Follow by Email

"Smart people (like smart lawyers) can come up with very good explanations for mistaken points of view."

- Richard P. Feynman, Physicist

"There is a danger in clarity, the danger of over looking the subtleties of truth."

-Alfred North Whitehead

November 29, 2010

Major Risk Factors for Criminal Behavior

When a judge sentences a defendant, one of the factors that he or she is required to consider and discuss on the record is the protection of the public from further criminal offenses. Judges are also required to consider the character of the defendant (or the character of the accused, as it is sometimes put—but we are beyond the accusation stage at sentencing). Why should judges consider the defendant’s character? Is it proper for a defendant to receive different punishments based solely on a judge’s assessment that the defendant is gluttonous, immodest, greedy, uneducated, lazy, unpatriotic, wasteful, or cowardly rather than generous, educated, modest, ambitious, frugal, patriotic, or brave? Should we care if a defendant is employed in industries that many in society consider immoral such as the exotic dancing industry and therefore that he or she possesses “bad character”?

I believe the proper reason to assess the character of the defendant when imposing criminal punishment is to assess only those character traits that impact the probability that the defendant will or will not re-offend. That character assessment is then used to make a judgment about the need to protect the public. Punishing a defendant for character traits that do not impact the need to protect the public from further criminal offenses is beyond the boundaries of proper criminal punishment. In a pluralistic, democratic society, a defendant is only rightly punished for a violation of the criminal law, and not for a violation of the sentiments of a judge or segments of society.

An assessment of the character of the defendant by a judge involves an examination of historical, static conditions, such as his or her criminal record, educational level, and employment history. Further, interviews by police officers or presentence report writers may also provide insight into the thought processes of defendants and therefore insights into what are considered dynamic criminal factors, or criminal needs.

What does the research say about which character traits are associated with the risk to re-offend? Andrews and Bonta (2010) reported the results of eight different met-analyses that calculated the correlation between various factors and criminal behavior.[i] These researchers found that four factors that were consistently more correlated to criminal behavior than others. They call these factors the “big four”, and they have correlation coefficients averaging .26.[ii]

The first of the “big four” is a history of antisocial behavior. This factor includes the number of times that the defendant was involved in antisocial behavior such as arrests, convictions, probation violations, rule violations within prison; the different types of antisocial behavior; and early onset of that behavior. The severity of the offenses is not part of this factor.[iii]

The second of the “big four” is the defendant having an antisocial personality pattern. Andrews and Bonta (2010) define it as “In everyday language: impulsive, adventurous, pleasure-seeking, generalized trouble (multiple persons, multiple settings), restlessly aggressive, callous disregard for others.”[iv]

The third of the “big four” is antisocial cognition. This factor includes thought processes that are favorable for committing a crime, such as rationalizing crime, such as the victim deserves it, the victim liked being assaulted, society is unfair, societal rules are stupid, the criminal justice system is corrupt. Also included in this factor are attitudes of defiance toward the rules and an identification with criminals.[v]

The fourth of the “big four” is antisocial associates. How closely is the person associated with other individuals who support the defendant’s criminal behavior? What is his social support for crime? Are his friends and family all involved in criminal activity? [vi]

These risk factors have been consistently identified as being associated with criminal activity. Other risk factors have been shown to be associated with criminal activity to a lesser extent than the big four. I will address those factors and some factors that have not been associated with criminal behavior in another blog entry.



[i] Andrews, D.A. and James Bonta, (2010), The Psychology of Criminal Conduct, 5th Edition, Mathew Bender and Company, New Providence, NJ, p. 65.

[ii] Ibid, p. 64

[iii] Ibid. p. 58

[iv] Ibid. p.58

[v] Ibid. p. 59

[vi] Ibid. p. 59

The views expressed in this blog are solely the views of the author(s) and do not represent the views of any other public official or organization.

November 22, 2010

The Effectiveness of Boot Camps in Reducing Recidivism

Correctional boot camps, sometimes called shock or intensive incarceration, that use a military boot camp paradigm, exist in many states throughout the U.S.. These programs often employ physical exercise, drills, ceremony, and uniforms, in an attempt to instill discipline and respect in criminal defendants.

As someone who, as a child, revered Vince Lombardi and was trained to be up to milk cows at 5:00 a.m., I can understand the political popularity of these types of programs. Shouldn’t individuals who committed crimes be taught the value of discipline and respect that we all learned as children? Despite the strong intuitive allure of these programs, are these programs effective in reducing recidivism?

A systematic review of 32 research studies was undertaken to answer this question.[i] These researchers could not find any statistical difference in recidivism between offenders who participated in a boot camp and those who did not participate in a boot camp. The offenders who had not participated in a boot camp did not have a lower recidivism rate than non-participating offenders, nor did they do have any greater recidivism. The researchers concluded that the military component (drills, exercise, uniforms, etc.) had no effect, one way or another, on recidivism.[ii]

The researchers also cautioned that when added to boot camps, effective treatment may make boot camps effective for treating criminal behavior. (Wisconsin’s challenge incarceration program involves alcohol and drug treatment.) The question yet to be answered is whether the boot camp setting for the treatment causes any further reduction in recidivism beyond the effects of the treatment alone.

These researchers further discuss that there may be other benefits of boot camps, such as reduced need for prison beds, improved pro-social attitudes, attachment to community or reduced impulsivity. These factors were not studied, but regardless, any positive changes in these factors resulting from boot camp did not result in less crime.[iii]



[i] Wilson, David B., Doris L. MacKensie, Fawn Ngo Mitchell, (2008), “Effects of Correctional Boot Camps on Reoffending.” Campbell Systematic Reviews, 2003:1, p. 12.

[ii] Ibid. p. 19.

[iii] Ibid. p. 20

The views expressed in this blog are solely the views of the author(s) and do not represent the views of any other public official or organization.

November 15, 2010

Juvenile Sex Offenders

Juvenile sex offenders pose a challenge to decision-makers in the criminal justice system. To ensure the protection of society, decision-makers want to know whether or not the offending behavior is a youthful indiscretion or the start of a life-long series of predatory sex offenses.

Sexual behavior that would not be criminal if the participants were adults becomes criminal when the participants are juveniles. The behavior is a malum prohibitum because of the age of the participants. Other juvenile sexual behavior may be exploitative, and in addition to being a malum prohibitum, is a malum in se sex offense. Sometimes, the line between these two types of offenses is blurry—especially when age differences between the participants exist.

Rehabilitation of a juvenile sex offender is rightly more central to decisions about consequences to the offender for the criminal behavior than it is in the adult system. Juveniles do dumb things because of their youth. But because they are young, and still developing physically, mentally, and emotionally, they may quickly grow out of the criminal behavior. An adjudication for some juvenile sex offenses may have negative life-long implications, such as a lifetime requirement to register as a sex offender—all working against successful rehabilitation.

A recent study done by a University of Wisconsin-Madison researcher took a look at recidivism rates of juvenile sex offenders.[i] This research was comprised of a literature review and meta-analysis of 63 data sets tracking recidivism of juvenile sex offenders. The results showed that for a mean follow-up period of 59.4 months, the mean sexual recidivism rate was 7.08%. (Less than the 13-15% for adult sex offenders at the five-year mark reported in my last blog entry.) The mean average age of the offenders was 14.8 years.[ii]

Caldwell (2010) further found that the recidivism rate for juvenile sex offenders was about 4 times higher for the period of time after their offense when they remained juveniles, than for the period of time after their offense when they became adults.[iii] Although, a diminishing marginal re-offense rate is typical when studying recidivism, this researcher thought the measured difference was inordinately high. He opined that juvenile sex offenders, like most juvenile delinquents, may stop offending in response to developmental maturity. In effect, they grow out of the behavior as they move through adolescence.[iv]

The result of this research is consistent with studies that have repeatedly shown that juvenile delinquents, in general, re-offend more often in their juvenile years, and then age out of the criminal justice system.[v] (Some have argued that this differential may also be the result of the deterrent effect of the adult system.)[vi]

Caldwell (2010) argues that studies of adult sex offenses may not be applicable to juvenile sex offenders. He writes: “Recent studies have also found that the overwhelming majority of adult sex offenses are committed by individuals that were not known to be juvenile sexual offenders and raised questions as to whether juvenile sex offenders pose a risk of adult sexual recidivism that is significantly different from that of other serious delinquent offenders.”[vii]

Caldwell (2010) concludes that his findings support treating juvenile sex offenders differently than adult offenders. He writes: “Cognitive changes related to brain development, hormonal changes related to the onset of puberty, the role of family and peer relationships, judgment, impulse control, bonds to school and other pro-social groups, and the response to social stressors such as child abuse could all play an important role in repeated adolescent sexual misconduct but may have little impact on persistent adult sexual offending.”[viii]

He states that targeting offenders during their teen years with interventions, when they are the highest risk to re-offend, appears to be the most likely successful intervention to prevent another offense.[ix]

This research supports the proposition that juvenile sex offenders should not necessarily be considered as adult sex-offenders in training, and for the criminal justice system to function properly, it must recognize this difference.



[i] Caldwell, Michael F., 2010, “Study Characteristics and Recidivism Rates in Juneile Sex Offender Recidivism, Int. Jour. Of Offender Therapy and Comparative Criminology, Vol. 54 No. 2, pp. 197-212.

[ii] Ibid p. 202.

[iii] Ibid. p. 204.

[iv] Ibid. p. 205.

[v] Ibid. p. 206.

[vi] See for example, Levitt, Steve D., 1998, “Juvenile Crime and Punishment,”, Journal of Political Economy, vol. 106 No. 6 pp 1156-1185.

[vii] Ibid. p. 207.

[viii] Ibid. p. 207

[ix] Ibid. p 207

The views expressed in this blog are solely the views of the author(s) and do not represent the views of any other public official or organization.

November 8, 2010

Sex Offender Recidivism Base Rates

We have all heard the claim that sex offenders cannot be rehabilitated and that they all will invariably reoffend if allowed back into the community. What does the research tell us about that? What are the recidivism base rates for sex offenders? Are there differences in the recidivism base rates for offenders with different types of sex offenses or that possess different traits? As sex offenses are among the most serious offenses in our society, these are important questions for community safety reasons and the treatment of sex offenders.

There are many different studies that attempt to calculate recidivism base rates for sex offenders. A base rate is the proportion of a group of sex offenders that will reoffend within a specified period following any incarceration period (where the assumption is they are not able to re-offend—an assumption which is not always correct).

One of the largest meta-analysis studies of this topic showed the following recidivism base rates for sex offenders.[i]

Follow-up period Recidivism rate (95% CI)

5 years 13-15%

10 years 19-21%

15 years 22-26%

20 years 24-30%

The marginal increase in the recidivism rates decline over time. In other words, the longer a sex offender has gone without reoffending, the lower the chances are that he will reoffend. That conclusion seems to have face validity.

Hanson et al (2003) found that sex offenders with an adult victim have similar recidivism rates for another sex offense as sex offenders with child victims. However, sex offenders with adult victims have higher recidivism rates for other nonsexual violent offenses than child sex offenders.[ii]

Offenders that offend against different victims appear to have different base recidivism rates. Offenders who have offended against unrelated boy victims have a recidivism rate approximately 10 % greater than the average rate at a ten year follow-up period (30% compared to 20%). Offenders who have offended against unrelated girl victims have a recidivism rate approximately 8 % lower than the average rate at a ten year follow-up period (12% compared to 20%). Incest offenders have a recidivism rate approximately 10 % lower than the average rate at a ten year follow-up period (10% compared to 20%). [iii]

These researchers discuss the problem that many sexual offenses are not reported to police and that these recidivism rates include only the reported offenses. These researchers believe that the reported rates for the 20 year follow-up period under-estimate the actual recidivism rate due to unreported offenses, and the rates should be increased by “at least” 10% to 15% with a proportionate increase for the other follow-up periods.[iv]

This research also looked at the correlation between possible risk factors and sexual offense recidivism rates. Sexual interest in children as measured by a phallometric assessment was the risk factor most highly associated with increased recidivism with a correlation coefficient of .32. All of the following factors were correlated with increased recidivism rates; any deviant sexual preference, a history of prior sexual offenses, having any stranger victims, early onset, any related victims, any boy victims, diverse sexual crimes, antisocial personality, any prior criminal history, age, never married, and treatment dropout. The correlation coefficients for these factors ranged from a high of .22 to a low of .10. [v]

(It is important to note that although these correlation coefficients show a relationship between a factor and a recidivism rate, the correlation coefficients are far less than one, the highest correlation coefficient. A correlation coefficient of one is defined as the data showing that the factor and the recidivism rate kept in perfect step, and that the presence of one factor was always associated with an increase in the recidivism rate by an identical percentage.)

These researchers also concluded that several factors were not related to increased recidivism. These factors include, victim empathy, denial of sex offense, unmotivated for treatment, general psychological problems, sexually abused as a child, and degree of sexual contact. [vi]

The above-described research explains the base-rates of sex offenders, what does the research say about identifying those more likely to re-offender. We will take a look at that question in a future blog entry.



[i] Hanson, Kelly R, Kelly E. Morton, and Andrew J.R. Harris, (2003) Sexual Offender Recidivism Risk, What We Know and What We Need to Know, Ann. N.Y. Acad. Sci. 989: 154-166, page 155.

[ii] Ibid, p. 155.

[iii] Ibid. p. 156.

[iv] Ibid. p. 157.

[v] Ibid. p. 157

[vi] Ibid. p. 158.

The views expressed in this blog are solely the views of the author(s) and do not represent the views of any other public official or organization.

November 1, 2010

Lies, Damn Lies, and Numbers

Science and numbers go hand-in-hand. As the foundation of science is the use of empirical research in a search for knowledge, to interpret and apply science, one requires at least a basic understanding of mathematics and its close cousin statistics. And because the use of numbers has been shown to be as equally persuasive in both science and propaganda, one needs to learn to evaluate the use of numbers with a critical eye—especially when using numbers to decide a legal issue. Early in my academic career, a statistics professor told me to never forget that statistics never lie but that only liars use statistics. Of course that was hyperbole, but the admonition has served me well through the years.

A couple of mathematically astute legal commentators have lamented the deficiency in mathematics in the legal profession:

“The fact that many judges suffer from an estrangement from, resistance to, and incapacity in mathematics should not be surprising. This condition, after all, afflicts most lawyers, as it does most Americans. One need not conduct a study to know that law students are typically smart people who do not like math. Law professors are of little help to their students, because legal academics tend not to have a background in, or use statistical analysis, or are unfamiliar with empirical data collection. Indeed, it is clear that there is a prevalent (and disgraceful) math-block that afflicts the legal profession.”[i]

There are a couple of good books I can recommend related to the use of mathematics in the law. The first is Calculated Risks-How to Know When Numbers Deceive You, by Gerd Gigerenzer. This is an excellent book with many examples related to the law. The other one is Proofiness – The Dark Arts of Mathematical Deception by Charles Seife. Seife covers the use of math in science, elections, and propaganda.

There is one illustration of the use of statistics that both books discuss which is called either the prosecutor’s or defense attorney’s fallacy—depending on which side is using it. The example that is referenced in both books comes from the OJ Simpson trial where a defense attorney attempted to preclude any evidence of former domestic violence on the grounds that it was irrelevant.

The argument went something like this. A past domestic violence perpetrator is on trial for now killing his victim. Statistics show that there is only a 1 in 5,000 chance that a defendant with a history of domestic violence with a victim will kill his victim. Therefore, the defense attorney argues, there is only a 1 in 5,000 chance that his or her client has killed his victim, making any evidence of domestic violence irrelevant.

The error in this example comes in confusing the odds that a defendant with a domestic violence history with his victim will murder his victim with the odds that given that his victim is murdered, that a defendant with a domestic violence history with the victim was the murderer. A murdered domestic-violence victim is relatively rare, but once a domestic violence victim has been murdered, the odds are much higher that the former domestic violence perpetrator was the murderer.

The way to look at this is that a 1 in 5,000 chance that a defendant with a history of domestic violence with his victim will kill his victim means that for every 100,000 victims of domestic violence, 20 will be murdered by individuals who had previously perpetrated a domestic violence incident on the victim. If we also know that an additional 5 women per 100,000 will be murdered by someone other than someone who had previously perpetrated a domestic violence incident on the victim, then we know 25 women per 100,000 (20 plus 5) will be murdered. Of the 25 women who are murdered, 20 will have been killed by a former domestic violence perpetrator, or 20 out of 25 women who are murdered (80%) are murdered by a former domestic violence perpetrator —a chance far higher than the 1 in 5,000 chance argued by the defense attorney.

Meyerson et al (2010) (which I also recommend), as well as the authors of the two books referenced above, argue that people, including judges, are quite easily swayed by mathematical arguments. Oliver Wendell Holmes Jr. once wrote that numbers “flatter that longing for certainty and for repose which is in every human mind. But certainty generally is illusion, and repose is not the destiny of man.”[ii]

Meyerson et al (2010) argue that judges’ misunderstanding of mathematics and statistics sometimes result in decisions that unwittingly favor one side of a dispute over another. [iii] These authors conclude:

“It is not necessary for judges to become ‘amateur mathematicians’ in order to reclaim their rightful role. However, they must be aware that the apparent objectivity of mathematics often masks subjective judgments, and not be fooled when ‘hard’ numbers are really based on little more than intuition and guesswork. Numbers can communicate important information. Judges just need to make sure that they are able to comprehend what those numbers are trying to say.”[iv]



[i] Meyerson, Michael I and William Meyerson, 2010, “Significant Statistics: The Unwitting Policy Making of Mathematically Ignorant Judges.” 37 Pepp. L. Rev. 771.

[ii] Ibid, p. 10., quoting Holmes’s The Path of the Law.

[iii] Ibid. p. 14. The article also discusses the problems of a court adopting science’s risk of error, which leads to tipping of the equities in favor of one party or the other..

[iv] Ibid. p. 33.

The views expressed in this blog are solely the views of the author(s) and do not represent the views of any other public official or organization.