73
Adolescent Risk and Vulnerability: Concepts and Measurement 4 Modeling the Payoffs of Interventions to Reduce Adolescent Vulnerability Martha R. Burt, Janine M. Zweig, and John Roman INTRODUCTION Public policy often has been blind to adolescents, except when it has focused on aspects of their behavior that trouble their elders. Too often, policy makers limit their attention to artificially narrow and isolated aspects of youth behavior. They consider only health, or only criminal, or only educational issues. In addition, the payoff of youth vulnerability and our failure to ameliorate it are rarely addressed. The few existing treatments of the cost of adolescent risk behaviors have likewise focused on single behaviors (e.g., teen childbearing—Burt, 1985, 1986; Burt and Levy, 1987) or narrowly defined patterns (e.g., being a career criminal—Cohen, 1998). A just-released report identifying important future research issues related to youth (Millstein et al., 2000) does not even mention cost, either as the cost of outcomes to society or the cost of interventions or approaches to produce better outcomes. The absence of cost concerns is even more striking as Millstein and her colleagues review and summarize a decade of published documents that in their turn summarize and integrate research on adolescence and make recommendations for future research. Compared to very young children and the elderly, adolescents suffer Although the authors are affiliated with the Urban Institute, the views expressed in this chapter are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders.

OCR for page 73
Adolescent Risk and Vulnerability: Concepts and Measurement from few conditions that will kill them while they are still young. The formation in adolescence of certain health habits with long-term negative consequences (such as smoking tobacco products, use of other addictive substances, or sexual activity without protection from STD and AIDS) often does not produce morbidity or mortality in adolescence itself. Rather the effects, and the payoffs, develop over a lifetime. Other behaviors such as school dropout, running away from home, or criminal involvement also exert their most powerful effects in adulthood. Thus, when societies face decisions about where to invest significant health and other supportive resources, programs for adolescents often receive short shrift. This is true despite the fact that after early infancy, adolescence is the period of greatest vulnerability, during which patterns and habits affecting a lifetime are established and solidified. In 1998, youth made up about one in every seven people in the U.S. population, whether the focus is on the younger end of the age spectrum (10–19 year olds were 14.3 percent) or the older end (15–24 year olds were 13.8 percent). These are the individuals on whom the future of this country rides. A strong argument can be made that we need all of our youth to develop into productive adults, with skills and attitudes ready to cope with twenty-first-century work, politics, and community and interpersonal relationships. The evidence suggests that for significant portions of our youth, seriously inadequate educational achievement, and life-threatening habits such as addictions, risky sexual behavior, involvement in crime and violence, and too-early childbearing foreclose the possibility that they will become contributing members of society. With respect to adolescents, the focus of attention is far too often on individual behavior, with far less attention being paid to context. But context is critical for understanding, and perhaps altering, the choices that youth make about their own behavior. For youth to make prosocial choices, it is essential that communities create increasingly broad and rewarding economic and social opportunities. There is an important interaction between economic opportunity and the readiness of today’s youth to take advantage of it. Without the realistic hope of getting ahead economically, there is little incentive for youth to invest in education or refrain from some of the less healthy, or less legal, habits they may acquire during adolescence. But without the expectation that there will be a qualified workforce to fill newly created jobs, many employers will send jobs overseas or fill them with people trained outside the United States, while the jobs that remain will be the least challenging, interesting, and rewarding ones. To the extent

OCR for page 73
Adolescent Risk and Vulnerability: Concepts and Measurement that the youth of today and tomorrow are not prepared for the future (and many are not), expectations for the country’s continued economic prosperity are open to question. We have choices to make. We can invest society’s resources in activities that will increase the odds that youth will become contributing members of society, or we can invest primarily in institutions such as health services or prisons designed only to compensate or protect society from the consequences of their negative behaviors. Given these choices, the payoffs from the former over the latter should make the policy choices clear. This paper is an exercise in designing an approach to illuminate the costs and opportunities of various policy choices with respect to investing in youth. Why We Need to Think About Payoffs (Costs and Benefits) Americans have a very strong belief in the efficacy of individual initiative and self-reliance. Far too often, and in too many arenas, this translates into policies that withhold support and investment in people until they fail, and then spend considerable sums on programs that try to protect society from the results or, on occasion, pick up the pieces. The earlier these policies are applied in people’s lives, the more global the ultimate effects. Failing to invest in securing productive futures for this nation’s most vulnerable youth has implications for everything from family formation to economic competitiveness. Yet public policy in this country related to people’s well-being rarely issues from considerations of “the big picture.” In part this is an inevitable aspect of how politics works in America, but in part it stems from lack of information, and information can sometimes make a difference to policy. To give one example, at the request of the (then) Center for Population Options, a research and advocacy organization, Burt (1985) developed a simple method that local jurisdictions could use to calculate the cost of first births to teenagers within their jurisdiction within a given year or for a given year’s birth cohort over 20 years. Many jurisdictions actually made these calculations and used them to lobby their legislative bodies for more resources to address the problem. One particularly telling example was a small rural jurisdiction in a conservative state, where it was very difficult to get any resources either for pregnancy prevention or to help teen mothers stay in school. After making the calculations for the 20-year projection, the jurisdiction realized that it was spending more than $1 million in welfare benefits for each and every birth cohort, without even knowing it and without

OCR for page 73
Adolescent Risk and Vulnerability: Concepts and Measurement helping anyone very much. The size of this inadvertent “investment” got the attention of local policy makers, and funding for more appropriate services followed. If we are able to create some viable models for estimating the payoffs of adolescent vulnerability, and compare them to investments in youth (always assuming that we can make the connection between the investment and desirable outcomes), we will be in a position to use these figures to influence policy. We do not want to make this endeavor seem too complicated, but we do not want to make it seem too simple either. During the past decades, a body of literature has been building to indicate the complexities of youth behavior patterns and the inadequacy of single-problem approaches to understanding risk and vulnerability (Catalano et al., 1999). Those complexities multiply when we begin to think about outcomes and associated payoffs, but only by considering the complexities are we likely to get within shooting range of a reasonable estimate of payoffs. The Approaches We Will Explore We will try to develop a hybrid approach to assessing payoffs of investing in youth that avoids the disadvantages of some classic economics formulations of cost-benefit analysis. We want to be able to identify the payoffs of youth risk behavior to the public purse, but we also want to capture the broader context that includes personal or private costs and benefits. The reasons for these preferences will be detailed later in this paper. Furthermore, we will examine the payoffs of patterns of youth risk behavior, rather than of a single type of risk behavior. The reasons for this approach should be obvious from the results of the past decades of research on youth risk behaviors and evaluations of programs taking a single-focus versus a holistic approach to promoting positive youth outcomes. Our approach involves modeling a conceptual framework containing three sets of transitional probabilities: (1) from antecedent risk factors to risk behavior patterns; (2) from risk behavior patterns to outcomes (pregnancy, addiction, suicide, jail, CEO of Fortune 500 company); (3) and from outcomes to payoffs (probability of using or contributing to public resources/well-being, private resources/well-being). The Structure of This Paper The remainder of this paper is structured to address the three compo-

OCR for page 73
Adolescent Risk and Vulnerability: Concepts and Measurement nents of our conceptual framework. The first component goes from risk/ vulnerability factors to risky behavior; that is, it should be able to model the transitional probabilities that certain behaviors or patterns of behavior will emerge, given the existence of certain antecedent conditions. We treat this component very lightly, as these issues have been the focus of a great deal of research. In addition, the paper by Blum, McNeely, and Nonnemaker in this volume summarizes these issues in sufficient detail. The second component goes from risk behaviors or patterns to outcomes, both positive and negative. We must determine the likelihood that any given behavior, repeated behavior, or pattern of behaviors will result in particular outcomes. Part of this task includes the important element of estimating co-occurrence or patterning of behaviors. This is essential because the synergies or interactions of certain behaviors in the presence of other behaviors may be more likely to produce costly consequences than if the focal behavior occurred in isolation. For instance, risky sexual behavior may lead to pregnancy, or to sexually transmitted diseases (STDs). Risky sexual behavior in combination with serious use of illegal drugs may add addiction, problems with a pregnancy, a child suffering the effects of fetal drug exposure, prison time for the mother, and a fractured family unit to the “simple” costs of pregnancy or treatment for STDs. Relatively little work of this type has been done to date, but some data sets exist that could be used to begin relevant analyses. The third component is even more challenging, and less explored, than the second one. That is to translate outcomes of risk behavior patterns into payoffs. Our presentation here will be almost totally speculative. It will cover the probability of using and/or contributing to public resources in various arenas (education, health, mental health, criminal justice, social services, cash benefits, and so on, as well as taxes paid, contributions to community well-being, becoming an employer of others, and other fanciful conceptions). It also will cover the probability of incurring private costs (e.g., costs of health insurance, income foregone) and/or reaping private benefits (e.g., earnings, long life, benefits to children of stable families). It will attempt to present models projecting over a person’s lifetime. It will attempt to meet various challenges such as “payoffs of adolescent risk behaviors to/for whom?” and “compared to what?” It will attempt to model ways to compare the cost of various investments that could be made in youth throughout their adolescence to their potential long-term effects on payoffs in adulthood. It will raise issues of who must make the decision to invest in adolescents versus who will incur the costs or reap the benefits of

OCR for page 73
Adolescent Risk and Vulnerability: Concepts and Measurement these investments in later years. It certainly will not succeed to everyone’s (anyone’s?) satisfaction, but it will be an interesting beginning. This Paper Is Hypothetical—Data Will Come Later Our task in this paper is to develop one or more frameworks for analyzing the payoffs of adolescent behavior and the outcomes that follow from it in adulthood. We are also to suggest the types of data we would need to gather if we want to estimate any of the models that we will suggest. We were not charged with actually doing any data analysis—just with thinking through and laying out what it would take to “do it right.” Readers may have their own ideas for modifying the models we present, or their own sources of data for beginning the work of estimating all or part of our models. If we succeed in stimulating a new spurt of activity modeling payoffs of investing in adolescents, this paper will have done its job. FROM VULNERABILITY FACTORS TO RISK BEHAVIORS Past research has identified a number of vulnerability factors that increase the likelihood that youth will participate in health risk behaviors. It has shown that many of the same vulnerability factors predict a variety of health risks and related outcomes, such as substance use, delinquency, violence, adolescent pregnancy, and dropping out of school (Catalano et al., 1999). Over the course of the past decades, researchers also have sought to identify protective factors that help prevent youth from taking risks. Two recent analyses have moved to the forefront of the discussion on predictors of adolescent risk-taking behavior. Using data from the National Longitudinal Study of Adolescent Health (Add Health), both Resnick and colleagues (1997) and Blum and colleagues (2000) found that demographic variables (race/ethnicity, family income, and family structure) are only weakly related to adolescent risk-taking behaviors such as substance use, risky sexual activity, and violence. Additionally, Resnick, Blum, and others have found that processes such as family connectedness, school connectedness, and time spent in structured activity work to reduce the amount of risky behavior among youth. Although the above research sheds light on predictors of risk-taking behaviors one at a time, it is not clear if the predictors hold when capturing the multidimensional nature of adolescent risk-taking behavior. Building on the seminal work of Jessor and Jessor (1977) on the co-occurrence of

OCR for page 73
Adolescent Risk and Vulnerability: Concepts and Measurement risk-taking behaviors, many researchers have documented links and patterns among various behaviors. These patterns of co-occurrence include aggression, substance use, and suicidal behavior (Garrison et al., 1993); substance use, sexual activity, and suicidal behavior (Burge et al., 1995); substance use and violence (Durkham et al., 1996); and substance use and sexual activity (Shrier et al., 1996). Jessor and colleagues (1977, 1991) speculated that youth risk taking comprises a single syndrome of problem behaviors, or as Elliot (1993) described it, a single health-compromising lifestyle. Pursuing this direction of inquiry further, Zweig et al. (2001a) decided to model the reality of adolescent risk taking. We attempted to capture the multidimensional nature of youth risk taking using Add Health data and cluster analysis. We found that youth participate in both health-enhancing lifestyles (Elliot, 1993) and a variety of different health-compromising lifestyles that we have called health risk profiles. We examined sexual activity, general alcohol use, binge drinking, cigarette use, marijuana use, other illicit drug use, fighting, and suicide for female and male students in grades 9 through 12. Four distinct profiles were identified for females and four for males (Figures 4-1 and 4-2). The four risk profiles for females included: (1) a low-risk, sexually active group (having used contraception during both their first and most recent sexual experiences, if sexually active); (2) a low-risk group, with higher levels of fighting and of suicidal thoughts and behaviors; (3) a moderate-risk group, with higher levels of substance use and risky sexual behavior; and (4) a high-risk group across all risk behaviors. The four risk profiles for males included: (1) a low-risk group across all behaviors; (2) a moderate-risk group with higher levels of alcohol use, binge drinking, cigarette use, and risky sexual behavior; (3) a moderate-risk group with higher levels of marijuana use and of suicidal thoughts and behaviors; and (4) a high-risk group with low levels of suicidal thoughts and behaviors. Once we identified adolescent health risk profiles, we too wanted to know about the vulnerability and protective factors related to each. Like our colleagues, we found that demographic factors such as age, race/ ethnicity, and family income did not distinguish the profiles in meaningful ways (Zweig et al., 2001b). Also like our colleagues, we found that other processes predicted differences in profiles, and we have been able to make clearer distinctions about what factors predict particular lifestyles. Youth in low-risk profiles and profiles distinguished by substance use and sexual activity reported higher levels of individual psychosocial adjustment, family

OCR for page 73
Adolescent Risk and Vulnerability: Concepts and Measurement FIGURE 4-1 Profiles of risk—Females grades 9-12. SOURCE: Zwieg, J. M., Lindberg, L. D., & McGinley, K. L. (2001). Used with permission of the Journal of Youth and Adolescence. FIGURE 4-2 Profiles of risk—Males grades 9-12. SOURCE: Zwieg, J. M., Lindberg, L. D., & McGinley, K. L. (2001). Used with permission of the Journal of Youth and Adolescence.

OCR for page 73
Adolescent Risk and Vulnerability: Concepts and Measurement connectedness, and school connectedness than students in high-risk profiles and profiles distinguished by suicidal thoughts and behaviors. The important message from our analysis is that teens in low-risk profiles and profiles distinguished by substance use and sexual activity are similar, and at relatively low risk—they consistently report lower levels of vulnerability factors and higher levels of protective factors than other teens. Some teens who have sex and use alcohol and tobacco have as few vulnerabilities and as many protective factors as teens who participate in little or no risk behavior. Teens in high-risk profiles and profiles distinguished by suicidal thoughts and behaviors are also similar—teens in both groups consistently report higher levels of vulnerability factors and lower levels of protective factors. Teens who are suicidal but do not report participating in any other risk behaviors are as vulnerable and unprotected as those who participate in all types of risk behaviors. FROM HEALTH RISK PATTERNS TO OUTCOMES Thus far we have discussed the evidence that youth participate in both health-enhancing and health-compromising lifestyles and that membership in groups based on these lifestyles can be predicted by vulnerability and protective factors operating in the lives of youth. Next we must establish the probability that these lifestyles will lead to particular outcomes and patterns of outcomes. To date, whenever we have thought about assessing the public burden of adolescent risk, it usually has been done with one risk behavior or one outcome in mind. Private payoffs largely have been ignored. But we know that youth participate in different lifestyles comprising various combinations of behaviors, some more risky than others. These health-compromising and health-enhancing lifestyles can lead to combinations of both negative and positive outcomes that can contribute to or help reduce the public burden or general social welfare outcomes of youth behavior. To understand the scope of outcomes youth may face as a result of their risk-taking behavior, we cannot examine one risk or one outcome at a time. Rather, we must keep their risk-taking patterns in mind, and attempt to link these to all possible related outcomes. By linking lifestyles to the many outcomes that may result, we will more realistically discuss adolescent risk taking, its outcomes, and its payoffs. So, how do we link adolescent lifestyles to outcomes, and thence to their associated payoffs? First we need to know what information exists (in the form of results from previous analyses or actual data that lend them-

OCR for page 73
Adolescent Risk and Vulnerability: Concepts and Measurement selves to the necessary analyses) that allows us to identify adolescent lifestyles and link these to possible outcomes associated with each. Then, once we establish relationships between lifestyles and outcomes, we can incorporate the known probabilities in our model estimating payoffs. National data sets may provide some of the answers when it comes to linking lifestyles and outcomes, but no one data set has all the necessary information. Some data sets can only be assessed for shorter term outcomes, while others can be assessed for both shorter and longer term outcomes depending on the length of the longitudinal study. Furthermore, some data sets are much richer with respect to some outcomes than to others (e.g., economic behavior versus sexual behavior versus criminal or violent behavior). Therefore, we will almost certainly have to use more than one data set to understand the full range of outcomes. This necessity leads in turn to the need to resolve a number of methodological issues. For example, different measures have been used across studies to assess adolescent risk-taking behavior, making it more or less difficult to model adolescent health-compromising and health-enhancing lifestyles. The lifestyles identified in one data set may not be comparable to those identified in another data set. In addition, when relying on older data sets to assess longer term outcomes, we must remember that the youth of interest were participating in risky behavior 20 years ago. The meaning of adolescent risk taking and its associated outcomes may have changed since then. More current data provide information about how risk behaviors have shifted over time. For example, recent trends in adolescent risk taking indicate decreases in some risk behaviors such as violence and sexual activity, and increases in others such as substance use (Boggess et al., 2000). Therefore, although we may be able to measure the same age groups, differences of cohort and time may make it difficult to compare results across data sets and tell a full story of the payoffs of adolescent risk (Baltes et al., 1977). Data options that may be relevant to the current effort are discussed in the following paragraphs. 1. National Longitudinal Study of Adolescent Health (Add Health: http://www.cpc.unc.edu/addhealth) Add Health was designed to examine adolescent physical, mental, emotional, and reproductive health. Add Health’s first wave of data collection was completed in 1994-95. That year, 90,000 youth completed in-school surveys about their background, friends, school life, school work

OCR for page 73
Adolescent Risk and Vulnerability: Concepts and Measurement and activities, and general health status. Of these youth, 21,000 also participated in an in-home survey about family and peer relationships, school environment, health risk behaviors (including sexual activity, violence, and substance use), psychosocial adjustment, physical health, and perceptions of risk. Wave II was completed in 1996, a year after Wave I. Wave III is currently in the works and will be collected in 2001, with youth now young adults approximately between the ages of 18 and 24. Add Health is an exceptional data set to identify lifestyles of youth risk taking, and indeed, we have already done this with Wave I data. Until Wave III is completed, however, little can be done to assess outcomes of these lifestyles given that Wave II was collected only one year after the first wave. The Wave III data are an excellent resource to help us understand the shorter term health-related outcomes of youth risk (such as teen pregnancy and STDs) and educational and work histories thus far. In addition, we may know about participants’ financial situations, health insurance, and use of public programs. Less will be known about participants’ criminal behavior and history, however, and we will also not know about the longer term outcomes of adolescent lifestyles given the length of the project thus far. 2. National Survey of Adolescent Males (NSAM: http://www.nichd.nih.gov) NSAM was designed to assess male adolescent risk taking and reproductive health. To date, it includes three waves of data collection, with Wave I completed in 1989 when males were 15 to 19 years old. Wave II was collected in 1990-91 and Wave III was collected in 1995. A new second cohort of males ages 15 to 19 were also added at Wave III. Participants were asked about their background, educational history and aspirations, sexual activity, substance use, attitudes about contraception and gender roles, and knowledge about sexual activity, contraception, and AIDS. Like Add Health, NSAM would be an appropriate data set to identify adolescent health-compromising and health-enhancing lifestyles, but also like Add Health, the participants were only followed through young adulthood, allowing assessment of only shorter term outcomes related to risk. In addition, use of social programs, violence, criminal behavior, employment, and suicide ideation are not identified as areas of focus for the study, so presumably we have less information on these issues. 3. National Longitudinal Survey of Youth (NLSY: http://www.bls.gov.nlsy) The NLSY began in 1979 to examine labor force participation and

OCR for page 73
Adolescent Risk and Vulnerability: Concepts and Measurement FIGURE 4-4 Modeling behaviors to negative outcomes, focusing on health risk effects and omitting effects of resiliency factors.

OCR for page 73
Adolescent Risk and Vulnerability: Concepts and Measurement FIGURE 4-5 Modeling behaviors to positive outcomes, showing direct effects of resiliency factors.

OCR for page 73
Adolescent Risk and Vulnerability: Concepts and Measurement driving, at the least) and pregnancy/STDs are also somewhat elevated. Interesting to us, as we tried to attach probabilities to Figure 4-5 for positive outcomes, is the relative lack of research documenting these, and we were forced to insert many question marks. The only fairly certain association is a negative one for school performance and educational attainment. With respect to the second profile illustrated in Figures 4-4 and 4-5, the strongest associations are for physical injury to self or others, with an equally strong expectation of current and continuing mental health problems. Associations of this profile with positive outcomes were fairly speculative, but we mostly expected them to be negative (compared to youth with low-risk profiles). We expected that this profile could experience a fair degree of lowered outcomes in the area of family relationships, and also might be somewhat lower on community involvement. We hope these profiles convey that behaviors occurring together in patterns may be expected to interact with each other to produce even more, or even less, of an outcome than would have occurred if one behavior occurred in isolation, as well as some outcomes that would not have occurred at all without both behaviors being present (e.g., babies born with fetal alcohol syndrome or crack addiction, in the case of the first profile). We did not include youth with the very highest risk profiles in these figures, basically because we could not fit in all of the very thick arrows we would have needed. However, we do expect that both boys and girls in these very high risk groups would exhibit very elevated levels of most of the negative outcomes and depressed levels of most of the positive outcomes. The important thing to note is that we are going from one pattern (for behaviors) to another pattern (for outcomes), rather than from single behaviors to single outcomes. With respect to the associations of health risk profiles with negative outcomes, space and layout on the page did not let us show in Figure 4-4 the moderating effects of resiliency factors (paths E in Figure 4-3), because we would have had to draw arrows from resiliency factors to every arrow in the figure. Nor did we show the direct effects of antecedents (path B in Figure 4-3). Many more complexities would have been introduced had we done so, such as the possibility that sexual activity and substance use might escalate to prostitution and homelessness in the presence of physical or sexual abuse in the home environment, or that strong attachments to adults with pro-social values might provide the motivation to avoid pregnancy and substance abuse. Figure 4-5 does show the moderating effects of resiliency factors for positive outcomes because we had enough room on the page to do so.

OCR for page 73
Adolescent Risk and Vulnerability: Concepts and Measurement Next we examine the types of payoffs that are most likely to be associated with particular outcomes (the final arrows in Figure 4-3). Table 4-3 shows the various positive and negative outcomes of our model as rows, and the various domains in which we can expect payoffs to occur as columns. Expectations for the intensity and direction of payoffs are indicated by plus and minus signs. Cells with a single minus sign indicate that we expect the outcome to produce net negative payoffs for that domain (e.g., pregnancy/teen childbearing/STDs in relation to family/community outcomes). Cells with a double minus sign indicate an expectation of strong negative payoffs. Conversely, cells with one or two plus signs indicate an expectation of positive payoffs. Cells without any sign indicate that we have no particular reason to expect unusual payoffs in that domain. Needless to say, Table 4-3 is vastly oversimplified. It is probably no exaggeration to say that at least 10,000 decisions would need to be made before we could attach real payoffs to real outcomes. First we would need to specify all the elements of each outcome, on the basis of at least some justifying evidence. Second, we would have to specify all of the different types of crime, health, education, and other payoff types and subtypes. Third, we would have to attach a value to each, again on the basis of some evidence. Fourth, we would have to determine the probability that some entity would actually incur the payoffs, given that the outcome pattern happened. This sounds seriously intimidating, but at some level it is certainly possible. Putting the Model Together with Interventions The last thing to depict in this paper is the various paths that would have to be estimated to test the payoffs of different models of intervention with youth. We started this paper considering what we would need to do to show that investing in youth has important benefits for society. Figure 4-6 provides a schematic diagram of every component in our model; basically, this is what we would have to estimate to achieve the demonstration we seek. Embedded in Figure 4-6 are two hypothetical “designs” for estimating payoffs. We spoke earlier of the traditional prevention approach and of the positive youth development approach, and specified in Table 4-2 how we expected payoffs to be distributed among the various recipients—youth, their community, the public sector, and the rest of society. One design, for an efficient (that is, an “indicated”) prevention model, is shown by the

OCR for page 73
Adolescent Risk and Vulnerability: Concepts and Measurement TABLE 4-3 Relationship of Outcomes to Payoff Domains   Payoff Domain Outcome Crime Education Employment Family/Community Health Other (Social Welfare) Negative Outcomes   Pregnancy/Teen Childbearing/STDs   —— — — —— — Injury/Death   — — — —— — Other Morbidity   — — — —   Addiction-AOD — — —   —— — Addiction-Cigarettes         ——   Mental Health — — — — —  

OCR for page 73
Adolescent Risk and Vulnerability: Concepts and Measurement Crime ——           Homelessness/Prostitution, etc. —     — — — Positive Outcomes   School Performance/School Attainment + + + + + +     Community Involvement +   + + +   + Family Relationships + + + +   + Attachment to Labor Force/Earnings +   + +   + NOTES:—= Negative payoffs/costs within a particular domain; + = Positive payoffs/benefits within a particular domain. Two signs (for example—) indicate a stronger relationship than a single sign. STDs = Sexually Transmitted Diseases; AOD = Alcohol and Other Drugs.

OCR for page 73
Adolescent Risk and Vulnerability: Concepts and Measurement FIGURE 4-6 Alternative models of intervention and their implications for calculating payoffs. NOTES: A = Point of intervention and payoff goals of typical “prevention” (tertiary attention/indicated intervention) program—reduce association between health risk profiles and negative outcomes, and reduce associated public costs. (Gray shaded boxes and the paths between.) B = Points of intervention and payoff goals of youth development approach—increase resiliency factors, reduce less healthy risk profiles, increase positive as well as reduce negative outcomes, and reduce negative and increase positive payoffs for youth themselves, their communities, and the rest of society, as well as reducing public costs.

OCR for page 73
Adolescent Risk and Vulnerability: Concepts and Measurement shaded boxes in Figure 4-6 and the two paths between them (labeled A). The direction of effect is shown by the signs, indicating that this prevention model tries to reduce the association between health risk profiles and negative outcomes, and thereby reduce the public costs associated with the negative outcomes. To see whether this approach “pays off,” one would add up all the costs of the intervention itself, and weigh these against the net value of the payoffs to the various sectors that could benefit or be harmed by the outcomes. The second design embedded in Figure 4-6, depicting a positive youth development approach, includes the same two pathways as for the indicated prevention approach, but also encompasses many other pathways and payoff recipients. Typical efforts of these programs start early and try to affect resiliency factors, behaviors, attitudes, relationships, and competencies leading to positive outcomes as well as reducing negative ones. The paths labeled “B” symbolize the goals of these programs—to increase payoffs for youth, communities, and the rest of society through creation of more positive outcomes, as well as to reduce public costs by reducing negative outcomes. In theory, to see whether this approach pays off, we follow the same tactics as we did for the indicated program. But obviously we have much more to identify, estimate, and calculate to achieve a full accounting of the payoffs of the second approach. The motivation to do so is that the payoffs potentially include much that is positive for communities and for society as a whole. IMPLICATIONS—“WHERE TO NEXT?” The “task” of justifying investment in youth, now that it is all laid out, seems quite enormous. But it also seems exciting, at least to the authors. Even thinking through what it would take, as skeletally as we have done it here, prompted many new thoughts and forced us to reconsider some ways we had thought about these issues before. It is important to realize that although we have developed the model in a mostly linear fashion, it does not have to be researched that way. Researchers can take some of the newer pieces and work on them simultaneously. Thus we can be using existing databases to develop increasingly sophisticated analyses of associations between patterns of behavior and patterns of outcomes, at the same time that we are assembling existing literature to document the costs of various outcomes to different sectors and the probability that various outcomes will indeed lead to those costs. And we

OCR for page 73
Adolescent Risk and Vulnerability: Concepts and Measurement can think about and try to collect new data that we will need to turn these models into reality. In addition, we can be doing more thinking about how to model the payoffs from different types of policy action. In the models presented here, we considered only “programs” involving fairly intensive face-to-face interactions among youth and others, including program staff, teachers, families, and others. We did not pay any attention to government actions such as pricing policies (raising the tax on cigarettes or alcohol, for example, as a deterrent to use). Nor did we consider the effects that changes in eligibility for benefit programs, such as the change from Aid to Families with Dependent Children (AFDC) to Temporary Assistance for Needy Families (TANF), might have on teen decision making about sexual behavior. Nor did we examine proposed “single bullet” solutions to certain problems such as “testing” (students, teachers, or both), “vouchers,” or reducing school class size. In part, we have not done so because we believe the findings of decades that making a difference for at-risk youth means major investments in fairly complicated, intensive, enduring interventions. We don’t think there are “single bullets.” We also think it is quite difficult to take a very complex policy change such as federal welfare reform and attempt to articulate its effects on a single behavioral domain of a small part of its target population. Also, many such policies have a single focus (e.g., reduce teen smoking). Although this is an important goal, it is not likely to change the lives of the youth who most need help, and we chose to concentrate on programs with a chance of doing that. But others may choose to model the payoffs of these types of policy changes, and such modeling efforts are sure to advance the entire enterprise of estimating payoffs, which can only be good. REFERENCES Baltes, P. B., Reese, H. W., & Nesselroade, J. R. (1977). Life-span developmental psychology: Introduction to research methods. Hillsdale, NJ: Lawrence Erlbaum Associates. Blum, R. W., Beuhring, T., Shew, M. L., Bearinger, L. H., Sieving, R. E., & Resnick, M. D. (2000). The effects of race/ethnicity, income, and family structure on adolescent risk behaviors. American Journal of Public Health, 90(12), 1879-1884. Boggess, S., Lindberg, L. D., & Porter, L. (2000). Changes in risk-taking among high school students, 1991-1997: Evidence from the Youth Risk Behavior Surveys. In Trends in well-being of America’s children and youth 1999 (pp. 475-488). Washington, DC: Department of Health and Human Services. Burge, V., Felts, M., Chenier, T., & Parillo, A. V. (1995). Drug use, sexual activity, and

OCR for page 73
Adolescent Risk and Vulnerability: Concepts and Measurement suicidal behavior in U.S. high school students. Journal of School Health, 65(6), 222-227. Burt, M. R. (1985). Teenage pregnancy: How much does it cost? Washington, DC: Center for Policy Options. Burt, M. R. (1986). Estimating the public costs of teenage childbearing. Family Planning Perspectives, 18(5), 221-226. Burt, M. R., & Levy, F. (1987). Estimates of public costs for teenage childbearing: A review of recent studies and estimates of 1985 public costs. In S. L. Hofferth and C. Hayes (Eds.), Risking the future: Adolescent sexuality, pregnancy and childbearing, Vol. II (pp. 264-294). Committee on Child Development Research and Public Policy, Commission on Behavioral and Social Sciences and Education. Washington, DC: National Academy Press. Catalano, R. F., Berglund, M. L., Ryan, J. A. M., Lonczak, H. S., & Hawkins, J. D. (1999). Positive youth development in the United States: Research findings on evaluations of positive youth development programs. Seattle, WA: Social Development Research Group. Cohen, M. A. (1998). The monetary value of saving a high risk youth. Journal of Quantitative Criminology, 14(1), 5-33. Cohen, M. A. (2000). Measuring the costs and benefits of crime and justice. In Measurement and analysis of crime and justice, Volume 4: Criminal justice 2000. NCJ 182410. Washington, DC: National Institute of Justice. Durkham, C. P., Byrd, R. S., Auinger, P., & Weitzman, M. (1996). Illicit substance use, gender, and the risk of violent behavior among adolescents. Archives of Pediatric and Adolescent Medicine, 150, 797-801. Elliot, D. S. (1993). Health-enhancing and health-compromising lifestyles. In S. G. Millstein, A. C. Petersen, & E. O. Nightingale (Eds.), Promoting the health of adolescents: New directions for the twenty-first century (pp. 112-145). New York, NY: Oxford University Press. Garrison, C. Z., McKeown, R. E., Valois, R. F., & Vincent, M. L. (1993). Aggression, substance use, and suicidal behaviors in high school students. American Journal of Public Health, 83(2), 179-184. Jessor, R. (1991). Risk behaviors in adolescence: A psychosocial framework for understanding and action. Journal of Adolescent Health, 12, 597-605. Jessor, R., & Jessor. S. (1977). Problem behavior and psychological development: A longitudinal study of youth. San Diego, CA: Academic Press. Miller, T. R., Cohen, M. A., & Wiersema, B. (1996). Victim costs and consequences: A new look? Washington, DC: Department of Justice, Office of Justice Programs, National Institute of Justice. Millstein, S. G., Ozer, E. J., Ozer, E. M., Brindis, C. D., Knopf, D. K., & Irwin, C. E., Jr. (2000). Research priorities in adolescent health: An analysis and synthesis of research recommendations, executive summary. San Francisco: University of California, National Adolescent Health Information Center. Resnick, M. D., Bearman, P. S., Blum, R. W., Bauman, K. E., Harris, K. M., Jones, J., Tabor, J., Beuhring, T., Sieving, R. E., Shew, M., Ireland, M., Bearinger, L. H., & Udry, J. R. (1997). Protecting adolescents from harm: Findings from the National Longitudinal Study of Adolescent Health. Journal of the American Medical Association, 278(10), 823-833.

OCR for page 73
Adolescent Risk and Vulnerability: Concepts and Measurement Roman, J. Woodard, J., Harrell, A., & Riggs, S. (1998). A methodology for measuring costs and benefits of court-based drug intervention programs using findings from experimental and quasi-experimental evaluations. Washington, DC: Urban Institute. Shrier, L. A., Emans, S. J., Woods, E. R., & DuRant, R. H. (1996). The association of sexual risk behaviors and problem drug behaviors in high school students. Journal of Adolescent Health, 20, 377-383. Zweig, J. M., Lindberg, L. D., & McGinley, K. L. (2001). Adolescent health risk profiles: The co-occurrence of health risks among females and males. Journal of Youth and Adolescence, 30(6). Zweig, J. M., Phillips, S. D., & Lindberg, L. D. (2001). Predicting adolescent profiles of risk: Looking beyond demographics. Washington, DC: Urban Institute. Paper prepared for the Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation.