Predicting Dangerousness - The Flaws

By Charles Hanson

With the so many and various type of risk assessments being used by probation officers and forensic psychologists and none more so than with offenders in the prison system one could be forgiven for thinking that here at last we have a scientifically reliable method of predicting future offending.

The Offender Group Reconviction Scale (OGRS) is a predictor of re-offending based only on static risks - age, gender and criminal history. It allows probation, prison and youth justice staff to produce predictions for individual offenders even when the use of dynamic risk assessment tools (e.g. The Offender Assessment System (OASys) or Asset) is not possible. It will form the basis of an improved static/dynamic predictor in OASys, and also assist researchers in controlling for expected levels of re-offending when comparing samples.

OASyS devised by academics working within the Home Office, is based on mathematical formula which measures the variables of previous and present offence details, social class, age, gender, age first convicted, social problems for example drugs or alcohol, employment history and time in last job, number of associates, previous breaches of parole, probation, supervision or bail, times between each prison sentence, whether single married or divorced and even whether living alone or with parents etc. the variables appear to be unlimited.

It is calculated that against each factor a score can be applied that when measured overall can determine risk, but how accurate is the method for there does exist the critics who argue that the whole area of risk prediction is flawed and unreliable?

The critics argue that a dangerous person is not a psychological entity nor is dangerousness a scientific or medical concept, neither is it necessarily associated with mental illness.

The notion of a 'dangerous person' as one with a propensity to inflict harm is empty of meaning until it is given social content.

There are considerable difficulties in defining dangerousness satisfactorily for legal purposes. The greater problem is in selecting dangerous offenders although psychologists and indeed the probation service continually strive to do just that through the often use of dubious methods.

The literature on predicting dangerousness is amassing all the time and yet no one has come up with a valid and accurate method of assessment, models of past behaviour seem certain to continue to be the criteria in assessing future risk irrespective of the outcome of any behaviour or psychological programmes.

Professor Norval Morris states:- "Since we cannot make reliable predictions of dangerous behaviour, considerations of justice forbid us to confine people against their wishes in the name of public safety for longer periods than we can justify on other grounds."

What is a dangerous person? "No such entity exists in the nosology of psychiatry" remarks H.L. KOzol.

Cocozza and Steadman in a 1976 study reported that those evaluated by psychiatrists as being dangerous were no more so than those evaluated as safe.

Two first hand studies were undertaken at institutions for the diagnosis and treatment of dangerous offenders in the U.S.A. (Maryland) 1973.

A number of inmates were released against the advice of clinical staff and followed up for a period of five years in one case and three years in another.

The assessments were thorough involving psychiatrists, psychologists, social, workers, law enforcement agencies etc. Half to two thirds of the judgments of dangerousness that were put to the test were NOT borne out by subsequent harmful behaviour on the part of the offender concerned.

Monahan (1973 AND 1977) in referring to this states:-

"Even under favourable conditions, the risk of unnecessary detention is likely to be considerable, and on the evidence of Kozol it is likely to be at least 50% and may be as much as 66% even when the offenders concerned have had records of serious crime accompanied by violence and the assessments have been carefully made."

If most people do not commit serious crime it is far easier to be wrong than right if you predict that someone will do so however carefully you make the assessment of his character and circumstances.

The prevailing tendency for the courts and executive authorities (the Ministry of Justice) to defer to psychiatrists and indeed psychologists should be checked, given that each side will use 'expert testimony' to promote their case without the benefit of being objective. The adversarial system of justice in the U.K. virtually compels the expert to be partisan to the side that calls him or her.

Protection of the public is a function of the Criminal Justice System but the technique at its disposal is punishment justly related to Past Conduct.

Predictive judgments of future conduct are out of place in systems of natural justice.

They are highly inaccurate but even if they could be made as accurate as judgments of past conduct are required to be they would not be acceptable, for preventative confinement preempts a man's future course of actions.

Once preventative measures are permitted against dangerous offenders the way is open in principle to the extension of measures to non-offenders.

A difficulty in predicting dangerousness is that a man must forfeit his right to be presumed innocent before his right to be proved harmless can be brought into question.

The right to punish for past wrong-doing is a pre-condition of the right to prevent future wrong-doing.

Though all penal and other assessments rely on the distinction between serious and other harm, the concept of seriousness is necessarily ambiguous in this connection for it has a moral as well as a factual dimension referring as it does to the wrongfulness of acts as well as to the injuriousness of their consequences.

In 'Sentencing in a Rational Society' (1972) Professor N.D. Walker recommended cutting out the ambiguity by abandoning the concept.

There is agreement in many quarters that the cost in resources and human suffering of imprisonment for any purpose is high, there is more reluctance to use it for the purpose of punishment for the harms people have actually done, we must be even more reluctant to use it to prevent them causing harm in the future.

We have to distinguish between the persistent or nuisance offender and those deemed 'dangerous'. Persistence refers to the past, dangerousness to the future. Persistence in offending can undoubtedly provide some evidence of 'dangerousness', morally however to deny human change is to deny human existence.

None of the legislative attempts have provided any substantive criteria for establishing who is a 'dangerous' offender, certainly there is no such offence as being 'an enemy of society'

The concept of 'dangerousness' in English criminal justice is elusive. It is not used with any precision, and the nature of the risk to which it refers is never defined in terms to make it contestable.

Jean Floud (1981) proposed that any court or body who receives reports from psychiatrists, psychologists, probation officers or social workers should provide for the defence or applicant to call their own 'expert' witnesses to make an assessment contestable and for reasons to be given for the imposition of a sentence of imprisonment or the grounds of refusal to release.

The idea of 'dangerousness' is often taken for granted. It is so often defined so as to be unhelpfully imprecise, circular, misguided or irrelevant for practical penological purposes. Moreover it raises anxiety and is therefore particularly open to abuse.

Cocozz And Steadman (1976) claimed that psychiatrists and psychologists under pressure assume to be experts in diagnosing 'dangerousness' to meet the expectations of society. Psychiatrists they claim pose as scientists but practice magic in the sense that through claiming 'special knowledge and being granted 'expert' status in law they make assessments of dangerousness which rely on empirically untested beliefs and represent an effort at control of the 'potentially harmful'.

In any case they have been allowed to exceed their powers, they represent an excellent example of professionals who have exceeded their so called 'expertise' and for whom society's confidence in their ability is empirically unjustified.

There is a good deal of evidence to suggest that as matters stand, psychiatrists, psychologists, probation officers, the Parole Board and other bodies are on average at best as likely to be wrong as right in thinking that the offenders they report on and decide to detain as 'dangerous' would actually do further harm if left at large or released.

Charles Hanson <charles.hanson@live.co.uk>