March 25, 2013

Miracle of the day: 80-year-old man recaptures long-lost youth

(Or: How committing a new sex crime can paradoxically LOWER risk on the Static-99R)

"How old is the offender?"

 Age is an essential variable in many forensic contexts. Older people are at lower risk for criminal recidivism. Antisocial behaviors, and even psychopathic character traits, diminish as criminals reach their 30s and 40s. Men who have committed sex offenses become at considerably lower risk for further such misconduct, due to a combination of decreased testosterone levels and the changes in thinking, health, and lifestyle that happen naturally with age.

Calculating a person's age would seem very straightforward, and certainly not something requiring a PhD: Just look up his date of birth, subtract that from today's date, and -- voila! Numerous published tests provide fill-in-the-blank boxes to make this calculation easy enough for a fourth-grader.

One forensic instrument, however, bucks this common-sense practice. The developers of the Static-99R, the most widely used tool for estimating the risk of future sexual recidivism, have given contradictory instructions on how to score its very first item: Offender age.

In a new paper, forensic evaluator Dean Cauley and PsyD graduate student Michelle Brownfield report that divergent field practices in the scoring of this item are producing vastly different risk estimates in legal cases -- estimates that in some cases defy all logic and common sense.

Take Fred. Fred is 80 years old, and facing possible civil commitment for the rapes of two women when he was 18 years old. He served 12 years in prison for those rapes. Released from prison at age 30, he committed several strings of bank robberies that landed him back in prison on six separate occasions.

At age 80 (and especially with his only known sex offenses committed at age 18), his risk for committing a new sex offense if released from custody is extremely low -- something on the order of 3 percent. But evaluators now have the option of using any of three separate approaches with Fred, with each approach producing quite distinct opinions and recommendations.

Procedure 1: Age is age (the old-fashioned method)

The first, and simplest, approach, is to list Fred's actual chronological age on Item 1 of the Static-99R. Using this approach, Fred gets a three-point reduction in risk for a total of one point, making his actuarial risk of committing a new sex offense around 3.8 percent.

Evaluators adopting this approach argue that advancing age mitigates risk, independent of any technicalities about when an offender was released from various periods of incarceration. These evaluators point to the Static-99R's coding manuals and workbook, along with recent publications, online seminars, and sworn testimony by members of the Static-99 Advisory Committee. Additionally, they point to a wealth of age-related literature from the fields of criminology and psychology to support their scoring.

Procedure 2: Reject the Static-99R as inappropriate

A second approach is not to use the Static-99R at all, because Fred's release from prison for his "index offenses" (the rapes) was far more than two years ago, making Fred unlike the members of the samples from which the Static-99R's risk levels were calculated. Evaluators adopting this approach point to publications by members of the Static-99 Advisory Committee, generally accepted testing standards and actuarial science test standards to support their choice to not use the test at all.

Procedure 3: The amazing elixir of youth

But there is a third approach. One that magically transports Fred back to his youth, back to the days when a career in bank robbing seemed so promising. (Bank robbery is no longer alluring; it is quietly fading away like the career of a blacksmith.) The last five decades of Fred's life fade away, and he becomes 30 again -- his age when he was last released from custody on a sex offense conviction.

Now Fred not only loses his three-point age reduction, but he gains a point for being between the ages of 18 and 34.9. A four point difference! The argument for this approach is that it most closely conforms to the scoring methods used on the underlying samples of sex offenders, who were scored based on their date of release from their index sexual offense. These evaluators can correctly point to information imparted at training seminars, advice given by some members of the Static-99R Advisory Committee, and sworn testimony by developers of the test itself. They can also point to an undated FAQ #27 on the Static-99 website to support their opinion.

Fred could rape someone to reduce his risk!

Back-dating age to the time of the last release from a sex offense-related incarceration allows for a very bizarre twist:

Let's say that after Fred was released from prison on his most recent robbery stint, back when he was a vigorous young man of 61, he committed another rape. Being 60 or over, Fred would now get the four-point reduction in risk to which his age entitles him. This would cut his risk by two-thirds -- from 11.4 percent (at a score of 5) all the way down to a mere 3.8 percent (at a score of 1)!

While such a scenario might seem far-fetched, it is not at all unusual for an offender to be released from prison at, say, age 58 or 59, but to not undergo a civil commitment trial for a couple of years, until age 60 or 61. Such an offender's score will vary by two points (out of a total of 12 maximum points) depending upon how the age item is scored. And, as Cauley and Brownfield describe, the members of the Static-99R development team have, at different times, given contradictory advice on how to score the age item.

By completely negating the very substantial body of research on age and crime, this technocratic method creates other very concerning -- and paradoxical -- implications, Cauley and Brownfield argue: As the risk estimate for a more persistent offender is lowered, the offender who does not reoffend is stuck with a risk score that is forever jacked up.

Back-dating an offender's age is also at odds with the research that generated the test itself, they say, because the offenders in the samples used to construct the Static-99R had finished serving their sentences on their index sexual offenses within two years of being studied. In other words, none of the offenders had been released many years earlier, and there was none of this curious time-travel business in regard to their ages. As the instrument's developers noted in a publication just last year, the Static-99 "was developed on, and intended for, sexual offenders with a current or recent sexual offense."

So, if you are evaluating an old geezer in the local pen and he tells you that he is only 30 years old, don't assume that he has a delusional belief that he has discovered the elixir of youth -- or that he's pulling your leg. He just might be reciting the age that he was just assigned by a technocratic Static-99R evaluator.

The paper, "Static-99R: Item #1 -- What is the Offender's Age? A lack of consensus leads to a defective actuarial," is available for download both HERE and HERE.

No comments: