Title: Perspectives on Crime and Justice: 2000-2001 Lecture Series Series: Research Forum Author: National Institute of Justice Published: March 2002 Subject: Criminal justice research 86 pages 213,000 bytes ------------------- Figures, charts, forms, and tables are not included in this ASCII plain-text file. To view this document in its entirety, download the Adobe Acrobat graphic file available from this Web site or order a print copy from NCJRS at 800-851-3420 (877-712-9279 for TTY users). ------------------- U.S. Department of Justice Office of Justice Programs National Institute of Justice Perspectives on Crime and Justice: 2000-2001 Lecture Series Research Forum Volume V March 2002 National Institute of Justice ------------------- U.S. Department of Justice Office of Justice Programs 810 Seventh Street N.W. Washington, DC 20531 John Ashcroft Attorney General Deborah J. Daniels Assistant Attorney General Sarah V. Hart Director, National Institute of Justice Office of Justice Programs World Wide Web Site http://www.ojp.usdoj.gov National Institute of Justice World Wide Web Site http://www.ojp.usdoj.gov/nij To find out more information about the National Institute of Justice, please contact: National Criminal Justice Reference Service P.O. Box 6000 Rockville, MD 20849-6000 800-851-3420 e-mail: askncjrs@ncjrs.org To obtain an electronic version of this document, access the NIJ Web site (http://www.ojp.usdoj.gov/nij). If you have any questions, call or e-mail NCJRS. ------------------- National Institute of Justice Perspectives on Crime and Justice:2000-2001 Lecture Series Alfred Blumstein Laurence Steinberg Carl C. Bell Margaret A. Berger March 2002 NCJ 187100 ------------------- National Institute of Justice Sarah V. Hart Director The National Institute of Justice is the research and development branch of the Department of Justice. NIJ's Perspectives on Crime and Justice Lecture Series is part of its Professional Conference Series, which supports exchanges between researchers and practitioners through conferences, workshops, planning and development meetings, and other support to the criminal justice field. The Perspectives series features speakers from a broad cross-section of the criminal justice field who are invited to speak on topics of interest to them. The lecture series was developed to foster dialogue among criminal justice professionals and to promote further research in the field. The Research Forum publication series, of which this publication is a part, was designed to share information about the Perspectives series and other forms with a larger audience. The opinions expressed in this publication are those of the presenters and other persons attending the Perspectives lecture forums. As such, they do not represent the official positions of the U.S. Department of Justice. The National Institute of Justice is a component of the Office of Justice Programs, which also includes the Bureau of Justice Assistance, the Bureau of Justice Statistics, the Office of Juvenile Justice and Delinquency Prevention, and the Office for Victims of Crime. ------------------- Preface The National Institute of Justice (NIJ) is the research and development arm of the U.S. Department of Justice. NIJ's Perspectives on Crime and Justice Lecture Series is an opportunity for policymakers and researchers to pause, reflect, and engage in discourse on the best research about crime and justice issues. Through this series, NIJ continues to present discussions by some of the Nation's most distinguished scholars representing criminology and related disciplines. In 2001, Alfred Blumstein discussed the perceived drop in our violent crime rate, which he noted may be the beginning of a continuing decline, the start of a new increase, or a plateau. Laurence Steinberg discussed concerns about the way the criminal justice system identifies and treats adolescent serious offenders. Carl Bell pointed out the varieties and types of violence and discussed basic strategies for violence prevention. Margaret Berger's lecture underscored the tremendous and far-reaching impact of DNA evidence. The lectures in this volume follow a rich tradition of inquiry and enlightenment on crime and justice issues. They are intended to raise the level of discourse and promote dialogue and interest in new areas of research. I hope you enjoy these thoughtful and thought-provoking discussions. Sarah V. Hart Director National Institute of Justice ------------------- Contents Preface Lecture 1: Why Is Crime Falling--Or Is It? Alfred Blumstein, Ph.D. Lecture 2: The Juvenile Psychopath: Fads, Fictions, and Facts Laurence Steinberg, Ph.D. Lecture 3: Violence Prevention 101: Implications for Policy Development Carl C. Bell, M.D. Lecture 4: Raising the Bar: The Impact of DNA Testing on the Field of Forensics Margaret A. Berger, J.D. ------------------- Why Is Crime Falling--Or Is It? Presentation by Alfred Blumstein Professor H. John Heinz III School of Public Policy and Management Carnegie Mellon University February 14, 2001 Washington, D.C. The Recent Crime Drop To those who worry about crime in the United States, the period from 1993 through 1999 was a welcome relief. We witnessed a steady drop in crime rates to a level lower than we have seen for more than 30 years. My presentation focuses on violent crime, primarily homicide, because it is so serious. It also is the most reliable and consistently measured crime and is highly correlated with many other aspects of crime. Between 1993 and 1999, the U.S. homicide rate dropped by an impressive 40 percent to a level of 5.7 per 100,000 population, a rate not seen since 1966. This almost brings the United States into the range of some of the countries in Western Europe. Exhibit 1 presents homicide and robbery rates from 1972 through 1999. These two types of violent crime track each other closely. Both homicide and robbery declined between 1993 and 1999. These current favorable trends, however, cannot continue indefinitely. We should try to identify the factors that contribute to the downward trend and, as those effects are saturated, determine whether the downward trend will flatten or, because of other factors, reverse. Whenever crime rates decrease, there are usually claims of both credit (e.g., "it's a result of my administration's policy of . . .") and explanation (e.g., "demographic shift"). Television newscasters always look for a single explanation and are particularly troubled when more than two mutually supportive factors come together. I recently co-edited with Joel Wallman The Crime Drop in America,[1] which addresses the multiple factors that together contributed to the crime drop, including the waning of crack markets, the strong economy, efforts to control guns, intensified policing (particularly in efforts to control guns in the community), and increased incarceration. Aggregate Trends Exhibit 1 shows that from the 1970s to the mid-1990s, the homicide rate oscillated in the fairly narrow range of 8 to 10 per 100,000 population. Only recently has this rate declined below 8. The peak in 1980 probably was a demographic peak because baby boomers reached the high-crime ages of the late teens and early 20s in the 1970s and moved beyond those high-crime ages in the 1980s. There was a trough in 1985, followed by the homicide epidemic in the late 1980s and early 1990s. Then, following a peak in 1991 and a slightly lower peak in 1993, the homicide rate declined to its lowest point in 1999, the latest year for which the Federal Bureau of Investigation (FBI) published data in the Uniform Crime Reports (UCR). Age Differences This aggregate picture requires various types of disaggregation if we are to isolate the important factor differences that are masked when looking at only the aggregate. For demographic factors, arrest information is studied because only the demographics of those who get arrested are known. Exhibit 2 is the first step in that disaggregation, where we begin to examine the homicide arrest rate trends for individual ages in the traditional peak range of 18 through 24. There was a clear rise among all ages between 1965 and 1970, followed by a period of reasonable stability from 1970 through 1985; the flatness of the 18 to 24 peak is reflected in the fact that the lines for the individual ages are mixed together and cross one another. That flat period runs through the 1980 peak in the aggregate rate shown in exhibit 1, again suggesting that the 1980 peak was primarily a consequence of changing demographic composition rather than changes within any particular age groups. Beginning in 1985, we start to see a major divergence across ages, even in this narrow age range of 18 through 24. By 1993, the 18-year-olds had more than doubled their rates, and 24-year-olds showed no growth at all. Those younger than 18 showed a growth pattern similar to that of the 18-year-olds, more than doubling their rates between 1985 and 1993. These annual growth rates were impressively high, in the range of 10 to 20 percent per year for all ages under 21. After the 1993 peak, we saw a comparably impressive decline in these ages; by 1999, they were roughly back to their 1985 levels. In contrast to this remarkable growth and decline in homicide arrest rates among the younger ages, the rates for people older than 24 showed no growth; in fact, their rates have steadily declined since 1975. These trends are summarized in exhibit 3, which shows two important lines. The upper line reflects the growth period of 1985 to 1993; the lower line reflects the decline or recovery period, which was from 1993 to 1999. Each line depicts for each age the ratio of the age-specific arrest rate for murder in 1993 (the upper line) and in 1999 (the lower line) to the rates that prevailed in 1985. Ratio values greater than 1.0 (denoted by the thick black line) represent an increase in the rates; points below 1.0 represent a decrease. The upper line portrays the ratio reached in the peak year, 1993, and the lower line portrays the degree to which the ratio had declined by 1999. The arrest rate for 15-year-olds in 1993 was triple the rate in 1985. The growth rate was less for older teens, but it was still more than double the 1985 rate for all those age 20 and younger. In contrast, for those age 30 and older, 1993 rates were actually about 20 percent lower than 1985 rates. The trend line depicting the 1999:1985 ratio is well below that for 1993, and the greatest decline occurred among younger people. For the first time since 1993, young people's rates in 1999 were roughly back to their 1985 levels--about 20 percent below for the 14- to 17-year-olds and about 20 percent above for the 18- to 22-year-olds. To the extent that the 1985 rates, which are those that prevailed from 1970 through 1985, represent a stable level that is not easily penetrated, that finding may suggest that the current approaches-- having finally undone the effects of the 1985 to 1993 rise--may have reached their limits, and we may need to consider different approaches if we are to move significantly below that level for these young people. The positive effects of current approaches--including expanded afterschool programs, conflict resolution, and job skill training--on young people may have reached their limit, and we may have to consider other prevention strategies. The homicide rates for older people have seen a continuing decline since the mid-1970s. By 1993, rates for people older than 30 had declined about 20 percent from 1985 levels, and by 1999 they had declined another 20 percent to a level that is about 40 to 60 percent of their 1985 level. The differing patterns for younger and older groups underscore the importance of examining each age group's role in explaining the trends in the aggregate homicide rate since 1985. The aggregate rates shown in exhibit 1 grew to their 1991 peak solely because the rates of the younger people were increasing faster than the rates for the older people were declining. Between 1991 and 1993, the rise for younger people had generally flattened out (as reflected in the pattern for the 18-year-olds in exhibit 2), and so the rate of decline seen in the older ages began to dominate the aggregate rates. Since the rates of both young and old were decreasing after 1993, the aggregate rate continued to fall. In sum, the increase in the level of homicide in the United States during the growth period of the late 1980s and early 1990s was due entirely to the trends in the younger age groups; homicide rates for those age 25 and older did not increase. However, the decrease since 1993 is due to both the recent sharp drop in offending among young people and the continuing decline in offending among older persons. Even though they commit homicide at lower rates, the contribution of the older age groups to the recent decline in the aggregate homicide rate may be appreciable because of their large numbers. For example, in 1980 and 1985, people older than 30 accounted for about 30 percent of the homicides, but by 1993, their contribution declined to only 22 percent, partly because of the growth in the contribution of the younger offenders and partly because older people were committing fewer homicides. By 1999, their contribution increased to 24 percent, driven largely by the sharp decline in the rates for younger offenders after 1993. Explanations of the homicide decline must differentiate between the factors that are responsible for the long-term fall in homicide rates among the older adults and the ones causing the post-1985 rise and the more recent drop in homicide offending by the younger groups. Those two explanations are likely to be different. The Role of Weapons Young people experienced a major growth in the use of handguns in homicide after 1985. Exhibit 4 displays the number of homicides--relative to the number of handgun homicides in 1985, which is set to an index of 100--in each year with three types of weapons: handguns, other guns, and weapons other than guns. The exhibit focuses on the weaponry used in homicides by youths between the ages of 18 and 24, using data from the Supplementary Homicide Reports (SHR), compiled by the FBI, of factors associated with individual homicide events. Before 1985, there was some oscillation, but no clear trend. But between 1985 and 1993, there was an increase of more than 130 percent in homicides committed with handguns, with no marked change in the number of homicides committed with long guns and about a 50-percent decrease with nonguns. This suggests that handguns were partly a substitute for nongun weapons (e.g., knives) and caused more homicides that, if handguns had not been used, may have been merely assaults. The decline started in 1994 and, by 1997, had decreased to about only a 50-percent increase over 1985. Handgun homicides committed by juveniles younger than 18 quadrupled between 1985 and 1993, with a doubling in the number of long-gun homicides and about a 20-percent decrease in the number of nongun homicides. By 1997, the number of handgun homicides had decreased sharply to about 80 percent below the 1993 rate. There was no such increase in the number of handgun homicides committed by 25- to 45-year-old adults; that age group displays a downward trend that accelerates after 1991 and reaches a level about 60 percent of the 1985 level in 1997. A major change occurred after 1985; young people were acquiring handguns in alarming numbers. Older people may have had more handguns during this period, but they appear to have exercised greater restraint in their use. It is widely recognized that teenage males are poor dispute resolvers; they have always fought to settle their disputes. When they fight with fists, the conflict evolves relatively slowly; the loser will eventually find a way to withdraw or a third party, observing the incident, has time to intervene. The dynamics are extremely different when a handgun is present; the conflict escalates well before anyone can retreat or intervene. Once handguns become prevalent in a neighborhood, each person who carries one has an incentive to make a preemptive strike before his adversary does. Between 1985 and 1993, the weapons involved in settling young people's disputes changed from fists and knives to handguns--and more recently, to semiautomatic pistols, which have much greater firepower and lethality. The growth in lethal weaponry is reflected in the changes in the weapons used in homicides committed by different race and age groups. Beginning in 1985, there was a sharp growth in the firearm homicide death rate among young people (those in their early 20s and younger; youths [ages18-24], especially juveniles [under 18 years old]; but not among adults [ages 25-45]) that changed a flat trend to a sharply rising one, with the rise sharpest for young ages. At the same time, the shift was much smaller for the number of homicide deaths due to means other than handguns. The decline in the number of handgun homicides almost mirrors the rise. Following the peak in 1993, the rate of decline was steepest for juveniles; it was less steep for youths. Adults, whose rate displayed no peaking, nevertheless showed a steady decline of almost 40 percent after 1993. These data end in 1997, two years before the age-specific homicide arrest rates reached the 1985 level, and one can speculate that their decline continued until at least 1999. The increase in suicide weapons-specific death rates before 1993 was similar to that of homicide death rates. Following a period of generally flat rates, the rate of suicide by firearms increased sharply after 1985, but the rate of suicide by other means did not change. This shift was especially marked in suicides of black youth and juveniles, whose suicide rate had previously been markedly lower than that of whites.[2] These observations suggest that the growth in homicide committed by young people was more attributable to the weapons they used than to the emergence of inadequately socialized cohorts of "superpredators," as some observers claimed during the period that saw such an increase in the number of homicides. If the cohorts were indeed more vicious, then one would expect to see a growth in homicides by all forms of weaponry rather than by only handguns. The findings strongly suggest that teenagers had disputes as they always had, but that the availability and lethality of handguns, and later semiautomatic pistols, resulted in an increase in homicides. The steady decline in handgun homicide rates after 1993 is consistent with the decline in youth-perpetrated homicide rates shown in exhibit 2. The pattern in handgun use is also reflected in exhibit 5, which depicts the time trend in the rates of arrests for weapon possession at various ages. This pattern is also similar to the homicide patterns depicted in exhibit 2. Weapon arrest trends show a distinct peaking in 1993, followed by a clear decline. Changes in the rates of weapon arrests result from a combination of changes in the illegal carrying of weapons and changes in police aggressiveness in pursuing illegal weapons. Exhibit 5 shows a considerable growth in weapon possession among young people during the late 1980s. Police also became more concerned about weapons, especially in the hands of young people. That combination is reflected in the rise in weapon arrests, which peaked in 1993. There is no indication of any diminution in police aggressiveness in pursuing young people with guns after 1993, so the decline after 1993 is likely due more to a reduction in the carrying of guns than to a slackening of police efforts to capture the guns. This reduction in carrying seems to have been an important factor contributing to the decrease in homicides after 1993. Thus, we have clear indications from SHR data on weapons used in homicides and weapon arrests that there was a significant decline in the use of handguns by young people after 1993. It is difficult to sort out all the factors that contributed to that. One important contributor was the aggressive stop-and-frisk tactics used by local police, especially in many large cities. Community groups in many cities also took an active role in negotiating truces among gangs and seeking to establish norms that precluded the carrying of guns. Important Federal initiatives also are likely to have contributed to the decline. The Brady Handgun Violence Prevention Act (P.L. 103-159), which requires a 5-day waiting period for a background check for any person who wants to buy a gun from a licensed dealer, became effective in 1994, the first year of the decline. The denial rate under the Brady Act has been reported at 2.4 percent of those who apply to purchase a gun.[3] Uncertain is the degree to which these individuals simply accepted the denials or resorted to one of the many loopholes left open by the Brady Act: purchasing a gun at a gun show, buying one from a private individual, hiring a straw purchaser to buy it, stealing it, or using any of the other means left open to a determined illegal purchaser. There are also approaches by the Bureau of Alcohol, Tobacco and Firearms (ATF) to identify dealers and individuals disproportionately involved in the sale or purchase of "crime guns." ATF tries to trace back to the original dealers guns seized by law enforcement. Such efforts may lead not only to deterring inappropriate handgun transactions but to making guns harder to obtain. All these efforts have a mutually reinforcing effect. A reduction in the carrying of handguns, because of either the threat of confiscation or the difficulty in acquiring them, would lead to a reduced incentive for others to carry, thereby reducing the likelihood of handgun homicides, especially among the young people for whom it was so deadly. The Role of Drug Markets One important factor that has affected criminality throughout the 1980s and 1990s has been the problem of drug abuse and drug markets. In a survey conducted in 1991, 32 percent of prisoners reported using cocaine or crack regularly and 15 percent used heroin or opiates regularly. At the time of the offense that led to their imprisonment, 14 percent were using cocaine or crack. These numbers were appreciably higher than those reported in a similar survey conducted 5 years earlier.[4] These are much higher rates than one finds in general population samples (e.g., the National Institute on Drug Abuse household surveys), which strengthens the importance of a connection between drug use and crime rates. Paul Goldstein developed a useful taxonomy of the drug-crime connection composed of three components other than the sale or possession of the drugs themselves: --Pharmacological/psychological consequences. The drug itself causes criminal activity (most notably, the connection between alcohol and violence). --Economic/compulsive crimes. Drug users commit crimes to get money to support their habit. --Systemic crime. Crimes are committed as part of the regular means of doing business in the drug industry (including violence as the accepted way to solve disputes between competing sellers or as retribution between a seller and a buyer as a result of reneging on a drug deal).[5] There is a fourth, more broad connection that should be considered: the community disorganization caused by the drug industry and its operations, including the manner by which the norms and behaviors in the drug industry, which can pervade some communities, influences the behavior of others who have no direct connection to that industry. For example, the widespread prevalence of guns among drug sellers can impel others in the community to arm themselves to similarly defend themselves, to settle their own disputes even if they do not involve drugs, or to gain respect. The problem of crack cocaine emerged in the early 1980s and accelerated significantly in the late 1980s. One indication of this growth lies in the rate of arrests of adults for drug offenses, which, especially for nonwhites (primarily blacks) started to increase in the early 1980s and accelerated appreciably after 1985 with the wide distribution of crack, especially in low-income urban neighborhoods. The steady growth in drug arrests of nonwhite adults compared with those of white adults is reflected in exhibit 6, which depicts the ratio of nonwhite-to-white drug arrests for both juveniles and adults. The trend for juveniles is strikingly different. Throughout the 1970s, the arrest rate for nonwhite juveniles was below that of whites (the ratio is less than 1:1). Starting in 1986, however, their rate grew rapidly, reaching a rate four times that of whites during 1989-92, then began a steep decline to about 50 percent above the white juvenile rate in 1999. This pattern shows that the major recruitment of nonwhite juveniles into the drug markets did not begin until the distribution of crack became widespread in about 1985. Exhibit 6 provides important information linking some earlier observations about the rise in homicides committed by young people and the role of guns in that rise. Three major increases--more than a doubling--occurred in the short period between 1985 (the beginning of the involvement of young people in drug markets) and 1993 (the peak year of youth violence): --Rates of homicides committed by youths age 20 and younger, with no growth for adults 25 and older (exhibit 3). --The number of homicides those younger than 25 commit with guns, with no growth in nongun homicides (exhibit 4). --The quadrupling of the arrest rate of nonwhite juveniles on drug charges compared with white juveniles (exhibit 6). One explanation for this dramatic combination of changes involves a process that is driven by illegal drug markets, which appear to operate in conjunction with the demand for drugs despite massive efforts during the past two decades to attack the supply side. In the late 1980s, the illegal drug trade recruited juveniles because they were willing to work more cheaply than adults, were less vulnerable to the punishments imposed by the increasingly punitive adult criminal justice system, and were willing to take more risks than adults. The rapid growth in the demand for crack required more sellers--many new users used crack because they could buy one hit at a time, unlike powder cocaine, which was not sold in small quantities--and encouraged the market to find its labor supply wherever it could. Furthermore, recruiting juveniles was the market's means of replacing the large number of adult drug sellers who were being incarcerated during the 1980s. The economic plight of young urban black juveniles, many of whom saw no other comparable route to economic sustenance at the time, made them particularly responsive to the lure of employment in the crack markets. Because crack markets were run as street markets, especially those operating in inner-city areas, the participants were especially vulnerable to attack by robbers who targeted their sizable assets, either the drugs or the money from the sale of drugs. Calling the police for protection was not an option, so participants in those markets, including recruited juveniles, were likely to carry guns to protect themselves and solve disputes. Once these juveniles started carrying guns, other teenagers who were not involved in the drug markets but went to the same schools or walked the same streets also were more likely to arm themselves. These teenagers felt they needed guns for their own protection, but they also may have believed that weapon possession was a status symbol in the community. This initiated an escalating arms race: As more guns appeared in the street, there was an increased incentive for individuals to arm themselves. In light of the much tighter networking of teenagers than of older people, that diffusion process could proceed quickly. The emergence of teenage gangs-- some involved in drug markets--in many cities at about this time contributed to that diffusion. In view of the recklessness and bravado that often characterize male teenagers and their low skill level in settling disputes other than through the use of physical force, many of the fistfights that would otherwise have taken place escalated into shootings as a result of the presence of guns. This escalation in violence can be exacerbated by the problems of socialization associated with high levels of poverty, high rates of single-parent households, educational failures, and a widespread sense of economic hopelessness. Not until they reach their mid-20s do they develop some prudence, become more cautious even if they are armed, and display greater restraint. This hypothesized diffusion process[6] has been tested further with city-level data on juvenile arrests for drugs and homicides, taking advantage of the fact that drug markets flourished at different times in different cities, such as in the mid-1980s in New York and Los Angeles and later in smaller cities. Daniel Cork[7] has shown the connection between the rise in handgun homicides and the recruitment of juveniles into crack markets. Using an epidemic model originally developed for marketing literature, Cork identified--in individual cities--the time when juvenile arrests for drugs began to accelerate and the corresponding time when juvenile homicide arrests increased. He found most typically a 1- to 3-year lag between the two, with homicides following involvement in drug markets. These results are consistent with the hypothesis that the rise in juvenile homicides was attributable to the diffusion of guns from young people recruited into drug markets to their friends and beyond. His analysis of individual cities also showed that crack markets generally emerged first in the largest coastal cities, especially in New York and Los Angeles, and then appeared in Middle America and smaller cities. Thus, the observed patterns in the rise of homicide committed by young people with handguns are highly consistent with explanations that assign central importance to the rise and decline of crack markets in the United States. The fall-off in the nonwhite/white drug-arrest ratio (exhibit 6) in the 1990s is a reflection of the changing tastes for crack, especially in urban neighborhoods. As recognition of its deleterious effects became widespread, word spread through the streets that crack was an undesirable drug, and this wisdom had a major effect on diminishing the number of new users.[8] This contributed to a major reduction in the need for street sellers. As a result, the nonwhite juvenile sellers, who had been important participants in those street markets, were no longer needed. Older users continued to be major crack consumers, but their demand could be served more readily by individual delivery, thereby diminishing the need for street markets. All these changes contributed to a decline in street markets, the recruitment of juveniles, and handgun possession by young people following the 1993 peak. One important contributing factor to the decline in violence as crack demand ebbed has been the strength of the U.S. economy during the past decade. If there were no legitimate jobs for young people, it is reasonable to anticipate that they would have found other criminal activity to provide economic sustenance. But the abundance of job opportunities, including those not requiring high skill levels, provided legitimate alternatives. Individuals in legitimate jobs have a strong incentive to conform and avoid criminal activity. This should indicate the desirability of finding approaches that bring young people into the legitimate economy through appropriate training to develop legitimate employment opportunities. Incarceration The United States has gone through a dramatic transformation in its sentencing policies and practices in the past 25 years. As shown in exhibit 7, the United States maintained an impressively stable incarceration rate (prisoners per capita) of about 110 per 100,000 population during the 50-year period from the early 1920s to the mid-1970s,[9] when it suddenly grew exponentially at a rate of about 6 to 7 percent per year. The rate is currently more than four times the previously stable rate. Various attempts have been made to correlate the rising incarceration rate with the crime rate. The most aggressive of those analyses use the period after 1991--they argue that the crime rate has been steadily decreasing because the incarceration rate is increasing. But such simplistic attempts to estimate the incarceration effects on crime are likely to be misleading. For example, the analysis must also account for the period in the late 1980s when crime was increasing at the same time the prison population was growing. Attributing the decline to incarceration is far more tenable if one focuses on older offenders, whose homicide rates have declined steadily since the mid-1970s. This group is the appropriate focus for estimating the incapacitative effect of incarceration (i.e., crime is reduced because offenders are removed from the streets). One can appreciate that the incapacitation effects were an important contributor to the continuing decline of violent crime rates among older people, especially for those over 30, who displayed about a 40- to 60-percent drop in homicide rates between 1985 and 1999 (exhibit 3). This connection is particularly appropriate because 32 is about the median age of State prisoners. One of the contributors to the growth in incapacitation is the large number of drug sellers who have been sentenced to prison in the past two decades. Exhibit 8 shows the growth in incarceration from 1980 through 1996 by crime type. The greatest growth--by a factor of more than 10-- was among drug offenders. Ironically, their incarceration did not have a major impact on the drug trade because others, particularly younger sellers, replaced them. But if they would have engaged in violence on the outside, their incarceration could have contributed to the decline in violent crime rates. The incapacitation effect, however, is at least partially negated by violence committed by the replacements. Indeed, because many replacements were young people, who have a greater propensity for violence, the net effect may have been an increase in violent crime, undoubtedly a factor that must be considered when discussing the rise in violence of the late 1980s. Even if drug offenders' incarceration contributed to the reduction in homicide, it is not clear whether imprisoning them was an efficient use of fiscal or prison resources. As the prison population grows, marginal offenders are likely to have a lower offending frequency (lambda) than those who were already incarcerated.[10] A sizable but unknown fraction of drug offenders who are incarcerated (comprising more than 20 percent of State prisoners and more than 60 percent of Federal prisoners) are predominantly entrepreneurs rather than generic criminals, and they are not likely to be violent. Incarceration effects are far less likely to have been a significant factor in the more recent decline in violent crime rates among teenagers and youths, and most likely were limited to older youths, whose risk of incarceration is greatest. In addition, levels of violence have fallen in the younger age groups in recent years even as their risk of incarceration has increased. The decline might have been less steep in the absence of the "get tough on juveniles" sentencing policies enacted in recent years. But it seems more likely that the other factors considered in this paper-- the reduction in the use of guns, changes in the drug markets, and the growing number of legitimate job opportunities--have had more dominant effects. In The Crime Drop in America, William Spelman and Richard Rosenfeld derived estimates of the contribution of incarceration to the crime drop of the 1990s. Spelman[11] used general elasticity estimates (percentage reduction of crime resulting from a 1-percent increase in the prison population) from the literature and estimated that the crime reduction is associated with steady growth in the prison population. Rosenfeld[12] used estimates of prisoners' offending frequency (lambda) based on homicide rates in the most disadvantaged neighborhoods of St. Louis and Chicago and estimated that prisoners would commit 150 homicides per 100,000 prison population. Both were crude estimates, and they used very different approaches, but both estimated that incarceration contributed about 25 percent of the crime drop, leaving 75 percent to other explanations. Changing Demographic Composition Much of the speculation about the recent decline in homicide rates attributes it to changing demographics. This may be a holdover from the realization that much of the decline that began in 1980 was attributable to a demographic shift as the baby-boom generation outgrew the high-crime ages.[13] Those same demographic effects were not at work in the early 1990s, since demographic effects do not always move in the same direction. The decline after 1980 was significantly affected by the shrinking size of the cohorts in the high-crime ages--late teens and early 20s. In the late 1990s and currently in the United States, those cohort sizes are growing. Exhibit 9 depicts the age distribution of the U.S. population in 2000. It is evident that the smallest age cohort under 40 is about 24 (those born in 1976). Each of the youngest cohorts is larger than its predecessor until the peak at age 9. Thus, if age-specific crime rates are to remain constant for teenagers, the aggregate crime rate should be increasing as a result of the larger cohort sizes. These age-composition changes are relatively small, with cohort sizes growing at a rate of about 1 percent per year. Even disaggregating the composition by race reveals a similar pattern: Both whites and blacks have their smallest cohort at about age 24 in 2000, but the rate of growth of the younger black population is about 2 percent per year. These demographic trends are small compared with the much larger annual swings in the age-specific crime rates, as much as 10 to 20 percent per year growth in the 1980s (16 percent per year for 18-year-olds from 1985 to 1991) as well as decreases in the 1990s (6 percent per year for 18-year-olds from 1991 to 1998). Some Observations The sharp rise in violence by young people during the late 1980s and the correspondingly sharp decline in the 1990s are striking. The increase in the aggregate homicide rate was due to escalating rates among juveniles and youths, predominantly (although not exclusively) by and against black males, particularly in larger cities and exclusively involving handguns. By 1999, the rate of homicide perpetrated by youths finally returned to the stable rate that prevailed from 1970 through 1985. Although the causes of the rise in violent crime are reasonably clear--homicides by young people with handguns, mostly as a result of diffusion out of drug markets--the factors contributing to the decline are more complex. Various forces are involved, some more salient in certain places. They include efforts by local police, communities, and Federal agencies to separate young people from their guns. Those efforts have been helped considerably by the waning of crack markets, especially the diminished participation of young people in those markets. As an alternative, the robust economy has provided legitimate job opportunities for them, which has created incentives to avoid illegal activities. The changes in drug markets also help account for the variation in the timing of the peaks and declines in rates of violence across cities. Large coastal cities such as New York and Los Angeles, where crack took hold earlier and violent crime rates peaked sooner than in other cities, were expected to--and did--experience a drop in homicide rates sooner than other cities. The effects of drug markets also directs attention to the population groups in which the changes in homicide were concentrated: youth, particularly black youth who did not have more attractive economic opportunities and became drug sellers at disproportionately high rates in inner-city crack markets. For older offenders, the growth in incarceration is an important component of the explanation, although other considerations seem to be relevant. These include the widespread availability of domestic violence services, which seem to have contributed significantly to the reduction in male victimization in homicide.[14] One final observation is somewhat provocative. The UCR reports for the first half of 2000 were released by the FBI in mid-December 2000. They are strikingly different from the previous 6 years in which annual decreases in crime rates of 6 to 8 percent were common. The new report estimates both crime and homicide drops of merely 0.3 percent. This could be an indication that the decrease in crime, which could not continue indefinitely, has finally flattened out. A precursor to this estimate occurred in 1998, when homicides dropped by only 2 percent in the large cities (populations of 500,000 to 1 million and more than 1 million) when the national aggregate drop was about 6 to 8 percent. Large cities have led the rise and decline in crime rates. We cannot be certain whether this flattening is an indication of one small disruption to a continuing decline, the start of a next increase, or a plateau from which changes will require particularly innovative approaches that are quite different from the actions that have taken us to this point. Regardless, we should take advantage of the current opportunity to better understand these processes and to pursue criminal justice and community-based policies to forestall the next increase as long as possible. As we look to the future, we should be concerned both about the possibility of a resurgence of active drug markets and any violence they may bring with them and about a downturn in the economy and the impact it would have in communities in which violence is most likely to reignite. Question-and-Answer Session Patrick W. Murphy, American Police Association, Alexandria, Virginia: Remember what the Crime Commission said about the police? How can you forget--about [being] ineffective, inefficient, fragmented, insular, unprofessional, and not dedicated to research? Community policing has brought a small revolution in policing. In a quick-and-dirty survey, the L.A. Times found that the chiefs of about a dozen major cities that had experienced significant declines in murder, credited community policing. Although the debate continues about New York, there are those who say community policing has had a significant impact there, even if it has not been given the same credit by city hall or police headquarters. Police have a long way to go, but I'm encouraged that some kind of a corner has been turned, with police focusing more on prevention and working more with the community. Would you comment on that? A.B.: Policing has been most notable in its development and its openness. Police chiefs are remarkably astute and policing has been impressive in its readiness to try new ideas. Community policing makes so much sense at an abstract level, but when you examine its implementation at the city level, it is difficult to evaluate just what is going on. Programs vary enormously. When the COPS program began, community policing became the byword for whatever police departments wanted to do so they could take advantage of new sources of funding. A 1999 paper coauthored with Allen Beck[15] examined arrests per crime over time; for all crime categories we looked at, except for drugs--for which we did not know the number of crimes--the arrest rate per crime was flat. So, at least it casts some question about the degree to which innovation in policing has contributed to significant improvements in solving crimes. The extent that it helps solve a variety of other problems, the extent that it contributes to much better relations between the police and more highly mobilized communities, is all to the good. I've been to visit the Compstat operations in the New York Police Department. Much has been made of the technology there, but I think it is secondary to the chiefs and "wise old hands" who put precinct captains on the spot and say, "What are you doing about all this?" and get them to address the emerging crime problems with a variety of challenges and ideas. These are the places where we can see a lot of innovation. I think police departments are ready to add operations research units, groups that will analyze the data that are now streaming in from 911 calls (which are not easily manipulated), with the data that stream in from other sources, figuring out improved tactics and improved approaches. I would hope that funding is available to provide this kind of link between the police, who are now anxious for interesting sophisticated analysis leading to better operations, and the resources to do it. Joanne Wiggins, U.S. Department of Education, Washington, D.C.: You talked about murder, you talked about robbery, you talked about drug arrests; you did not talk about rape, sexual assault, or domestic violence. I would issue a challenge that when we talk about "crime dropping," it is not true to say that crime is down unless you can also say that crimes against women are down. I think we need to be careful how we make these statements because the realities of women's lives may not be exactly like the charts you have been showing and research needs to take these disparities into account. A.B.: I think that is an important issue. Unfortunately, our data sources, on rape particularly, are not terribly reliable in the sense that in the victimization survey, the number or rape victim sample is too low to get reliable estimates, and in police records there is too much "un-founding" (i.e., the alleged incident did not occur) going on. The category of reported rapes is subject to significant shifts in women's willingness to report over the past couple of decades. One interesting study that came out of the work of the National Consortium on Violence Research was a dissertation by Laura Dugan that compared the association of intimate-partner homicide across cities with the degree of services (e.g., legal services, counseling services). Dugan found that the more services a city provided for domestic violence, the more good it did to reduce intimate-partner homicide, but only of one gender. And the gender that benefited from the reduction was not women, but men. The interpretation has been that increased victim services provided alternatives for women who might otherwise have killed their victimizers. So the major drop in intimate-partner homicide victimization has been for men. Dugan is revisiting the issue of domestic violence itself to gain a better understanding of the situation. Devon Brown, Office of the Corrections Trustee, Washington, D.C.: I was delighted recently to learn that your colleague, the esteemed John DiIulio, has recently "found religion," inasmuch as he has changed his view on the prediction that so-called "superpredators" would take over our cities. Could you share with us your view on that? A.B.: When we saw this rise in juvenile violence, rhetoric started to flow. The "superpredator" theory argued that we are now seeing a breed of kids that is far worse in socialization, conceivably in genetics, than previously seen. A major thrust of everything I said about that rise in violent crime was that it was not different kids, it was the same kids doing what they had always done, with more lethal weaponry. The handgun became the major source of the problem when it got into the hands of irresponsible people. Violent crime was reduced not because we had changes in socialization after 1993, but because the nucleating role of the drug markets diminished. It declined because kids no longer had to carry guns and we saw a general disarmament. Another reason for disarmament was that the police were posing larger threats--taking the guns away and imposing other punishment. As we saw fewer guns in the street, the incentive to carry them was diminished. So we saw this gradual dropping away. John DiIulio has acknowledged that "superpredator" was really an inappropriate characterization. I think the essence of the data I presented today shows that it wasn't different kids; it was the weapons those kids were carrying. These factors should stimulate everybody to work to prevent handguns from getting into the hands of irresponsible people. There are many ways we can do that without significantly inconveniencing the large number of responsible people who have every right to have handguns legally. These include tracing guns, as ATF does to see where those guns are coming from; identifying dealers who are major sources of guns used in crimes; and restricting gun purchases to one gun per month to inhibit aggressive marketers. Federal laws will be necessary to ensure that one State is not vulnerable to neighboring States that do not enforce gun laws. As long as we see a clearly interstate commerce in guns, Federal intervention will be needed to identify the source of the problem and identify minimally intrusive methods to ensure that guns don't get in the hands of statutorily defined irresponsible people (youth, felons, and individuals who have been involuntarily committed to mental institutions). A number of States' mental health departments, for example, have set up registries of people who have been committed to mental hospitals who should not have access to guns. A gun dealer can query that registry to determine whether a particular customer is prohibited from purchasing a gun. The response gives no detail on the person's mental illness, but merely provides a yes/no answer on prohibition. Stephen Rickman, Executive Office for Weed and Seed, U.S. Department of Justice, Washington, D.C.: As a consequence of this increase in incarceration rate over the past 20 years, you have also had an increase in the number of people who are coming out of prison. And I think there was a study done over at the Bureau of Justice Statistics (BJS) estimating more than 500,000 people released. Given the fact that you have had this diminution in postrelease supervision and services with this population, how do you factor this into your trends and how it may affect crime in the future? A.B.: That's a good question, because it really pulls together a number of important issues. Number one, we are keeping people incarcerated longer. We are keeping people longer partly because our sentences are increasing, partly because of mandatory minimums, and, most important, because of parole violations (increasingly for technical violations). We are increasing the probability that these people coming out are coming out well past their criminal careers--and we must think seriously about criminal careers. There was a period when NIJ sponsored some very important research on criminal careers, looking at the duration of a criminal career, how long people stay active. Research found that the duration of the residual criminal career goes up through the 20s, is fairly flat through the 30s, and then falls off in the 40s. So people in their 30s are the ones most likely to continue--if they are active in their 30s. The issue of postrelease management is a complex one. It involves a mixture of providing services and exercising control by sending the person back to prison. The trends lately have emphasized the control aspect, often at the expense of services. Indeed, many States are finding that more of their admissions are now composed of parole violators than new court commitments. Service needs are complex, and perhaps the most essential is drug treatment, which everyone acknowledges is important. In addition, there has to be help in finding and keeping a job, treating mental illness, and various other forms of counseling. The whole notion of parole has became politicized because parole officials rather than prison officials were the ones who made release decisions: "We will stop the release decisions by moving to determinant sentences" without attention to the handholding and the guidance and the counseling needed and without attention to the rate of reincarceration. Parole officials took the political heat for being "soft on crime" at the time when everyone else was being "tough," so they began sending violators back to prison on the least provocation without dealing with the issue of the optimum policy for dealing with somebody who now is drug positive. The parole issue very much needs rethinking. Parole recommitments have been a major factor in the growing incarceration rate over the past 5 years or so. Obviously crime rates are declining. We are not doing a lot more on drug enforcement because the drug markets have thinned out and have become more surreptitious. They now look more like powder cocaine markets--more like pizza delivery rather than street markets. Ted Gest, University of Pennsylvania, Criminal Justice Journalists (based in Washington, D.C.): While your presentation related to government responses to crime, you concentrated almost totally on what you would call enforcement or incarceration remedies. Could you describe the field of so-called crime prevention in the past 10 years or so? I think it's fair to say there has been an increase in programs dealing with juveniles and drug treatment--certainly not so much as was contemplated in the 1994 crime law,[16] but a lot of programs. Does the absence of crime prevention from your comments indicate that either we don't know what any of these programs contributed or that you think they have had a marginal or insignificant effect on the phenomenon? A.B.: I think that's an important question. There is so much I didn't talk about but I certainly didn't mean to slight prevention. I still think the efforts that governments, particularly the Federal Government, put into issues of prevention are quite minimal and the issue is compounded by the fact that we have so little research, between the Office of Juvenile Justice and Delinquency Prevention (OJJDP), National Institute of Justice, and BJS, to have any accurate assessment of what works best. The programs we get data on and that have been a major cornerstone of both Federal and State policies have been in the incarceration area. Prevention efforts are still strikingly minuscule in comparison to what is needed, particularly the need to combine those efforts with evaluations targeted at what looks like the most promising opportunities. OJJDP has been a major leader in, for example, the longitudinal research in tracking kids through emerging and eventually terminating criminal careers--and the factors associated with getting involved in crime--because that fundamental research helps to identify what kind of interventions are best for whom. We went through this horrible period in the late 1970s and early 1980s when Bob Martinson's "nothing works" theme emerged out of a variety of evaluation studies that tried to find the "silver bullet" and tested individual technologies to estimate their effectiveness in reducing recidivism. Any such treatment cannot be universally applicable. Treatments have to link the individual offender or potential offender and his or her needs, the treatment provider and his or her skills in delivering various kinds of treatment, and the environment in terms of what kinds of crime that person might be getting involved in. We have to experiment with many approaches and evaluate them, but budgets in the order of tens of millions of dollars are inadequate to deal with this problem. We continue to be impressively ignorant about the effects of any of our interventions within the justice system. And cutting that back would be the height of folly at a time when the establishment is ready to be open and interested in getting research findings. But we need much more research and evaluation to track the changes that are going on in this phenomenon-- because they are changing. Clinton G. Turner III, U.S. Capitol Police, Washington, D.C.: My primary interest is victims of crime, especially those victimized at night. What about the enhancement of victim/witness support nationwide? At the U.S. Capitol Police, we are very good at dealing with victims of crime, primarily crime that occurs around 6:00 p.m. and later. Youths commit a high percentage of these crimes. If we enhanced victim programs in the various police departments, I believe we would have even fewer. Is there anything being done nationwide to enhance the victim-witness programs in police departments? A.B.: I believe there are; I just haven't looked very carefully at that. I'm sure there is someone here who could answer that much more fully than I. I'm sure there are many here who might want to comment. Pat Murphy, do you have a sense of the degree to which police departments are developing victim service programs? Patrick W. Murphy: From what I read in the newsletters and other sources, more departments are getting into victim services. But, you are so right about how chintzy we are about research. It's crazy that we are not spending 10 times as much on research as we are, especially on policing. Earlier, I was too polite. The basic problem about crime in this country is that we do not have a national police system, and we do not have democratic policing in the inner cities where we need it. You know from your murder rates how much higher the murder rates are in cities with populations over 100,000 than they are in the rest of the country. Fifty percent of the murders in the United States are in cities with populations over 250,000. Within those cities, crime is concentrated in the inner cities, just as 100 or 150 years ago it was in the ghetto neighborhoods. On one block in 1870, 365 murders were committed at "Five Corners" in New York. Both the perpetrators and the victims were "nice people." But we do not have democratic policing. Democratic policing is self-policing. The people are supposed to be policing themselves with police assistance. Well, for 100 years the police have been telling the people that they can protect them, and the good people don't even know that they are supposed to be policing themselves with police assistance. Now, that is happening in some cities, which will lead to more drug treatment and more victim programs. About 30 percent of the police officers in the country today have college degrees; this is an encouraging sign, moving toward professionalization of the major departments. And I think as we see more of that, we will see more social programs being supported and actively involved with the urban police. Notes 1. Blumstein, A., and J. Wallman, eds., The Crime Drop in America, New York: Cambridge University Press, 2000. 2. Blumstein, A., and D. Cork, "Linking Gun Availability to Youth Gun Violence," Law and Contemporary Problems 59 (1) (1996): 5-24. 3. Gifford, L., D.B. Adams, and G. Lauver, Background Checks for Firearm Transfers, 1999, Bulletin, Washington, D.C.: U.S. Department of Justice, Bureau of Justice Statistics, 2000, NCJ 180882. 4. Innes, C.A., Drug Use and Crime: State Prison Inmate Survey, 1986, Washington, D.C.: U.S. Department of Justice, Bureau of Justice Statistics, 1988, NCJ 111940, and Beck, A., et al., Survey of State Prison Inmates, 1991, Washington, D.C.: U.S. Department of Justice, Bureau of Justice Statistics, 1993, NCJ 136949. 5. Goldstein, P., "The Drug/Violence Nexus: A Tripartite Conceptual Framework," Journal of Drug Issues 15 (1985): 493-506. 6. Blumstein, A., "Youth Violence, Guns, and the Illicit-Drug Market," Journal of Criminal Law and Criminology 86 (1) (1995): 10-36. 7. Cork, D., "Examining Space-Time Interaction in City-Level Homicide Data: Crack Markets and the Diffusion of Guns Among Youth," Journal of Quantitative Criminology 5 (4) (1999): 379-406. 8. Johnson, B., A. Golub, and E. Dunlap. "The Rise and Decline of Hard Drugs, Drug Markets, and Violence in Inner-city New York," in The Crime Drop in America, ed. A. Blumstein and J. Wallman, New York: Cambridge University Press, 2000: 164-206. 9. Blumstein, A., and J. Cohen, "A Theory of the Stability of Punishment," Journal of Criminal Law, Criminology, and Police Science 63 (2) (1973): 198- 207. 10. Canela-Cacho, J.A., A. Blumstein, and J. Cohen, "Relationship Between the Offending Frequency (lambda) of Imprisoned and Free Offenders," Criminology 35 (1) (1997): 133-175. 11. Spelman, W., "The Limited Importance of Prison Expansion," in The Crime Drop in America, ed. A. Blumstein and J. Wallman, New York: Cambridge University Press, 2000: 97-129. 12. Rosenfeld, R., "Patterns in Adult Homicide: 1980-1995," in The Crime Drop in America, ed. A. Blumstein and J. Wallman, New York: Cambridge University Press, 2000: 130-163. 13. Blumstein, A., J. Cohen, and H. Miller, "Demographically Disaggregrated Projections of Prison Populations," Journal of Criminal Justice 8 (1) (1980): 1-25. See also Blumstein, A., J. Cohen, and H. Miller, "Crime, Punishment, and Demographics," American Demographics (October 1980): 32-37. 14. Dugan, L., D. Nagin, and R. Rosenfeld, "Explaining the Decline in Intimate Partner Homicide: The Effects of Changing Domesticity, Women's Status, and Domestic Violence Resources," Homicide Studies 13 (1999): 187-214. 15. Blumstein, A., and A. J. Beck, "Population Growth in U.S. Prisons, 1980-1996," in Prisons, ed. M. Tonry and J. Petersilia, volume 26 of Crime and Justice, Chicago: University of Chicago Press, 1999: 17-61. 16. Violent Crime Control and Law Enforcement Act of 1994, P.L. 103-322. ------------------- The Juvenile Psychopath: Fads, Fictions, and Facts Presentation by Laurence Steinberg Distinguished University Professor of Psychology Temple University Director, John D. and Catherine T. MacArthur Foundation Research Network on Adolescent Development and Juvenile Justice March 20, 2001 Washington, D.C. This morning, I'd like to talk about a worrisome practice that I believe is becoming increasingly common in prosecutions of serious juvenile offenders: the characterization of the juvenile offender as a "psychopath," a "budding psychopath," or an individual with "psychopathic tendencies."[1] I do not have hard data to validate my suspicion that this practice has in fact become more common in recent years,[2] but my work on adolescent development and juvenile justice brings me into regular contact with prosecutors, defense attorneys, and judges, and the frequency with which I am asked questions about psychopathy and its diagnosis among juveniles has undoubtedly increased. Whether this reflects a genuine trend may depend as much on one's definition of psychopathy as it does on one's definition of what constitutes a trend. As some of you know, to a social scientist, one case constitutes an anecdote; two cases, data; three, a pattern; and four, a trend. At least four practitioners have asked me about this issue in the past year, so we may be talking about a trend. Regardless of whether labeling some youngsters as true or incipient psychopaths affects 4, 40, or 4,000 juveniles a year, there is, nevertheless, good reason to worry about the practice, and there are many reasons to sound a note of caution within the juvenile and criminal justice systems about the potential overuse of psychopathy as a diagnostic label when applied to juveniles. Juveniles who are branded as psychopaths are more likely to be viewed as incorrigible, less likely to receive rehabilitative dispositions, and, if it is an option, more likely to be transferred to the criminal justice system to be tried as adults and face the possibility of adult sanctions, including incarceration in adult jails and prisons. We are not sure if the construct of psychopathy makes sense when applied to adolescents and children, we are not sure that measures of psychopathy are stable over the course of adolescence or between adolescence and adulthood, and we have no data on the predictive utility of the diagnosis when applied to juveniles. We are not ready--if indeed we ever will be--to base transfer or other dispositional decisions on diagnoses of psychopathy among juveniles. Understanding Psychopathy Let me begin with a brief introduction to the construct of psychopathy and its uses and misuses in clinical and forensic practice. First, I want to distinguish between two similar words with different meanings for those who do not have a background in psychology or psychiatry. "Psychopathology" refers to any sort of psychological disorder that causes distress either for the individual or for those in the individual's life. Depression, schizophrenia, attention deficit hyperactivity disorder, alcohol dependency, conduct disorder, and bulimia are all forms of psychopathology. Most researchers use the term "psychopathology" loosely to refer to a continuum of problems that range from mild discomfort to full-blown psychosis. Psychopathology can be ascertained through the administration of standardized tests or questionnaires or through some form of diagnostic clinical interview. "Psychopathy," in contrast, refers to a very specific and distinctive type of psychopathology. Psychopathy is a type of personality disorder defined chiefly by a combination of antisocial behavior, callousness, and emotional detachment. As one set of writers recently noted-- Psychopaths are typically charismatic individuals who readily manipulate others and engage in risky behaviors designed to satisfy their own personal needs. They are undeterred by pangs of conscience and have little or no concern for the welfare of others. Their relationships tend to be shallow and they often meander from one opportunistic setting to another without much concern for the future.[3] Although psychopathy is often treated as a unidimensional construct, current thinking indicates that what we call psychopathy is actually composed of two related, but independent, components. Factor I reflects a cluster of affective and interpersonal features best described as callous emotional detachment (e.g., glibness, egocentricity, superficial charm, and shallow affect), whereas Factor II represents the chronic unstable and antisocial lifestyle (e.g., irresponsibility, impulsivity, criminality, and proneness to boredom) associated with psychopathic individuals. The fact that psychopathy is defined by two factors--one having to do with emotional detachment and the other having to do with antisocial behavior--is very important in understanding why it may or may not be useful in predicting future dangerousness among juveniles. As it is presently defined, psychopathy is very similar to what psychologists call a personality disorder--indeed, although it is not listed in the Diagnostic and Statistical Manual-IV (DSM) as a personality disorder, it has many characteristics in common with antisocial personality disorder, and some have argued that psychopathy should be viewed as a subtype of antisocial personality disorder.[4] Personality disorders are unlike other forms of psychopathology, which are viewed as potentially treatable (or at least manageable) conditions. In contrast to such disorders as depression or conduct disorder, psychopathy is presumed to represent a pattern of personality and behavior that is deep seated, originating in early experience and/or biological functioning, relatively stable over time, and resistant, if not absolutely immutable, to change. Psychopathy is generally assessed via structured interviews. In studies of adults, the gold standard from which most other measures derive is the revised version of the Psychopathy Checklist or PCL, which was developed by Robert Hare, the North American authority on psychopathy and its assessment. A slightly revised version of this measure, the PCL Youth Version, is now available for use with adolescents. The PCL is based on a series of ratings a trained expert makes on the basis of the individual's interview responses and information taken from the respondent's file or official records. The ratings are made on 20 different dimensions (see exhibit 1). The rater assigns a score of 0, 1, or 2 for each of these 20 dimensions. The field generally uses a cutoff score of 30 to designate an individual as a psychopath. Before turning to questions about the appropriateness of using the psychopathy diagnosis for making dispositional decisions about juvenile offenders, a few words are in order about the use of this construct in making judgments about adults. First, there seems to be little disagreement about the utility of the construct "psychopath" in describing certain adult criminals. There is a substantial literature suggesting that criminals who score high on the PCL can be distinguished from their nonpsychopathic counterparts in theoretically meaningful ways on other personality measures and on measures of information processing and decisionmaking. Studies of neuropsychological functioning suggest that at least some portion of individuals who score high on the PCL may be "wired" in a way that could conceivably contribute to their criminality. Second, there is near consensus that the PCL is a valid and reliable way with which to measure psychopathy among adults, although several writers have suggested that there is room for improvement[5] and others have argued that there are three, not two, underlying factors.[6] Reliability here refers to agreement between raters who are assessing the same individual. Estimates of the reliability of PCL scores over time are more difficult to calculate, because at least some of the score an individual receives is based on facts that are in the individual's record. Because these facts do not change, certain elements of the individual's PCL score remain fixed and therefore artificially inflate estimates of the reliability of the measure over time. Nonetheless, by conventional social scientific standards, the PCL is a more than acceptable psychological assessment tool. Third, and most important, the PCL is unrivaled in its ability to predict future antisocial behavior; it is especially good at predicting future violence. For example, a recent meta-analysis of findings from several large-scale followup studies reported modest but significant correlations between scores on the PCL and general recidivism, violent recidivism, and sexual recidivism.[7] Psychopathic individuals (i.e., those obtaining PCL scores of 30 or more) were approximately four times as likely to commit a future violent crime than were nonpsychopathic offenders. Moreover, the PCL "has been found in many studies to perform as well as (and in some cases better than) statistically derived actuarial measures designed specifically to predict future violence."[8] In the MacArthur Violence Risk Assessment Project, the PCL was the single best predictor of future violence out of a set of 134 predictors.[9] This does not mean that the PCL is fantastically accurate in predicting future criminal behavior, only that the PCL does a better job than other instruments designed for the same purpose. Whether this level of accuracy is acceptable for making decisions about an individual is a more complicated matter, and one that, as experts in risk assessment like John Monahan[10] have pointed out, is best determined by legal practitioners, not social scientists. One factor that must weigh heavily in any decision regarding the use of an imperfect risk assessment instrument is whether the consequences of a false positive classification-- that is, the classification of an individual as high risk when he in fact is not--carry unintended harmful effects. If, for example, the outcome of being classified as a psychopathic individual is long-term incarceration, it would not be advisable to accept a high false-positive rate. If, on the other hand, a classification of this sort is used simply to hold an individual for a few days pending further evaluation, the same false-positive rate may be acceptable in light of the potential benefits to community safety derived from identifying the small number of genuinely dangerous individuals. My concern is that assessments of juvenile psychopathy are not being used to recommend further evaluation but are instead forming the basis for definitive dispositional decisionmaking. The fact that assessments of juvenile psychopathy are being used to make decisions about the transfer of young offenders into the adult system--decisions that necessarily imply judgments about the likelihood of individual rehabilitation and that effectively determine whether any attempt will be made to rehabilitate the young offender--makes any false-positive problem especially worrisome. In sum then, at least when speaking about adults, we can say the following: --It makes sense to characterize some criminals as psychopaths. --It is possible to do so using the PCL. --Knowing whether an individual scores high on the PCL is useful information to have when trying to predict an individual's risk for future criminality or violence. It is against this backdrop that I want to consider the use of the PCL and other instruments designed to assess psychopathy to make decisions about juvenile offenders. Juvenile Psychopathy Assessment in Historical Perspective Psychopathy has become in recent years an attractive notion to those interested in serious juvenile offenders. It is instructive to examine the current interest in the juvenile psychopath in historical perspective because there are striking parallels between today's situation and that which existed in the early decades of the 20th century during the beginning years of the juvenile justice system. Historian Michael Willrich has written a marvelous description of what he calls "eugenic jurisprudence" in the early part of the 20th century,11 and it is fascinating to read this while pondering the juvenile psychopathy question before us today. It is the alignment of four factors today, as was the case 80 years ago, that is driving the contemporary search for the nascent Hannibal Lecter. These four factors are: (1) a problem; (2) a crisis; (3) a theory; and (4) a diagnostic tool. The Problem One of the most difficult problems facing practitioners in the juvenile justice system today is differentiating between serious offenders who are at risk of reoffending and those who are not. This is an especially important challenge to those making transfer recommendations, both because we want to protect the community from individuals who have a high risk of reoffending (especially if the risk is for violent reoffending) and because we do not want to expose relatively low-risk young people to the likely iatrogenic effects of adult sanctions. If it were possible to differentiate between juveniles who are at high risk of reoffending and juveniles who are not, we could, in theory, make far better decisions regarding transfer to the adult system, the decertification of juveniles charged as adults, and so forth. The problem facing legal practitioners in the early 20th century was much the same: how to draw meaningful distinctions among offenders who had committed similar offenses. As Harvard Law School professor Roscoe Pound noted early in the 20th century, "Criminals must be classified as well as crimes."[12] In 1909, the first court-affiliated psychiatric clinic was founded in Chicago, attached to the Nation's first juvenile court, and designed to "assist judges in devising a disposition or 'treatment' appropriate for each offender and to conduct policy-shaping clinical research into the 'root causes' of crime."[13] Coincidentally, and ironically in light of today's discussion, the clinic was named the Psychopathic Institute; a sister clinic, the Psychopathic Laboratory, served the municipal court. As the chief justice of the municipal court explained-- [T]he idea [of the Psychopathic Laboratory] marks a turning point from the traditional policy of society of treating the delinquents as a single large class without consideration of the various individual characteristics which distinguish them, and are now seen to divide them into a number of sharply differentiated classes, each with its separate proclivities, potentialities, and prognostic characteristics.[14] The Crisis Each problem noted above was made more pressing by either a crisis or the public's perception of one. Today's sense of urgency over the need to determine which offenders are genuine psychopaths has its origins in the now infamously wrong prediction about the coming wave of superpredators made by prognosticators like John DiIulio.[15] DiIulio has now softened his views on the inevitability of an epidemic of juvenile violence, but the legacy of the superpredator lives on today in the label of the psychopath. If DiIulio had a counterpart in the early days of the juvenile court, it may well have been Harry Olson, the chief justice of Chicago's municipal court. Olson's speeches about the growing "menace of the feeble-minded"--a term that was used to refer not only to the mentally retarded, but to individuals characterized by any number of mental deficiencies--sparked the same sort of panic and ill-informed changes in policy and practice as DiIulio's incendiary op-eds about superpredators[16] would do some 75 years later. Indeed, as Willrich points out, "by the mid-teens, the discourse about 'the menace of the feeble-minded' had spilled over from the welfare circles and professional journals into popular culture"17 in much the same way that the inside-the-Beltway musings of DiIulio and other conservative pundits eventually found their way into Time magazine and onto the evening news. There are parallels in practice between then and now as well. Just as today's prosecutors propose to use the diagnosis of psychopathy as grounds for the transfer of juveniles into the adult system, their forefathers advocated using feeblemindedness as grounds for commitment. In neither time period do practitioners pay much attention to the problem of false positives. The Theory Although warnings about the wave of superpredators in the 1990s and the epidemic of feeblemindedness in the 1920s were based more on rhetoric than reality, more on politics than precise evidence, the translation of rhetoric into policy and practice necessitated two other ingredients: a salable theory about the origins of the condition and an available diagnostic test to distinguish those who had the condition from those who did not. Interestingly enough, during both eras the theoretical perspective that carried the most weight emphasized the organic bases of criminal behavior. During the early part of the 20th century, the work of the neurologist William James Hickson, whom Judge Olson recruited to direct the municipal court's Psychopathic Laboratory, was especially influential, at least in Chicago, and as juvenile justice practice went in Chicago at the turn of the century, so it went in the Nation. Hickson, who had studied in Europe with Emil Kraepelin and Eugen Bleuler, two of the most influential thinkers in the history of psychiatry, and in the United States with Herbert Goddard, who popularized the use of the Binet-Simon intelligence test in this country, created "a fascinating synthesis of organic psychiatry and eugenic criminology."[18] Hickson believed that the central cause of criminality was not low intelligence (the conventional definition of feeblemindedness and at the time one of the presumed causes of criminal behavior) but an "affective defect" characterized by apathy, lack of remorse, and impulsivity, not unlike notions of "moral insanity" that had been popularized during the mid-19th century,[19] and remarkably similar to the callous emotional detachment that defines one of the core components of psychopathy as it is assessed by the PCL. Hickson believed that this affective defect was inherited in the form of "dementia praecox," a catchall diagnosis for serious mental disorder popularized by Kraepelin. Hickson also believed that it was possible to pinpoint the anatomical site of the defect: the lower brain. Although little is written today about the genetic basis of psychopathy, there is no shortage of references in current writing about antisocial behavior, or about psychopathy in particular, to the possible organic bases for the disorder. The notion that a subset of chronic offenders is biologically different from other offenders is central to the dominant theories of the development of antisocial behavior today, such as that put forth by Terrie Moffitt,[20] whose widely cited model of juvenile offending distinguishes adolescent-onset criminality (which presumably is environmental in origin) from life course-persistent criminality (which, in addition to its earlier onset and later offset, is presumed to have a strong biological basis). The notion is also prominent in the works of leading writers on the subject of psychopathy, including Joseph Newman,[21] Paul Frick,[22] Adrian Raine,[23] and, of course, Robert Hare.[24] Psychopathy has been attributed in recent years to low serotonin levels, low autonomic reactivity, unusual patterns of physiological arousal, and prefrontal cortical dysfunction. It has become increasingly common at professional meetings to see "pictures" of criminal brains; whether this is genuinely informative or simply high-tech phrenology remains a matter of some controversy. What's important, though, is that at both the beginning and the end of the 20th century, a theory or set of theories suggested that there exists a subpopulation of serious offenders whose criminality is caused by an affective defect, based in an organic brain deficit, and deserving of special treatment within the justice system. In each historical period, the scientific legitimacy of the theory provided a foundation on which a change in practice and, by extension, policy was fashioned. The Diagnostic Tool The confluence of a problem, a crisis, and a theory about serious offenders might influence debate, but to influence day-to-day practice it is necessary to not only lead practitioners to the water but provide them with the cup from which to drink it. Nothing succeeds in this task so well as a test. Indeed, as Thomas Grisso, a member of our MacArthur Network and expert on the forensic assessment of juveniles, has argued, we typically think of practice as being driven by policy, but in many instances, the reverse is true. As new assessment tools become widespread, changes in practice made possible by the introduction of new measures or assessment instruments often lead over time to changes in policy. In the 1920s, the search to identify criminals with the affective defect of dementia praecox was greatly facilitated by the availability of the Binet-Simon intelligence test, whose use had become increasingly widespread in the years following World War I. One would not think an IQ test would be very helpful in the search for affectively defective individuals, but in this case, necessity was the mother of misuse. Hickson, eager to please the judge who had appointed him (a judge who firmly believed that the menace of feeblemindedness was epidemic), to build his own reputation as a theorist and diagnostician, and to promote the Psychopathic Laboratory, began using the Binet, which had been designed solely to measure certain aspects of intelligence for purposes of identifying children who needed special educational programs, as a means of identifying individuals with the affective defect presumed to underlie intractable criminality. On the basis of individuals' scores on one subtest of the Binet--a visual memory test-- Hickson and his colleagues made diagnoses and recommendations to judges that were used to justify the incarceration of certain offenders. These psychiatric workups were especially important in cases involving juveniles and females because it was in these cases that the courts were most interested in assessing the amenability of the offender to rehabilitative intervention, which was precisely what Hickson believed he could assess.[25] All of the elements that led to interest in the diagnosis of dementia praecox among juvenile offenders in the early decades of the 20th century- -the problem of prediction, the perceived crisis of epidemic feeblemindedness among inner-city youth, and the theory of organically based criminality-- were present in a more contemporary form at the end of the 20th century. And just as the existence of a diagnostic tool (in the 1920s, the Binet-Simon test of intelligence) concretized these elements into a practice that influenced legal decisionmaking early in the 20th century, the growing and, in some senses, well-founded popularity of the PCL as a diagnostic instrument in assessments of adults more recently has made the use of versions of the PCL (or adaptations of it designed for the assessment of juveniles) all but inevitable in the evaluation of juvenile offenders. Whether the use of these youth-friendly PCL instruments--now, there's an oxymoron if there ever was one--for purposes of making transfer decisions today is as problematic as was the use of the Binet-Simon intelligence test for making commitment decisions 80 years ago is the topic I turn to next. The Utility of Juvenile Psychopathy Assessment Using a measure based on a well-validated and reliable instrument designed to assess psychopathy among adult offenders to perform a similar task among juveniles is not the same sort of problem as using a test of visual memory to diagnose intractable criminality. I am not suggesting that researchers and practitioners who are using currently available measures of psychopathy with juvenile populations are committing the same sort of egregious mistakes that Hickson was making by using the Binet to make diagnoses of affective defect. However, the translation of an instrument from one appropriate for use with one age group to one appropriate for use with another is far from worry free. The fact that a problem, a perceived crisis, and a theory about juvenile offending encourage us to grab onto an available diagnostic instrument should not make us ignore some very real dangers in doing so. Until certain fundamental questions about the use of the PCL and instruments derived from it are answered through systematic scientific research, practitioners should not be using such measures in forensic practice. Nonetheless, there is evidence that practitioners are using PCL-derived instruments for making transfer, sentencing, and decertification decisions, despite the cautions (some of them transparently half-hearted) raised by some of the researchers who are heavily invested in the use of these measures. In my view, there are three questions that we must answer before encouraging courts to consider assessments of juvenile psychopathy in making sentencing, transfer, and decertification decisions: 1. Do measures of psychopathy mean the same thing when used in adolescent populations as they do when used in adult populations? This is the question of construct validity. 2. Are scores on measures of psychopathy derived during adolescence correlated with scores on measures of psychopathy derived during adulthood? This is the question of stability. 3. Do scores on measures of psychopathy derived during adolescence predict antisocial behavior and violence during adulthood? This is the question of predictive utility. Research by Elizabeth Cauffman and Jennifer Skeem is about to be launched to examine these questions. When or until this work is completed, if the answer to any of these questions is "We don't know," the use of psychopathy assessments to make important decisions about juvenile offenders is premature. If the answer to any of these questions is "No," the use of psychopathy assessments to make important decisions about juvenile offenders is bad--perhaps even unethical--practice. To look at these questions of construct validity, stability, and predictive utility more closely, first consider the conceptualization and measurement of psychopathy among adults. As noted earlier, the construct of psychopathy is predicated on the notion that psychopathic individuals have a deep-seated, unchanging pattern of personality and behavior that has its roots in early experience and/or biology. The pattern is characterized by two broad factors: callous, emotional detachment, which includes such traits as grandiosity, manipulativeness, shallow affect, and failure to accept responsibility for one's own actions; and an unstable antisocial lifestyle, which includes such traits as impulsivity, irresponsibility, delinquency, and proneness to boredom. Callous, antisocial individuals are presumed to be at higher risk for continued antisocial behavior and violence. Construct Validity Let me begin with the question of construct validity. Do the indicators of psychopathy when assessed in adolescence mean the same thing as they do when assessed in adulthood? A different way of asking this is to ask whether the correlates of psychopathy are the same during different developmental periods. We know that it is not always the case that traits assessed at one point in development have the same meaning as they do at a different point in development. Consider, for example, an individual who refuses to cooperate with other people. During early childhood, this may indicate healthy independence; during adulthood, this may indicate selfishness or antisocial behavior. Or consider this: Imagine that we separate a mother from her child for 10 minutes and then we reunite them. Clinging behavior in a young child indicates a strong and healthy attachment, which is correlated with a variety of measures of positive functioning; the same clinging behavior in an older child may indicate insecurity and would likely be correlated with indicators of emotional and behavioral problems. One worry I have about the assessment of psychopathy among adolescents is that many behaviors we associate with normal adolescent development are the same behaviors we associate with psychopathy in adults. To be sure, some of the items used to judge psychopathy likely apply to younger as well as older offenders. These include pathological lying, manipulativeness, shallow affect, and poor anger control. For the most part, I think that these constructs mean similar things when observed in an adolescent as they do when observed in an adult. (The appropriateness of these items in assessments of preadolescent children, where they may in fact be problematic, is not addressed here.) But many items that are central to the definition of psychopathy among adults may erroneously be viewed as indicators of psychopathy in an adolescent population and confused with normative adolescent development. These items include grandiosity, proneness to boredom, lack of remorse or guilt, impersonal sexual behavior, goallessness, impulsivity, irresponsibility, failure to accept responsibility for one's actions, and unstable interpersonal relationships. I say this both as the parent of a nonpsychopathic 16-year-old and the coinvestigator of a longitudinal study that tracks serious juvenile offenders over time. Because my son does not permit me to discuss our personal life in public (owing to the fact that he is glib, grandiose, egocentric, and impulsive), I will comment only on our experiences in the research. The MacArthur Foundation Research Network on Adolescent Development and Juvenile Justice, which I direct, is conducting a number of studies aimed at understanding the ways in which information about normal and atypical adolescent development can improve the quality of decisionmaking concerning young offenders in the juvenile and criminal justice systems. Among these projects is a large-scale study of adolescents' competence to stand trial, a comparison of the impact of juvenile versus adult sanctions on serious offenders, several studies of the ways in which adults judge adolescents' criminal culpability and blameworthiness, research on legal socialization (how young people acquire attitudes about the law), and a longitudinal study of "pathways to desistance." (More information about the Network and its projects is available on our Web site, www.mac-adoldev-juvjustice.org.) In our study of pathways to desistance, we are tracking 1,200 juvenile felons in Philadelphia and Phoenix to understand how they are affected by their experiences in the juvenile and criminal justice systems. Among the measures of individual functioning we use in this study is the Youth Version of the Psychopathy Checklist (PCL-YV), which is considered the state of the art in the assessment of psychopathy among adolescents. At our site in Philadelphia, we struggle often over how to interpret the behavior of the 14- to 17-year-olds in our sample in light of the items that compose the PCL-YV. Does a 14-year-old who blames his friends for dragging him into a robbery qualify for a rating of failing to accept responsibility for his actions, or is he accurately describing a heightened susceptibility to peer pressure that is characteristic of teenagers at this age? Is a 15-year-old who does not know what he wants to do with his life goalless, or is this indicative of an individual in the midst of an adolescent identity crisis? Is a 16-year-old who has a new girlfriend every other week someone who has unstable interpersonal relationships, or is he merely trying to figure out who he is and what he wants in a romantic partner? Is a 17-year-old who believes that he can drive while high grandiose or suffering from adolescent egocentrism? Even Cleckley, whose groundbreaking description of the psychopathic personality formed the basis for the PCL worried about this: Confused manifestations of revolt or self-expression are, as everyone knows, more likely to produce unacceptable behavior during childhood and adolescence than in adult life. Sometimes persistent traits and tendencies of this sort and inadequate emotional responses indicate the picture of the psychopath early in his career. Sometimes, however, the child or the adolescent will for a while behave in a way that would seem scarcely possible to anyone but the true psychopath and later change, becoming a normal and useful member of society. Such cases put a serious responsibility on the psychiatrist.[26] In the same way that Moffitt[27] argues that we cannot distinguish between adolescence-limited offenders and life-course persistent offenders by observing them only during adolescence, because their observable behavior may be identical, I wonder whether we can distinguish between psychopathic and nonpsychopathic teenagers on the basis of their scores on the PCL. The problem is that some items on the PCL describe characteristics that, while potentially indicative of problems among adults, may indicate normative development (or at least development within the normative range) among adolescents--transitory characteristics that disappear in most individuals by young adulthood. However, by definition, psychopathy is not something that individuals grow out of. Which brings me to the issue of stability. Stability Assertions about the presence of psychopathy among the young are necessarily based on the presumption that scores on measures of psychopathy are stable over time. To my knowledge, we do not really know if this is true among adults, but the fact that we do not have data at all on the stability of PCL scores among juvenile offenders over time is particularly problematic, since adolescence is inherently a time of change. I should note that the DSM does not permit the diagnosis of antisocial personality disorder, the closest diagnosis in the psychiatric lexicon to psychopathy, among individuals under the age of 18. It is not clear why the diagnosis of antisocial personality disorder among adolescents is problematic but the diagnosis of psychopathy is not. Furthermore, it is not simply that psychopathy is presumed to be stable over time; it is also presumed to be resistant to change. Unfortunately, the stability of traits over time is generally studied under constant contextual conditions. Assertions that aggression or impulsivity or antisocial behavior are stable traits are generally derived from longitudinal studies that do not consider whether the individual's social environment remained unchanged during the same time period. For obvious reasons, indicators of stability under constant conditions may not provide accurate estimates of stability under changing conditions. Therefore, saying something is stable under natural conditions (i.e., if nothing else changes) is not the same as saying it is immutable. To research this, we need more than studies that simply measure psychopathy at two points in time. Before we use assessments of adolescent psychopathy to draw inferences about the amenability of young offenders to rehabilitation, we need experiments that assign putative juvenile psychopaths to treatment and attempt to change them. We have some evidence that this is difficult to do in adults, but no evidence either way in adolescents. Predictive Utility My final worry concerns the predictive utility of the PCL in juvenile populations. The question is whether scores on the PCL that are derived during adolescence are predictive of antisocial or violent behavior in adulthood. At first glance, this appears to be the easiest of the three questions to answer, because it requires only that we assess psychopathy during adolescence and antisocial or violent behavior during adulthood and see if the first predicts the second. Indeed, at least one team of researchers has done this, and the results indicate that adolescents' scores on the PCL are predictive of the number of times they are convicted for violent offenses before age 21.[28] It is not quite so simple, however, to judge the predictive utility of the psychopathy assessments. The reason for this difficulty inheres in the two-factor structure of the PCL. (Factor I refers to the emotional and interpersonal aspects of psychopathy and Factor II refers to the antisocial aspects of psychopathy.) We know from a vast literature on antisocial behavior that the single best predictor of future antisocial behavior is past antisocial behavior and that the single best predictor of future violence is past violence. At issue here, then, is whether psychopathy in adolescence per se is an especially good predictor of future antisocial or violent behavior. Demonstrating that individuals who have offended in the past are at greater risk of offending in the future is not exactly the stuff that Nobel Prizes are made of. To the extent that the predictive utility of the PCL or any other measure of psychopathy inheres solely in the well-established link between past and future antisocial behavior, we might as well just assess past antisocial behavior and forget about trying to distinguish between psychopathic and nonpsychopathic individuals. And if this is the case, high scores on the PCL tell us nothing about individual amenability, which is the sort of inference that is being drawn from the PCL. Another way of thinking about this is to ask whether there is any "value added" to be gained by assessing the elements of psychopathy that are not indicators of current or prior antisocial behavior, such as those items that make up the factor defined by callous, emotional detachment. Unfortunately, extant research does not provide a clear-cut answer to this question. Although the connections between current psychopathy and later offending are real, the links are due largely to the predictive significance of the antisocial factor, which demonstrates what we know all too well: that earlier offending is predictive of later offending. In samples of adults, the links between the non-antisocial variables--grandiosity, glibness, shallow affect, and so on--and subsequent offending are statistically significant but very modest in size.[29] Before we can recommend the use of psychopathy assessments in dispositional decisionmaking regarding juveniles--especially in dispositional decisionmaking that is based on assumptions about amenability--we need more research that looks specifically within the adolescent population at the predictive utility of those elements of psychopathy that are not themselves indicators of current antisocial behavior. Questions about the stability and predictive utility of the PCL and related measures can be answered with well-designed longitudinal studies. Although we do not have such research yet and, although such studies are time-consuming and expensive to conduct, they are clearly within the realm of scientific possibility. In closing, I want to return to what I think is the most difficult of the three questions I raised: whether high scores on the PCL during adolescence indicate genuine psychopathy or something that is less worrisome. Many years ago, Anna Freud, the pre-eminent psychoanalytic theorist interested in adolescence, wrote what now is considered to be a classic article, "Adolescence as a Developmental Disturbance."[30] Although few experts in adolescent development continue to hold the view that we should see the period as one of temporary insanity, at least some of what Freud implied in this article is worth heeding. Many behaviors that adolescents engage in are behaviors that, if demonstrated by an adult, would indeed be indicative of psychopathology, if not necessarily psychopathy. Psychopathy is, by definition, not something that individuals grow out of. Adolescence, by definition, is. It is important that we do not confuse one with the other. Question-and-Answer Session Jenni Gainsborough, Senior Policy Analyst, The Sentencing Project, Washington, D.C.: I actually have three quick questions about the way the PCL is being used. First of all, is the PCL being widely used to decide whether juveniles should be transferred into the adult system? Second, you talked about a youth version of it; does the youth version take into account the specific characteristics of adolescence that are problematic? And third, we know that a lot of the fear about "superpredators" was racially based, and also we know that an extraordinarily high percentage of juveniles being transferred into the adult system are African-Americans; do these tools look at racial differences in any way? L.S.: Yes, no, and no. First, the PCL is being very widely used in Canada now. Robert Hare and his associates are Canadian and the popularity of the instrument has always been greater north of the border. It is becoming more widely used in the United States. No one has any data on this; but I receive calls regularly from public defenders asking if, in fact, this is a valid test. As to the second question, the attributes characteristic of normal adolescent development are identical to the youth version of the PCL (exhibit 1). The translation of the adult version to the youth version simply substituted words like "relationships" for "marriage" or minimized the importance of "work" and substituted "long-term goals." It is virtually the same instrument. The third question on race is a very important one. Some studies suggest that the instrument performs differently among African-American and white offenders in adult samples. That is to say that it is somewhat less useful in studies of African-Americans than in studies of white offenders. We do not know the answer to this question for the juvenile version. It has not been studied to my knowledge. Devon Brown, Deputy Trustee, Office of the Corrections Trustee of the District of Columbia, Washington, D.C.: The PCL is being used here in this country. In fact, just north of our border at the Patuxent Institution, its use is widespread. But my comment has to do with Robert Hare. The strength of his instrument, when applied to adults, is also its weakness when used with juveniles. Hare argues that sociopathy is due to a developmental lag. In both his manual and his research findings, he bases that instrument on this theory--that the characteristics (as you point out) are typical of adolescent behavior. It's just that they are demonstrated within an adult population. So in terms of your presentation, you need to be aware of that. L.S.: Thank you. Then I find it very curious that Hare is one of the coauthors of the youth version of the Psychopathy Checklist and is one of the team that is marketing it as a measure that can be used with juveniles. Either he has changed his theory or he is doing something that is inconsistent with what he has written. Lisa Greenman, Staff Attorney, Mental Health Division, Public Defender Service, Washington, D.C.: I wanted to ask you to comment on two things that you have mentioned during your talk. One has to do with the extraordinary nature of the marketing of these instruments. And actually the last question reminds me that in addition to the youth version there is now a Hare P-Scan version. I would like to hear more about the very aggressive marketing. My other question has to do with the significance of a very high false positive rate in the use of the PCL-R with adults, which is where it makes its strongest showing. I know some of your colleagues like Ed Mulvey have described it as the strongest in a field of weaklings. I wondered if you could expand on this subject. My first exposure to the construct of psychopathy and the use of the PCL-R for predicting future violence was in death penalty cases, where prosecutors were arguing that high PCL scores should be used by a jury to sentence an individual to death rather than to life without parole. In that context, I learned that the false positive rate (among people who were high scorers on the PCL) was, in many studies, shown to be less effective than a coin toss in predicting future violence. In some studies it was slightly more than 50 percent effective. But the false positive rate was right around 50 percent and sometimes even greater. I would like you to comment on the rate of false positives and how it affects your recommendations on how much courts should rely on this instrument in a forensic context. L.S.: First, there are, in fact, other measures being marketed for use in identifying juvenile psychopaths, some of which their creators market as usable with people as young as 10. These are sometimes questionnaire versions of a lengthy clinical interview administered to the respondent either directly or through parents or teachers. I think that if you look into this, you would find that the individuals who are developing these measures give--as I noted in my talk--a kind of half-hearted caution that these measures should be used for research purposes only but, at the same time, urge practitioners to "send your check to this company and we will gladly send you the measure." What happens next is that the measures find their way into the hands of practitioners who may or may not read the fine print that this is a research tool and not a proven clinical instrument. They then use it to make a diagnosis; they present it in court to a judge who doesn't know, and perhaps shouldn't be expected to know, the difference between a research instrument that is still in development and a diagnostic tool that has clinical validity. I think that if the people who were developing these instruments were in the room they would argue that there is a caution on the cover of their instrument that states it is a research tool. But nevertheless, by selling it to other individuals they are leaving open a very, very wide door for prosecutors and other legal practitioners to walk through. As to the problem of false positives, as I mentioned before, it depends on what you do when somebody scores positively. Any instrument that measures anything, even forecasting the weather, will have false positives associated with it. If the danger of a wrong forecast is that you carry your umbrella to work when you didn't need to, then that false positive is not especially problematic. If the danger of a false positive in the PCL assessment is that we are going to want to further assess this individual because we think he might be psychopathic, then I can live with a 50 percent false positive rate. If the danger of a false positive assessment is that the person gets a capital sentence, then I can't live with a 50 percent false positive rate. So, I think you can't separate questions about whether the false positive rate is high or low without asking what the consequences of a false positive assessment will be. I have argued the same point about the increasing use of Mosaic and other kinds of profiling programs to identify at-risk kids in schools because these assessments also have tremendously high false positive rates. The issue really is, what are we going to do with a kid who is rated at risk for violence? If the consequence of that assessment is suspension from school, that false positive rate is not acceptable. If the consequence is to talk to this kid a little more and learn a little more about him, then I can live with a false positive rate. So I think that the answer is: Yes, the PCL is the strongest measure in a field of "weaklings"--we are very bad at predicting violent behavior. But people need and want to make predictions and this is the best tool currently out there. Robert Stephenson, Director, Division of Workplace Programs, Center for Substance Abuse Prevention, U.S. Department of Health and Human Services, Washington, D.C.: You had not identified substance use or abuse as any one specific measure in your discussion. Could you briefly discuss the relationship of substance use or abuse in the PCL and pyschopathy, both in a youth version and an adult perspective? And are there any specific drugs or patterns of use that might be more predictive than others? L.S.: Drug use, in and of itself, is not an item that is used to predict psychopathy. There is an item on the list, you may recall, for juvenile delinquency (and in the adult version for past criminality) and drug use, particularly criminal drug use. I am not aware of any research that suggests that either drug use in general or the use of specific drugs is indicative of psychopathy. We do know, of course, that individuals who use and abuse drugs and alcohol are more likely to be involved in a range of antisocial activities. But to my knowledge it does not make them at heightened risk for psychopathy. Joanne Wiggins, Program Analyst, Planning and Evaluation Service, U.S. Department of Education, Washington, D.C.: You mentioned racial differences in the use of the instrument. Can you say something about differences between males and females? L.S.: There are very, very few females who score high on the PCL. It's not a completely male disorder, but it is not very often found among females and, to my knowledge, I don't know any studies that would allow us to really make a statement about its differential utility in males and females. I was asked to comment on what judges ought to rely on if they can't rely on this assessment tool. I think that individualized assessments of serious juvenile offenders--assessments done by skilled, developmentally sensitive forensic and clinical evaluators--can be very useful in helping judges formulate appropriate dispositions or sentences for juvenile offenders. But there is a danger when we begin to rely on standardized instruments that may become widely used in practice by individuals who don't have a great deal of training and who simply will turn over a score to a judge and say, "This person scored a 32; therefore, he is a psychopath." I certainly believe that psychology and psychiatry should play a role in helping courts formulate sentencing and dispositional decisions; I just worry that handing out a test and using people's scores on a test that has unproven validity in this population is not wise practice. Denise M. Juliano-Bult, Chief, Systems Research Program, Services Research and Clinical Epidemiology Branch, National Institute of Mental Health, Washington, D.C.: Could you say anything briefly about inroads to effective treatment for adolescents who do score high on the PCL? L.S.: I don't think we know anything about it. You know there have been literally a handful of studies of kids who scored high on the PCL. To my knowledge, there has never been a study done that has looked at interventions designed for kids with high PCL scores to see how that affects their behavior. My point, today, is not to provide definitive answers about this, but to raise cautions and questions about a practice that is just beginning to burgeon, but one that I think we need to be thoughtful about. Barbara T. Roberts, Senior Policy Analyst, Office of National Drug Control Policy, Washington, D.C.: You mentioned cautioning people against looking at a score and making certain cursory diagnostic determinations. What ethical responsibilities do you think should be imposed on people who may be in the field and still purport to make such recommendations? How do you propose to curb that? L.S.: Well, I guess for starters, I don't think it's wise to be distributing--in fact selling-- unvalidated instruments. Assessing psychopathy in juveniles may or may not turn out to be useful. I am not saying here today that I am certain it is not useful; I am saying we simply don't know. We need the kind of longitudinal research that I discussed earlier to look at how this tracks over time, to look at what it predicts, to look at how stable it is, before we move from research into practice. The fact that it is being used in practice now indicates to me that the individuals who developed these measures are not being appropriately cautious. Notes 1. I am grateful to Elizabeth Cauffman, Thomas Grisso, Daniel Seagrave, and Jennifer Skeem for graciously making available to me prepublication drafts of manuscripts on closely related topics. I draw heavily on their insightful analyses here. Thanks as well to Elizabeth Cauffman, John Monahan, and Jennifer Skeem for their comments on an earlier draft of this paper. 2. Edens, J., J. Skeem, K. Cruise, and E. Cauffman, "The Assessment of Juvenile Psychopathy and Its Association With Violence: A Critical Review," Behavioral Sciences & the Law 18 (2000). 3. Poythress, N., J. Skeem, S. Lilienfeld, and J. Edens, "Personality Features and Antisocial Behavior," unpublished NIMH grant proposal, 2000. 4. Ibid. 5. Harris, G.T., T.A. Skilling, and M.E. Rice, "The Construct of Psychopathy. Crime and Justice: An Annual Review of Research," unpublished manuscript, 2000. 6. Cooke, D.J., and C. Michie, "Refining the Construct of Psychopathy: Towards a Hierarchical Model," Psychological Assessment 13(2) (2001). 7. Hemphill, J.F., R.D. Hare, and S. Wong, "Psychopathy and Recidivism: A Review," Legal and Criminological Psychology 3 (1998): 139-170. 8. Edens, et al., 2000. See note 2. 9. Steadman, H., E. Silver, J. Monahan, P. Appelbaum, P. Robbins, E. Mulvey, T. Grisso, L. Roth, and S. Banks, "A Classification Tree Approach to the Development of Actuarial Violence Risk Assessment Tools," Law and Human Behavior 24 (2000): 83-100. 10. Monahan, J., The Clinical Prediction of Violent Behavior, Washington, DC: National Institute of Mental Health, 1981. 11. Willrich, M., "The Two Percent Solution: Eugenic Jurisprudence and the Socialization of American Law, 1900-1930," Law and History Review 16 (1998): 63-111. 12. Ibid, 82. 13. Ibid. 14. Ibid, 82. 15. Bennett, W., J. DiIulio, and J. Walters, Body Count, New York: Simon & Schuster, 1996. 16. DiIulio, J., "The Coming of the Super-predators," Weekly Standard (November 27, 1995): 23. 17. Willrich, 90. See note 11. 18. Willrich, 89. See note 11. 19. Maughs, S., "A Concept of Psychopathy and Psychopathic Personality: Its Evolution and Historical Development: Part I," Journal of Criminal Psychopathology 2(2) (1941a): 329-356; Maughs, S., "A Concept of Psychopathy and Psychopathic Personality: Its Evolution and Historical Development: Part II," Journal of Criminal Psychopathology 2(3) (1941b): 465-499. 20. Moffitt, T. E., "Adolescence-limited and Life-course-persistent Antisocial Behavior: A Developmental Taxonomy," Psychological Review 100 (1993): 674-701. 21. Newman, J., "Psychopathic Behavior: An Information Processing Perspective," In D. J. Cooke, A. E. Forth, and R. D. Hare, eds, Psychopathy: Theory, Research, and Implications for Society, Dordrecht, The Netherlands: Kluwer (1998): 1-11. 22. Frick, P., Conduct Disorders and Severe Antisocial Behavior, New York: Plenum, 1998. 23. Raine, A., and J.H. Liu, "Biological Predispositions to Violence and Their Implications for Biosocial Treatment and Prevention," Psychology, Crime and Law 4 (1998): 107-125. 24. Hare, R. D., "Performance of Psychopaths on Cognitive Tasks Related to Frontal Lobe Functions," Journal of Abnormal Psychology 93 (1984): 133-140. 25. Willrich. 26. Cleckley, H., The Mask of Sanity (5th edition), St Louis: Mosby, 1976. 27. Moffitt, T.E. See note 20. 28. Forth, A., and H. Burke, "Psychopathy in Adolescence: Assessment, Violence and Developmental Precursors," In D. Cooke, et al., eds, Psychopathy: Theory, Research and Implications for Society, Boston: Kluwer Academic (1998): 205-229. 29. Skeem, J., and E. Mulvey, "Psychopathy and Community Violence Among Civil Psychiatric Patients: Results From the MacArthur Violence Risk Assessment Study," unpublished manuscript, Law and Psychiatry Program, University of Pittsburgh, 2001. 30. Freud, A., "Adolescence as a Developmental Disturbance," Writings 7 (1969): 39-47.Kluwer: 1- 11. ------------------- Violence Prevention 101: Implications for Policy Development Presentation by Carl C. Bell President/C.E.O. Community Mental Health Council, Inc. Director of Public and Community Psychiatry Professor of Psychiatry and Public Health University of Illinois April 20, 2001 Washington, D.C. Violence Prevention 101: Implications for Policy Development Violence prevention rests on some basic principles that support systemic changes in health behavior. The discussion focuses on two basic premises of violence prevention that can inform public policy. The presentation begins with a broad look at violence within a variety of social sectors and age ranges. It then narrows the discussion to an examination of the school age population, which reflects how early implementation of violence prevention principles among these age groups can have a dramatic effect not only on violence reduction, but also in very real terms on individual outcomes. The first premise of violence prevention, often overlooked by practitioners when they design violence prevention programs across age and ethnic/racial groups, is that there are different types of violence (see "Types of Violence"), and each type requires different prevention, intervention, and followup strategies.[1] A second premise, similarly neglected, is that different racial/ethnic groups experience different types of violence at different rates, and each calls for different prevention strategies. The following comparisons exemplify these different rates. Most murder cases involve a victim and offender of the same race or ethnicity. Among groups, homicide rates for African-Americans are 6 times and Latino rates are 2.3 times the Native American, Asian-American, and white rates of homicide. Other crime categories also reflect differences. The difference in violence rates between African-Americans and whites reflects the vast differences in poverty and social infrastruture between these two groups.[2] In other violence comparisons among groups, interpersonal altercation and domestic homicide are major forms of violence in the African-American community. But because battered women now can rely on support from women's shelters to escape their batterers, the rate of homicide of African-American men by African-American women no longer exceeds that of African-American women by African-American men.[3] Gang-related violence is prevalent in the Latino community. Street violence is also common among Latinos: gun violence predominates, but knifings also are common.[4] Compared with other groups, however, Latino men are less likely to perpetrate domestic homicide. Interpersonal altercation is also common in the Native American community. Forty percent of Native American murder victims are killed by non-Native American offenders, and in 33 percent of the cases, the offender was white.[5] Suicide is a major form of violence in the white community. Whites represent most of the offenders motivated by anger or revenge, and in domestic- or romantic-related mass murders in which the offender also kills himself.[6] In anger/revenge mass murders in which specific persons were targeted, 66.5 percent of the perpetrators were white and 66.7 percent of these perpetrators attempted suicide. In anger/revenge mass murders in which a specific place was targeted, 76.5 percent of the perpetrators were white and 38.9 percent of perpetrators attempted suicide. In similar mass murders with diffuse targets, 80 percent of perpetrators were white and 30 percent of perpetrators attempted suicide. In domestic/romantic-related mass murder, 57.1 percent of the perpetrators were white and 28.6 percent attempted suicide. Research shows that among adolescents, suicide is extremely rare for those who lack a major mental disorder.[7] Research has also shown that mass media coverage of suicide may trigger copycat behavior in vulnerable adolescents who have major mental disorders.[8] In summary, if we are serious about effecting policy decisions concerned with violence prevention and systemic changes surrounding public health, two premises must be taken into account when designing violence prevention programs: 1) There are different types of violence, and 2) violence rates and prevention vary by ethnic groups. Once we have addressed these fundamental premises, we must cast a sufficiently wide net that incorporates certain principles that are not always closely linked to violence prevention. Violence Prevention Principles The basic principles of violence prevention as I see them are-- --Rebuilding the village. --Providing access to health care. --Improving bonding, attachment, and connectedness dynamics. --Improving self-esteem. --Increasing social skills. --Reestablishing the adult protective shield. --Minimizing the residual effects of trauma. These principles are interdependent and are key guiding principles for effecting large, systemic changes in health behavior.[9] Rebuilding the Village Social disorganization theories of deviance suggest that poverty, lack of job opportunities, single-head households, isolation from neighbors, and weakened community networks and institutions lead to reduced informal and formal social control,[10] which, in turn, promotes violence. Research demonstrates that communities need protective factors, such as an infrastructure and social fabric, to deliver effective public health interventions. Unfortunately, communities have varying levels of these protective factors.[11] In communities where they are lacking, a facilitator might be needed to help create community partnerships.[12] For example, evidence indicates that school-based interventions that shift how high-risk children are managed--by increasing parental involvement in school and collaboration with school personnel--can reduce risk, particularly for in-school violence.[13] Another violence prevention strategy for child and adolescent populations might be to pair schools with community-based secular and faith-based organizations to develop activities to reduce violent and disruptive behavior by and against youths in the schools and surrounding communities. By emphasizing a shared vision among cooperating groups, a facilitator can encourage a mission-driven philosophy that transcends the goals of individual agencies in favor of a more overarching vision. This can be accomplished by emphasizing the interdependency among diverse elements in a community, encouraging affiliation among these elements, promoting development of "systems" thinking, and providing leadership to community organizations--religious, business, social service, health, educational, civic, and social--to enable them to synergize their efforts to promote healthy development of youth. Several examples of the "rebuilding the village" concept are occurring in Chicago. One of these is the Community Mental Health Council/University of Illinois/Juvenile Court Juvenile Delinquent Assessment Program, which provides forensic assessments of youths in juvenile court. It will also provide the foundation for collaboration among three African-American and three Latino community-based organizations so that needed services will be available to youths and their families upon release from court custody. Over a 3-year period, a coalition composed of the Chicago Crime Commission, the University of Illinois, the Chicago Police Department, Chicago Public Schools, the Community Mental Health Council, the Illinois Department of Children and Family Services, Allstate Insurance Company, and community leaders led to development of the Chicago Crime Commission Community Youth Program. Additionally, there are plans for a behavior disorders clinic at the University of Illinois that will share a child psychiatrist with the Community Mental Health Council and address clinical needs of children identified by the previous collaboration. Through collaborations to assess the nature and size of the violence problem, these institutions are better able to facilitate "rebuild the village" efforts. These initial accomplishments convinced involved organizations that to effect positive change, they needed to work together to maximize their pooled resources. Providing Access to Health Care The second principle involves raising awareness among healthcare professionals of factors that affect individual health, particularly their mental health. A better understanding of the many factors affecting health can improve diagnoses, reduce gaps in delivery of services, and make management of treatment strategies more appropriate. The following examples underscore the value of more comprehensive medical screening for determining the causal links between health and behavior. Research supports, for example, the hypothesis that neuropsychiatric disorders among adolescents and children may predispose individuals to violence; specifically, neurophysiological brain impairments acquired since birth, i.e., not genetic biology, can lead to difficulty in bonding/attachment, poor social skills, and lack of impulse control. Environmental factors may also create physiologic responses that influence behavior. Research, for example, indicates that children with high levels of exposure to lead may be predisposed to violence.[14] Among male juvenile delinquents, at least one-third have an impairing mental disorder, aside from conduct disorder, that has not been adequately treated. Many of these children are also substance abusers. Two of the more common mental disorders for this group--ADHD and depression--are treatable. Children with conduct disorders--with and without accompanying ADHD--also can be adequately treated with medication.[15] As seen by these examples, timely and proper diagnosis and treatment programs are critical to improve overall health and individual outcomes in a variety of areas such as school. Juvenile offenders may also be at high risk for exposure to multiple types of trauma as a result, for example, of family and peer