5
Adolescent Vulnerability: Measurement and Priority Setting

Baruch Fischhoff and Henry Willis

INTRODUCTION

Adolescents face many threats to their health, safety, and well-being. Some are shared by their society as a whole (e.g., war, many diseases, crime). Others are unique to, or at least accentuated by, teens’ transitions to arenas beyond the control of their guardians. Many adults devote much of their lives to reducing these vulnerabilities. There are school, community, and religious programs. There are medical screening, treatment, and educational efforts. There are lectures, remonstrations, and rescues by parents. There are special laws governing adolescent driving and status offenses. There are summits and conferences, some with teen representation, some without.

Teens often are described as living in a fog of exaggerated personal invulnerability (Millstein and Halpern-Felsher, this volume; Quadrel et al., 1993). However, both the scientific evidence and direct discussion show teens as having many legitimate concerns on their minds (Blum et al., this volume; Fischhoff et al., 1998, 2000). They wonder if and how they’re going to get through this stage of their lives, with the world that they hope for reasonably intact. Chronic diseases are one part of that burden, especially when they induce moments of legitimate panic, like diabetes or asthma. Violence is another part, especially when teens feel as though they never know which minor incident (or sideways glance) is going to spin out of control. Fear about the continuity of the larger world is yet another part of the burden. It might weigh especially hard on teens attuned to signs of



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 109
Adolescent Risk and Vulnerability: Concepts and Measurement 5 Adolescent Vulnerability: Measurement and Priority Setting Baruch Fischhoff and Henry Willis INTRODUCTION Adolescents face many threats to their health, safety, and well-being. Some are shared by their society as a whole (e.g., war, many diseases, crime). Others are unique to, or at least accentuated by, teens’ transitions to arenas beyond the control of their guardians. Many adults devote much of their lives to reducing these vulnerabilities. There are school, community, and religious programs. There are medical screening, treatment, and educational efforts. There are lectures, remonstrations, and rescues by parents. There are special laws governing adolescent driving and status offenses. There are summits and conferences, some with teen representation, some without. Teens often are described as living in a fog of exaggerated personal invulnerability (Millstein and Halpern-Felsher, this volume; Quadrel et al., 1993). However, both the scientific evidence and direct discussion show teens as having many legitimate concerns on their minds (Blum et al., this volume; Fischhoff et al., 1998, 2000). They wonder if and how they’re going to get through this stage of their lives, with the world that they hope for reasonably intact. Chronic diseases are one part of that burden, especially when they induce moments of legitimate panic, like diabetes or asthma. Violence is another part, especially when teens feel as though they never know which minor incident (or sideways glance) is going to spin out of control. Fear about the continuity of the larger world is yet another part of the burden. It might weigh especially hard on teens attuned to signs of

OCR for page 109
Adolescent Risk and Vulnerability: Concepts and Measurement eroding faith in government, assaults on the natural world (and on animals, with which many young people feel a special affinity), turmoil in racial relations, or growing income inequality. Even with the recent economic boom for some, many teens must worry about having a decent career (not to mention a meaningful one). These external concerns notwithstanding, teens obviously do not always act in ways that serve their own best interests, even in terms of the goals they set for themselves (which need not correspond to the goals that adults set for them). Worrying about life in general is not incompatible, with teens sometimes underestimating the risks posed by particular behaviors (e.g., unsafe sex, drinking and driving). Nor need teens’ critical decisions be driven entirely by calm deliberation. Of course, adults, too, often have exaggerated feelings of control over life events and, occasionally, let emotion carry them away (Loewenstein, 1996; Weinstein, 1987). However, they may face a lower rate of fateful decisions than do young people, who are trying to set up their lives—including how they will deal with work, drugs, driving, drinking, and intimacy, among other things. Thus, teens themselves create risks that compound those that the world imposes on them. THE NEED FOR INDICATORS To deal effectively with these vulnerabilities, teens and adults need to know how big the threats are and how much can be done about them. That means knowing how big the overall burden of adolescent vulnerability is, in order to decide what personal and societal resources to devote to threats to adolescents (relative to other priorities). It means knowing the relative size of specific threats, and of the expected costs and benefits of opportunities for risk reduction, in order to identify the “best buys” in risk reduction. Where these questions cannot be answered confidently, better research is needed, for each link in the analytical chain. Systematic uncertainty reduction is the goal of research focused on patterns of problem behavior and predisposing conditions, creating either vulnerability or resilience (Blum et al., this volume; Jessor et al., 1991). Where even the best buys are not very attractive, then social investments (including research) are needed to make better options available for youth. The shift from problem-focused interventions to positive youth development ones is a response to feelings of fundamental inadequacy in what we offer young people (Burt et al., this volume). A sweeping change in

OCR for page 109
Adolescent Risk and Vulnerability: Concepts and Measurement policy requires a comprehensive look at the evidence, expressed in some common and relevant terms. Realizing this, both national and international bodies have called for routine reporting of comparable statistics on critical indicators of youth welfare (e.g., Department of Health and Human Services, 2000; Federal Interagency Forum on Child and Family Statistics, 1997; United Nations, 1989). Suitably chosen indicators provide targets for social action and allow tracking of changes over time. Identifying the critical indicators is also a necessary condition for communications focused on the facts the teens, adults, and policy makers most need to know (Fischhoff, 2000; Millstein and Halpern-Felsher, this volume). Without such analysis, people may be denied guidance for effective action. They may have their time and trust wasted by streams of irrelevant communications. They may be faulted for failing to know facts that were hardly worth knowing, yet found their way onto someone’s improvised test of lay understanding. The resulting disrespect undermines respect for citizens and contributes to their disenfranchisement. It perpetuates a vicious circle, leading citizens to mistrust these dismissive experts, who fail to provide viable solutions or even needed information. However, even the best data alone do not set priorities among threats to adolescents (or the natural environment or economic opportunity or anything else). Those priorities require value judgments regarding the relative importance of different outcomes. For example, Burt et al. (this volume) raise a not-so-hypothetical choice between two competing programs. One, focused on the most serious problem behaviors, could prevent “several of those ‘worst youth’ from fulfilling the worst, most costly, expectations for the outcomes of their behavior.” The other, focused on positive youth development, could prevent many less challenged youth from failing to fulfill their potential (“graduate from high school, go on to college or into the labor market, and lead productive lives”). In a world of finite resources, such choices are inevitable. They face not only agencies with limited budgets, but also parents with limited time, energy, and interpersonal credibility (with their offspring). Parents must decide whether to focus on their teens’ driving, drinking, diet, drugs, exercise, hygiene, studies, friends, sports, volunteering, moods, allergies, or physical safety, among other things. Within options potentially under their control, parents, too, must decide whether to invest in problem-focused interventions (e.g., grounding, curfews, driver education) or youth development ones (e.g., home schooling, family activities, religion).

OCR for page 109
Adolescent Risk and Vulnerability: Concepts and Measurement Overview The choices that policy makers and parents make or advocate reflect some amalgam of their values (about what matters) and beliefs (about what works). This paper casts these youth-specific choices in the general terms of priority-setting research and practice. One goal of these general approaches is increasing the expected value of invested resources. A second is clarifying the roles of social policy and social science in decision making, both for choices that have become norms and for new proposals. A third goal is revealing the value assumptions embedded in ostensibly objective analyses, clarifying the extent to which their conclusions are predetermined by their framing. For example, analyses focused on problematic end states (e.g., risk behaviors, adverse health outcomes) can divert attention from common sources, which contribute to multiple end states without being the primary determinant of any (e.g., low literacy, low birthweight). End-state analyses also divert attention from any value that programs have, independent of their effects on risk outcomes, such as making a social statement or contributing to those who implement them. Abstinence programs and Drug Abuse Resistance Education (D.A.R.E.), for example, might be rationally justified on those grounds, even if they had little direct effect on teens’ sexuality or drug use. Whether they should be depends on what one values. The next section, “Structuring Prioritization,” introduces some general concepts and nomenclature. The following section, “Social Mechanisms for Priority Setting,” contrasts two general approaches to determining priorities, differing in how explicitly they address value issues. The next section, “Deliberative Mechanisms for Priority Setting,” considers ways to determine the relevant values, with particular reference to analogous processes developed for setting environment priorities, over the past generation. The “Conclusion” speculates on the circumstances under which deliberate prioritization might and should occur. STRUCTURING PRIORITIZATION Trying to Separate Facts and Values Implicitly or explicitly, any policy regarding adolescent welfare embodies some notion of the overall burden that teens bear and its various expressions. These notions are reflected in the overall resources that teen issues receive and their allocation across problems. Pursued deliberately, the

OCR for page 109
Adolescent Risk and Vulnerability: Concepts and Measurement risk-assessment process has two stages: (1) characterizing the set of relevant adolescent vulnerabilities and (2) deciding what importance (or “weight”) to assign to each threat (see Chapter 5 Annex). The first stage is largely a matter of scientific fact, the second largely a matter of values. This fact/value distinction was central to the National Research Council’s (1983) “red book,” a founding document of risk assessment. Research and experience have shown life and analysis to be more complicated than this seemingly tidy separation suggests (e.g., Crouch and Wilson, 1981; Fischhoff et al., 1981; Institute of Medicine, 1998, 1999; National Research Council, 1996). Nonetheless, it is a point of departure for translating adolescent concerns into risk-based terms. These terms may have value in their own right, as a way of clarifying the structure of choices (complementing comprehensive analyses, such as Blum et al., this volume, and Burt et al., this volume). They may also help to make the case for youth when health and policy debates are cast in risk terms (as may happen increasingly). In the first stage, conventional scientific procedures are used to estimate the impacts on teens associated with different conditions. The application (and review) of these procedures should follow accepted scientific practice. However, doing so inevitably requires making value-laden assumptions, when the terms of the research are specified and its results are interpreted. These assumptions need to be determined explicitly, lest the values be hidden under a guise of analytic objectivity, or buried even more deeply in priorities arising from unstructured group processes or individual ruminations. The formalisms of risk assessment are intended to accomplish this task by making all steps in the prioritization process explicit and subject to external review. Nonetheless, any procedure, formal or otherwise, affords an advantage to those having greater fluency in its application. Indeed, much of the opposition to risk-based decision making in other areas reflects a fear that the promise of openness will not be realized. Rather, a new cadre of technical specialists will interject themselves in the process. Risk analyses can, in principle, consider a broad set of considerations without the sometimes-controversial monetization required by economic analyses (the primary current form of integrative approach). However, that promise will not be realized if the analyses are impenetrable to nonspecialists. One hope of this exposition is to clarify the assumptions made in prioritization, however it is accomplished.

OCR for page 109
Adolescent Risk and Vulnerability: Concepts and Measurement What Might Matter? The first of those assumptions is which things to consider. Box 5-1 shows three widely distributed sets of measures, translated from the originals so that all indicators are formulated negatively. The first list, from Healthy People 2010, has primarily health effects and (fairly proximal) predisposing conditions. The former are relatively uncontroversial, as outcomes that any society would want to avoid—even if there are disagreements about the completeness of the set and the weight to assign its members. The latter are more problematic. These conditions could be justified as indicators because they lead to adverse outcomes, a scientific claim. If those outcomes are also on the first list, then including the predisposing conditions would represent double counting. On the other hand, these conditions might efficiently represent a suite of concerns that are hard to assess directly (e.g., the variety of respiratory effects associated with airborne particulates). If so, then they might both avoid double counting and draw needed attention to problems with diffuse effects. However, placing a predisposing condition on the list also may reflect a value judgment, in the sense of its being considered bad, regardless of any associated health effects. For example, “irresponsible” sexual behavior may be treated as offensive, even if it does not lead to sexually transmitted diseases or undesired pregnancies. Such values should be reflected in the weights assigned to the different measures. Continuing the example, irresponsible sex should receive extra weight from individuals who are offended by the act, as well as being worried about the health outcomes it can cause.1 Thus, even this simple list could reflect rather different rationales. The reference document (Department of Health and Human Services, 2000) describes the extensive consultation process that led to selecting these indicators (11,000 public comments are still available at http://www.health.gov/healthypeople/), as well as the comprehensiveness of its view (467 objectives, organized into 28 focus areas). This very sweep led to a search for leading indicators that would focus attention. That selection process was guided by the indicators’ “ability to motivate action, the availability of data to mea- 1   Depending on the intent of the list’s compilers, everything but violence and injury could be considered a predisposing condition, in the sense of increasing the risk of some health problem. Indeed, even these two entries could serve that role, as when violent injuries (e.g., sexual assault) precipitate mental health problems.

OCR for page 109
Adolescent Risk and Vulnerability: Concepts and Measurement sure their progress, and their relevance as broad public health issues” (p. 24). Thus, the task force considered both science (what will work) and values (what matters). The report does not say how to resolve conflicts when initiatives directed at different problems compete for limited funds. Being on the list is, therefore, necessary, but not sufficient, for securing resources. The document assigns a “key role [to] community partnerships” for setting actual priorities (and implementing them) (DHHS, 2000, p. 4). However, limited guidance is provided for how such partnerships are to reach those priorities. As a result, prioritization is left to group (or political) processes: who gets to the table; who controls the agenda; who summarizes the proceedings. Stopping at this point may be entirely appropriate for these topics and the role of a federal agency. However, it leaves the process incomplete. Some of the approaches described here may be useful to those empowered to complete the work. Deliberately Embedding Values in a Method One place in which Healthy People 2010 does attempt to direct the process is in measuring those outcomes that a prioritizing group decides to value. It makes “eliminate health disparities” one of its two overarching goals, on a par with “increase the quality and years of healthy life.” It supports that focus by representing disparities in some of its measures (e.g., access to health care among different populations). Aggregate measures do not distinguish who suffers from a problem or benefits from a solution. Arguably, a life is a life and a cough is a cough, regardless of who suffers. However, ethical cases have been made for various forms of differential weighting. One common proposal assigns added weight to improvements benefiting individuals exposed to risks involuntarily (Lowrance, 1975; Starr, 1969). Those individuals might have been born with a problem or have had no political or economic influence over the conditions that created it. Involuntarily assumed risks also may have fewer compensating benefits (compared to risks that people chose to bring on themselves). Weighting involuntary risks more heavily provides a way to address such inequities. It is also possible to value the people affected by risks differentially because of who they are, rather than what they have done—or have had done to them. Some such weighting inevitably is embedded in the procedures of any priority scheme. For example, mortality risk may be measured in terms of probability of death from each source being considered, or in

OCR for page 109
Adolescent Risk and Vulnerability: Concepts and Measurement BOX 5-1 Alternative Indicators of Adolescent Vulnerability Healthy People 2010: Leading Health Indicators (DHHS, 2000) Outcomes Tobacco use Substance abuse Mental health problems Injury and violence Predisposing Conditions Overweight and obesity Physical inactivity Irresponsible sexual behavior Environmental pollution Lack of immunization Limited access to health care America’s Children Outcomes Poor health Chronic health conditions limiting activity Mortality Child bearing Cigarette smoking Alcohol use Substance abuse Victim of violent crime Abuse and neglect Predisposing Conditions Poverty Food insecurity terms of lost life expectancy arising from those deaths. Considering the number of years lost with each death puts a premium on deaths among young people. Using it focuses attention on threats that disproportionately affect them, such as accidents, relative to diseases of the aged, such as arte-

OCR for page 109
Adolescent Risk and Vulnerability: Concepts and Measurement Housing problems Parental employment insecurity Lacking health insurance Difficulty speaking English Lacking math and reading proficiency Neither working nor in school UN Convention on the Rights of the Child Outcomes Nondiscrimination Survival and development Name and nationality Preservation of identity Contact with parents Freedom of expression, thought, conscience, religion, and association Privacy Health Standard of living adequate for physical, mental, spiritual, moral, and social development Protection from drug abuse, sexual exploitation, abduction, torture, and armed conflicts Leisure Predisposing Conditions Decisions made in the best interests of the child Access to information Special protection for refugees, disabled, adopted, without families, and minorities Health and social services Education developing personality, talents, and mental and physical abilities Age-appropriate justice, promoting sense of dignity and worth riosclerosis. Of course, focusing on deaths raises the profile of risks such as auto accidents relative to ones that cause mostly morbidity and misery (such as drugs). Whatever unit is used, it represents a value (even if that choice is made unwittingly).

OCR for page 109
Adolescent Risk and Vulnerability: Concepts and Measurement The second list, created by the Federal Interagency Forum on Child and Family Statistics (1997), also includes both outcomes and predisposing conditions. Compared with Healthy People 2010, it has a larger set of health outcomes, while still not subsuming the previous list (e.g., mental health problems, unintentional injury). One could ask whether the compilers of the first list were not interested in activities limited by chronic health conditions (a value question) or believed that these outcomes were predicted from others in their list (a scientific question). As with the first list, the Predisposing Conditions also could be viewed as negative ends in their own right. Were that the case, then the second list would represent a broader definition of the conditions that our society owes its citizens. If not, then including these additional conditions reflects an alternative view of the facts regarding predisposing causes, with a larger role assigned to social and economic factors, such an employment and housing status.2 Evidence-Driven Criteria The third set of criteria is taken from an international document, the United Nations (UN) Convention on the Rights of the Child (signed by all member countries except Somalia, which lacks a central government, and the United States). One obligation of signing countries is to compile statistics reporting on the state of their children, reflecting these concerns. Perhaps the most striking difference between this list and its predecessors is the emphasis on political rights. In Box 5-1, some of these are cast as outcomes, others as predisposing conditions (a distinction that we imposed on the Convention’s list). In the former role, these criteria are ends in themselves; in the latter, they are means to other ends. Reasonable individuals could disagree about these roles, and about the kinds of evidence needed to evaluate the importance of each. For example, one might consider any discrimination to be wrong or only discrimination that could be linked to end states, such as survival and development. In the latter case, the weight assigned to discrimination would depend on the strength of the demonstrated connection (as determined, perhaps, by the sort of root-cause analyses demonstrated by Blum et al., this volume, and Burt et al., this volume). 2   Their omission from the first list could reflect a value judgment, to the effect that these conditions are predictors of the health outcomes, but not ones that should concern anyone other than the individuals involved.

OCR for page 109
Adolescent Risk and Vulnerability: Concepts and Measurement Neglecting discrimination, in the absence of such evidence, need not reflect indifference to this aspect of young people’s fate. Rather, the ties with direct effects may seem sufficiently strong that it is better to measure them than discrimination. Doing so avoids double counting (both causes and effects). Effects may be more observable and less controversial. One also may feel that discrimination is a separate effect, but belongs to some other jurisdiction, and hence is not an aspect of adolescent health and safety. The impact of that claim depends on whether the other jurisdiction actually assumes responsibility for assessing, and addressing, discrimination—and on whether it is, in fact, a problem. The UN Convention criteria are meant to serve the interests of young people in widely varying circumstances around the world. Problems that are egregious in some countries may be minor in other, more fortunate ones (e.g., in which few children are denied names or nationalities). At least two of the UN criteria should discourage the adoption of measures that obscure disparities when looking at overall performance. One is discrimination, which might predict such disparities. The second is special protection for several inherently vulnerable populations. Without those protections, one might presume variation in the achievement of other criteria, even without assessing it. Another apparent difference in the UN Convention criteria is the inclusion of such “positive” criteria such as education developing personality, talents, and mental and physical abilities. Like nondiscrimination, these criteria might be treated as ends or means. A society may be held to fail its children, if they fail to achieve their full potential. Or, the lack of effective investment in development may provide a predictor of other valued criteria. Like nondiscrimination, such education may be ignored because it belongs to another jurisdiction or because it is too hard to measure. Doing so requires an explicit theory for how various kinds and quantities of education achieve desired results. Where such measures of positive contribution are lacking, one might have to revert to the deficit model underlying most criteria. Criteria for Criteria The empirical constraints on measurement feature centrally in the selection rules described as guiding the choice of measures in America’s Children:

OCR for page 109
Adolescent Risk and Vulnerability: Concepts and Measurement minders distinguish these issues, asking participants to focus on importance. These instructions acknowledge that large risks may be neglected if nothing can be done about them. However, recognizing that fact is important, especially if it reveals too little investment in developing solutions. Conversely, small risks may be addressed if there are efficient solutions. However, that might mean they have received disproportionate attention in the past. When participants raise issues related to solutions, those are duly noted, both to acknowledge their eventual importance and to help participants make the conceptual distinction. Thus, the procedure allows participants to triangulate group and individual perspectives, as well as holistic and analytic ones. It also allows policy makers to use results in different ways. They can take initial values or concluding ones, group values or individual ones (collected in private). Policy makers can consider the change between initial and final values, individuals’ agreement with other group members, the degree of consensus on particular risks (in absolute terms or relative to the general level of consensus), and the coherence between holistic and analytic values. That interpretation should depend on the circumstances. For example, a group’s consensus may mean little unless its membership has some policy significance (e.g., an identifiable interest group, accustomed to resolving such issues together). Otherwise, it was just a vehicle for exposing individuals to diverse views. In conclusion, participants evaluate the process, including how well they communicated their views, as measures of its success (and legitimacy). The summary sheet ranks the risks by individual attributes. Although presented as effort saving, these rankings also show simple policies that participants could choose to adopt. One also could present rankings that reflect other, more complex principles, saving the more complex mental arithmetic that each requires. Those principles might be derived from the professional literature, ethical analyses, citizen interviews, or government regulations. For example, they might present the estimated (public or private) economic burden of each risk (to the extent that it can be calculated). Presenting them reduces the risk of participants missing perspectives that they would value or executing them poorly. It increases the risk of biasing expressed preferences, if the offerings are unbalanced. CONCLUSION Although this chapter makes the case for setting priorities systematically, it also shows the challenges that such exercises face. Recognizing these

OCR for page 109
Adolescent Risk and Vulnerability: Concepts and Measurement challenges and possible ways to address them should improve the process. However, one still should ask whether the best possible systematic prioritization is advisable. It could fail a cost-effectiveness test, in the sense of being a poorer investment of management energy than the best possible systemic prioritization (focusing intently on whatever risks happen to draw one’s attention). It could fail a cost-benefit test, in the sense of leaving one worse off than without any systematic analysis. Many factors affect the relative efficacy of spreading a given amount of decision-making resources over the broad set of risks (ensuring that each gains some attention) or focusing it on the few risks that seize public (or agency) attention: How well is the overall world of risks understood? If relatively few risks have drawn any concerted attention, then it is more likely that resources have been misallocated, and a systematic review will be informative. How much can be learned from a relatively quick look at individual risks? If a serious examination is required to learn very much, then it is harder to justify a broad review. How likely is it that some risks have been systematically over- or underestimated (e.g., due to flawed reporting or analytical methods that emphasize particular concerns, perhaps ones that are quantified most easily)? Such suspicions increase the expected value of looking hard at those specific risks, rather than assuming that things are generally in order. How much precision is needed to move from risk ranking to option ranking? If regulatory constraints or political inertia require strong evidence, then focusing on specific risks becomes essential—even if a broader look might show that they are not the most important targets for that focus. How are risks prioritized—by a best guess or by a worst case estimate of their magnitude? A broad look might do more to shift the tails than the central tendencies of probability distributions over possible risk levels. Bendor (1995) and Long and Fischhoff (2000) offer formal models for characterizing particular situations and simulating the expected yield of different strategies for prioritizing their risks. These models reflect concerns about the limits to analysis identified by Lindblom (1959), Simon (1957), and others. Even without running simulations, thinking about the formal properties of these situations should clarify what one wants, and can hope to get, from them. That assessment can be performed for the yield from

OCR for page 109
Adolescent Risk and Vulnerability: Concepts and Measurement both conventional procedures and more innovative ones. For example, our risk-ranking procedure is intended to increase the feasibility of systematic evaluation by using the time and energy of risk rankers more efficiently. Whether used on many risks or a few, such a procedure should increase the accountability of rankers by showing what evidence and factors have been considered (even if the integrative decision rule is embedded in their holistic judgments). If prioritization means anything, it should be capable of changing resource allocations. Individuals concerned with teens’ overall welfare should welcome an improvement in their ability to track the problems faced by teens (as a whole and by target subgroups). Such data should help to mobilize and allocate program resources. On the other hand, however valid the procedures, prioritization will tend to be opposed by individuals whose programs and concerns are relatively well supported—and to be endorsed by those who feel neglected. Analysis also can be used to frustrate and misdirect actions. “Further study” can be a ruse for protecting the status quo. Showing “better buys” in risk reduction is meaningless, or even disingenuous, unless there is a real opportunity to move funds from worse causes to better ones. When funds are not fungible, such comparisons can lead to canceling worthwhile programs without increasing support for better ones.6 Finally, some supporters and detractors of prioritization may be less concerned with adolescents than with how the choice of policy-making procedure affects civic governance. Policy-making procedures can range from direct democracy to having specialists act in the public’s name without any consultation—arguing that they not only have a better command of the facts, but also a better understanding of what the public really wants. Toward the latter extreme, one finds metrics like QALYs (quality-adjusted life years) (Tengs and Wallace, 2000), which represent citizens’ values by the views expressed by a one-time sample. Our own procedure lies further toward the former extreme, insofar as it allows the continuing involvement 6   Kelman (cited in Kolata, 2001) recalls a meeting with EPA and NIH officials regarding the regulation of lead levels. “From my standpoint as a scientist, I realized that well nourished kids absorb less lead. So, being pretty naive, I said, ‘Why not take the money that the EPA is talking about for lowering lead levels in drinking water and putting it into nourishing inner city kids?’” The EPA said it didn’t feed children; the NIH said it didn’t have the money. “It was a classic federal impasse . . . At which point I figured I’d better sit down and shut up.”

OCR for page 109
Adolescent Risk and Vulnerability: Concepts and Measurement of actual citizens, and not just summaries of their views. Thus, values shape both the priorities that we set on teens’ welfare and the procedures that we use to reach those priorities—just as they, in turn, shape our future society. ANNEX SETTING PRIORITIES BY WEIGHTING ATTRIBUTES The essence of priority setting is to identify the issues that matter, decide how important each is in the focal context, and then evaluate each option, considering how it stacks up on each issue, weighted by the relative importance of those issues. Multiattribute utility theory formalizes this logic (Fischhoff et al., 1984; Keeney and Raiffa, 1976; vonWinterfeldt and Edwards, 1986). In it, the issues are called attributes and relative importance is represented by weights. Although many sophisticated applications are possible, a weighted sum is adequate for characterizing options in many situations (Dawes, 1979). In the case of adolescent well-being, the multiattribute degree of concern evoked by a source of vulnerability might be expressed as: where j is the source of vulnerability, i is an attribute, n is the number of attributes, wi is the weight for attribute i, xij represents how source j performs in terms of attribute i, and ui is the utility attached to that degree of attribute i. This appendix illustrates how this approach might be applied to setting priorities. The rows of Table 5-1 list 12 attributes that might be considered when evaluating threats to the health and safety of students in a school. They include aspects of both mortality (number of deaths per year, average chance of death, highest chance of death for any student, and greatest number of deaths in a single episode) and morbidity (number of more and less serious cases of long- and short-term injuries and illnesses per year). The attributes also include features that often have been found to affect risk perceptions (e.g., Fischhoff et al., 1978; Morgan et al., in press; Slovic, 1987). These are the time between exposure and health effects, the quality of scientific understanding, the uncertainty regarding the outcomes, and the ability of students or parents to control exposure. The columns of Table 5-1 show four weighting schemes that might be applied to these attributes. Set A reflects a person concerned only with the

OCR for page 109
Adolescent Risk and Vulnerability: Concepts and Measurement TABLE 5-1 Four Possible Sets of Weights for 12 Attributes of Adolescent Vulnerability   Importance Weighting Scheme Attribute A B C D Number of deaths per year 0.050 0.050 0.050 0.150 Chance in a million of death per year for the average student 0.600 0.250 0.100 0.100 Chance in a million of death per year for the student at highest risk 0.300 0.250 0.100 0.100 Greatest number of deaths in a single episode 0.050 0.050 0.050 0.150 More serious long-term injuries or illnesses (cases per year) 0 0.200 0.150 0.025 Less serious long-term injuries or illnesses (cases per year) 0 0 0.200 0.025 More serious short-term injuries or illnesses (cases per year) 0 0.200 0.150 0.025 Less serious short-term injuries or illnesses (cases per year) 0 0 0.200 0.025 Time between exposure and health effects 0 0 0 0.100 Quality of scientific understanding 0 0 0 0.100 Uncertainty regarding death, illness, and injury 0 0 0 0.100 Ability of student/parent to control exposure 0 0 0 0.100 Sum of weights 1.000 1.000 1.000 1.000 probability of death. Set B corresponds to an individual concerned with serious illness and injury, as well as death. Set C weights also consider less serious illness and injury.7 Finally, set D also pays attention to the “qualitative” aspects of the risk in the final four rows. Each set of weights has been normalized to total 1.0; they correspond to wi, in the formula for concern. 7   It may seem counterintuitive to assign greater weight to less serious effects (injury or illness) than to more serious ones. However, if there is much greater variability in less serious consequences (e.g., because serious ones hardly ever occur from any of the threats under consideration), then that attribute might deserve more attention.

OCR for page 109
Adolescent Risk and Vulnerability: Concepts and Measurement Table 5-2 characterizes each of 5 possible threats to adolescents in terms of these 12 attributes. The values are taken from an elaborate test bed created to study prioritization processes at a hypothetical Centerville Middle School. The values were assigned to reflect circumstances that might be found in a typical U.S. school, and internally consistent, considering the specifics of this hypothetical school (i.e., size, location, age).8 These values are represented by xij in the formula for concern. In the interests of simplicity, the utility assigned to each level of that attribute was set equal to the level, normalized to range from 0-1.0, across the five sources of vulnerability. Combining attribute weights (wi) with the estimates of outcomes (xi) produces scores for overall concerns. Table 5-3 ranks the sources, from best (or least bad) to worst, for individuals with the four sets of values appearing in Table 5-3. For individuals focused on mortality (Set A), self-inflicted injury (i.e., suicide) draws the greatest concern and lead poisoning the least (in a school where lethal doses are impossible). If serious injury and illness also are important (Set B), then the less common infectious diseases at the school become the worst threat and the more common ones become more important. Giving weight to less serious injury and disease as well (Set C) further increases concern over common infectious diseases, and reduces that over intentional injury (whose nonfatal consequences at Centerville are rare). When weight is assigned to the qualitative attributes (rows 9–12 in Table 5-1), common infectious diseases drop in importance. They are understood very well, have immediate effects, and afford some measure of controllability (e.g., vaccination). As a result, they evoke little of the dread and discomfort associated with the less common infectious diseases or self-inflicted injury. Thus, under the circumstances of this hypothetical school, relative concern over some of these sources of vulnerability varies considerably, depending on the weight given to the different attributes. On the other hand, lead poisoning merits relatively little concern, whatever the weighting scheme. Although its consequences can be terrible, in this (relatively new) school they are not that much of an issue. Lead poisoning might rank much higher in priorities set at an aging, urban school or in national priori- 8   The project description is at: http://www.epp.cmu.edu/research/risk_ranking.html. Summary sheets describing the risks can be found at: http://www.epp.cmu.edu/research/risk-summary-sheets/risk1.html.

OCR for page 109
Adolescent Risk and Vulnerability: Concepts and Measurement TABLE 5-2 Estimates of the Performance of 5 Sources of Adolescent Vulnerability on 12 Attributes   Sources of Vulnerability Attribute Common Infectious Diseases Intentional Harm Lead Poisoning Less Common Infectious Diseases Self-Inflicted Harm Number of deaths per year 0.067 0.233 0 1 1 Chance in a million of death per year for the average student 0.071 0.286 0 1 1 Chance in a million of death per year for the student at highest risk 0.008 0.333 0 0.117 1 Greatest number of deaths in a single episode 0.04 0.107 0 1 0.107 More serious long-term injuries or illnesses (cases per year) 0.0005 0.01 0 1 0.02 Less serious long-term injuries or illnesses (cases per year) 0.2 0.05 1 0.2 0.3 More serious short-term injuries or illnesses (cases per year) 1 0.075 0 1 0.25 Less serious short-term injuries or illnesses (cases per year) 1 0.006 0 0.010 0.001 Time between exposure and health effects 0.5 1 0 0.5 1 Quality of scientific understanding 0 0.5 0 0 0.5 Uncertainty regarding death, illness, and injury 0.318 0.636 1 0.318 0.409 Ability of student/parent to control exposure 0.5 0 1 0.5 0

OCR for page 109
Adolescent Risk and Vulnerability: Concepts and Measurement TABLE 5-3 Risk Rankings of 5 Sources of Adolescent Vulnerability Given Sets of Weights on 1 Set of 12 Attributes and Estimates from a Hypothetical Middle School     Set of Weights Rank   A B C D Best 1. Lead poisoning Lead poisoning Intentional injury Common infectious disease   2. Common infectious disease Intentional injury Lead poisoning Lead poisoning   3. Intentional injury Common infectious disease Self-inflicted injury Intentional injury   4. Less common infectious disease Self-inflicted injury Common infectious disease Self-inflicted injury Worst 5. Self-inflicted injury Less common infectious disease Less common infectious disease Less common infectious disease ties that considered such schools. In that case, Centerville Middle School might have a mandate, and perhaps resources, to deal with a problem of relatively little local concern. REFERENCES Bendor, J. (1995). A model of muddling through. American Political Science Review, 89, 819-840. Bentkover, J. D., Covello, V. T., & Mumpower, J. (Eds.). (1985). Benefits assessment: The state of the art. Dordrecht, The Netherlands: D. Reidel. Breyer, S. (1993). Breaking the vicious circle: Toward effective regulation. Cambridge, MA: Harvard University Press. Budescu, D. F., & Wallsten, T. S. (1995). Processing linguistic probabilities: General principles and empirical evidence. In J. R. Busemeyer, R. Hastie, & D. L. Medin (Eds.), Decision making from a cognitive perspective (pp. 275-318). New York: Academic Press.

OCR for page 109
Adolescent Risk and Vulnerability: Concepts and Measurement Bureau of Labor Statistics. (1998). NLS1997 handbook. Washington, DC: U.S. Department of Labor. Carnegie Mellon University. (2001). Engineering and public policy risk analysis and risk communication. Available: >http://www.epp.cmu.edu/research/EPP_risk.html>. [August 31, 2001]. Cohen, B., & Lee, I. S. (1979). A catalog of risks. Health Physics, 36, 707-722. Crouch, E. A. C., & Wilson, R. (1981). Risk-benefit analysis. Boston: Ballinger. Davies, J. C. (Ed.). (1996). Comparing environmental risks. Washington, DC: Resources for the Future. Dawes, R. M. (1979). The robust beauty of improper linear models. American Psychologist, 34, 571-582. Dawes, R. M., & Hastie, R. (in press). Rational choice in an uncertain world (2nd ed.). San Diego: Harcourt Brace. Department of Health and Human Services. (2000). Healthy people 2010. Washington, DC: Author. Federal Interagency Forum on Child and Family Statistics. (1997). America’s children: Key national indicators of well-being. Washington, DC: Author. Fischhoff, B. (1995). Ranking risks. Risk: Health Safety & Environment, 6, 189-200. Fischhoff, B. (1999). Why (cancer) risk communication can be hard. Journal of the National Cancer Institute Monographs, 25, 7-13. Fischhoff, B. (2000). Need to know: Analytical and psychological criteria. Roger Williams University Law Review, 6, 55-79. Fischhoff, B., Bostrom, A., & Quadrel, M. J. (1997). Risk perception and communication. In R. Detels, J. McEwen, & G. Omenn (Eds.), Oxford textbook of public health (pp. 987-1002). London: Oxford University Press. Fischhoff, B., Downs, J., & Bruine de Bruin, W. (1998). Adolescent vulnerability: A framework for behavioral interventions. Applied and Preventive Psychology, 7, 77-94. Fischhoff, B., Lichtenstein, S., Slovic, P., Derby, S. L., & Keeney, R. L. (1981). Acceptable risk. New York: Cambridge University Press. Fischhoff, B., Parker, A., Bruine de Bruin, W., Downs, J., Palmgren, C., Dawes, R.M., & Manski, C. (2000). Teen expectations for significant life events. Public Opinion Quarterly, 64, 189-205. Fischhoff, B., Slovic, P., Lichtenstein, S., Read, S., & Combs, B. (1978). How safe is safe enough? A psychometric study of attitudes towards technological risks and benefits. Policy Sciences, 8, 127-152. Fischhoff, B., Watson, S., & Hope, C. (1984). Defining risk. Policy Sciences, 17, 123-139. Florig, H. K., Morgan, M. G., Morgan, K. M., Jenni, K. E., Fischhoff, B., Fischbeck, P. S., & DeKay, M. (in press). A test bed for studies of risk ranking. Risk Analysis. Institute of Medicine. (1998). Scientific opportunities and public needs: Improving priority setting and public input at the National Institutes of Health. Committee on the NIH Research Priority-Setting Process. Health Sciences Section. Washington, DC: National Academy Press. Institute of Medicine. (1999). Toward environmental justice: Research, education, and health policy needs. Committee on Environmental Justice. Health Sciences Section. Washington, DC: National Academy Press.

OCR for page 109
Adolescent Risk and Vulnerability: Concepts and Measurement Jenni, K. (1997). Attributes for risk evaluation. Unpublished doctoral dissertation. Department of Engineering & Public Policy, Carnegie Mellon University. Jessor, R., Donovan, J. E., & Costa, F. M. (1991). Beyond adolescence. New York, NY: Cambridge University Press. Kahneman, D., Slovic, P., & Tversky, A. (Eds.). (1982). Judgment under uncertainty: Heuristics and biases. New York: Cambridge University Press. Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47, 263-281. Keeney, R. L., & Raiffa, H. (1976). Decisions with multiple objectives: Preferences and value tradeoffs. New York: John Wiley. Kolata, G. (2001, April 8). Putting a price tag on the priceless. New York Times, Section 4, p. 4. Kubey, R., Larson, R., & Csikszentmihalyi, M. (1996). Experience sampling method applications to communications research. Journal of Communication, 46(2), 99-120. Lerner, R. M. (in press). Adolescence: Development, diversity, context, and application. Upper Saddle River, NJ: Prentice-Hall. Lichtenstein, S., Slovic, P., Fischhoff, B., Layman, M., & Combs, B. (1978). Judged frequency of lethal events. Journal of Experimental Psychology: Human Learning and Memory, 4, 551-578. Lindblom, C. (1959). The science of muddling through. Public Administration Review, 79-88. Loewenstein, G. (1996). Out of control: Visceral influences on decision making. Organizational Behavior and Human Decision Processes, 65, 272-292. Long, J., & Fischhoff, B. (2000). Setting risk priorities: A formal model. Risk Analysis, 20, 339-351. Lowrance, W. (1975). Of acceptable risk. San Francisco: Freeman. Masten, A. S. (2001). Ordinary magic: Resilience processes in development. American Psychologist, 56, 227-238. McFadden, D. (1999). Rationality for economists? Journal of Risk and Uncertainty, 19, 73-105. Macintyre, S., & West, P. (1993). What does the phrase “safer sex” mean to you? Understanding among Glaswegian 18 year olds in 1990. AIDS, 7, 121-126. Morgan, K. M., DeKay, M. L., Fischbeck, P. S., Morgan, M. G., Fischhoff, B., & Florig, H. K. (in press). A deliberative method for ranking risks: Evaluating validity and usefulness. Risk Analysis. Morgan, M. G., Fischhoff, B., Bostrom, A., & Atman, C. (2001). Risk communication: The mental models approach. New York: Cambridge University Press. Morgan, M. G., Fischhoff, B., Lave, L., & Fischbeck, P. (1996). A proposal for ranking risks within federal agencies. In C. Davies (Ed.), Comparing environmental risks (pp. 111-147). Washington, DC: Resources for the Future. National Institutes of Health. (1998). Setting research priorities at the National Institutes of Health. Washington, DC: Author. National Research Council. (1983). Risk assessment in the federal government: Managing the process. Committee on the Institutional Means for Assessment of Risks to Public Health. Commission on Life Sciences. Washington, DC: National Academy Press.

OCR for page 109
Adolescent Risk and Vulnerability: Concepts and Measurement National Research Council. (1996). Understanding risk: Informing decisions in a democratic society. Committee on Risk Characterization. P. C. Stern & H. V. Fineberg (Eds.). Commission on Behavioral and Social Sciences and Education. Washington, DC: National Academy Press. Office of Technology Assessment. (1995). Risks to students in school (OTA-ENV-633). Washington, DC: U.S. Congress. Available: <http://www.ota.nap.edu/pubs.html>. [August 8, 2001]. Quadrel, M. J., Fischhoff, B., & Davis, W. (1993). Adolescent (in)vulnerability. American Psychologist, 48, 102-116. Schriver, K. A. (1989). Evaluating text quality: The continuum from text-focused to reader-focused methods. IEEE Transactions on Professional Communication, 32, 238-255. Schwarz, N. (1999). Self reports. American Psychologist, 54, 93-105. Simon, H. A. (1957). Models of man. Cambridge, MA: MIT Press. Slovic, P. (1987). Perceptions of risk. Science, 236, 280-285. Starr, C. (1969). Social benefit versus technological risk. Science, 165, 1232-1238. Tengs, T. O., Adams, M. E., Pliskin, J. S., Safran, D. G., Siegel, J. E., Weinstein, M. C., & Graham, J. D. (1995). 500 lifesaving interventions and their cost-effectiveness. Risk Analysis, 15, 369-390. Tengs, T. O., & Wallace, A. (2000). One thousand quality of life estimates. Medical Care, 36, 583-637. United Nations. (1989). Convention on the rights of the child. New York: Author. vonWinterfeldt, D., & Edwards, W. (1986). Decision analysis and behavioral research. New York: Cambridge University Press. Weinstein, N. D. (1987). Unrealistic optimism about susceptibility to health problems. Journal of Behavioral Medicine, 19, 481-500.