location: publications / books / Justice Behind the Walls / Sector 4 / Chapter 3 The Arbour Report: The Indictment of a System / Postscript: "The Most Comprehensive Empirical Review . . ."

The structured interview portion of the testing posed five questions to be answered on a 10-point scale. The questions were, how are you feeling today? How important do you think it is to have friends? Do you have any problem with sleeping? Do you have any problem with appetite? Do you have any problem with concentration? The prisoners were also asked questions on suicide ideation. Prisoners who completed all three sessions were asked if they ever thought of committing suicide and at each session prisoners were asked if they had thought of committing suicide within the last week. The initial testing of the session lasted about two hours and the follow-up assessments at 30 and 60 days lasted approximately one hour. The shorter time for the latter was due to the fact that a number of the measures that were not thought to change over time were administered only once.

In total 83 segregated and 53 non-segregated offenders participated in the study. However, complete data for all three sessions (covering 60 days) was only available for 23 segregated and 37 non-segregated offenders. The loss of participants from the segregated group was primarily due to releases to the general prison population or transitional units or transfers to other institutions. For non-segregated prisoners, the loss of participants was mainly due to transfers to other institutions and placement in segregation. Of the 23 segregated prisoners who completed the study the full 60 days 10 were involuntary and 13 voluntary segregation cases.

Based upon his research Dr. Zinger found that "The hypothesis that the mental health and psychological functioning of segregated inmates would deteriorate over a period of 60 days in segregation was not supported . . . This study is somewhat encouraging because it provides evidence that segregation for 60 days as currently administered in Canadian penitentiaries does not negatively affect prisonersí mental health and psychological functioning." (at 75-6) Zinger suggested but did not choose between three interpretations for finding no deterioration:

(a) segregated prisoners generally adapted and coped well with the conditions of segregation;

(b) the segregated offenders did not perceive the conditions of their confinement as threatening or stressful and therefore were not affected by them; or

(c) the environment of the general population of segregated offenders came from was such that it was viewed more negatively than that of segregation. (at 75)

In reviewing the individual results on the eight measures administered Zinger found that "overall both segregated and non-segregated prisoners reported better mental health and psychological functioning over time." (at 74) This finding is explained by Zinger with reference to the "practice effect" phenomenon, where participants lose interest in answering repeatedly to identical questions and tend to report less problems over time.

The editors of the Canadian Journal of Criminology in publishing the article based on Dr. Zingerís research also arranged for the publication of three commentaries, including one by myself. The first commentary was written by Dr. Julian Roberts and Dr. Robert Gebotys. In response to Dr. Zingerís claim that his study is "the most comprehensive empirical review of the psychological effects of administrative segregation" they suggest that "there are in fact more than a few snakes on the loose in this methodological Garden of Eden." (Julian Roberts and Robert Gebotys, Prisoners of Isolation: Research on the Effects of Administrative Segregation, (2001) 43 Canadian Journal of Criminology 85 at 86) A number of Robertsí and Gebotysí criticisms relate to the sample size -- only 23 prisoners remained in the sample at the end of 60 days -- and the inclusion within the sample of prisoners who had requested protective custody and segregation -- who constituted 13 of those 23 prisoners. Another criticism is the lack of statistical power of the tests utilized by Zinger. A third relates to the high-rate of attrition within the segregated sample; only 10 of 55 involuntary prisoners remained at the end of the study compared to 13 from the original sample of 32 voluntary cases -- a percentage loss of 80 per cent compared to 60 per cent. Roberts and Gebotys suggest that this raises

The ominous (to the integrity of the experimental design) possibility that those who could not cope with isolation found ways to secure their transfer: the prisoners who remained were the individuals who could cope with the experience of segregation . . . Unfortunately for the researchers, the higher drop-out rate is in the direction that creates an alternate hypothesis to explain the findings: the prisoners who had problems with the experience of isolation were less likely to be in the study at the end. (at 92)

A further problem raised by Roberts and Gebotys, and one not dependent upon statistical analysis, was that almost all prisoners, segregated or not, had previous experience with segregation. The sample of prisoners who were segregated reported on average, having been isolated on 12 previous occasions and Zingerís analysis showed they had been segregated twice as often as the non-segregated prisoners. Roberts and Gebotys suggest that "Testing the effects of segregation on prisoners with a great deal of experience is akin to testing the effects on rates of seasickness, of a new design of ship, using sailors with years of experience at sea." (at 91)

In addition to the problems raised by Roberts and Gebotys Kate OíBrien, a law student at UBC Law School, in a paper prepared for my Penal Policy seminar, has also raised some questions relating to the validity of Zingerís findings of no negative effects. From reading the full thesis (as compared to the revised summary published in the Canadian Journal of Criminology) she discovered that the average elapsed time between a prisonerís placement in segregation and the first testing session was 3.6 days. She suggests there are two problems with this that are not mentioned by Zinger. The first is that the literature shows that crisis typically occurs within the first 72 hours of segregation (see Gondreau and Bonta, "Re-Examining the Cruel and Unusual Punishment of Prison Life" at 361). Thus by the time most prisoners were assessed by Zingerís researchers, they had probably already weathered the most difficult aspect of adjustment. It is possible that these prisoners started out with the same mental health levels as seen in the general population, were placed in administrative segregation and then deteriorated quickly within the first day or two so that by the time that they were initially tested their mental health and psychological functioning appeared lower than that of the general population, but nevertheless stable. This explanation is consistent with Zingerís results yet in no way supports the conclusion that administrative segregation has no detrimental effects on prisoners. Far from a "deep freeze," the favoured description of segregation for the "no negative effects" researchers and also used by Zinger, segregation may be more like a microwave, heating things up very quickly. The second problem with the first assessment session at 3.6 days is that it closely precedes the 5 day segregation review. After dealing with the initial despair, or, perhaps in the case of voluntary segregated prisoners, relief of being removed from the general population, prisoners by this time are probably turning their minds to their review. Involuntary segregated offenders in particular may be hopeful that they will be recommended for release and might show fewer negative effects at this time. There is a similar problem with the 30 and 60 day sessions as well: they also coincide with mandatory statutory reviews.

The other problem which I had asked Kate OíBrien to investigate was the anomalous finding by Zinger that all prisoners, both segregated and non-segregated, reported an improvement in their mental health and psychological functioning over the course of 60 days on all measures tested. Because the rate of improvement of both groups was the same Zinger concluded that the observed improvement was not the result of an actual improvement but simply an artifact of multiple testing, the phenomenon known as "practice effects." Textbook explanations of this phenomenon explain that this is usually associated with the test-re-test method where people take the same test on two different occasions, usually separated by a few weeks or months and then their scores are correlated with each other. An improvement on the second testing occasion because of familiarity with the test is related to the so-called practice effect. Textbooks often use the example of an intelligence test because intelligence is expected to remain constant over a period of time. To measure intelligence the subject is asked a battery of questions, such as "what is the next number in this sequence: 3, 9, 81, ?," and the faster he or she comes up with the correct answers the higher his or her score will be. Practice makes perfect, so if speed or performance of a complicated task is the key to doing well on the test it stands to reason that the more times you take the test the better you will perform.

Page 2 of 4