|A New Behavior Change Program Using Psilocybin|
|Written by Timothy Leary|
|Tuesday, 05 January 2010 22:29|
A New Behavior Change Program Using Psilocybin
TIMOTHY LEARY, RALPH METZNER, MADISON PRESNELL,
|TESTS||SIX WEEKS||TESTS||SIX WEEKS||TESTS|
The actual results will be presented separately for the three sets for data—personality tests, behavior ratings and return rates.
A. Personality Tests. Table I shows the MMPI scores from pre- and post-tests for the pilot and experimental groups. It should be remembered that the schedule and time-period intervening between pre- and post-tests for the pilot subjects was somewhat longer and more variable than for the experimental subjects. The number of subjects involved in the analysis is less than the total sample owing to incomplete test-protocols or subject dropout. All means are based on T-scores, except for the ego-strength scale, which is based on raw scores. It will be seen that although on several of the scales (D, Pd. Sc, Ma) there are changes in the expected direction, these are not significant. The drop in the psychopathic deviate scale is almost significant at the .05 level. The significant decreases in the two validity scales (L and F) indicate that subjects are less likely to put on a good front or answering at random after the treatment program.
Table II shows the MMPI scores for the Control group before the control period (I), at the end of the control period and before treatment (II), and after treatment (III). Three comparisons were made, all using Wilcoxon signed-rank tests and 2-tailed probability levels. Comparing I and II no difference was predicted and none was found. Comparing the changes from I to II with the changes from II to III it was predicted the latter would be greater. However, the differences were not significant on any of the scales. It will be noted that on several of the scales (Hy, Pd. Pt. Sc) there is a U-shaped trend, with a tendency for scale scores to rise on the third testing after the treatment. This might reflect some peculiar effects due to repeated testing. The fact that the mean F-score did not change also indicates that the validity of these post-treatment scale scores is somewhat doubtful. When I and III, i.e., first and last tests are compared, only the D-scale shows a significant decrease (p < .02 ) .
Table III shows the mean scale scores from the CPI (California Psychological Inventory) for the pilot and experimental groups. Again, significance of change was estimated for the experimental group alone and for pilot and experimental groups combined. Significant increases are shown on 12 out of 18 scales, the most marked being on Sociability, Sense of Well-Being, Socialization, Tolerance and Intellectual Efficiency. The socialization-maturity scales (Re, So, Sc) and the achievement scales (Ac, Ai, Ie) as a group all show significant increases. Table IV shows pre- middle- and post-tests for the control group. None of the changes from I to II are significant except an increase in Good Impression ( p< .01), which occurs also in the experimental group. Except for this scale then, the treatment group changed significantly more than the control group. For the final testing (III) only 7 subjects were available, hence the comparison of changes from I to II with changes from II to III, as well on the overall change from I to III becomes very unreliable. In fact, none of the differences are found to be significant, although they are in the predicted direction. There is however a significant increase from I to III on the Dominance scale for the control group (p < .05).
A different method of analysis is to compute change-scores for each subject and compare mean change-scores in the experimental and control groups, using the Mann-Whitney U-Test. This is a direct estimation of the significance of difference in change, rather than the indirect method of comparing pre- and post-test scores in the two groups. On the MMPI only the decreases on the F and K scales are significantly greater in the experimental than the control group (p. 10). On the CPI the increase in Tolerance (To) for experimental subjects (N = 20) is significantly greater than for control subjects (N = 12); the two-tailed p-value is <.05. The increase in Achievement via Conformity (Ac) is significantly greater for experimental subjects (N = 12) than for the control subjects (N' = 12); the two-tailed p-value is <.10.
Interpretation. There are two problems which make the interpretation of these data difficult. One is artificial, one is substantive. The artificial problem is that the final test scores (III) of the control group are not apparently very reliable, and hence the within-group comparison cannot be made adequately. The substantive problem is that given the existence of significant changes after the treatment program on the CPI it is not known which features of the program are responsible for the changes. Firstly, the fact that most participants are about two to three months removed from parole introduces some ambiguity. Wheeler ( 1961 ) has shown that there is increasing conformity to staff and community norms in prisoners' attitudes just prior to release. However, two facts make this explanation of the data implausible: (1) the changes observed by Wheeler occurred in the last six months prior to parole—there is no evidence to suggest that this trend is continued linearly over such short periods as 6 weeks; (2) this general norm-shifting effect should apply to the control group also, which, as we have seen, does not change. A second factor is the possibility that the feedback of results has simply made the subjects "test-wise" and this can account for all the variance. Although the tests were taken honestly i.e., there was no possibility of memorizing all the items of a scale and their direction, it is true that we do not know whether the feedback alone could have produced these results. We can only say now that this requires further analysis. For example, eventually, it will be possible to correlate success on parole with changes in personality tests. Ferdinand (1962) reports similar CPI changes in a group of juvenile offenders treated by "milieu therapy."
The Sentence Completion Test used here (cf. Watt and Maher, 1958) consists of the responses to 41 sentence stems, coded according to a five-point scale from "extreme disapproval, non-conformity, very negative" (1) to "praise, approval, conformity, extremely positive" (5). The content of the stems covered a variety of social institutions, e.g., law, family, sports, arts, business. etc. Three scores were computed for each person: (1) mean across all items, (2) frequency of very positive responses4's and 5's, and (3) frequency of very negative responses1's and 2's. Table V shows the means for two groups—the pilot plus experimental and the control group. On the mean "positive attitude" score the experimental and the controls increase, but the experimental group more. When only the extreme positive or negative 'responses are counted the experimental group changes (positively), the control group does not. Again the results from the third test are bedeviled by sample shrinkage. Since this test was not used in the feedback program, it is not subject to the same confounding variables as the MMPI and CPI. It would seem that a concomitant of the program is a decrease in cynical and hostile attitudes towards a variety of social institutions.
B. Behavior Ratings. Two types of behavior ratings by independent observers were collected, but both are subject to many sources of unreliability. (1) The regular quarterly institutional work reports, by the inmate's work-instructor are six rating scales covering work competence, industry, co-operativeness, etc. When July, September and December work reports were examined there were not consistent trends on any of the scales over the period of the program; nor were there any significant differences between experimental and control groups. However, the samples in this analysis are very small (4 to 9) and the ratings are made by different observers. Thus the absence of discernible trend is not really surprising. (2) A special rating sheet for officers was constructed on eleven areas of interaction and behavior (see Appendix). The "experimental group" here consists of 13 men still in the prison who had gone through the program; the "control group" are a matched sample, selected according to the same criteria, but never having had any contact with the project at all. Again, there is the problem of the effects of different raters, and their possible bias about the project affecting their judgments. With these reservations in mind, Table VI provides some suggestive evidence. The results are given in mean ratings (the scales were either four- or five-point) and the significance of differences was computed by means of chi-square. None of differences between the groups are significant, but 4, 5. 6. (which approach significance) very tentatively suggest that participants are seen as less excitable and as getting along better with officers and with other inmates.
C. Return Rates. One and a half years after termination of the project (18-26 months after release from prison). The recidivism rate in this project does not differ from the expected rate derived from base-rates for the Concord Reformatory as a whole (Metzner & Weil. 1963). In that study 56% of the 311 men released from Concord during 1959 had returned two and a half years later. Out of the 32 men involved in the project, four are still in prison and one escaped. These must therefore be omitted. Of the 27 men released, 11 are still on the street and 16 have returned, a return rate of 59%.
Expected Rate of Return by Type of ReturnIn the base-rate study half of the recidivists were returned for parole violations and half for new offenses. These two types were then combined for further calculation of predictive categories. In other words we would expect 28% of the released men to be returned as parole violators and 28% as new offenders. When we look at the figures actually obtained, we see that only two out of the 27 men (7%) were returned for new offenses, while 14 out of 27 (52%) were returned as parole violators. This discrepancy has a probability of less than .01 of occurring by chance, using the binomial distribution. In other words there is a significant reduction in the rate of new crimes and a significant increase in the rate of parole violations. This dual effect accounts for the lack of difference when the overall rate of return is considered.
One may speculate about the reasons for the rise in parole violation. Perhaps the men on the psilocybin project received an extra careful degree of parole supervision. The project had aroused a lot of interest in the Department of Correction, and it was impossible to prevent the parole officers from knowing which of their charges had been involved in it.
Expected Rate of Return by Prognostic CategoriesIn the base-rate study referred to above, expectancies were computed for six different sub-classes of offenders. The categories were obtained empirically on the basis of their predictive efficiency. Thus for example, men with no prior arrests and no prior commitments have an expected return rate of 22%. Men with prior commitments, who committed offenses against a person (but not sex offenders) or against property and who are non-white, have an expected return rate of 86%. Thus, these categories enable one to obtain a more precise expectancy for any particular sample than simply the overall rate.
Table VII compares for each category the percentage returning in the experimental and base-rate samples.
Although the experimental subsamples are too small to make statistically valid comparisons, the figures indicate a reduced return rate in groups (1) and (6) and an increased rate in group (2). It should be remembered that these figures for new offenses and parole violations are combined and therefore do not enable one to specify in what category a significant reduction of new crimes occurred.
Conclusion. Of the three types of evaluation, the most important is the rate of return. It is a completely objective behavioral index, not subject to any of the distortions of personality tests and clinical impressions. The main conclusion can be stated as follows: One and one half years after termination of the program the rate of new crimes has been reduced from 28% to 7%, although if parole violations are counted the overall return rate has not changed. It is proposed that these results warrant further research into the potentials of the methods used, especially since no other method of reducing the crime rate exists.
FUTURE DEVELOPMENTSFrom our experience in this project we would offer the following suggestions for an improved rehabilitation program designed to decrease the recidivism rate of offenders with relatively short sentences.
If the core of the rehabilitation or change process is some form of intense group experience designed to bring about insight then it is essential that the environment in which this insight takes place is supportive of applying such insights to behavior. The ideal solution to this problem is to involve the entire institution, officers, psychologists, as well as inmates, in a joint change process, as in the Herstedvester Center in Denmark (Stürup, 1959). We have attempted to tackle this problem by placing some responsibility for stimulating behavioral change on older, more experienced inmates.
It is highly undesirable to have an inmate return to the same frustrating environment after experiencing an internal liberation. An alternative would be to have the group experience (whether it involves drugs or not) occur outside of the prison, immediately after release, in a special transitional center. This would serve both as a sort of retreat for internal change and as a halfway house to prepare the convict for regular life on the streets.
The second suggestion concerns the importance of the follow-up period. Many convicts are reluctant to get involved in middle-class activities. The doctor-patient model, in which a client regularly visits the office of a professional, is simply not applicable. In practical terms, the "therapist" must be prepared to visit his clients at all times of the day or night in bars or homes, to help find employment, to lend money, etc., because these are the accepted "tests" of a trusting relationship. This is not to say that there should be no structure at all to the relationship, but the structure should come from a definite contractual agreement about the purposes of contacts—and not from arbitrarily imposed space-time limits. For further elaboration of these two ideas see the discussions by Leary (1961), Schwitzgebel (1961) and Eissler (1950) .
Footnotes1. The help, advice and encouragement of the following persons are gratefully acknowledged: Edward W. Grennan, Superintendent of Massachusetts Correctional Institution, Concord; David C. McClelland, Ph.D., Director of the Center for Research in Personality, Harvard University; Norman A. Neiberg, Ph.D., Director of Psychological Research in the Division of Legal Medicine; and David Houghey, Ph.D., Director of Psychological Research in the Department of Correction; Bernard Dee, Institutional Parole Officer; William P. Ryan, head Correctional Social Worker; Cornelius Twomey, Chairman of the Parole Board and Martin Davis, Chief of Parole Division The following graduate students contributed actively to this program: Stephen Berger, Jonathan Clark, Don Fowles, Rudolf Kalin, David Kolb, George Litwin, Jonathan Shay, James Uleman; and Michael Hollingshead. (back)
2. Grateful acknowledgment is also made to Sandoz Pharmaceuticals, and its director, Carl Henze, M.D., for supplying us with psilocybin and for his continued interest and cooperation in our program. (back)
3. The pilot group usually had three sessions; some of the later group members also participated as assistant group leaders, hence the number of psilocybin sessions is not constant. The actual distribution is as follows: 15 had two sessions, 9 had three, 4 men had four and 5 had five. One man, a chronic alcoholic and multiple parole violator, was given the drug once outside of the regular program; he is not included in the main sample. We feel that alcoholics present a special problem and have not attempted to cope with it here. Other investigators, however, (Chwelos. et al., 1959) have reported considerable success using LSD-25 with alcoholics. (back)
Tables(back to text)
|Pilot Group (N=8)||Experimental Group (N=11)||Significance*|
Col. B: Significance of difference from pre- to post-test for Pilot and Experimental Goups combined
* All significance estimates made with Wilcoxon signed-rank test. Probability levels are two-tailed.
(back to text)
(back to text)
|Pilot Group (N=8)||Experimental Group
|Scale||pre||post||I—pre||II—post||Exp. gr.||Pilot & Exp.
|Sy||53.4||58.5||47.3||56.0||p< .02||p< .01|
|Sp||57.5||62.4||52.4||59.5||p< .02||p< .01|
|Wb||46.1||54.5||43.0||52.5||p< .01||p< .01|
|Re||34.4||47.0||33.2||41.0||p< .05||p< .01|
|So||32.6||41.8||31.9||40.4||p< .02||p< .01|
|Sc||42.8||48.9||38.9||46.3||p< .05||p< .01|
|To||42.2||52.1||41.8||49.5||p< .02||p< .01|
|Gi||45.5||53.4||40.2||48.5||p< .01||p< .01|
|Ie||48.1||57.4||43.5||52.3||p < .05||p< .01|
(back to text)
|Pilot & Experimental Groups (N=19)||Control Group (N=11)|
|I vs II||II vs III|
**Wilcoxon sign rank test, 2-tailed
(back to text)
|Question||Exp. X(N=12)||Control X(N=12)|
|4. Not excitable||2.8||2.4|
|5. Getting along with officers||3.0||2.7|
|6. Getting along with inmates||3.1||2.5|
|8. Talk with officers||1.5||1.8|
|10. Parole success||2.5||2.6|
|11. Estimate of proiect||2.0||1.8|
|Description||Base-Rate Sample||Psilocybin Sample|
|N||% Return||N||% Return|
|1. No prior commitments; no prior arrests||23||22||2||0|
|2. Some prior commitments sex offender or parole violator
whose age at last commitment was more than 24
|3. No prior commitments but some prior arrests||59||37||3||33|
|4. Sex offenders or parole violators with prior
commitments aged 24 or less at last commitment
|5. Some prior commitments; offense against person (except sex),
against property or combination; whites
|6. Some prior commitments; offense against person (except sex)
against property or combination; other ethnic group
ABRAHAMSON, H. A. "Lysergic acid diethylamide (LSD-25): III. As an adjunct of psychotherapy with elimination of fear of homosexuality," J. Psychol., 39, 127, 1955.
BLEWETT, D. B., & CHWELOS, N. Handbook for the therapeutic use of lysergic acid dietltylamide-25: Individual and group procedures, 1959. CAMPBELL., D. T. "Factors relevant to the validity of experiments in social settings." Psychol. Bull. 54, 297-312, 1957.
CARTWRIGHT, D. Achieving change in people. Hum. Rel., 14, 38l-392, 1951.
CHEWELOS, N., BLEWETT, D. B., SMTTH, C., & HOFFER, A. "Use of LSD-25 in the treatment of chronic alcoholism." Quart. J. Stud. Alcohol., 20, 577, 1959.
CUTNER, M. Analytic work with LSD-25. Psychtiat. Quart. 33, 715-757, 1959.
DUCHE, D. J., & LAUT, J. "Effets de la psilocybine dans un cas d'hysterie. Annales Medico-Psychologiques, 119 (3), 590, 1961.
EISSLER, K. Ego-psychological implications of the psychoanalytic treatment of delinquents. Psychoanal. Study of the Child, 5, 97-121, 1950.
EYSENCK, H J. The effects of psychotherapy. In Handbook of Abnormal Psychology, ed. H. Eysenck, New York: Basic Books, 1961.
FRANK, L. K. Change through group experience. Social Welfare Forum, Proc. Nat. Conf. Soc. Welfare, Chicago, New York: Columbia, 1958. Pp. 225-268.
FERDINAND, T. N. An evaluation of milieu therapy and vocational training as methods for the rehabilitation of youthful offenders.J. Crim. Law, Criminol., Police Sci., 53, 49-54, 1962
FOX, V. Group methods in criminology. Group Psychother. 15, 40-45, 1962.
GOFFMAN, E. Characteristics of total institutions. In Symposium on Preventive and Social Psychiatry, 43-84. Washington, D.C.: Walter Reed Army Institute of Research, 1957.
HEARD, G. Training for a Life of Growth Santa Monica: Wayfarer Press, 1959.
JENKINS, D. H. Feedback and group self-evaluation. J. Soc. Issues, 4, 50-60, 1948.
KONOPKA, G. Group Work in the Insitution. New York: Morrow, 1954.
KONOPKA, G. The generic and the specific in group work practice in the psychjiatric setting. H.B. Trecker, (ed.) New York: Whiteside, 1956.
LEARY, T. How to change behavior. In: G. Nielsen (ed.) Intern. Congr. of Appl. Psychol., Copenhagen, 1961. (Vol. IV)
LEARY, T., LITWIN, G., & METZNER, R. Reactions to psilocybin in a supportive environment. J. nerv. ment. Dis. 1963 (in press), 137, 567-578.
LEVIT, G., & JENNINGS, H. H. Learning through role playing. In: Planning of Change. Bennis, W. G., Benne, K. D. & Chin, R. (eds.), New York: Holt, 1961. Pp 706-710.
Massachusetts Commissioner of Correction. Statistical Report., 1959.
McCORKLE, L. W., ELIAS, A. & BIXBY. F. L. The Highfields Story: An Experimental Treatment Project for Youthful Offenders. Next York: Holt, 1958.
METZNER, R., & WEIL, G. Predicting Recidivism: Base Rates for Massachusetts Correctional Institution Concord. J. Crim. Law, Criminol., Police Sci., 1963, 54, 307-316.
MORENO, J. L. Psychodrama. In: S. Arieti, (ed.), American Handbook of Psychiatry. New York: Basic Books, 1959. Pp. 1375-1396.
O'HALLORAN, W. Group psychotherapy and the criminal---an introduction to reality. Am. J. Correction 1961.
ROSOW, H. M. Some observation on group therapy with prison inmates.Arch. Crim. Psychodyn. 1955.
SANDISON, R. A. Psychological aspects of the LSD treatment of the neuroses. J. Ment. Sci., 100, 508, 1954.
SCHWITZGEBEL, R. Critical analysis and evaluation of the experimenter-subject role relationship in the reduction of known male adolescent crime. Ed.D. Dissertation, Harvard University, 1962.
SLATER, P. E. Displacement in groups. In: Planning of Change. Bennis, W. G., Benne, K. D., & Chin, R. (eds.), New York: Holt, 1961. Pp. 725-786.
STÜRUP, G. Group therapy with chronic criminals. Acta Psychother. 7, (Suppl.), 377-385, 1959.
SUTHERTLAND, E. H. Principles of Criminology, Philadelphia: Lippincott, 1955.
SZASZ, T. The Myth of Mental Illness. New York: Hoeber-Harper, 1961.
TENNENBAUM, B. Group therapy with LSD-25. Dis. Nerv. Syst., 22, 459-492, 1961.
WATT, N., & MAHER, B. A. Prisoners' attitudes toward home and the judicial system. J. Crim. Law., Criminol., Police Sci., 49, 327-330, 1958.
WHEELERS S. Socialization in correctional communities. Amer. Sociol. Rev. 26, pp. 697-712, 1961.
|< Prev||Next >|