Date of Degree


Document Type


Degree Name





Deidre Anglin

Committee Members

Geraldine Downey

Paul Wachtel

Elliot Jurist

Diana Punales

Subject Categories

Clinical Psychology


Mass Incarceration; Rehabilitation; Narrative Therapy; Resilience; Compassion for Self and Others


The Longtermers Project is a fifteen-session group therapy intervention that has been run in three prisons in New York state, two men’s facilities and one women’s, since 2010. The Coming to Terms curriculum, designed specifically for work with this population by Kathy Boudin and her partners at the Osborne Association, a non-profit organization that provides assistance to formerly and currently incarcerated persons in New York state, asks participants to think, write, and speak about their lives in general and particularly about the incident crime for which they were incarcerated. Now that the program has run for several years and a significant number of participants have completed the program and subsequently secured release from prison, the time is opportune to compare program participants to otherwise similar participants who did not complete such a program. This is a hard test of the program – more than a year after participation, do participants look different from otherwise similar non-participants on major outcomes of interest: resilient coping, generativity, self-efficacy, compassion for self, compassion for other, and life narrative themes?

The present study investigated these questions by matching eight Coming to Terms participants who had since been released from prison with twelve formerly-incarcerated persons who did not complete the program but who did receive post-incarceration services from the Osborne Association or a parallel organization. Participants were matched on the variables of total time incarcerated and length of time elapsed since release from prison. While groups were not explicitly matched on other criteria, control and intervention groups evidenced similar patterns of age, racial and ethnic representation, and severity of crime. Participants completed six instruments shown to have good psychometric properties: the Connor-Davidson Resilience Scale (2003), the Self Compassion scale (Neff, 2003; Raes, Pommier, Neff, & Van Gucht, 2011), the Compassion for Others scale (Pommier, 2011), the Loyola Generativity Scale (McAdams & de st. Aubin, 1992, the New General Self Efficacy Scale (Chen, Gully, and Eden, 2001), and the Guilt and Shame Proneness Scale (Cohen, Wolf, Panter, & Isko, 2011) in addition to a demographic questionnaire developed by the researcher. Subjects likewise completed a structured, qualitative narrative protocol (see McAdams, 2006; McAdams & Guo, 2015; Ehlman & Ligon, 2012) that was assessed for the salience of certain themes, such as generativity and redemption, found in previous research to correlate with indicators of more adaptive coping. Narratives were also assessed qualitatively via a purposeful selection technique that compared the most adaptive - that is the most resilient, generative, and self-efficacious - cases in the sample to the least adaptive.

The study found that Program Participants demonstrated greater levels of generativity than did controls. Additionally, while compassion for self and compassion for other were significantly and strongly correlated constructs for intervention subjects, they were unrelated for control participants. These results provided preliminary support for the efficacy of the intervention. There were no statistically significant differences between Program Participants and controls on the dimensions of compassion for self, compassion for others, resilience, and self-efficacy, though p-values were relatively low (ranging from .15 to .26) for all variables except for compassion for self, suggesting that these variables may have assumed statistical significance given a larger sample size. There were also no statistically significant differences across treatment condition in the prevalence of six identified narrative themes. Though quantitative analysis did not find statistical differences in the prevalence of narrative themes by treatment condition, qualitative analysis found distinct patterns of redemptive and contamination themes when more adaptive – that is more resilient, generative, and self-efficacious - cases were compared to less adaptive cases. Additionally, when all twenty narratives were coded for salience of in-prison educational programming, the vast majority of subjects were found to have spontaneously identified education as an important and positive aspect of their life trajectories; a slightly higher proportion of intervention subjects than control subjects identified education in this way.