Research Paper:Policy Issue and Solution Assignment

9 Early Intervention Saving Children From a Life in Crime

David P. Farrington

Cambridge University

Author of Saving Children From a Life in Crime

Fat Katie. There was a crisis in the Cullen household. Jordan, Cullen’s daughter, had a birthday on the immediate horizon. It was thus time to invite all the children in Jordan’s kindergarten class to this celebratory gathering. Jordan dutifully compiled her roster of invitees. Everyone was on the list except Fat Katie. When queried about this glaring omission, Jordan said that Fat Katie pushed her and other kids down and that nobody much liked her.

Cullen was about to lecture his typically kind-hearted daughter about the inappropriateness of calling anyone Fat Katie—especially since as a child he had been tormented with the taunt of Fat Franky by his older brother. But a larger issue beckoned. In the Cullen household, Jordan was told, nobody—not even Fat Katie—would be excluded from a class event. Either Fat Katie was invited or there would be no birthday party. With the prospect of treats, attention, and gifts of 13 Barbie dolls hanging in the balance, Jordan allowed her nicer sentiments to guide her to the compassionate decision: Fat Katie was invited. Somewhat sadly, she did not show up.

Fat Katie. As early as kindergarten—age 5 to 6—Jordan and her classmates had identified one child, Fat Katie, who was different. She was big, pushed kids down, and was rejected by her classmates. She manifested developmental problems, at least in this setting. By the next year, she had departed the school. In all likelihood, she has since blossomed into a wonderful young adult. A change in context—a place where other kids did not label her as Fat Katie—might have been all that was needed. If any real problems had existed, Fat Katie’s affluent parents undoubtedly would have noticed them, had them diagnosed by professionals, and ensured that an appropriate intervention took place.

But not all Fat Katies will be this fortunate. They will be born into more disadvantaged settings—perhaps to homes headed by young, single, impoverished mothers with substance abuse addictions. When they manifest behavioral problems in early childhood, everyone will know they exist. Fellow classmates, teachers, and parents—all those interacting daily with these Fat Katies—will experience and thus know that trouble is brewing. Nothing, however, will be done. There will be no change in school or expert professionals called in to save these Fat Katies. As their lives develop, they will not blossom into wonderful young adults. Too many of them will endure repeated peer rejection, fail at school, flock together with other wayward Fat Katies, and confront a life of increasing cumulative disadvantage. One bad thing after another—one more criminogenic risk factor after another will pile up; one on top of another, on top of another, on top of another. These Fat Katies—or Fat Frankys—will move steadily away from a conventional life course and become trapped on a developmental trajectory headed straight into crime. The train will have left the station, with prison as its destination.

When these Fat Katies and Fat Frankys are age 6 or perhaps age 10 or 11, we—the good and decent people of the world—will look at them and shake our heads. We will see that these kids, through no choice of their own, are destined for a life of hardship and crime. We will feel fortunate that our children do not face the array of personal and social pathologies that these children face. We will see them as victims of being born with the wrong traits and in the wrong place. We will lament their horrible luck of the draw—they have won the lottery for a life in crime—and we will feel sorry for them. But by the time they reach age 16 or perhaps 21, our language will change. We will see these very same children—now larger and stronger and meaner—as super-predators to be feared. We will experience a sense of relief when they are locked away. “Get the dangerous off the street. Incapacitate them,” we will chant. “Hold them responsible for the choices they make. After all, they are adults, aren’t they?”

We will forget that today’s super-predator was yesterday’s wayward child. Today, we see choice, harm, and accountability. Yesterday, we saw no choice, a child being harmed, and an inevitable future in crime. As Gustave de Beaumont and Alexis de Tocqueville reminded us long ago, those who become delinquent “have been unfortunate before they became guilty” (as cited in Allen, 1981, p. 12). Somehow, however, we disconnect these links in the developmental chain into crime. Doing so allows us to focus, undistracted by sympathy, on the very real threat posed by the adult offender standing before us. It also allows us to absolve ourselves for why none of us cared enough to redirect these offenders, while they were still helpless youngsters, into a more productive and civil life.

The great lesson of life-course criminology, which we will discuss in a moment, teaches us that criminals come from somewhere. Most often, children develop in predictable ways into juvenile delinquents and then into adult criminals. Such a simple insight, huh? But it is an insight pregnant with important policy implications. What we now know is that the vast majority of serious, violent offenders do not suddenly spring into action in the teenage or adult years. Some do, of course, and we call them late onset offenders. If they commit a heinous offense, they end up in the front page of the newspaper, with all of us wondering how such a bad thing could have been done by such a good person. But the prototypical pathway into serious, violent criminality starts in childhood—perhaps in the womb where the brain develops and temperament is acquired. Again, most criminals are not late bloomers. The seeds of their criminality were planted and nurtured in the beginning stages of life.

If the origins of crime start early in life, then what lesson should we draw? At present, it is largely to do nothing until a wayward child turned super-predator is standing before the court and before us—society. Cullen and Jonson think that is tragic—stupidity and neglect run amok. We see the failure to intervene early in life as needlessly consigning many youngsters to a life in crime and—in what is often forgotten—as needlessly consigning many innocent citizens a few years hence to a preventable victimization. Any sensible approach to crime control would make early intervention with at-risk children a high priority (Farrington & Welsh, 2007).

Saving wayward children from a life in crime is really a policy that extends correctional rehabilitation to the earlier years in life. It is based on the sound criminological insight that we should not sit by and wait for troubled youths to commit illegal acts that place them behind prison bars. Crime has causes, and we should knife them off as early as possible. Thus, corrections should be implemented as part of a seamless set of interventions that seeks, from birth into adulthood, to prevent further criminality. As Friedrich Lösel (2007) notes, it is never too early and never too late to intervene. Put a different way, the stuff in the last chapter and in this chapter are part of the same paradigm.

Fortunately, there is an increasing recognition of the wisdom of this view. In this chapter we thus will reveal the case for expanding early intervention programs. We will show that such a policy would be based on sound criminology. We will show that there is a range of programs available that have proven, at least in some instances, to be effective in preventing youngsters from developing into offenders. These programs are also cost effective. And we will show that the American public is virtually unanimous in supporting efforts to save children from a life in crime.

Lessons From Child Criminology

The Discovery of Childhood

Students majoring in psychology typically are required or encouraged to take a course in Child Psychology (Cullen did so a few decades ago when he was an undergraduate studying psychology). Criminal justice majors, however, do not enroll in Child Criminology because, alas, this course does not exist! In retrospect, this curricular omission is rather odd. After all, psychologists have known for ages—even for decades before Cullen, old as he is, went to college!—that humans develop. They study childhood because this is where—you guessed it!—that human development begins. Until recently, criminologists have pretty much ignored this reality. Instead, they skipped childhood and created courses in Juvenile Delinquency.

Did criminologists skip childhood because they were stupid? Well, in a way they were. They would not come out and say that the first stages of life were inconsequential; rather, they simply ignored what occurred prior to adolescence. In all fairness, they had two good reasons for doing so—even if these justifications do not obviate the fact that ignoring childhood led to stupid criminology.

First, as criminologists, they were interested in crime. With rare exceptions, children are not seen as able to form criminal intent and do not do things that get them arrested. Yet, in a way, kids are quite “criminal.” From very early in life, they hit, bite, kick, and shove one another down—and do so a lot (Nagin & Tremblay, 2005; Tremblay, 2006; Tremblay et al., 1999). This is why parenting can be exhausting; children always seem to be doing something that requires correction or discipline. But even when our children bloody someone’s nose or knock another kid down, we do not interpret this act as an assault, call the police, and seek to give them a prolonged time out behind prison bars. In short, children are not seen as law-breakers and thus it is easy to dismiss them as falling outside the subject matter of criminology.

Second and related, there is something called the age–crime curve. You likely know what a bell curve is: a curve, shaped like a bell, that is low at its left and right tails and high in the middle. The bell curve is typically used to show the normal distribution of IQ scores, but much human behavior can be plotted on such a curve. When it comes to crime, the bell curve’s tail is low in childhood whereas the peak of the curve is around age 17 or 18 (a bit later for violent crimes like homicide). The curve then gradually descends on a downward slope, showing that crime decreases with age.

Practically speaking, this age–crime curve conveys the view—correct in some ways—that crime is low in childhood, rises to a peak in the teenage years, and then decreases as youths move into adulthood. Importantly, the age–crime curve thus suggests that something peculiar must occur during adolescence that transforms youngsters from cherubic children into juvenile delinquents. And when this storm of criminal influences subsides, these same teenagers move into the adult roles of work, spouse, and parent and stop offending. This process is sometimes called maturational reform (see Matza, 1964).

As might be anticipated, criminologists focused on this transitional, turbulent stage in life. Albert Cohen (1955) and Richard Cloward and Lloyd Ohlin (1960) wrote famous books on the origins and nature of gang delinquency. In Causes of Delinquency, Travis Hirschi (1969) showed how self-report data from junior high and high school students could be used to test competing delinquency theories, including his own classic social bond perspective. Following in Hirschi’s large footsteps, scholars have conducted hundreds of studies on juveniles, all with the goal of deciphering the criminogenic features of adolescent life.

Terrie Moffitt (1993) perhaps has done the most to challenge this brand of research. In her developmental theory of antisocial behavior, Moffitt observed—consistent with traditional criminology—that most youths engage in a lot of crime only during the teenage years. She called this group adolescence-limited offenders (also known to criminological insiders as the ALs). In essence, she was arguing that the standard developmental pathway for youths obeys the age–crime curve: okay until adolescence, run haywire for a while, then back on a prosocial track as adulthood beckons. Many readers might well be on this pathway. Not too long ago, some of you might have done some deviant things—knocking down mailboxes, shoplifting, needless fights, even trying to tip over cows—that now you would not do. If so, your delinquency was adolescence-limited. You were an AL!

But Moffitt was brilliant enough to understand that the age–crime curve masked a second developmental pathway: life-course-persistent offenders (known as LCPs). When the age–crime curve was disaggregated—a fancy word for broken down more carefully—it was apparent that a smaller group of youngsters also existed who were hidden by the bell curve pathway that most kids—the ALs—followed. This second group, the LCPs, started to manifest conduct problems very early in childhood. Remember Fat Katie? Well, these kids fall into the Fat Katie category. Both in the preschool years and in the primary grades, they have lots of difficulties. Parents, teachers, and other students all know that “something is wrong with them.” These youngsters graduate from conduct problems to delinquent acts and then to adult criminality.

Their age–crime (or age-antisocial behavior) curve is more like a line that jumps up in very early childhood and stays high for years to come. Admittedly, few actually are holding up banks at age 80. Feel relieved that you do not have to ask grandma and grandpa if they are packing guns and scoring some crack today! However, these offenders remain entrenched in a criminal career deep into adult life (e.g., into their late 30s and 40s). This is why Moffitt called them life-course-persistent offenders (Moffitt, 2006).

Cullen and Jonson like Moffitt’s theory because it is simple: two groups with names and acronyms that are easy to remember because they make intuitive sense. Adolescence-limited offenders’ crimes are limited to adolescence. Life-course-persistent offenders persist in crime across the life course. Brilliant! Alas, criminologists cannot resist complicating things! They have used fancy statistical techniques to map out developmental trajectories. They find that there may be another two or three groups of offenders whose trajectories in crime are variants of the two groups (ALs and LCPs) identified by Moffitt (see, e.g., Nagin & Tremblay, 2005; Piquero, Farrington, & Blumstein, 2007). But we can put this more sophisticated stuff aside for another time. The main point to grasp here is that the source of serious, life-course offending typically does not start in the teenage years but in childhood. Yes, it looks like we need a course in Child Criminology after all!

Beyond Adolescence-Limited Criminology

Moffitt’s work had profound implications (see also Sampson & Laub, 1993). It made clear that scholars could no longer practice adolescence-limited criminology (Cullen, 2011)! For if the origins of serious criminality can be traced to childhood, then we must explore what it is about the early years of life that place a youngster on one developmental pathway rather than another.

A dramatic revelation advanced by Moffitt is that the fork in the road—one path leading to crime and the other to conformity—may start in the womb (see also Wright et al., 2008). Traditionally, criminologists have been skeptical about biological explanations of crime. They worry that they are a mask for blaming crime on individuals when the true fault lies in the awful life conditions into which too many future inmates are born. But the science is now irrefutable that pre-natal experiences affect brain development and that everyone enters the world with a disposition rather than as a blank slate. Indeed, think of your siblings. Do you ever wonder how you came from the same parents? Biology matters.

When Moffitt made this point, criminologists surprisingly did not dispute it—as they normally would any other biosocial theory of crime. Why not? Well, we suspect that it was because everyone liked her and knew that she was smarter than the rest of us. But it also was because Moffitt was shrewd enough not to claim, as Cesare Lombroso did long ago, that offenders were born criminals. Rather, she said that these unfortunate kids suffered from neuropsychological deficits. Scholars assume that inborn traits are immutable and cannot be changed; deficits, however, sound like they can be replenished and taken away. In short, Moffitt’s theory did not seem to consign those with biosocial pathology to a life in crime. She was pointing to a deficit that might be amenable to intervention. She was a biosocial theorist with a hopeful message.

As Alper and Beckwith (1993) point out, it is common for people to erroneously believe in genetic fatalism, which is the fallacious view that traits or behavior rooted in biology cannot be changed. But the notion that biology is destiny is simply false. Eyesight is inherited but it can be adjusted through the use of glasses. Attention Deficit Hyperactivity Disorder (ADHD) has a high heritability component, but its impact on behavior—including crime—can be modified through medicine (e.g., Ritalin) and cognitive-behavioral therapy. And so on.

For Moffitt, neural or brain development can be compromised during the pre-natal period due, for example, to mothers’ drug use or poor nutrition. These problems in turn affect psychological traits such as executive functioning (e.g., self-control or impulsivity) and verbal development. These neuropsychological deficits do not automatically lead to antisocial conduct. But children with deficits tend to be difficult to handle, and they can evoke harsh and erratic discipline from parents who may share similar deficits from their own childhood. When they enter school, they are likely to fail and be rejected by other children. In response, their antisocial propensity deepens, provoking more harsh reactions and failure. Over time, these individuals experience the cumulative continuity of these criminogenic conditions, causing them to become increasingly entrenched in an antisocial life course. They become, in short, life-course-persistent offenders.

In any event, criminologists’ warm acceptance of Moffitt’s theory was important because it legitimated two views about crime. The first view, as mentioned, was that the roots of chronic offending extend to the earliest moments of life. Child development thus was essential to understanding criminal development. The second view was that, as Moffitt (1993) put it, “continuity is the hallmark of the small group of life-course-persistent antisocial persons” (p. 697). Unlike adolescence-limited offenders, this group does not change by jumping into and out of crime. Rather, the lives of LCPs are marked by stability of behavior across the life course. Thus, they bite and hit others at four years of age, they skip school and pilfer from stores at age 10, they sell drugs and steal cars at age 16, they rape and rob at age 20, and they engage in fraud and child abuse at age 30 (p. 695).

This is not the place to evaluate the merits of every aspect of Moffitt’s theory (see Moffitt, 2006; Sampson & Laub, 2005). However, it is essential to note that there is considerable support for Moffitt’s claim that among those who become the most serious and chronic offenders, a high proportion manifest conduct problems early in life (Farrington, 2003; Murray & Farrington, 2010; Wright et al., 2008). Early antisocial behavior is the strongest predictor of involvement in delinquency (Lipsey & Derzon, 1998). Not all problem children become criminals, but the risk for future difficulties is clearly elevated. This is especially the case among youngsters who (1) engage in high rates of misbehavior, (2) are in trouble across social settings (e.g., home, school), (3) are involved in many different types of antisocial behavior, and (4) show a particularly early onset of problems (Wright et al., 2008, p. 27). Although late onset into crime occurs—some individuals turn to crime later in life—most of those found in prison for serious offenses experienced troubled childhoods and lives. According to Lee Robins (1978), “adult antisocial behaviour virtually requires childhood antisocial behaviour” (p. 611, emphasis in the original; see also Robins, 1966).

To summarize:

· Having early childhood conduct problems does not inevitably lead to a life in crime.

· However, there is considerable stability in behavior, so that troubled kids are likely to be troubled teens.

· Further, adults who are antisocial and are imprisoned typically are drawn from the pool of children who manifest strong antisocial propensities early in life. Again, these are Moffitt’s life-course-persistent offenders.

What does all this discussion of stability in behavior have to do with early intervention? Well, we suspect that readers already have anticipated the punch line: If many serious offenders start their criminal development early in life, then it is logical that efforts to knife off a life of persistent offending should occur early in life. Child criminology thus provides the rationale for—indeed, it demands—early intervention (see also Farrington & Welsh, 2007).

The Need for Early Intervention

Child criminology is really part of a broader attempt to study offenders from womb to tomb. This theoretical and research paradigm is called life-course or developmental criminology. As might be expected, to follow offenders across their lives, it is necessary to employ a particular kind of study design. Do you know what that is? Yes, it is a longitudinal study. A cross-sectional study examines people at one point in time—often with the advantage of great depth but with the disadvantage of not being able to measure developmental processes. Longitudinal studies are expensive because they necessitate keeping tabs on people over many years and interviewing them multiple times. Even so, this is the only way to capture how criminal careers unfold over time, including revealing how what occurs in childhood matters later in life.

Much has been learned from longitudinal studies, but too many suffer from an odd omission: They do not measure whether the sample members have received a treatment intervention for their antisocial problems. Efforts are at times made to record whether people have been in prison and, if so, for how long. But few studies have reliable data on whether at-risk youngsters receive any early intervention or simply are allowed to develop into criminals. Why have researchers not incorporated this potentially important source of behavioral change into their investigations? Cullen and Jonson believe that the most likely answer is that they simply did not think about it. When most of the great longitudinal studies were initiated, early intervention was not on most scholars’ radar screen. And you cannot measure what you are not thinking about.

One notable exception to this rather remarkable omission is the Pittsburgh Youth Study, known by its acronym PYS (do not sound this out!). Under the direction of Rolf Loeber and Magda Stouthamer-Loeber, this longitudinal project followed three samples of boys attending Pittsburgh public schools who, at the beginning of the study, were in Grades 1, 4, and 7. For each grade, 500 boys were selected, 250 of whom were assessed by the researchers to be at high risk of wayward conduct. Many studies have been published using PYS data. We are interested in two studies that examined the neglected topic of whether troubled youngsters in the PYS received quality intervention. The focus was on the help-seeking behavior of the boys’ “caretakers,” who in over 90% of the cases were their mothers.

In the first study, Stouthamer-Loeber, Loeber, and Thomas (1992, p. 161) explored all three groups of boys (called, you guessed it, “youngest, middle, and oldest samples”). The PYS addressed the issue of whether youngsters, by age group, received help for their problem behaviors. We will focus mainly on whether they received help—intervention—for their delinquencies. Here are the study’s main findings:

· Younger or “early onset” youths who manifested a problem were less likely to receive help than older youths who manifested the problem.

· Caretakers were more likely to seek help for aggression and oppositional behaviors—especially Disruptive Behavior Disorder—than anything else. This may have been because they had to “deal” with this behavior at home on a regular basis.

· However, especially for younger children, many caretakers (almost 50%) did not seek help. “This suggests that a substantial portion of caretakers struggle unaided with serious child behavior problems” (Stouthamer-Loeber et al., 1992, p. 175).

· Seeking help for delinquents—even serious delinquents—was less common. “Between 50% and 70% of the caretakers of the most seriously delinquent boys had never sought help [for various problems], and only about a quarter had ever sought help from a mental health professional” (p. 173).

· When caretakers did seek help, about a quarter of the time they had only one or two contacts.

· Bottom line: A substantial proportion of delinquent youths did not receive adequate intervention.

Now on to the second study, authored by Stouthamer-Loeber, Loeber, van Kammen, and Zhang (1995). This work focused only on the oldest sample and attempted to see how interventions may have differed for delinquent youths who had contact with the juvenile court versus those who did not. Again, here are the major findings:

· Compared to delinquents who did not go to court, youngsters who appeared in juvenile court had similar histories of offending (onset, frequency) but they were involved in more serious delinquent acts.

· For court delinquents—who were only in the eighth grade when surveyed—the average interval or time between the onset or beginning of problem behaviors and court contact was four years; more than half of this group exhibited problems for more than five years. The interval between serious problem behaviors and court contact was two years.

· Those who had not had contact with the court also exhibited delinquent problems for a similar length of time. The point: Youngsters can be involved in delinquency for some time before coming into contact with the court.

· Despite having “careers of disruptive problems of many years’ standing . . . only 41% of the caretakers had ever sought help for the boys’ problems” (p. 248).

· Those delinquents who received help tended to show problems for a longer period of time and that were more serious. Those delinquents who did not receive help, however, “were not without problems; on average they had exhibited problems for 5.75 years” (p. 248).

· Bottom line: “Thus, the disruptive problem careers of many of the delinquent boys were left unchecked” (p. 248).

It is possible, of course, that the findings from the PYS are somehow biased. But in all likelihood, any “Steeler . . . er . . . Pittsburgh” effect on the findings is not great. A similar longitudinal study conducted in Denver reported comparable results. The Denver Youth Survey (DYS) was comprised of 1,527 youngsters (806 boys and 721 girls) drawn from at-risk, socially disorganized, high-crime neighborhoods in—you guessed it—Denver! The DYS participants were purposely sampled at ages 7, 9, 11, 13, or 15. As Huizinga, Weiher, Espiritu, and Esbensen (2003) discovered, the percentage of parents who sought help for their kids was only 20% for children and 30% for adolescents—though the percentages increased as their offspring’s delinquency and psychological problems grew more serious. As Huizinga et al. (2003) concluded, “for the majority of delinquent and serious delinquent youth, help has not been sought” (p. 84; for more details, see Espiritu, 1994).

This state of affairs bothers Cullen and Jonson. Year after year, children are manifesting psychological, behavioral, and criminological problems. Their careers in crime are blossoming, and nothing is being done. Part of the problem, of course, falls on the shoulders of parents. But the reality is that many parents of at-risk youngsters are at risk themselves. They lack the health insurance, money, education, transportation, and ability to take time off work to haul a resistant child off to a help provider on a weekly basis. They also are living in at-risk neighborhoods where effective services may not be close at hand (Guerra, 1997). In this context, the real culprit is that we lack an overarching agency whose job it is to identify troubled youths and to coordinate a coherent intervention (see also Farrington & Welsh, 2007). Meanwhile, the future super-predators of America are allowed to fall between the cracks—until they victimize, earn an arrest, and eventually are sent off to prison. Based on the PYS, Loeber and colleagues (2003) leave us with this sobering reality that we ignore at our own peril:

In summary, the development of disruptive and delinquent behaviors was largely left unchecked by parents and helping agencies. These findings have important implications for planners of preventative interventions and policy makers. Preventative interventions should take place in the relatively long time window between the onset of early problem behaviors of a minor kind and the first contact with the juvenile court. Policy makers should realize that eventual index [or serious] offenders have often had the unchecked opportunity to commit delinquent acts for many years. (p. 127, emphasis added)

Five Programs That Work—At Least When Done Right

Okay, there is a need to intervene with at-risk youths. But do any programs exist that work, that are effective in knifing off their criminal development? Fortunately, the answer to this question is a resounding “yes.” Based on rigorous, experimental evaluations, the evidence is clear that we have a growing repertoire of interventions that can save children from a life in crime (Farrington & Coid, 2003; Farrington & Welsh, 2007; Greenwood, 2006; Morizot & Kazemian, 2015; Welsh & Farrington, 2012). In this section, we review five famous interventions. They are exemplars for how to design an effective system of early intervention.

Still, this is not a time for hubris—for getting so overconfident or giddy that we assume that child saving is an easy task. The most daunting challenge is what is known as technology transfer, which is how to take knowledge about what works and spread it into new settings (see Becker, David, & Soucy, 1995). As Welsh, Sullivan, and Olds (2010, p. 115) point out, this involves “going to scale.” To use an analogy, companies develop products all the time and then test them in a limited way. But in the end, they must move from achieving positive results in the test market to selling the product profitably nationwide. In the same way, once an effective program is designed, it remains to be seen if it can “go to scale” and be used successfully across diverse contexts. In this dissemination process, some “discounting” or drop in the overall success rate is virtually inevitable (Welsh et al., 2010). The key is anticipating this possibility and taking steps to minimize it.

Remember that the original programs are typically invented and implemented by researchers. These folks are bright, well funded, and devoted to ensuring that their program is effective. As a result, the founding program works to reduce problem behavior. But when others try to transfer this intervention technology to another setting, the same positive results may not be forthcoming. This sometimes is called a researcher or investigator effect. This reality does not mean that the interventions that researchers design cannot be used effectively by other people. But it does mean that unless those who borrow the program do what the researcher did—unless they achieve therapeutic integrity—then the intervention is unlikely to be effective. Showing fidelity to the original program design can be difficult for a host of reasons: lack of resources, lack of organizational talent, hostile judges, a resistant staff, and so on. Even so, these are obstacles and not insurmountable barriers. Programs can be established effectively in multiple settings; it just is a challenge to do so. In short, we know what works—if it is done the right way.

Beyond the issue of ensuring fidelity to the program design when implementation occurs, Cullen and Jonson are going to tell you something else about why certain early intervention programs work: They adhere to the principles of effective treatment. Wait a minute! Didn’t we cover that in  Chapter 7  on rehabilitation when we talked about the Canadians, eh? Yes, but Cullen and Jonson have realized that the principles that guide effective correctional interventions also guide effective early intervention. To be sure, those who invented the exemplary early intervention programs had no idea that they were obeying principles made up by a bunch of Canadians. Nonetheless, in order to change human behavior—especially problem human behavior—programs sort of have to follow the same principles.

The Canadians’ approach is sometimes referred to as the RNR paradigm. Each letter stands for a core principle: R = risk principle; N = need principle; R = responsivity principle. Let us reiterate the key points:

· The risk principle says that interventions should be focused on high-risk individuals. This is because there is more to change and more future problems to prevent.

· The need principle says that interventions should target the known predictors of the problem behavior. These should be dynamic risk factors (also called criminogenic needs) in that they must be changeable. This also means that we must know what the risk factors are if we are to target them for change.

· The responsivity principle says that we should use interventions that are capable of changing the risk factors causing the problem behavior.

Guess what? Virtually all effective early intervention programs adhere to the RNR paradigm. They focus on high-risk cases; they identify risk factors and target them for change; and they use interventions that are responsive to—capable of altering—the risk factors that they have identified. This approach seems to make sense, doesn’t it? It is like medicine: Focus on people who are really sick and will not get better on their own; know what it is you are trying to cure; and then select medicine or surgery that can fix the problem. The RNR paradigm is thus a powerful strategy for saving adults and kids from a life in crime.

Now it is time to turn to five programs that have rightly earned national attention for their effectiveness. As we take a tour across these programs, keep an eye out for how they typically, if implicitly, employ the RNR paradigm. Also keep in mind what it might take to transfer these programs to settings across the nation and have them be effective.

Nurse Home Visitation Program

After graduating from Johns Hopkins University in 1970, David Olds started working in a day care center in West Baltimore for underprivileged children. Despite his best efforts, Olds felt that some of the youngsters were so troubled that it was too late—even by ages four and five—to reverse the damage. In a common story, Olds discovered that one “boy’s speech was so severely delayed because his mother was a drug addict and alcoholic, and had been using throughout her pregnancy” (Goodman, 2006, p. 7). Frustrated, Olds decided to learn more about children and to pursue a doctorate in developmental psychology at Cornell University. His mentor was the famous scholar Urie Bronfenbrenner.

Olds was anxious to make a difference in the lives of children. In 1975, he jumped into a part-time job at an organization—located in Elmira, New York—that “conducted programs intended to prevent health and developmental problems in young children” (Goodman, 2006, p. 8). Under the auspices of this organization, Olds had the opportunity to try to save the kind of kids he could not impact in West Baltimore. He understood that youngsters’ problems could be traced back to the earliest time in life—to the pre-natal and post-natal periods. This was especially the case for those born to at-risk mothers. A pressing need thus existed to intervene at this stage in life. But how? Who could establish trust and work with these moms? Olds had a remarkable insight: nurses!

Nurses were an ingenious choice for two main reasons. First, nurses are not threatening, judgmental figures. Rather, they are part of a helping profession that would legitimately provide expectant moms something they would like to have: health care. Second, given this inroad into the lives of at-risk women, nurses have the skills to do things that could change early predictors of crime and other wayward behaviors. In his blueprint, the nurses would work with first-time mothers, go to the moms’ homes, and start visiting early in pregnancy (hopefully in the first trimester). The intervention was intended to improve the health of the mothers’ pregnancy, to teach them how to parent a newborn, and to help them to envision a future that involved few pregnancies, schooling, and employment.

From a criminological standpoint, Olds’s approach is important because he was targeting for change pre-natal and post-natal criminogenic risk factors among high-risk mothers (Olds, 1998, 2007; Olds, Hill, & Rumsey, 1998). According to Olds et al. (1998), “to prevent youth crime and delinquency, it is important to understand how antisocial behavior develops and to design programs to interrupt that developmental pathway” (p. 1). He wished to impact “three important risk factors associated with early development of antisocial behavior” (p. 1). In his words, this is what those factors are:

· Adverse maternal health-related behaviors during pregnancy associated with children’s neuropsychological deficits.

· Child abuse and neglect.

· Troubled maternal life course.

The gist of his thinking was that one challenge was pre-natal: to get mothers to do healthy things (e.g., eat correctly) and to stop mothers from doing unhealthy things (e.g., smoking, taking drugs). This would mean that a child would have a better chance of being born without neuropsychological deficits. The next challenge was post-natal—after the child was born—and involved two parts. First, it would be essential to teach mothers how to parent newborns and toddlers so that they would be more caring and not react to their offspring by hitting or ignoring them. Second, mothers would be more likely to remain effective parents and good role models if they could improve their lives through schooling and/or employment. They needed to be helped to make this transition.

The key was to develop an intervention that was responsive to—that is, that could change—these risk factors experienced by high-risk moms and their children. Again, here is where the nurses came in! Pregnancy is a crisis for many women. The nurses would see the expectant mothers once a week to build up a bond of trust so that the moms would follow their advice to take steps to ensure a healthy pregnancy. As Olds (2007) notes:

During pregnancy, the nurses helped women complete 24-hour diet histories on a regular basis and plot weight gains at every visit; they assessed the women’s cigarette smoking and use of alcohol and illegal drugs and facilitated a reduction in the use of these substances through behavioral change strategies. They taught women to identify the signs and symptoms of pregnancy complications, encouraged women to inform the office-based staff about those conditions, and facilitated compliance with treatment. They gave particular attention to urinary tract infections, sexually transmitted diseases, and hypertensive disorders of pregnancy (conditions associated with poor birth outcomes). They coordinated care with physicians and nurses in the office and measured blood pressure when needed. (pp. 212–213)

The nurses would continue seeing the moms every other week until the baby’s birth when they would resume weekly visits for six weeks. Eventually, the visits would be spread out to every other week and then to once a month. They would end at the baby’s second birthday (Goodman, 2006). Overall, then, the program would last about two and a half years. During this time, the home visiting nurses “would focus simultaneously on the mother’s personal health and development, environmental health, and quality of caregiving for the infant and toddler” (Olds et al., 1998, p. 2). Put more simply, the nurse would help the mother to stay healthy, to give the right kind of care to her child, and to make good choices in her life.

The name of Prenatal and Early Childhood Nurse Home Visitation Program—that is a mouthful!—has since been shortened by Olds to the Nurse–Family Partnership or NFP. Programs based on the Olds model are now operating in 250 counties nationwide and serve over 20,000 mothers (Goodman, 2006; Olds, 2007). Many other variants of this home visitation program, not specifically designed by Olds but based on the same general idea, are in operation as well. By one estimate, “as many as 200,000 children and their families” are currently involved in home visiting programs (Gomby, Culross, & Behrman, 1999, p. 1). Cullen and Jonson think that this attempt to save children is pretty amazing! But with so many lives at stake, the next question only increases in its significance: Does this model actually work to improve the health and lives of mothers and, in particular, to reduce future antisocial behavior, including crime, among the children?

Based on reviews of the evidence, the answer appears to be “yes” (Farrington & Welsh, 2007; Greenwood, 2006; Piquero, Farrington, Welsh, Tremblay, & Jennings, 2009). Greenwood and Turner (2009) include the NFP in their list of “preferred programs,” which earn this designation because the extant research suggests that they are “proven, effective models” (p. 369). Still, we say appears to be “yes” because Gomby et al. (1999) have concluded that the effects of home visitation programs are mixed and modest. Part of the problem may be that many programs do not follow the Olds model fully. To save money, some interventions have visitations conducted not by nurses but by paraprofessionals—often high school graduates trained to intervene with the at-risk mothers. But lack of fidelity to the original design has a steep price: Evidence exists that nurses are an essential ingredient in program effectiveness (Goodman, 2006; Olds, 2007). Another issue to consider is that mothers and kids who finish the visitation programs do not receive any aftercare that builds on gains made early in life. More work needs to be done to explore how early gains can be sustained as at-risk youngsters negotiate at-risk environments.

With these qualifications, Olds has provided experimental evidence across three sites—Elmira, Memphis, and Denver—that his intervention is effective with mothers and with their children (Olds, 1998, 2007; Olds et al., 1998). Let us take as one example the evidence from Olds’s initial program in Elmira, New York. This evaluation included 400 pregnant mothers—200 randomly assigned to the treatment condition and 200 randomly assigned to the control condition (who also received some pregnancy services but no nurse visitations). In most areas examined, the mothers engaged in healthier behavior and more nurturing parenting. Most important, when evaluated at 15, the children who were visited by nurses—especially those of mothers who were poor and were unmarried—had fewer arrests, convictions, and violations of probation (Olds, 2007).

We want to mention one other program outcome recently reported by Olds and his colleagues (2014): the impact of nurse visitation on maternal and child mortality. Yes, we are talking about death deferred! Olds et al. explain why this result merits investigation. “Mortality,” they assert, is “an unequivocal outcome and its prevention aligns with the goals of the program and earlier program effects” (p. 804). They note further that “nurses are charged with improving maternal and child health by helping activate and support women’s motivations to protect children and themselves” (p. 804). Based on a two-decade follow-up (1990–2011), they do indeed find that nurse home visitations reduced both mothers’ mortality and what they call “preventable deaths” among children, which included “sudden infant death syndrome, unintentional injuries, and homicides” (p. 802). More broadly, these findings capture one of the important benefits of early intervention programs: They can have diverse, often-unanticipated positive effects on participants’ well-being across the life course.

High/Scope Perry Preschool Program

In the early 1960s, David Weikart was the director of special education in the Ypsilanti, Michigan, school district. Weikart and his colleagues were concerned that many youngsters experienced academic failure upon first entering the educational system. It seemed obvious that the need existed to enrich these children’s intellectual functioning. Trying to achieve reform across the school district seemed too complex. Instead, they decided to try to intervene before at-risk youngsters would enter the formal system. To do so, they selected youngsters who would eventually attend Perry Elementary School to participate in a preschool program (Schweinhart, 2007). The Perry Preschool is no longer in existence. In 1970, however, Weikart established the High/Scope Educational Research Foundation, which continued to follow children in the program for nearly 40 years (Parks, 1998; Schweinhart, 2007). This is how this intervention became known as the High/Scope Perry Preschool Program.

Weikart wanted to break the cycle of poverty. When underprivileged children experienced educational failure, they were less likely to complete high school and be economically successful. This poverty would then be transmitted to their offspring, who would enter the cycle of school failure, economic failure, and poverty. To knife off this intergenerational cycle of poverty, it thus was essential to ensure that children do well in elementary school, complete high school, and move into decent paying jobs (Schweinhart, 2007).

But what should be the components of an effective preschool intervention? Weikart and his buddies were really bright. They created a program that adhered to the RNR principles years before these ideas were set forth by the Canadians (Andrews & Bonta, 2010). First, they selected for the program African American preschool children ages three and four who were from low socioeconomic families and who had IQ scores “between 70 and 85, the range for borderline mental impairment” (Parks, 1998, p. 2). In short, the children were at high risk of educational failure. Second, they targeted for change the youngsters’ intellectual development. Of course, this leads to the third component: developing a responsive intervention. What did they come up with?

Well, from what Cullen and Jonson can determine, Weikart realized that the intervention had to be intensive. Similar to curing a serious disease, treating a serious human deficit needs high-dose treatment. A few flimsy tutoring sessions or some happy playgroup meetings would not do the trick. Weikart thus devised the following intervention (Parks, 1998; Schweinhart, 2007):

· Children would be in the program for two years.

· Five mornings a week, the kids would attend preschool for two and a half hours.

· The teachers would visit each child’s parents in the families’ homes one afternoon a week for one and a half hours.

· Parents were required to attend a monthly group meeting of parents that the staff facilitated.

· There was a teacher–student ratio of between 5 and 6.5—or four teachers assigned to every 20 to 25 children.

The next issue is the nature of the treatment. In various places, the program is described as being “based on an active learning model that emphasizes participants’ intellectual and social development” (Parks, 1998, p. 2). Okay, but what the heck does that mean? Part of the process is that teachers try to develop the kids’ “intellectual and social skills through individualized teaching and learning” (Schweinhart, 2007, p. 147). They do this, it seems, through a process that guides the children to internalize the ability to learn on their own. Schweinhart (2007) has done a nice job of explaining the program’s educational approach, noting that it is based on:

the natural development of young children. It emphasizes the idea that children are intentional learners, who learn best from activities that they themselves plan, carry out, and review afterward. Adults introduce new ideas to children through adult-initiated small- and large-group activities. Adults observe, support, and extend the children’s play as appropriate. Adults arrange interest areas in the learning environment; maintain a daily routine that permits children to plan, carry out, and review their own activities; and join in children’s activities, asking appropriate questions that extend their plans and help them think about their activities. They add complex language to the discussion to expand the children’s vocabulary. Using key developmental indicators derived from child development theory as a framework, adults encourage children to make choices, solve problems, and engage in activities that contribute to their intellectual, social, and physical development. (pp. 148–149)

This all sounds wonderful, but Weikart and the Perry Preschool Program would have lapsed into historical oblivion if not for one consideration: data! Weikart did not intend simply to build a program but rather wanted to develop a program that worked. Having confidence in his ideas, he decided to test his program and get good data through an experimental design. To do this, he and his staff recruited 128 African American male and female children to participate in the study. Half were randomly assigned to the treatment or preschool group and half to the control group. Due to attrition, the study lost five cases. Remember that folks like Weikart who put their ideas to the test have, well, guts. Data are dangerous; they can show that one’s ideas are pretty stupid and that years of well-intentioned child-saving efforts were a waste of time.

Of course, Cullen and Jonson would not be discussing Weikart and his intervention if the data had shown that the Perry Preschool Program was a bust! In fact, the outcomes were amazingly positive. Compared to the control group, program participants were less likely to be placed in a special education program and were more likely to graduate from high school. They had lower rates of births out of wedlock and of welfare assistance. And they earned more money and were more likely to own a home (Parks, 1998; Schweinhart, 2007). Still, from our perspective, the key consideration is whether any of these is relevant to crime. Again, as you may have imagined, the answer is “yes.”

Human beings tend not to be very compartmentalized. If they do well in one area of their development, they tend to do well in other areas. Unfortunately, if they tend to have problems, these difficulties tend to pervade their lives. This is what criminologists call the generality of deviance or the co-morbidity of problems. Thus, those who hit and steal from others also tend to smoke, abuse drugs and alcohol, become sick, be accident prone, get a lot of tattoos, fail at school, lose jobs, and experience unstable relationships. The important insight here is that if intervention programs succeed in fixing kids so that the youngsters succeed at school and thereupon in other domains of their lives, then they also tend to have the collateral effect of reducing involvement in juvenile delinquency and adult crime. Alas, although not intended as a crime prevention strategy, the Perry Preschool Program diverted a lot of participants from criminal activities across their life course (Parks, 1998; Schweinhart, 2007).

Schweinhart (2007) has presented data based on a follow-up of the Perry Preschool kids at age 40. All readers should stop and realize, as Cullen and Jonson have done, how darn amazing this is. Can you imagine keeping track of study participants for nearly four decades (the kids started the program at age 2 or 4)? Such a follow-up is the gold standard of a longitudinal study. It does not get much more longitudinal than this!

In any event, by age 40, the program participants had experienced a lot fewer legal entanglements (Schweinhart, 2007, p. 150). This is not meant to imply they were free from system contact. As we know, for a host of reasons, disadvantaged African Americans—such as those in the Perry Preschool Program—are prone to justice system processing. In fact, in the preschool treatment group, only 29% had never been arrested. Even so, compared to non-participants who were assigned to the control group, their odds of an arrest over their lifetime were 46% lower. And if arrested, they experienced fewer numbers of arrests. Further, the percentage never arrested for major crime categories was consistently higher for those from the program: violent offenses (67% vs. 52% for the control group), property offenses (64% vs. 42%), and drug offenses (86% vs. 66%). Only 28% of the preschool kids ended up spending time in jail or prison; the statistic for the control-group kids was 52%. Finally, the difference in criminal participation between the preschool and control groups became wider as the sample aged. This finding is surprising because in most instances program effects tend to weaken as time passes and participants are farther away from the original intervention.

Beyond intellectual development (IQ is a predictor of criminal involvement), Schweinhart (2007) offers the interesting thesis that the Perry Preschool Program may also have lowered the participants’ impulsivity. Recall the description of the program presented above. It may be that the program’s focus on kids’ planning, decision making, and cognitive awareness instilled (much like a cognitive-behavioral program) more self-control. Impulsivity or a lack of self-control (including ADHD) is an empirically established predictor of crime, largely, it is believed, because this trait prompts people to act before they think—in particular, before they think of the consequences of what they are about to do (Farrington & Welsh, 2007; Gottfredson & Hirschi, 1990). As Schweinhart (2007) suggests, if inculcated with the capacity to handle impulsivity, those in the program group may have been “more purposeful in avoiding” criminal choices (p. 157).

There is general agreement that the High/Scope Perry Preschool Program is an effective intervention and that similar early enrichment programs have shown success (Farrington & Welsh, 2007; Greenwood, 2006; Greenwood & Turner, 2009; Reynolds, Temple, Ou, Arteaga, & White, 2011; Schindler & Yoshikawa, 2012). In fact, high-quality findings tend to have “robust” effects across a variety of outcomes (Farrington & Welsh, 2007, p. 119). Again, however, the key qualification here is that the programs must be high quality. Do things the right way, and it seems that prosocial development will be promoted and meaningful savings in criminal involvement will be achieved (Schweinhart, 2007).

Cullen and Jonson find two other aspects of the Perry Preschool Program and similar enrichment programs noteworthy. First, James Heckman, a Nobel Prize-winning economist from the University of Chicago, has become a prominent advocate of early intervention programs, including his book, Giving Kids a Fair Chance (2013). Using sophisticated statistical techniques (something that economists do for a living!), Heckman and colleagues examined the “rate of return” of the Perry Preschool Program (Heckman, Moon, Pinto, Savelyev, & Yavitz, 2010a, 2010b). They noted that the cost-benefit ratios were not as high as those found by previous analyses using less rigorous statistical controls. That said, they reported that the results remained statistically significant in favor of program participants. In more concrete language, they noted that the annual return for the program was between 7% and 10%. This figure surpassed the post-World War II “annual return to equity” in the stock market, which is “estimated to be 5.8 percent before the 2008 meltdown” (Heckman, 2011, p. 35). Based on these findings, Heckman (2013, p. 31) now touts early childhood education as a means of addressing the “economics of inequality.” In particular, he argues for what he calls “predistribution—improving the early lives of disadvantaged children”—rather than trying to redistribute incomes for people later in life (Heckman, 2012, p. 11; emphasis in the original). “Predistribution policies,” he argues, “are both fair and economically efficient” (2012, p. 11).

Second, the early enrichment interventions—such as the Perry Preschool Program and Head Start—hoped to boost at-risk kids’ IQ and standardized test scores. As Duncan and Magnuson (2013) report, the initial gains achieved by these programs tend to decay gradually and to disappear after 10 years. But here is the catch—actually, here is the good news. While these intended intellectual consequences peter out, the programs have a lot of unintended consequences in various aspects of participants’ lives that seem to persist deep into adulthood! Duncan and Magnuson (2013, p. 120) call this “the puzzle—academic fade-out, but long-term benefits.” These include “beneficial impacts on a broad set of later-life outcomes like high school graduation rates, teen pregnancy, and criminality” (p. 120). Understanding why this occurs is challenging. One possibility is cumulative advantage: Program kids enter school and succeed, which in turn leads them to eventually graduate from high school and avoid drugs and crime. Another possibility is that the intervention inadvertently builds other personal skills—such as self-efficacy, the capacity to pay attention, and emotional intelligence—that are needed to succeed in various life domains. Regardless, one reason why early intervention programs are valuable—and cost effective—is that they have the potential to improve kids in ways that have multiple and cascading benefits.

Functional Family Therapy

It is commonly believed that if you grow up in a troubled family, you are likely to live a troubled life. To a large extent, this wisdom turns out to be, well, wisdom. Scholars have reviewed many studies, and they tend to agree that family factors are consistent predictors of early misconduct and of subsequent offending. It also appears that the more bad family factors children are exposed to, the more likely they are to be wayward. This effect is called cumulative risk (Farrington & Welsh, 2007). As Farrington and Welsh (2007) conclude with regard to families and offending, “the strongest predictor is usually criminal or antisocial parents. Other quite strong and replicable predictors are large family size, poor parental supervision, parental conflict, and disrupted families” (p. 74; see also Lipsey & Derzon, 1998; Petrosino, Derzon, & Lavenberg, 2009).

These findings point to a potential avenue for early intervention. How about if we teach parents to be better parents? Will that make a difference? One complication is that some research suggests that parenting style—how kids are parented—is not a strong predictor of conduct problems and offending (Wright et al., 2008). Scholars thus challenge the nurture assumption—the idea that kids enter the world as a blank slate and that their destiny is determined by how their parents socialize them. Instead, they argue that it is mainly through nature or genetics that parents really transmit advantages—and problems—to kids (Harris, 1998). For example, impulsive parents with ADHD may raise their children in a lousy way. When their kids then get into trouble, why is this? Is it because of the lousy parenting they have received or because their parents’ ADHD and impulsivity have been inherited, which then causes them to act waywardly?

This is not the place to disentangle this hotly debated issue, but we can make an important point: Even if bad parenting is not the cause of problem behavior, this does not mean that good parenting is not the solution to decreasing such conduct. Recall from earlier in the chapter the idea of genetic fatalism. Let us assume, again, that a child has ADHD and gets into trouble by acting impulsively. If mom and dad are taught how to parent in a way that helps the child to cope more effectively with impulsivity, then the child may develop increased impulse control and avoid making bad decisions. That is, parents may learn how to employ cognitive-behavioral techniques to manage their youngsters in a way that curbs criminal propensities—regardless of whether these propensities are due to social or biological causes. In fact, there is now fairly substantial evidence that interventions that teach parental management of children work to reduce offending (Farrington & Welsh, 2003, 2007; Greenwood, 2006; Greenwood & Turner, 2009; Petrosino et al., 2009; Piquero et al., 2009). Thus, in their meta-analysis of the impact of parental management training programs on antisocial behavior and delinquency, Farrington and Welsh (2003) reported a mean effect size of .395. This meant that the programs achieved a “20% reduction in antisocial behavior/delinquency (e.g., from 50% in a control group to 30% in an experimental group)” (Farrington & Welsh, 2007, p. 127).

Teaching parents how to manage their children more effectively is important because, as Farrington and Welsh (2007) state, “crime runs in families” (p. 57). To be sure, there are families where there is the so-called black sheep—the single bad child in an otherwise good family. In fact, those asserting that parenting has no effect on any youngster’s behavior use this empirical reality as evidence in support of their perspective. After all, if two kids share the same family environment yet one turns out good and the other bad, how can parenting be said to cause both outcomes? Cullen and Jonson have no insightful response to that conundrum. But the other reality is that in at-risk families, as opposed to families from across the spectrum, it is often the case that crime is highly concentrated—it runs in families. That is, most of the family members—parents, child, siblings—are all troubled. An important implication of this fact is that families such as these are inviting targets for early intervention.

As is likely apparent, in this chapter we often make a statement and then cite some research by David Farrington. This is because he is very bright and publishes a zillion articles—all of them good ones! Google him or go look him up on the website of Cambridge University—the one in England! Well, here we go again. In this area, Farrington and his colleagues conducted two studies that show the extent to which offending is concentrated in at-risk families—one based on his Cambridge Study in Delinquent Development and the other on the PYS (Pittsburgh Youth Study) data. Both studies show virtually the same thing, which gives us confidence that the results hold across social context (London and Pittsburgh) and types of samples (mostly White and a majority African American).

The Cambridge Study in Delinquent Development is a longitudinal study of 411 males born in the working-class section of South London. They were drawn from 397 families. Most of the sample members were born in 1953, and then first contacted by a research team led by David West in 1961 or 1962. (David Farrington joined the research team in 1969 and has been director of the Cambridge Study since 1981.) These lads were followed from as young as age 8 to the age of 32, during which time they were interviewed on multiple occasions. Data from records were collected on them from ages 10 to 40 (Farrington, 2003; Piquero et al., 2007). In the study of interest, Farrington, Barnes, and Lambert (1996) examined how the convictions of these males “relates to the convictions of their biological fathers and mothers, full brothers and sisters, and wives” (p. 47).

The key issue is whether offending is spread evenly across families or is concentrated in a subset. To start with, nearly two thirds of the families (64%) had some family member convicted of a crime. But here is the key finding: Only 5.8% of the families accounted for 49.9%—or about half—of all the convictions experienced by the study’s families. Further, just over one tenth of the families (10.3%) accounted for 64.3% of the convictions. To hit the point home a bit more, here are other findings quoted from Farrington et al. (1996, p. 47):

· Convictions of one family member were strongly related to convictions of every other family member.

· About three quarters of convicted fathers and mothers had a convicted child.

· In both generations, the majority of convicted mothers mated with convicted fathers. (Note from Cullen and Jonson: This is called assortive mating—or birds of a feather flocking together!)

Well, it could be that there was something screwy about males born in 1953 in South London. Maybe if they drank water with fluoride, as we do in the United States, crime would be dispersed and not so concentrated—and they would have healthier teeth! Maybe if they rooted for Manchester United and not Chelsea, their lives would be more in order. Just some thoughts from Cullen and Jonson! But, alas, this is not the case. When Farrington, Jolliffe, Loeber, Stouthamer-Loeber, and Kalb (2001) examined data from the PYS—a Pittsburgh sample first contacted in 1987–1988—similar results were found. In fact, they discovered that the “concentration of offenders was greater in Pittsburgh than in London” (p. 586).

Cullen and Jonson think that these findings are immensely important. If such at-risk families are targeted for intervention, it might be possible to reduce the offending of multiple children in the household. It also might be possible to channel these youngsters along a life course where they would eventually select prosocial mates, which in turn would block the intergenerational transmission of criminality. Let us put it another way. The concentration of offending means that some families are really messed up and pose a daunting threat to society. But if an intervention were to work, then it could do a great deal of good—for the families and for public safety.

In this regard, we again reiterate that parental management interventions have been found to be effective. Here, we want to focus on one with a good track record: Functional Family Therapy, also known by its acronym FFT. Cullen and Jonson think that it should have been called Dysfunctional Family Therapy since, after all, the families getting fixed are dysfunctional! But those in the treatment business are an optimistic bunch and prefer to give their programs nice-sounding names. In this case, the program inventor is James Alexander. He founded the program in 1969 at the University of Utah where the Department of Psychology had a Family Clinic. He chose the name family because this was the core unit or focus of the intervention. He chose the name functional—positive guy that he is!—because of “the overriding allegiance to positive outcome” (Alexander, Pugh, & Parsons, 1998, p. 7). Today, FFT is in 220 sites, including Europe, New Zealand, and nearly every state in America (FFT, 2010).

Alexander and his colleagues were quite shrewd in how they developed FFT. They realized that a viable program had to be practical. In this regard, they decided—as all good programs do!—to focus on at-risk antisocial youths. In this case, they targeted youngsters ages 11 to 18 who were at risk of institutionalization. They also understood that lengthy psychotherapy was a non-starter when funding was coming from the public treasury. Freud might not be happy, but nobody wants to pay for troubled families to spend a couple of years in psychoanalysis trying to figure out who is envious of what. Rather, they designed the program to be completed in three months. Staff would meet with families in one-hour sessions that numbered from 8 to 12 for “mild cases” and from 26 to 30 for youths in family situations that were “severely dysfunctional” (Alexander et al., 1998, p. 8; see also Onedera, 2006).

FFT is extremely well planned. It is divided into five stages (Alexander et al., 1998). The first two, engagement and motivation, are designed to ensure that at-risk families will enter and develop favorable views about the intervention. In the engagement stage, the therapist’s job is to engage the family and the antisocial youth and to take steps to retain them in the program. The main strategy is to create a belief that positive change can occur. This occurs in the first session. In the motivation stage, the special challenge is to focus on things like negativity, hopelessness, and blaming within the family. If strides can be made, this shows everyone involved that change is possible. If so, then these positive experiences may increase the desire to stay in the program and make a commitment to further change. As Alexander has pointed out in an interview, without “engagement and motivation, all the rest of the phases, even if they are done perfectly, are basically almost doomed to failure” (Onedera, 2006, p. 308). Put simply, before you can fix people, you gotta get them involved!

The third stage, assessment, is when the therapist tries to understand what the heck is going on in the family. In figuring this out, the therapist does not look aimlessly but uses the program’s conceptual framework; this is like putting on FFT glasses that channel the therapist’s vision to look for some things but not others. The challenge here is to determine what needs to be targeted for change. Based on the empirical literature, FFT has identified a range of child, intrafamilial, and contextual risk and protective factors that are likely to affect outcomes (Alexander et al., 1998, pp. 16–17). As an intervention based on crime-producing family dynamics, the main focus is on detecting poor parenting skills and negative (or belittling and blaming) communication or, conversely, on detecting any positive parenting skills and communication that might exist and can be built upon. Based on this assessment, a plan for change individualized to the needs of the family and child is developed.

Getting back to the RNR model, FFT is meant for high-risk families and youths (one R) and focuses on sources of criminality (the N). But what about the second R—responsivity? Well, this is undertaken in the fourth stage of FFT, which Alexander calls behavior change. So, what do you think Alexander came up with? If the sources of problem behavior involve the inability of parents to communicate effectively and to parent their child, what should the intervention do? What would be a treatment that would be responsive? You got it! The FFT therapist works with the family to train them to communicate in a positive way (e.g., how to listen and to interact in a direct, concrete, and optimistic way) and to train adults in effective parental management strategies (e.g., how to use praise, to ignore irrelevant acts, to set limits, to monitor, and to set up behavioral contracts where rewards are tied to achievable good conduct).

In essence, FFT employs a cognitive-behavioral approach, which as you know from  Chapter 7  is pretty darn effective in addressing problem behavior. As Alexander notes, “people have belief systems, and they have ways of thinking. They also have behavior patterns. They also have emotional reactions. You need to address all of those, and you need to address them differently based on different people” (Onedera, 2006, p. 309). FFT is oriented toward giving people the “tools” to transform how they think, behave, and feel. According to Alexander, “these include emotional tools, cognitive tools, and behavioral tools to initiate and maintain the short-term and then long-term behavior changes that are necessary for them to be more adaptive within the family and outside the family” (p. 309).

By following the sequence from stages 1 to 4, FFT hopefully has achieved effective change within the family. But as the intervention—again, which lasts three months—starts to wind down, the therapist must enable the family and problem youth to negotiate the context in which they live. Alexander calls this the generalization stage. Some risk factors in this context cannot be altered, such as the presence of gangs in the neighborhood. Still, this is a time when the therapist can collaborate with community systems to undercut potential risk factors and to take advantage of opportunities for support. For example, the therapist might work with teachers to address learning deficits, direct a youth to participate in prosocial community recreational centers, or actively assist family members to secure treatment in self-help groups for substance abuse problems.

Most noteworthy, FFT has been shown to reduce a range of problem behaviors. Across 13 experimental or quasi-experimental studies, with outcomes ranging from offending to foster care placement, consistent positive effects have been reported. Compared to control groups (random or matched samples), problematic outcomes have been reduced between 25% and 60%. One study also discovered that positive effects extended to siblings. More studies are ongoing (Alexander et al., 1998; FFT, 2010).

Seattle Social Development Project

To be effective, early interventions need to operate in those social domains or systems where kids reside. One of these is the school. It is estimated that from Grades 1 to 12, youngsters spend 18% of their waking hours within educational institutions (Gottfredson, Wilson, & Najaka, 2002). The other reality is that for youths, a “disproportionate amount of crime occurs in or around school buildings” (Gottfredson et al., 2002, p. 149). Bullying is another form of student victimizing behavior that can have damaging physical and psychological consequences.

When kids have trouble at school, it can cause them to have trouble outside of school. Put another way, some important risk factors for crime include academic failure and weak bonds to the school (Gottfredson et al., 2002). Further, it appears that certain kinds of schools—those marked by poor classroom management, distrust between teachers and students, and inconsistently enforced rules—produce higher rates of misconduct inside and outside the school grounds (Farrington & Welsh, 2007).

Fortunately, there is mounting evidence regarding the effectiveness of school-based interventions to reduce kids’ antisocial conduct. Some do not work particularly well (such as the D.A.R.E. program), but others do work well (Baldry & Farrington, 2007; Catalano, Arthur, Hawkins, Berglund, & Olson, 1998; Catalano, Loeber, & McKinney, 1999; Farrington & Welsh, 2007; Gottfredson et al., 2002; LeMarquand & Tremblay, 2001; Mytton, DiGuiseppi, Gough, Taylor, & Logan, 2002; Najaka, Gottfredson, & Wilson, 2001; Wilson, Gottfredson, & Najaka, 2001; Wilson & Lipsey, 2007; Wilson, Lipsey, & Derzon, 2003; see also Gottfredson, 2001). Consistent with the risk principle, evidence exists that the program achieves greater reductions with higher-risk youth (Wilson et al., 2003). Importantly, Gottfredson et al. (2002, pp. 176–182) have highlighted programs that seem to be most effective:

· Programs that improve school and discipline management. These can be multifaceted, involving efforts to bolster school climate, to increase achievement, to target at-risk students for special assistance to avoid academic failure and behavioral problems, and to enhance the consistency of student discipline.

· Programs that attempt to establish norms and expectations about acceptable behavior. This might include instructing students that bullying will not be tolerated and supervising students to ensure that this victimizing conduct does not occur. Or it might include correcting misconceptions that substance abuse among classmates is widespread (when the silent majority does not engage in serious abuse) and teaching students how to resist peer pressure to consume alcohol and drugs.

· Programs that use instructional strategies or cognitive-behavioral interventions to improve students’ coping and decision-making skills. This might involve teaching students how to exercise self-control, to manage their anger, to consider alternatives to deviant ways of solving problems, to think about possible consequences that actions might have, and to be more sensitive to their classmates.

One exemplary school-based intervention strategy is the Seattle Social Development Project, also known by its acronym SSDP (Hawkins, Smith, Hill, Kosterman, & Catalano, 2007; Hawkins et al., 2003). The SSDP is directed by J. David Hawkins, with Richard F. Catalano serving as the project’s associate director. As with all effective programs, the inventors are smart people. Or, to put it another way, they are really good criminologists.

Hawkins and Catalano realized two things early on. First, they knew that “to be effective in stopping crime before it happens, crime prevention efforts must address factors that predict crime” (Hawkins et al., 2003, p. 270). Duh! Of course! But until the 1980s—about the time they started the SSDP—this knowledge did not exist, programs were largely blind in what they targeted for change, and thus few school-based interventions were found to be effective (Hawkins et al., 2007). Second, Hawkins and Catalano knew that schools are invaluable settings in which to locate a prevention project. They comprise the one place that gives nearly universal access to youngsters in a community. Further, things done in and through schools can change students’ lives. According to Hawkins et al. (2007), “many of the factors that predict problem outcomes are accessible through schools” (p. 162).

The SSDP started as an intervention with first graders in five Seattle public schools in 1981 and then was expanded into a longitudinal study four years later. The sample included 808 students, who were age 10 and in the fifth grade in 1985. They attended 18 Seattle elementary schools that served high-crime neighborhoods. The sample members were followed into early adulthood and beyond (Hawkins et al., 2007; Hawkins et al., 2003).

Cullen and Jonson think that the neat thing about the SSDP is that it was based on a coherent theoretical model. Hawkins and Catalano did not just go out and start intervening aimlessly. Rather, they built their program around what they believed were the two most empirically verified theories of their day: differential association/social learning theory and social bond theory. Many readers will recognize these theories; you probably have gotten multiple-choice questions correct on these perspectives in your introduction to criminology course! The first theory, popularized by Edwin Sutherland and Ronald Akers, says that crime is learned through interaction with others and is repeated when reinforced. The second theory, popularized by Travis Hirschi, says that the more closely bonded to society individuals are, the less likely they are to break the law (for summaries, see Lilly et al., 2015). These theories are typically seen as rivals, but Hawkins and Catalano understood that learning and bonding can happen at the same time. Duh! What’s more, they added four important twists to create what they called the social development model.

First, in advancing one of the earliest life-course theories, they noted that differential association/social learning and social bond theories had a common weakness: They ignored childhood and were, in essence, adolescence-limited. But Hawkins and Catalano asserted something that, while obvious today, was not obvious back in the 1980s: That these perspectives had to be extended back into childhood and placed within a developmental framework. Ergo, the name they gave to their theory: the social development model. Second, they knew that kids did not begin a developmental pathway as blank slates. Accordingly, they identified three sets of factors that affect whether children’s social development gets off on the right or wrong foot: (1) individual constitutional factors or traits; (2) a child’s location in the social structure (e.g., race, class); and (3) something called external constraints, which seems to mean how prosocial or antisocial their parents and peers are.

Third, they then said that youngsters were likely to leave the train station of childhood on one of two developmental tracks: prosocial or antisocial. Their constitutional, social structural, and external constraint factors increased their opportunities to begin life with an abundance of either prosocial or antisocial interactions. These interactions would shape children’s involvement in activities, social bonds, and moral beliefs. Importantly, if kids could be directed early in life into prosocial activities, they would likely develop prosocial bonds and prosocial moral beliefs and, voilà, we have a prosocial life course! Of course, the opposite pathway had the same components, except that they were antisocial in nature and led to criminological hell rather than to prosocial heaven. Thus, one pathway brought exposure to protective factors, whereas the other brought exposure to risk factors.

Fourth, Hawkins and Catalano had one final crucial insight. In order to keep on a prosocial pathway, youngsters had to find the experience rewarding; that is, they had to be positively reinforced. But in order to receive reinforcements, they had to have the appropriate skills for interaction. If they had deficits and ended up failing at school or aggravating and being rejected by other kids, they might decide that all this prosocial stuff is worthless and decide to flock over to the dark side with the antisocial birds, so to speak; here they could be successful messing up and be accepted for it. Most theories of crime simply do not talk about the kinds of skills needed to take advantage of prosocial opportunities; Hawkins and Catalano’s did. Perhaps not surprisingly, subsequent empirical tests of their model have been mostly favorable (Hawkins et al., 2003).

Now on to the SSDP intervention! Given their theory, Hawkins and Catalano decided to save kids in three ways—again, all of which were school-based. First, parents who volunteered were given classes to increase their skills in behavior management, in supporting their children’s academic performance, and in helping their kids to exercise self-control and to contribute to the family. Second, for the children, efforts were made to enhance their social and emotional skills, including communication, problem solving, conflict resolution, and learning how to refuse to do bad things. This was accomplished through a cognitive and social skills curriculum. Third, teachers were trained in classroom management techniques, how to engage in more interactive instruction with students, and how to encourage cooperative learning among their youngsters (Hawkins et al., 2007, p. 171; Hawkins et al., 2003).

Again, these programs were oriented toward undercutting risk factors and increasing the likelihood that kids would find prosocial activities and relationships rewarding. Phrased differently, Hawkins and Catalano carefully designed interventions that were responsive to criminogenic needs (or risk factors)—then delivered to kids attending elementary schools in high-risk neighborhoods (although, due to integration, not all kids were from the local area). RNR sort of revisited, once again.

The youngsters finished the program by the sixth grade. Some had been receiving the intervention since the first grade, some only in the fifth and sixth grades (and some of those only the parenting part). The control group, of course, received no intervention. The key issue is whether any of this would do any good many years later. Well, the gist of the findings is as follows: For those who received the full intervention—first through sixth grade—the results were pretty impressive. Here are some key accomplishments for the full intervention group at age 18 (Hawkins et al., 2007, pp. 176–177):

· Greater school attachment and commitment. Higher grades and less misbehavior at school.

· Less reported violent behavior in their lifetime and less heavy alcohol use in the past year.

· Less sexual activity and sexual partners.

By age 21, the treatment group still achieved “broad significant effects on functioning in school and work, on emotional and mental health, and on risky sexual practices and adverse health outcomes” (p. 177). Differences in crime and substance use narrowed by age 21, “although those in the full intervention group were significantly less likely than the controls to be involved in a variety of crimes, to have sold drugs in the past year, or to have received an official court charge in their lifetime” (p. 177).

Multisystemic Therapy

Multisystemic therapy—also known by its acronym MST—is one of the most popular early intervention strategies. It is estimated that MST programs serve approximately 10,000 youngsters in 30 U.S. states and in 11 nations (Jonson & Cullen, 2011). That’s a lot of kids in a lot of places! Where did MST come from? What does it involve? And does it work? Each of these questions is addressed below.

MST is the invention of Scott Henggeler. The ideas first coalesced in 1978 when Henggeler and his psychology students at Memphis State University developed a juvenile diversion program. (He is now at the Medical University of South Carolina in Charleston.) Recall that this was not long after Martinson’s (1974) nothing works doctrine emerged and was dominating thinking about corrections. However, Henggeler was a psychologist, not a criminologist, so he was not about to believe that kids were beyond redemption. After all, a lot of psychologists would be out of jobs if they agreed that therapy didn’t work!

Indeed, Henggeler thought that he knew the real nature of the problem. It was not that antisocial youngsters could not change but rather that they were being treated with the wrong medicine, so to speak (Cullen, 2005). At the time, the traditional psychotherapeutic model required troubled youths, and perhaps their parents, to come weekly to an office for an individual session lasting an hour. The length of treatment was open ended and could last months, if not years, on end. Recalcitrant youngsters were candidates for placement in a residential facility. This approach often proved costly, was ineffective in reducing antisocial conduct, and often failed to reach inner-city youths.

This individual psychotherapeutic model typically had a psychodynamic, rather than a sociological, orientation. It tended to look inside a youth’s mind to unravel what was wrong with him or her. But most kids in the justice system not only were personally troubled but also were drawn from troubled contexts or systems. As part of a community psychology movement, Henggeler was aware of these realities and was convinced that these multiple systems had to be addressed if kids were to be effectively treated. More specifically, he embraced the social-ecological model, which “depicts the process of human development as a reciprocal interchange between the individual and ‘nested concentric structures’ that mutually influence one another” (Henggeler, 1999, p. 2). That is, youths not only have individual traits but also are affected by such social systems as families, peers, schools, and community. In short, they live in multiple systems. Cullen and Jonson hope readers are getting the point! The concept of multiple systems has been italicized twice because it is the core of multisystemic therapy. If the risks for crime come from multiple systems, the intervention darned well better take that into account!

Empirical research has demonstrated that along with individual traits, each of these systems can expose a child or adolescent to criminogenic risk factors that increase the likelihood of law-breaking conduct. For example, a family risk factor might include ineffective parental discipline; a peer risk factor might involve contact with antisocial friends; and a school risk factor might be low achievement. For each troubled youth, the particular set of risk factors underlying his or her misconduct is likely to be unique. The intervention thus must be aware of this fact and be flexible enough to address risk factors drawn from different social systems.

So far, so good. But what exactly does MST involve? Similar to other inventors of classic intervention programs, Henggeler is a really clever guy—with clever being used in the best sense of the word! It makes little sense to ask, for example, a single mother—who might be a working mom—to try to get her at-risk son (or daughter) and perhaps a couple of siblings to board a bus and travel downtown to some psychotherapist’s office where, like Cullen (but fortunately not Jonson), the therapist can stroke his beard and say, “Do I hear you saying you want to beat someone up? Do you want to go with those feelings?” Instead, Henggeler understood that he would bring the intervention to the troubled youngster. Go to the home—or, on occasion, to some close-by community setting, such as the kid’s school.

To provide meaningful therapy, each MST clinician has a small caseload that is limited to about five families, each of which receives 2 to 15 hours of intervention each week. Recognizing the uncertain timing of crises, clinicians are available 24 hours a day, seven days a week. The MST intervention is intended to be intensive but also time limited (about four to six months), a factor that contributes to its cost effectiveness (Henggeler, 1997, 1998, 1999).

Cullen and Jonson again believe that, as with other successful early intervention programs, the RNR model implicitly informs Henggeler’s MST. His intervention focuses on high-risk kids (ages 12 to 17). In his words, these youths possess “serious behavior disorders,” and include those who are “violent and chronic juvenile offenders” (1997, pp. 1, 6) and who are “at high risk of out-of-home placement” (1998, p. 3). Clients are assessed individually, with the focus on their criminogenic needs or with an “emphasis on addressing the known causes of delinquency” (p. 6). Finally, treatment interventions are selected that are responsive to—or capable of changing—these targeted risk factors. In Henggeler’s (1999) words, “MST addresses the known determinants of clinical problems” using services “with high ecological validity” and “targeting processes directly in home, school, and neighbourhood contexts that are linked with identified problems and that can serve as protective factors” (p. 4).

In more concrete language, when initiating the intervention, the MST clinician and supervisor take the responsibility to engage or hook the youth and family into the treatment. (This is similar to the engagement and motivation stages of Alexander’s FFT.) An important task is to diagnose the risk factors or problematic relations that are contributing to the youngster’s misbehavior. This might involve persistent conflict with parents, drinking with friends, and truancy. The intervention attempts to build on individual (e.g., high IQ) and system (e.g., a supportive teacher or coach) strengths in a positive way. Services are individualized to meet the unique needs of each troubled youngster. The specific interventions used are adapted from “pragmatic, problem-focused treatments that have at least some empirical support. These include strategic family therapy, structural family therapy, behavioural parent training, and cognitive behaviour therapies” (Henggeler, 1999, p. 3).

Henggeler (1997, 1998, 1999, 2011, 2015) has amassed a fair amount of experimental evidence showing that MST is effective in reducing antisocial behavior. We should note that Julia Littell (2005) stirred up some controversy when she published a meta-analysis suggesting that the effects of MST were suspect when complex methodological issues were taken into account. However, other scholars have reported more favorable results. Thus, in their meta-analysis, Farrington and Welsh (2007) concluded that MST was “the most effective family-based approach” (p. 135). The mean effect size was .414 (p. 107). Drake, Aos, and Miller (2009) assessed 10 MST studies that focused on offenders in the juvenile justice system. They calculated a shrinkage in crime recidivism of 7.7%. Further, in a meta-analysis of 11 studies conducted by Curtis, Ronan, and Borduin (2004), MST was found to significantly lower criminal behaviors. MST was associated with a reduction in the number of arrests for all crimes, the number of arrests for substance abuse crimes, seriousness of arrests, number of days incarcerated, and self-reported delinquency and drug use. Similarly, Borduin and his colleagues (1995) explored whether MST would have long-term (four years) preventative effects on crime, including violent offending. When comparing juvenile offenders placed in MST versus individual therapy, Borduin et al. determined that MST was more effective in reducing rearrrests. Specifically, 71.4% of the individual therapy youths were rearrested within four years, whereas only 26.1% of the MST youths were rearrested. Finally, van der Stouwe, Asscher, Stams, Deković, and van der Laan (2014) recently published perhaps the most sophisticated meta-analysis, including 22 independent studies. The results fell in between Littell’s (2005) assessment and those reaching more promising conclusions. The effects of MST on general delinquency were modest but meaningful (d = .233). Several variables moderated the effect. For example, MST effects were larger for youths who were offenders and under 15 and when the control group received a “single and non-multimodal control treatment type” (van der Stouwe et al., 2014, p. 472).

Cullen and Jonson decided to find out for themselves how well MST worked. We realized that obtaining all evaluation studies on MST—published and unpublished—would be a lot of work. So, we got Jennifer Lux to take on this task, guiding her efforts along the way (i.e., she did all the work and we rode her coattails!). Lux’s (2010) results are mostly good news for MST. Lux analyzed 21 studies that had delinquency or problem behaviors as the dependent variable; 16 of the studies used a random experimental design. The control groups typically received the usual services given to kids in trouble. This is important, because MST is not often compared to doing nothing. Although not all differences were statistically significant, the general pattern was that MST consistently outperformed the control group in manifesting less antisocial behavior. Further, when Lux assessed a larger group of 45 studies, MST evidenced positive effects on a range of other outcomes (e.g., mental health, family functioning, diabetes, and HIV/AIDS measures) (see also Cullen, Lux, & Jonson, 2012).

Two More Reasons to Support Early Intervention

Let’s pretend that you are a policy maker—maybe a governor, mayor, city council member, or director of a social services agency. If asked by a constituent whether you intend to expand early intervention programs, what would you say? Well, Cullen and Jonson have already given you two good reasons for supporting child-saving programs. The first reason, based on life-course criminology, is that today’s highly troubled kids are tomorrow’s serious delinquents and life-course-persistent offenders. If we do not intervene early—which we typically do not—then the future will be dismal for these children and hazardous for those whose property they will steal and bodies they will victimize. It is simply irrational to sit back and let criminal lives unfold.

The second reason, based on the section just concluded, is that we know how to divert kids either off of, or from ever entering, this antisocial developmental pathway. The knowledge now exists to intervene in the womb, in the post-natal period, in pre-kindergarten, in families, in schools, and across multiple systems. Cullen and Jonson are not naïve. We already have said that transferring intervention technology from program inventors to real-world settings administered by typical people is daunting—and will not always be successful (Welsh et al., 2010). Still, regardless of these challenges, the fact remains (1) that non-intervention, or doing nothing, is a guaranteed recipe for failure and (2) that treatment programs—based on rigorous experimental evaluations—have been shown to achieve meaningful reductions in future criminal behavior and enhancements in life outcomes (e.g., education, employment, health).

So, to be direct, we now know that early intervention is needed and that it works! But here we give to you, Mr. or Ms. Policymaker, two more reasons to support early intervention: These programs are cost effective and the American public supports child saving.

Cost Effectiveness

These days, any attempt to use public money to help someone might get you accused of being one of those Obama socialists who want to take money from good people (us) and give it to bad people (them). In fact, attempts to save kids could be the first step down the slippery slope to death panels to decide which kids live and which will die. You can never be too careful! Not to worry, our fellow Americans: Capitalism is on the side of early intervention! As it turns out, investing in programs can actually save money. Early intervention is cost effective.

Americans really love it when you can say that something is cost effective. It appeals to three of our central, if unspoken, values:

· Rationality: We should do things that have utility.

· Accountability: “Cost effective” implies that we have a standard—that we will keep doing what pays and not do what does not pay.

· Commodification: The idea that anything in American society can be reduced to a price—to a matter of dollars and cents.

Now, Cullen and Jonson are a bit skeptical of cost-benefit analysis. No, we are not in favor of cost ineffective policies! We are not that stupid or un-American. Rather, the problem is two-fold. First, money has a way of crowding out other considerations. However, should we base all public policy decisions on money? Can we commodify everything? What about doing what is right? Should other values play a role? Such as justice? Humaneness? Let us assume, for example, that keeping very old people alive on life support in the intensive care unit, where they have virtually no quality of existence, is not cost effective (which it is not). Should we simply euthanize them or forbid doctors to prolong their time on earth? Or are you willing to say that life trumps money on this occasion? Well, with a child destined to a life in crime, money should never be the sole consideration.

Second, cost-benefit analysis is predicated on the assumption that it is feasible to measure accurately all possible costs and benefits. However, it is almost impossible to get such precise and complete information. For example:

· How do you measure how much crime is saved? What if only official data (e.g., arrests) are available? We know that if self-report data were used—which reveal all the offenses not detected by the police—then the count of crimes saved would be much higher. Ergo, the estimated benefits might skyrocket. Based on evaluation data from the Stop Now and Plan program for boys ages 6 to 11, Farrington and Koegel (2015, p. 263) computed what this difference would be. They found that for every dollar invested, the benefits based on convictions ranged from $2.05 to $3.75. But when the estimates were scaled up to include “undetected offenses,” the savings jumped to between $17.33 and $31.77.

· What costs of crime do you measure? Do you include only things that are tangible, such as the medical bills or court expenses that flow from a specific crime? What about more intangible savings, such as victim suffering? (We will return to this shortly.)

· What range of factors needs to be included when trying to compute an intervention’s benefits? How many years are sample members followed? What areas of a person’s life are taken into account? Only crime? Or education, employment, marriage, and health? Are impacts on siblings or on the next generation considered?

Our point here is not that cost-benefit analysis is worthless and should not be conducted (Greenwood, 2006). If the assumptions on which the analysis is based are specified, then the meaning of what has been computed can be considered in its appropriate context. What Cullen and Jonson most fear, however, is that that cost effectiveness will be granted sacred status and be accepted uncritically. As often occurs in capitalism, when it comes to claims of cost effectiveness, it is buyer beware!

Except in this instance, because, as it turns out, quality early intervention programs are cost effective (Aos, Phipps, Barnoski, & Lieb, 2001; Cohen, Piquero, & Jennings, 2010; Drake et al., 2009; Farrington & Welsh, 2007; Greenwood, 2006; Greenwood et al., 2001; Welsh, 2003; Welsh, Farrington, & Gowar, 2015)! Cullen and Jonson like this finding, so forget what we said about treating such data with caution. Trust us. The data we like are sacred and should not be questioned! We are, of course, only kidding. Skepticism is the hallmark of a quality scientific mind.

Still, for most major early interventions, the inventors or their followers are very conscientious to present cost-benefit data. They take this step because they realize that it is a powerful way to convince policy makers to implement their programs (see, e.g., Henggeler, 1997; Olds, 2007; Schweinhart, 2007). More independent evidence on the cost effectiveness of the programs reviewed in the previous section—the five programs that work—has been presented by Steve Aos et al. (2001; see also Drake et al., 2009; Greenwood et al., 2001; Welsh, 2003).

Aos et al. (2001), who are located at the Washington State Institute for Public Policy, were mandated by the Washington state legislature to conduct a cost-benefit analysis of a range of programs, from those focused on adult offenders to those focused on early intervention with young and future offenders. In essence, they were engaged in an evidence-based approach to assist legislators in allocating resources (see also Drake et al., 2009). The results of their cost-benefit analysis for the five programs are presented in  Table 9.1 .

A few explanatory comments about the columns in  Table 9.1  are in order. Column 1 presents the cost (in dollars) of the program for each participant. Column 2 presents how much money Aos et al. (2001) estimated would be saved by taxpayers because of lower criminal justice expenditures (e.g., costs not incurred by the police, courts, corrections). Column 3 is simply column 2 divided by column 1. The goal is to determine how many benefits are realized for every dollar spent. In concrete terms, if the dollar amount in column 1 exceeds the dollar amount in column 2, the program is not cost effective; the number in column 3 will be below $1.00, which is the break-even point. So, if I spend $4,000 on a program participant (column 1) and save an average of $2,000 per participant (column 2), I lose money. For every dollar I spend, I save only 50 cents (column 3). Of course, a program is cost effective if I save more than I spend. Thus, if I spend $2,000 (column 1) and save $4,000 (column 2), then for every dollar I spend, I save 2 dollars (column 3).

Now things get a touch complicated. When a crime occurs, a victim has monetary costs (e.g., medical expenses, property loss, wages not earned) and quality of life costs (e.g., pain and suffering) (Drake et al., 2009, p. 179). These are sometimes called tangible and intangible costs, respectively. Tangible costs are more easily measured because they affect physical things—such as your car is gone or court costs go up. Intangible costs are fuzzier, because they involve pain and suffering. The pain and suffering are real—especially if you are the one in pain and the one suffering! The trick, however, is to figure out what dollar amount to assign to such things. How much compensation is due to someone who is forced to hurt—or, alternatively, what would it be worth to you not to feel such things? Importantly, scholars have estimated intangible or quality of life costs based on the awards given to victims by juries in civil cases for pain and suffering. All this is complex, and there are debates over the best way to measure victims’ costs (for a helpful discussion, see Greenwood, 2006). In any event, Aos et al. (2001) use a standard estimate of this sort to assess, when a crime does not occur, how much savings to victims are realized.

SOURCE: Adapted from Aos, Phipps, Barnoski, and Lieb (2001, pp. 154–155).

NOTE: Cost-benefit estimates made per participant in each program.

Okay, back to  Table 9.1 . Here, column 4 adds together the savings or benefits per participant that come from the (1) criminal justice costs and (2) victim costs that are not incurred because a given intervention has prevented offenses from occurring. Finally, column 5 divides column 4 (total benefits per participant) by column 1 (total costs per participant). The result is the cost-benefit ratio. If more money is saved than spent, then the figure in column 5 exceeds 1 dollar. So, column 5 is where we see what programs are cost effective.

Whew, that was a mouthful, huh? Our eyes glazed over just writing this. But the gist of the matter is this: When a program is effective, it prevents crime. Each crime that is prevented lowers criminal justice and victim costs. Programs that prevent a lot of crime—that work—thus save a lot of costs. The money that is saved is, in other words, the program’s benefits. But programs also cost money to deliver, especially high-quality ones. The money it takes to run the intervention is the program’s cost. To be cost effective, the program’s benefits must be higher than the program’s costs.

Perhaps you can understand why designers of interventions want to target high-risk kids and to treat them in programs that last a limited time (e.g., three months). Thus, if a program is successful with a high-risk kid, a lot of crime will be saved and the program will have lots of benefits (see Cohen, 1998). Indeed, a recent sophisticated analysis estimates that the “costs imposed by a career criminal” range from $2.1 to $3.7 million (Cohen et al., 2010; see also Cohen & Piquero, 2009). And if the program can be delivered in a shorter rather than a longer time period, it will not cost as much. So, treating high-risk kids keeps benefits up, whereas a time-limited program keeps costs down. Of course, none of this matters if the intervention does not work. Only if crime is saved can something be cost effective.

So, now to the punch line; let’s look at  Table 9.1 . Three conclusions are apparent:

· MST and FFT seem to be really cost effective.

· Just based on criminal justice expenses saved, three of the programs are not cost effective (see column 3).

· However, when victim costs are added in, all five of the programs are cost effective (see column 5). Again, the issue of victim costs is somewhat controversial because pain and suffering are intangible. But to the extent that crime control policy is supposed to be victim-oriented, such benefits cannot be dismissed.

A few final considerations are important. First, we would caution not to treat the cost-benefit estimates reviewed here as though they were sacrosanct. Second, we have chosen to focus on five major programs because they are well researched and an abundance of information about them is available. However, the five programs reviewed are a subset of a wide array of interventions that have been found to be both effective in reducing antisocial behavior and cost effective (Aos et al., 2001; Drake et al., 2009; Farrington & Welsh, 2007; Greenwood, 2006; Welsh et al., 2015). Policy makers thus now have available a rich menu of successful interventions from which to choose.

Third and perhaps most important, do not fall into the trap of assuming that the only standard for judging cost effectiveness is if 1 dollar invested in a program yields more than 1 dollar in benefits. The crucial issue with programs is effectiveness, not cost effectiveness—that is, whether children are saved and crime is reduced. Who is to say that one dollar in crime benefits—given what is at stake—is not worth a two-dollar investment? Of course, it certainly makes sense to allocate dollars to those interventions that yield the highest returns. Nonetheless, lack of complete cost effectiveness should not be used as an excuse for doing nothing. If no intervention takes place, taxpayers may have a few more dollars, or cents, in their pockets. But the cost will be that a lot of kids will remain entrenched in an antisocial pathway. And their future victims will await, unsuspecting, these youngsters’ inevitable criminal maturation.

Public Support

Okay, early intervention is needed, it is effective in reducing crime, and it is cost effective. But will the American public ever support this loony, bleeding-heart idea of trying to save kids from a life in crime? Well, as it turns out, the answer is “yes”—overwhelmingly so. Every opinion poll—including those conducted by Cullen—has reached this conclusion.

Americans, it seems, are softies when it comes to their wayward youngsters. Why is this so? Cullen and Jonson think it is because child saving meshes with three core beliefs that Americans embrace:

· The belief that youths are malleable and can change.

· The belief that youths are less responsible for the choices they make.

· The belief that children are victims of their circumstances.

And we might add in one more: rationality. Again, it makes no sense to let a troubled youngster grow up into a serious offender when we could intervene with a proven program and stop that from happening. After all, not to do so is a form of child neglect. Who would be in favor of allowing kids that we know are on an antisocial developmental pathway to reach a deviant destination? Only a fool, say Cullen and Jonson!

Now let’s get on to the data! We have claimed that there is a bunch of evidence that child saving is as American as apple pie and, as was once the case, General Motors! We can start, therefore, with surveys on whether the justice system should mainly punish or try to rehabilitate juvenile offenders. The research is clear in showing that the public embraces rehabilitation (Cullen et al., 2000; Moon, Cullen, & Wright, 2003; Moon, Sundt, Cullen, & Wright, 2000; Nagin, Piquero, Scott, & Steinberg, 2006). For example, in a 2001 national study, 80.1% of the sample said that rehabilitation should be the main emphasis of prisons that hold juveniles. The comparable figures for the other options were 8.3% for punishment, 7.7% for protect society, and 3.8% for not sure. When asked how important rehabilitation is as a goal of juvenile prisons, over 97% said that it was either very important (72.6%) or important (24.5%). Only 2.2% selected a little important or not important at all (0.6%) (Cullen, Pealer, et al., 2002, pp. 137–138). Similarly, in a 2005 statewide survey in Pennsylvania, nearly three in four respondents (74.2%) answered that it was true that “juvenile offenders are more likely to become adult criminals if they are sent to jail than if they get rehabilitation in juvenile facilities” (Piquero, Cullen, Unnever, Piquero, & Gordon, 2010, p. 195; see also Piquero & Steinberg, 2010).

But now let’s go to surveys that deal directly with the issue of early intervention—of the kinds of community-based programs we have been talking about in this chapter. Well, public support is, if anything, even stronger (Cullen, Vose, Jonson, & Unnever, 2007; Cullen et al., 1998; Moon et al., 2003). One question, which always gets the same results in surveys, is to ask respondents whether they would rather use their tax dollars to control crime (1) by building more prisons “so that more criminals can be locked up for longer periods of crime” or (2) by expanding early intervention programs “that try to prevent crime by identifying delinquent youths early in life and rehabilitating them so that they do not grow up to be criminals” (Cullen, Pealer, et al., 2002, p. 139). In our 2001 national survey, 86% of the sample chose the early intervention option compared to 14% that favored the prison option (Cullen, Pealer, et al., 2002).

In other research, Cullen and his associates have asked people about the extent to which they supported a range of specific early intervention programs even if that might mean raising taxes (Cullen, Pealer, et al., 2002; Cullen et al., 1998; Moon et al., 2003). In all surveys, large majorities of the public embraced a diverse range of programs aimed at improving parental management, early intellectual development, conduct disorders, school retention, and school after-care. In the national study by Cullen, Pealer, et al. (2002), the level of support for seven programs ranged from a low of 89% to a high of 95%. What other social policy in today’s political climate—a time of Red States and Blue States, of culture wars between Right and Left—would enjoy such universal support? Based on these and a wealth of similar findings, Cullen et al. (2007) have thus concluded that child saving is a “habit of the heart”—a fundamental, unshakable cultural belief that we Americans should do whatever we can to rescue troubled youngsters and to place them on a healthy developmental pathway (see Bellah, Madsen, Sullivan, Swidler, & Tipton, 1985).

Conclusion: Beyond Adult-Limited Corrections

Books on corrections do not typically include a discussion of early intervention. This omission is understandable. Works on corrections are focused mainly on adults, with perhaps a chapter on institutionalized juveniles as a special population. This adult-limited orientation makes sense because it is mainly adults that enter the correctional system and that we lock up in our jails and prisons. By contrast, many of those served by early intervention are in the beginning stages in life, well before they would fall into the clutches of the legal system.

Still, just because something is understandable does not make it intelligent. In fact, Cullen and Jonson think that it is pretty stupid for corrections to be adult-limited. As we have seen, most individuals involved in crime do not suddenly awaken one day at age 25 and say, “Oh, I think I’ll become a super-predator. It should be more fun than being a teacher.” Instead, as we have seen, they enter the gates of our correctional institutions only after embarking on a life course, at times while still in the womb, in which they are exposed to multiple and accumulating criminogenic risk factors. To ignore this stubborn criminological fact requires an ostrich-like capacity to stick our heads in the sand.

As a field, then, corrections needs to broaden its subject matter to include the study of youngsters who are destined to replenish prison populations a decade or two down the road. Corrections must transform itself—much as policing is doing by embracing problem-oriented approaches—from a field that is reactive and into a field that is proactive. That is, corrections should see itself not only as a system that processes those sent by the courts to its doors but also as a system that should advocate for the prevention of offending. Knifing off criminal careers can occur among adult offenders through effective rehabilitation programs. But prevention must be early as well, seeking to head off the development of troubled youths into incarcerated felons (Howell, 2003).

Put another way, corrections must consider the wisdom of becoming a profession that is involved in preventing crime from womb to tomb. The goal should be to develop a range of interventions that are age-graded—that is, that are capable of attacking the unique criminogenic factors that arise at each stage in the life course. If this daunting challenge is undertaken, it might be possible to create a continuum of care in which multiple attempts are made to redirect at-risk individuals—when they are children, juveniles, and then adults—from life-course-persistent antisocial and criminal conduct. Corrections thus would truly embrace the mission that it is never too early and never too late to use criminological knowledge and evidence-based interventions to save the wayward among us (Lösel, 2007).