Measuring Risk of Recidivism Among Detainees - Or Not

Although the Obama Administration has recently pulled back from its earlier emphasis on a criminal justice model of fighting ‘extremism', it is still emphasizing the need to use tools from both criminal justice and the military/national security apparatus.  It is strange therefore to see the Administration seeming to overlook a basic tool of criminal justice.

At the end of May, the Administration released the Guantanamo Review Task Force's January 22, 2010 Final Report.  The Task Force recommended prosecuting some detainees, transferring most to third countries, and detaining the rest without trial.  Some of those detained without trial were approved for transfer if and when a suitable transfer destination could be found. 

The report justifies continued detention without trial on the basis of the Congressional Authorization for Use of Military Force passed after September 11, 2001.  This is presumably on the basis that the detainees are enemy combatants although that is not stated.  Presumably the same basis would justify holding all enemy combatants.  The report acknowledges that for a "small handful" of detainees the Task Force found little evidence of their involvement, but declares that for the rest it had such evidence.  Nevertheless, the report recommends transferring 126 of the remaining 248 detainees.

The recommendations were made with the aid of "Guidelines" developed for the purpose.  In deciding whether or not to transfer a detainee, the Task Force evaluated the threat each posed under a "non-exclusive list of factors" identified by the Guidelines.  The list is not set forth as such, but in the section discussing those detained without trial the following factors are identified:

  • - Significant organizational role within al-Qaida, the Taliban, or associated forces.
  • - Advanced training or experience.
  • - Expressed recidivist intent.
  • - History of associations with extremist activity.

The report emphasized that the decision to transfer "does not reflect a decision that the detainee poses no threat or no risk of recidivism."  Instead, it "reflects the best predictive judgment of senior government officials" that the threat posed by transferred detainees can be handled by "appropriate security measures in the receiving country."

Given the Administration's often-touted application of criminal justice standards to fighting the war on what it calls ‘extremism', it may be instructive to see how the Guantanamo Guidelines' "non-exclusive list of factors" compares to factors used to predict the risk of recidivism among ‘ordinary' criminals. 

Currently, the federal criminal justice system applies a calculation called the Criminal History Category to sentencing decisions.  But as acknowledged by the U.S. Sentencing Commission, which developed the Criminal History Category for its Sentencing Guidelines, "the goal of the CHC [Criminal History Category] encompasses more than recidivism prediction.  Its second core philosophy involves... just punishment and deterrence."  p.3

By contrast, the indicator used before the Sentencing Guidelines were adopted, the U.S. Parole Commission's Salient Factor Score ("SFS") [ p.73], is designed solely "to assess the likelihood of recidivism."  p.3  It is also, according to the Sentencing Commission, "a better predictor of recidivism than is the CHC [Criminal History Category]." p.12

In evaluating risk of recidivism among offenders eligible for parole, the SFS considers these factors:

  • - Number of prior criminal commitments.
  • - Number of prior criminal commitments lasting at least 30 days.
  • - Whether the current offense was committed while the offender was under a sentence/supervision.
  • - Whether at least 3 years have elapsed between the current offense and the offender's last commitment.
  • - Offender's age at time of current offense (under-41s having the strongest likelihood for recidivism).

In effect, the Salient Factor Score measures risk of recidivism based on factors related to an offender's criminal history.

Of the factors identified in the Final Report, the only one that might consider factors similar to those in the SFS is the fourth one, history of associations with extremist activity, but not enough information is given to determine whether that factor attempts to include similar criteria in determining risk based on history.  Certainly it includes several factors which are not part of criminal history, like family ties to extremist organizations.

Strangely, in discussing the Transfer Decision, the Final Report says little about history although it does note that the detainees approved for transfer were "low-level" fighters (i.e., they did not have a significant organizational role within al-Qaida) and were "uneducated and unskilled" (i.e., they did not have advanced training or experience).  Yet the ordinary criteria by which criminal offenders' risk of recidivism is measured every day is not mentioned.  This lack is notable given that the Final Report emphasized it relied on "standards used by federal prosecutors across the country to determine whether to charge a case, as set forth in the United States Attorneys' Manual." pp.7-8

Admittedly the Salient Factor Score would need to be adapted to apply to detainees.  In its current form, it evaluates recidivist potential mostly on the basis of criminal convictions -- how many, how recent, whether the offender was still serving a prior sentence at the time of the current offense of conviction, how old the offender was at the time of the offenses.  Detainees considered for transfer are not serving sentences for current convictions.  Many of them have spent their ‘careers' in countries where membership in al Qaida and related jihadist activities are not illegal; they may have a history of involvement with violent jihadi activities but have no criminal convictions.  So, the Task Force could have adapted the SFS for its review by replacing formal criminal commitments with known jihadi activities (what the Final Report calls "engaging in extremist activities") as the standard measure. 

Although information about detainees' histories is imperfect, the government does know a fair amount.  The Final Report acknowledges that "for most of the detainees approved for transfer, there were varying degrees of evidence indicating that they were low-level foreign fighters affiliated with al-Qaida or other groups operating in Afghanistan...  For many of the detainees approved for transfer, however, the review participants found there to be reliable evidence that the detainee had engaged in conduct providing a lawful basis for his detention"  pp.16-17  If there were any attempt to use this information in a coherent fashion when gauging detainees' future dangerousness, it is not apparent from the Final Report. 

It is possible that the Task Force relied on a separate tool measuring risk of recidivism among ‘extremists'.  Recidivism prediction is not a one-size-fits-all endeavor.  For example, studies have shown that recidivism among sex offenders "is best predicted by a different set of factors than those that predict general or nonsexual violent recidivism."  The report does note, "review participants also were kept informed of intelligence assessments concerning recidivism trends among former detainees." p.17  But it does not state that the "trends" were measured in any scientific or statistical fashion, only that the Task Force was "kept informed" of "trends". 

Which begs the question -- why?  Suppose the Task Force determined that as with recidivism among sex offenders, recidivism among extremists was "best predicted by a different set of factors" than those used for ordinary criminals.  What are those different factors?  In order to devise and apply a proper tool to measure recidivism, the Task Force would have to identify the characteristics that make extremists more likely to recidivate.  The more precisely those factors can be identified, the more useful a recidivism indicator is likely to be.  Conversely, the less precisely they are identified, the more difficult it would be to devise a measure that would give useful information.  In light of the current Administration's refusal even to identify Islamism as the ideological foundation motivating detainees, see here, here, and here [ at 6:05], it's hard to imagine that the Task Force could have articulated a precise list of recidivism factors.   

And what would be the effect of a more accurate measure?  The end result of the failure to use some sort of criminal history is likely to understate detainees' likelihood of recidivism, making them more attractive candidates for transfer to another country that would monitor them and, perhaps, try to rehabilitate them.  Between the Administration's difficulty identifying adherence to a violent interpretation of Islam as a crucial motivating factor, and general First Amendment [ pp.13-14] concerns about government distinguishing between legitimate and illegitimate religious expression, the Administration may realize it could not possibly hope to rehabilitate detainees.  Certainly it could not offer a program of "religious reeducation taught by Islamic clerics." p.10.  Limited evidence available to date suggests that reeducation without clerical participation has failed.  The Administration may be betting that other countries here [] and here less squeamish about drawing religiously-based distinctions can do a more effective job of de-programming radicals.  If it needs to tilt the scales a bit in order to justify transferring detainees considered lower-level to the custody of other countries, it will do so.

The author is an attorney specializing in criminal law. 
If you experience technical problems, please write to