2007 Speeches

WHAT WORKS?  Reducing Crime and Delinquency

March 22, 2007

Dr. Roy L. Austin
United States Ambassador
American Chamber of Commerce
Grande Riviere Room, Courtyard by Marriott

It is not widely understood that there are times when crime and delinquency resist valiant efforts to reduce them, sometimes trending upwards despite our exertion for a contrary result. I shall demonstrate the validity of this statement by showing you some U.S. homicide data.

Although the U.S. homicide rate started climbing about 1962, a downward trend did not begin until 1994. By 1972, the rate per 100,000 (9.4) was twice the rate in 1960 (4.7). From 1972, the remainder of the 1970s, 1980s and the early 1990s showed a fluctuating rate with no deep valleys or high mountains, rates remaining between 9.4 and 10.7 (1980) per 100,000.

In Trinidad, The number of reported homicides per year from 1999 through 2006 in chronological order is 92, 119, 150, 172, 229, 260, 386, and 369, there being a continuous upward trend to 2005. Using questionable estimates of the population size for T&T in 1999 and 2005, I calculated homicide rates per 100,000 for these two years as 13.68 and 30.02, respectively. That is, during that six-year period, the rate had increased 2.2 times. Also, for the periods compared, T&T’s homicide rate increased 2.18 times as fast as that of the United States.

What accounts for the relatively rapid increase in the homicide rate in T&T compared with that in the U.S.? There has been much speculation about the T&T increase, mere speculation. For the U.S., I have conducted research on juvenile delinquency data covering 1971 through 1986. I examined the relationship of three demographic variables (the percentage of 15- to 17-year olds, female headship, and median family income) to the eight most serious offenses used to indicate the seriousness of crime, separately and combined.

My results showed that, except for age-composition with burglary, there was no significant effect of the causal variables on black delinquency. However, as the proportion of white 15- to 17-year olds increased, the white juvenile arrest rate for robbery, aggravated assault, burglary, larceny-theft, and all offenses combined increased significantly. Also, among white juveniles, burglary and the totality of serious offenses increased significantly as median family income decreased over time. Notably, neither among blacks nor whites was there a significant relationship between female headship and any delinquent offense.

I have also conducted an empirical analysis of longitudinal T&T crime data and demographic data between 1970 and 1984. My goal then was to determine whether increasing petro-dollars in the 1970s may have caused an increase in serious crime rates. While there was an upward trend in GDP beginning in 1976, measures of homicide and murder separately showed no sustained upward trends. Also, while there was an upward trend in burglary rates for the relevant period, these rates took off prior to the GDP take-off. Thus, economic growth cannot be seen as causing an increase in homicides in T&T.

Of course, one notable change in T&T during the recent (1999-2006) upsurge in homicide rates is another economic boom from energy dollars. This rapid economic improvement may be occurring in a different context than that in the 1970s. Therefore, the expansion in homicide may have economic expansion as a cause in this recent period even if these variables were not causally related in the 1970s. A systematic study is required in order to provide a convincing answer to this question.

Another study of T&T data collected by UWI’s Centre for Criminology and Criminal Justice for our Embassy’s International Narcotics and Law Enforcement (INL) program yielded some interesting findings. Among Standard 3 students, frequency of delinquency did not differ by socioeconomic status or by family structure; but Afro-Trinidadian students were more frequently delinquent than Indo-Trinidadians and Mixed-Trinidadians. Trinidadian data on the racial composition of prisons do lead us to expect the racial difference. However, frequent claims that social class and father-absence cause crime are not supported by these T&T data; and it is noteworthy that the T&T and American data agree that father absence is not related to delinquency.

Although the UWI Centre’s study provides a point of entry into the discussion of crime-reduction programs that work, I shall postpone that conversation in favor of informing you about the many points of intervention for amelioration of crime rates. Figure 2 says basically that more criminally motivated persons are more likely to commit crimes than persons less criminally motivated. Additionally, where there are better opportunities to commit crime, crime is more likely to occur. Furthermore, when criminally motivated persons perceive that there are good opportunities to commit crime with impunity, they are more likely to attempt to commit crime than under contrary conditions. Moreover, Figure 2 suggests that many different modes of intervention may be effective in reducing crime. However, experience should tell us that many programs which seem as if they should be effective have proven to be ineffective. I shall quickly describe one such program before moving to some for which there is empirical evidence of effectiveness.

Juvenile Aversion Programs

Many other similar programs followed the original. Regrettably, independent scientific evaluation of the original and other programs yielded results showing that this kind of intervention was ineffective; and some data even suggested that it might be harmful. I turn next to what may arguably be the most effective program for preventing criminality.

The High/Scope Perry Pre-School Program

The subjects for this study were 123 African Americans born in poverty and at high risk of failing in school. At ages 3 and 4, they were placed into a special preschool group and a no preschool group. Data have been collected on these groups at least seven times from age three through 39-41. One inquiry describes the results of the study thus: “The study shows that a high-quality program for young children living in poverty, over their lifetimes, improves their educational performance, contributes to their economic development, helps prevent them from committing crimes, and provides a high return on taxpayer investment” (Schweinhart 2003; Paper presented at the Meeting of the Society for Research in Child Development).

A publication of the U.S. Department of Health and Human Services is more specific about the programs benefits. At age 27, the former High/Scope program group had 63% fewer habitual criminals (five or more lifetime arrests); 68% fewer arrests for drug use; 26% fewer adult welfare or other social service recipients; nearly twice as many homeowners; significantly higher achievement and literacy scores; and other desirable characteristics.

The Chicago Child-Parent Centers Study has also reported evidence that their preschool programs helped to reduce crime. At the same time the Abecedarian study did not find evidence that its pre-school program helped to prevent crime.

Some of you may be thinking that that the results of American programs are unlikely to be replicated in Trinidad and Tobago. It just happens that Servol’s Early Childhood Care and Education (ECCE) programs may have benefited from an association of at least one key staff member with the High/Scope program. A Servol document states that “before initiating its graduate training program in 1980, the coordinator of the servol project was sent to HIGH/SCOPE for special training through the courtesy of Dr. David Weikhart, Director, who was a personal friend of SERVOL’S chairman.”

Also, there is some evidence that Servol’s ECCE is more effective than other ECCEs and no ECCE programs in T&T in preventing delinquency. By Standard 3, the UWI study mentioned earlier showed that Servol students reported being less likely to use force to get something than other ECCE students. Servol students also reported being less likely to be in a fist fight than other ECCE students and students who did not attend ECCEs. The same held for threatening to use a weapon.

Other School Interventions

1.     Structured Playground Activities. Organized games for kindergarten through 2^nd grade students who arrived at school early have been reported to reduce aggression by 53 % in one school.
2.     Behavioral Monitoring. Intervention staff and teachers met regularly to discuss students’ tardiness, class preparedness, performance and behavior of 7^th grade students. Staff also met in small groups with students to review their school behavior, Students earned points for positive behaviors which they later redeemed for special trips. Staff also routinely informed parents of their children’s progress. The performance of the monitored students was superior to that of a non-monitored group on several measures. The program was continued for two years; and one and one-half years afterwards the monitored students reported less illegal drug use and criminal behavior than non-monitored youths.

Community Interventions

1.     Situational interventions such as placing steering locks on vehicles reduced the rate of car theft. Success also occurred when retail stores used numerous surveillance techniques such as merchandise tagging.
2.     Mentoring also had a desired effect when mentors were properly trained on behavior management and youths received several types of reinforcement for appropriate behavior.
3.     Mandatory sentencing laws for felonies involving firearms seem to prevent homicides involving firearms.

Other Programs

Morash and Rucker report that research shows that correctional work programs which enhance personal skills and develop interpersonal skills reduce recidivism. Also, hard physical labor, often intended primarily as punishment, failed to reduce recidivism. Furthermore, antisocial behavior increases when authority figures provide aggressive models for behavior.

A Los Angeles County gang programusing highly mobile street teams to mediate gang conflicts reduced gang-related homicides. The teams strove for high visibility in selected target areas; and they were supported by schools, PTAs, churches and other community specialists (Needle and Stapleton).

Early identification and stricter sentencing of serious habitual offenders seems to have been effective in redirecting young persons to acceptable ways of behaving. Indicators of serious habitual offending include family disruption, serious school problems, drug and alcohol abuse, and two or more contacts with law enforcement.

Differential Treatment

Attempts to rehabilitate lawbreakers often fail. When programs are successful, the differences between treatment and control groups although statistically significant are substantively minor. This consideration alone should be sufficient to help you understand why T&T has a high recidivism rate, claimed in a UWI study to have been 56% between January and May 2002. A somewhat different operational definition of recidivism has yielded for U.S inmates released in 1994 recidivism rates of 44% within one year, and 67% within three years.

If the UWI and U.S. definitions were comparable, you would have to conclude that Trinidad does a better job of rehabilitating its prison inmates. Indeed, one social scientist was so discouraged after surveying a large body of American rehabilitation research literature that he pessimistically declared in the title of a publication that “Nothing Works.” Yet America can boast of possessing many well-educated criminologists, several with a correctional specialization. I now draw your attention to a treatment approach, differential treatment, that may partially explain the dismal record of failure in trying to rehabilitate lawbreakers.

In the medical field, doctors do not use the same drugs to treat typhoid, malaria, and tuberculosis. Likewise, some correctional professionals argue that lawbreakers are different in many important respects, therefore, a single method of treatment should not be expected to succeed with rehabilitating all. That is, do not expect one size to fit all. My dissertation research and further research on the same body of data show that the argument of these professionals is valid. Higher Interpersonal Maturity inmates housed in a correctional facility had a lower rate of recidivism from placement in a relatively permissive environment than in an authoritarian environment; but lower Interpersonal Maturity inmates benefited more from an authoritarian environment.

Differential treatment with Interpersonal Maturity levels requires valid and reliable classification of offenders and treatment staff, and proper training of the staff so that appropriate treatment environments can be established. In short, it is not an easy program to implement and manage. The approach was first used in California, but may later have been more widely used in Canada.

I have drawn your attention to differential treatment in order to help you understand that there may be an underlying weakness in many of our crime-reduction efforts. We make “one size fits all” assumptions when we should be making distinctions between types of people to be treated, types of treatment staff, and types of treatment environment; and combining these types in ways to obtain optimum results. As with medicine, the wrong dosage can also be given in these programs. I found that it was only in a unit where the evidence showed that treatment was verifiably “intensive” that expected results were obtained.

Execution and Deterrence

During the period February 7^th to 10^th, 2007, the UWI/ANSA McAL Psychological Research Centre polled T&T residents about support for the death penalty and the effect of the death penalty (Sunday Guardian, March 18^th, 2007). Eighty percent of respondents approved of the death penalty and 14% disapproved. Also, 74% believed that the death penalty would reduce serious crimes while 22% did not so believe. Americans also tend to favor the death penalty, 71% expressing this opinion in 1996 if the offender is convicted of murder while 21% were opposed. Yet, the weight of empirical evidence from U.S. data indicates that the death penalty lacks a general deterrent effect.

However, in the early 1990s, I recognized that more recent evidence was more frequently reporting a deterrent effect, that more recent data might be more reliable, and that more sophisticated statistical techniques and novel methodological approaches were available. Therefore, I began research on the question, adding new data up to 1998 as they became available.

Only for the earliest of five periods, and utilizing a method employing change in the value of variables over time did the evidence support the proposition that legal execution deters murder. But using the more traditional methodological approach of relating murder rates (rather than change in murder rates) to previous rates of other variables (rather than changes in these rates) yielded no support for the proposition. Instead, this method favored the contrary proposition that American states engaging in legal executions had a higher murder rate than those with no legal executions. One may be tempted to claim that this body of evidence shows that executions cause an increase in murder; but my analyses cannot rule out an interpretation that reverses the causal order, high murder rates causing states to enact capital punishment laws.

There are two other findings worth noting. First, consistent with a common belief, states that possess more economic resources as measured by poverty rates, have lower murder rates. Secondly, murder rates are higher in states where other violent crime rates are higher than in states that are relatively free of other violent crimes. This second finding supports the culture of violence theory of homicide.

A quick examination of recent Trinidad homicide data also suggests that execution is not a deterrent to homicide. Many of you are probably aware that Dole Chadee and eight members of his gang were executed in June 1999. Also, in July of 1999, Peter Briggs was executed. Yet 1999 is the year in which T&T’s homicide rate took off; and the homicide rate continued to climb rather than decline, as would occur if deterrence were operating. However, some persons may argue that we should not expect to find support for the execution deters homicide proposition in this situation. For punishment to deter, it must be relatively certain to follow the behavior it is supposed to deter. The execution which preceded those of the Chadee gang was Glen Ashby’s in 1994. Therefore, it appears that certainty has by no means been a feature of executions in T&T in the 1990s or the present decade.

Understandably, T&T citizens expect significant crime relief immediately. I cannot promise you such; but the situational interventions, mandatory sentencing laws, correctional work programs, the L.A. County gang program, and the serious habitual offender program  are likely to bring faster relief than those involving  pre-school attendees and other pre-teens. There are other programs intended to reduce criminal opportunities that provide the promise of quicker fixes than those directed at restricting the development of motivated offenders. This is the reason authorities put much effort into increasing the capability of guardians. The newly announced government initiative entitled “Policing for People” is one such program. But its effectiveness will surely be limited if citizens refuse to  provide information to the police or to give evidence in court; if police officers do not perform professionally; if investigations and evidence collection are not aided by modern technology; if courts are inefficient; and if space in correctional institutions is in short supply.

However, the motivated offenders side of the causal diagram (Figure 2) cannot be ignored. The early childhood education program mentioned earlier shows that in Trinidad and Tobago, as early as age nine, girls are less frequently delinquent than boys on five out of six offenses. Using illegal drugs is the single exception. Today, most Trinbagonians will admit that boys’ educational performance is worsening relative to that of girls, and perhaps in absolute terms. This observation suggests that the gap between girls’ and boys’ delinquency is widening and may have been for several years. Yet, girls’ delinquency may also have increased.

Children learn values that control their behavior in interaction with parents, teachers, peers, adults in general, and community institutions. Therefore it is to proper socialization that we must look for a solution to delinquent tendencies. Furthermore, we must remember that in the absence of these tendencies, children may withstand the temptation of criminal opportunities.

Finally, recently, a parliamentarian in T&T created a furor when he was perceived as recommending that abortion and sterilization of members of certain segments of  Trinbagonian society would reduce crime and delinquency. Interestingly, this prescription  can be deduced from Levitt and Dubner’s claim that abortion is the greatest crime-lowering factor in American history, being primarily responsible for the decrease in crime in the U.S. in the 1990s. In their book, Freakonomics, they argue that in the U.S., legalized abortion after Roe vs. Wade resulted in fewer unwanted children among lower income, poorly educated, unmarried teenagers; and “unwantedness leads to crime.

I have argued elsewhere that the unstated corollary in Levitt and Dubner’s argument is that rearing children in a loving environment reduces the likelihood that they will develop criminal tendencies. Also, we can reduce teenage pregnancies through parental, school, and community intervention. As importantly, persons in positions of authority can contribute to the reduction of delinquency by being good role models. It is our failure to instill desirable values in youths that encourages the recommendation of  such desperate measures  as abortion and sterilization to curb criminality.