A COMPUTERIZED DECISION-SUPPORT SYSTEM

A COMPUTERIZED DECISION-SUPPORT SYSTEM

IN THE JUVENILE PROBATION SERVICE

Levi Eden
Director,
Juvenile Probation Service,
Ministry of Labor and Social Affairs

The Juvenile Probation Service (JPS) is a social-therapeutic service operating nationwide from within the Division of Youth Development and Correctional Services in the Ministry of Labor and Social Affairs. The Service's function is to treat those juveniles referred to it by the Israel National Police (INP) as suspected of a criminal offense and to rehabilitate them. The service also treats juveniles who are the subject of a court treatment order, e.g. a probation order. "Juveniles" are aged from 12 to 18 but occasionally one may remain in the care of a probation officer up to age 21. All probation officers are professionals with a B.A. or M.A. degree in social work, who have taken a special orientation course in probation work.

The service's work is governed by laws and regulations that lay down, inter alia, three guiding principles for the probation officer:

1. Any juvenile who is the subject of a criminal investigation by the INP must be referred to a probation officer as soon as there is sufficient evidence to put him on trial;

2. Once the juvenile has stood trial and been found guilty, the court must order a probation officer's report;

3. The probation officer's report must include: the juvenile's biography; his familial, social and economic circumstances; his physical-medical state; and a recommendation as to the most suitable form of care.

With the referral of a juvenile to the JPS, a probation officer will invite him for an intake interview during which the officer will obtain from him and his parents preliminary information on his and his family's situation and probe their attitude to the offence for which the juvenile has been referred.

Subsequently, the probation officer makes it his object to get to know the juvenile more deeply, his family and immediate environment too, including school or place of work. He may refer him for medical, psychological or other examination so as to reach a more exact assessment of his personality. This stage the psycho-social investigation has two aspects: it is a therapeutic process in itself and also serves as the basis for the report and recommendation as to rehabilitative treatment which the officer will present to the court, if and when the juvenile is brought to trial for the court to determine his responsibility for the offense. The probation officer also has the power to recommend to the police to close the criminal case against the juvenile without trial, should he be convinced that the juvenile would be best treated without one. The offense may have been a one-time act with no fear of a return to delinquency, or perhaps enough has already been done, during the psycho-social investigation, to rehabilitate the youth.

The Decision-Support System

The computerized decision-support system has been developed by Dr. Monica Shapira of the Hebrew University School of Social Work. It is Dr. Shapira's contention that, since more than 80% of the recommendations made to the court by a probation officer are adopted by the court, it is essential that the professional judgement that the officer uses to decide on the content and manner of his recommendation, be as consistent, as reliable and as correct as possible.

Using statistical regression methods, Dr. Shapira found that probation officers may collect many items of information but rest their final recommendation on eight variables only, namely:

1. The number of offenses committed by the youth
2. Instances of anti-social behavior
3. His level of functioning in school or place of work
4. His relations with his parents
5. Pathological phenomena in the family
6. The youth's age
7. The gravity of the offense
8. Whether the offense was committed alone or with others.

Shapira accorded each variable a weight and used statistical modeling to determine the influence of each on the probation officer's final decision. Her final linear model permits the officer's professional decision policy to be characterized in terms of weighted variables, by affixing pre-set weights to the eight individual variables.

Shapira further states that the probation officer, in making his judgment

(i.e. his clinical decision his prediction and estimate of future human behavior), the officer will employ predictor variables (i.e. the information about the juvenile) that stand in a conditionally monotonic relation to the criteria (i.e. the response of the officer to the predictor variables).

The second important feature of the probation officer's decision-making process, the feature that makes it structurable by regression equations, is the premise that the manner in which he arrives at his judgment and recommendation as to the juvenile's method of care (basing himself on character data and the circumstances of the juvenile's life) can be behaviorally defined as a cyclical process of problem solving and information processing.

Decisions to be Made by the Probation Officer

In deciding on his recommendation as to the further handling of the juvenile's criminal file, the probation officer faces two choices to recommend to the police to file criminal charges in the Juvenile Court or to close the file without a trial. If the juvenile is brought to trial a range of recommendations are available to the officer, from the severest imprisonment, actual or conditional, if the youth is convicted through further treatment orders without a criminal conviction, and the lightest discharge of the juvenile without any court order.

If the recommendation is to send the youth for treatment (without criminal conviction) then the probation officer has eight options for further recommendation:

1. To transfer the youth to the care and supervision of a qualified person who is not one of the parents;

2. To have the youth put on probation;

3. To have the youth or his parents stand guarantee for his future good behavior;

4. To have the youth attend a day center;

5. To commit the youth to an open or closed residential institution;

6. To issue some other directive as to the youth's future behavior, e.g. a period of public service or a prohibition on driving a vehicle;

7. To oblige the youth or his parents to pay a fine or the court costs;

8. To oblige the youth or his parents to pay compensation to the person injured by the youth's offence.

With respect to all these possible recommendations, the probation officer arrives at his decision by a process of probing and combining the items of information collected during the psycho-social investigation, and his recommendation is usual fateful for the youth's future.

Although Shapira found that probation officers base their decisions on eight items of information alone, this did not prevent her also finding important differences between the decisions of two different officers about youths with similar records. These differences stem from the personalities, education and general approach of the officers who come from varying backgrounds and cultures. The common training and supervision system does not succeed in getting them to employ a uniform process of judgment formation.

It is this problem that the computerized system is intended to help the probation officer overcome by standardizing the data, and by creating pre-determined data combinations and a universal decision-making model. This model, founded on the accumulated practice of past officers, is intended to increase the uniformity and equality of officers' recommendations with respect to juveniles with similar psycho-social backgrounds and delinquency record.

How the System Works

Once an officer has completed his decision-making process with respect to a particular youth and arrived at a firm recommendation, he opens a "dialogue" with the computerized program (running on specially developed software). He taps in the youth's identity number and the computer proceeds to ask him a series of 73 questions, to which the officer replies by keying the appropriate code.

The questions are divided into four groups:

1. Data on the youth's family: e.g. parents' employment status and state of health, relations between the youth and his parents, the family's delinquency record, signs of pathology;

2. Data on the youth himself: e.g. age, his place in the sibling order, where he lives, physical development, state of health, occupations, school record, behavioral problems at home and school and the strength of his motivation to rehabilitate himself;

3. Data on past and present delinquency by the youth: e.g. category and gravity of his offense(s), the injury or damage caused, the youth's attitude to the current offense and the intention or not to commit it, previous offenses and previous treatment in the JPS.

4. The therapeutic relationship between the probation officer and the juvenile and the former's assessment of the chances of rehabilitation: e.g. the relations between the youth and the officer, results of previous treatments (if any), the youth's abilities and his response to the treatment.

The dialogue between the probation officer and the computer program lasts ten to fifteen minutes. Only after he has given answers to all the questions and keyed in his recommendation, does the program display its own recommendation on the screen. If the two recommendations are the same then the probation officer may forward his report to the court or his professional opinion to the police. Should the program's recommendation be severer or milder than the officer's, then the latter has two options to accept or reject the computer's recommendation but, in either case, he must record his reasons for so doing.

The computerized system stores the probation officers' new recommendations as well as its own in a databank and this accumulated material is used to monitor the efficiency of the system. The arguments of the officers for rejecting the program's recommendations are used to update, expand and improve it. In this way accumulating experience reveals subjects and issues, both known and new, on which the Service has yet to lay down policy or where policy is either unclear or inconsistent. Thus, the system in the Jerusalem District Office has been updated three times since its installation and that in the Tel Aviv District Office once, and probation officers have added to their list of possible recommendations. The most important of these additions is a recommendation to the court to defer continuation of the legal process pending the receipt of new information on the youth or his transfer to a Diagnostic Center for observation.

The Influence of the System on the Officers and on the Organization

One of the changes anticipated as a consequence of the installation of the system in the JPS is that the officers will feel that have undergone a professional and emotional readjustment.

34 officers with at least one year's seniority, 12 in Jerusalem and 22 in Tel Aviv, have filled out satisfaction and attitudinal questionnaires with regard to the system. They were requested to respond to 51 clustered statements, each cluster representing a facet of their relation to the system:

* Satisfaction with their use of it;
* Its contribution to equality, fairness, etc.;
* Its effect on the officer's organizational environment;
* Its effect on the officer's professional performance.

Despite the difficulties of analyzing the meaning of the responses due to the high number of statements (51) and small number of respondents (34), the organizers of the survey report a number of findings:

* The majority of respondents agree that the system does not infringe on their professional judgment, but allows them full discretion in weighing their recommendation against that of the program;
* The system helps the officer keep track of all his decisions and to test their validity;
* The officers dismiss any fear that use of the system might dehumanize their work or over-emphasize its technical aspects;
* Disagreement between the officer's and system's conclusion leads the officer to re-examine and re-think his decision;
* Satisfaction with the decision to participate in the innovative project was unanimous;
* The respondents' recommendation that the system be installed in other welfare services is further evidence of its positive influence on their work;
* The officers do not feel threatened by the system and their relations with their supervisors have not been changed by it.

It is the expectation that the installation of the system in all the Service's units will permit management to set a policy of supervision and control that, on the one hand, will meet the personal needs of each probation officer and, on the other hand, will conduce to more equal and consistent recommendations uninfluenced by individual predilections.

Probation officers are supposed to reach their professional decisions independently, relying on their understanding, moral sense, experience and knowledge. Within this context, the supervision and guidance exercised by the Service should be such as to cope with three sorts of professional behavior: invariable rejection of the computer's recommendation, invariable acceptance of it, and patternless decision making in response to irrelevant criteria. The computerized system is an extra tool to enable the probation officer/social worker achieve and maintain quality of judgment and it is management's responsibility to see that it functions as neither more nor less than this

Summary and Conclusions

The computerized decision-support system supports the probation officer's decision-making process by providing him with computer-generated rules for assessing and combining the separate items of information that he collects. These rules are formulated linearly, on the premise that the relation between the independent and dependent variables can be represented by a straight line uninfluenced by the scores of other dependent variables. In other words, officers can characterize the juveniles in their charge by a weighted combination of variables.

The probation officer and the computer "collaborate" in reaching the decision in each individual case by combining the pieces of information on the youth and his surroundings with the subjective assessments of the officer. This process of combination is executed by the computer in conformity with the overall policy of the Service or of one of its regions which, in turn, is based on the accumulated practice of all past and present officers.

The system permits the officer to diverge from the computerized model's recommendation whenever he feels justified in so doing, because that recommendation is no more than an extra item in the bank of information at his disposal. Any one recommendation by the computer reflects the average of all past judgments in like cases: it is not binding on the probation officer in the present. The distribution of the arguments against the computers' conclusions serves as important feedback for modifying and upgrading the software and permits the adaptation of the "average judgment" to new norms generated by current practice, at both regional and national level. Although the system has brought about changes in officers' perceptions by exposing them to a fusion of technology with therapeutic methods, experience to date does not yet permit the management of the Service to commit itself to long-range conclusions. In our estimation, however, a decision-support system should find its place in other social services that make fateful decisions about persons in their care so as to raise the level of their professional performance and advance them to new standards of quality of care.

* This article is based on one appearing in: Vozner, Y., Golan, M. and Hovav, M. (Eds.) 1994. Delinquency and Social Work: Knowledge and Intervention. Ramot-University of Tel Aviv, Israel. (in Hebrew).