This commentary (1998) by Ralph Metzner reflects on the Harvard Psilocybin Project and the Concord Prison Experiment, concluding that the original positive findings regarding reduced recidivism were likely erroneous and no better than chance. He suggests that while psychedelics can induce insight, lasting behavioural change requires sustained rehabilitation and support.
From the second paragraph of the paper: It is disconcerting, of course, to discover 35 years after the fact that a research project I was involved in and wrote about made quantitative errors and reported erroneous conclusions. As I read Rick Doblin's findings, and re-read our original papers, it did give me occasion to reflect on that period, and what was called the Harvard Psilocybin project-and to come to the depressing conclusion that none of it did any better than chance (as far as one could tell from the tests).
Metzner situates this paper as a retrospective reflection on the Concord Prison Project, a psilocybin-assisted programme for convicts affiliated with the Harvard Psilocybin Project in the 1960s. He frames the work in relation to other follow-up studies conducted by Rick Doblin and notes that Doblin's re-analysis prompted recollection and reassessment of earlier reports. The introduction emphasises that, although those involved—including Tim Leary, graduate students and Metzner himself—believed they were conducting rigorous research, later scrutiny reveals quantitative errors and conclusions that do not withstand re-analysis. The stated gap motivating the reflection is twofold: first, uncertainty about the original quantitative findings concerning recidivism after the intervention, and second, a broader question about whether profound psychedelic-induced personal insights translate into durable behavioural change without sustained post-release support. Metzner indicates the purpose of the present paper is to revisit the project record, acknowledge errors, and consider what the corrected interpretation implies for psychedelic-assisted interventions with criminal populations.
Metzner recounts core elements of how the Concord Prison Project was conducted, as presented in the original reports. The intervention involved psilocybin sessions delivered to incarcerated men at Concord Prison, accompanied by psychotherapy; personality tests were administered before and after therapy. The investigators regarded recidivism—the rate at which released parolees returned to prison—as the principal, objective behavioural index of lasting change. Secondary outcomes included psychometric measures such as the California Psychological Inventory (CPI), the Minnesota Multiphasic Personality Inventory (MMPI), and behavioural ratings. To establish an appropriate comparison, project members assembled a recidivism base-rate by reviewing archival prisoner files at the State Department of Corrections—Metzner and a colleague spent a summer extracting those data. Follow-up periods referenced in the extracted text include a 10-month follow-up figure reported in some accounts and a 30-month base-rate used elsewhere. The extracted text does not clearly report sample sizes, exact dosing regimens, session frequency, psychological support model details, or the statistical methods originally used; those specific procedural and analytic details are not present in the provided extraction. Metzner also describes subsequent re-analyses and commentaries by others (notably Rick Doblin) that revisited the original outcome comparisons and base-rate choices. These re-analyses examined classification of returns for parole violations versus new crimes and the appropriateness of the control rates used in earlier reports.
According to Metzner's account of the original publications, the 1965 paper reported no difference between the programme group and the base-rate in overall recidivism (the rate of return to prison). The same paper, however, reported significant changes on several scales of the CPI and on one MMPI scale, as well as some inconclusive behavioural rating changes. Metzner reiterates that these psychometric shifts were observed despite the absence of a demonstrable reduction in recidivism. Metzner recounts that later accounts by Leary presented a more favourable interpretation. In particular, a 10-month follow-up recidivism figure of 32% was reported in one venue and was compared against a 30-month base-rate of about 30% in another, producing an apparent significant reduction. Metzner states that this comparison was erroneous: using a shorter follow-up figure against a longer base-rate inflates the impression of benefit. Rick Doblin's re-analysis, as described here, further showed that many prisoners returned for parole violations had also committed new crimes, so the distinction Leary drew between returns for parole violation and returns for new crimes was largely an artefact. Metzner acknowledges that the positive-looking findings publicised by Leary arose from mistakes in choice or application of control figures and from the "halo effect"—the tendency for investigators to see their data in a positive light. Interviews conducted roughly 30 years later with two programme participants are reported to document the human side of the outcomes and to illustrate that profound experiences alone did not produce durable behavioural integration without external supports. The extracted text does not provide precise numerical results beyond the cited 32% and the approximate 30% base-rate, nor does it give sample sizes or confidence intervals for the effects.
Metzner interprets the corrected record as supporting a nuanced conclusion: psychedelic-assisted sessions can occasion profound insight and personality change, but converting those inner changes into sustained behavioural change—specifically reduced criminal reoffending—requires structured rehabilitation and community support upon release. He reflects on the practical difficulties the project faced in helping participants reintegrate, describing Leary's personal efforts to find work, housing and companionship for parolees, and noting that the intervention lacked a reliable post-release support system. Metzner candidly attributes the original overly optimistic presentations in part to the halo effect and investigator enthusiasm rather than to clear evidence of intentional data manipulation. He states that he cannot determine whether Leary's mistakes were conscious falsifications or unconscious carelessness but expresses a preference for the latter explanation because the project record contains inconsistencies rather than clear evidence of deliberate fraud. Metzner emphasises that making mistakes is part of the scientific process so long as errors are acknowledged and corrected once discovered. On the question of ethical standards, Metzner questions Doblin's assertion of a need for "higher" or "highest" standards to regain regulators' trust, arguing that existing norms of honesty and complete reporting within the scientific community are adequate; those norms should be applied to ensure accurate observation and reporting. He concludes by expressing appreciation that Doblin's re-analysis led to a more complete and accurate understanding of the Concord Prison Project, while reiterating the primary substantive lesson about the limits of psychedelic experience alone to change entrenched criminal behaviour without systemic aftercare.