Trusting the Untrustworthy: The Social Organization of Trust Among Incarcerated Women

Abstract

Although the benefits of trust are well documented across a variety of settings, little empirical attention has been dedicated to trust in carceral settings, particularly among incarcerated women. Knowing how individuals in prison establish relationships of trust with one another is crucial for understanding how individuals adjust to conditions of confinement. Using data from 133 incarcerated women in a Pennsylvania prison unit, this study adopts a network approach to examine the role of individual and structural determinants of trust using exponential random graph models. Findings provide weak support for the claim that individual determinants (e.g. age, religious affiliation) shape whether women are more likely to trust someone to support them during an argument or a dispute. Instead, our findings show that structural determinants are the primary drivers of trust relationships. Trust is deeply entwined with friendship relations among women who get along with each other. Our approach paves a new path for the examination of trust in correctional settings and other criminological contexts.

Keywords:

• Trust
• women
• incarceration
• network
• ergm

Disclosure statement

No potential conflict of interest was reported by the authors.

Acknowledgments

The content is solely the responsibility of the authors and does not necessarily represent the official views of the respective agencies. We thank Elaine Arsenault, Ted Greenfelder, Nicole Harrington, Sade Lindsay, Alex Russo, and Michaela Soyer for their interviewer assistance. We also thank Scott Duxbury for assistance with modeling and Audrey Puckett for assistance with editing. We also thank three anonymous reviewers and the editor for their improvements to the manuscript.

Notes

1 We chose not to ask about friendship itself given that some prisoners are adamant about not having “friends” in prison, only “associates” (Crewe, 2009).

2 Specifically, we first construct a matrix representing the number of two-paths from i to j where a two-path is defined as a tie from i to k and a tie from k to j in the trust network. This matrix is constructed by squaring the trust network. This two-path matrix is then multiplied by the transpose of the get along with matrix. The product matrix is the network embeddedness measure.

3 Models are estimated and goodness-of-fit is evaluated using the ergm package (Handcock et al., 2019; Hunter, Handcock, Butts, Goodreau, & Morris, 2008) in R. Goodness-of-fit indices are available in the online supplement.

4 Average marginal effects and predicted probabilities are calculated using the ergMargins package (Duxbury, 2019).

5 We thank an anonymous reviewer for suggesting we calculate the conditional probabilities to better communicate the substantive effects of our results.

6 To control for the tendency of some individuals to receive more ties and send more ties, we take into account the indegree distribution and the outdegree distribution, as research indicates that failure to do so may overestimate effects of attributes on receiving ties (Lusher & Ackland, 2011). We parameterize the degree distributions using the geometrically weighted indegree (GWIDEGREE) and outdegree (GWODEGREE) effects (Hunter, 2007). Two-path counts the number of configurations where there is a tie from i to k and a tie from k to j in the trust network. Triadic closure captures the likelihood of an i→j tie, conditional upon the number of k prisoners for which i→k and k→j ties also exist. Cyclical closure captures the likelihood of an i→j tie, conditional upon the number of k prisoners for which k→i and j→k ties also exist. We parameterize triadic closure and cyclical closure using the geometrically weighted edgewise shared partner (GWESP) effects (Hunter, 2007; Hunter & Handcock, 2006).

7 The graph correlation between the trust network and the get along with network is only 0.4, suggesting that there are a nontrivial amount of dyads for which either a get along with tie or a trust tie exists, but not both.

8 It may be the case that women were interpreting our measure of support as reflecting physical violence. As such, we might expect that physique would predict trust. To check whether this was the case, we examined whether BMI, a measure of exercise frequency, and an interviewer report of how muscular the respondent appeared. These measures were not associated with trust nominations.

Additional information

Funding

This research is supported by funding from the National Institute of Justice (2016-MU-MU-0011).

Notes on contributors

Jacob T.N. Young

Jacob Young is an Associate Professor in the School of Criminology and Criminal Justice at Arizona State University. He is also the Associate Director for the Center for Correctional Solutions and Curriculum Coordinator for the Institute for Social Science Research. His research examines how the structure of social networks influence behavior across a variety of contexts as well as identifying the mechanisms by which social networks develop and evolve. His published research on these topics has appeared in the American Sociological Review, Criminology, Justice Quarterly, Social Networks and the Journal of Quantitative Criminology. He is a founding member of the Prison Incarcerated Persons Network Studies (PINS), an ongoing, multi-site study of the social networks among incarcerated men and women and the implications of network position on re-entry. His research has been funded by the National Science Foundation, the National Institutes of Health, and the National Institute of Justice. He received his PhD in Sociology from the University of Washington in 2011.

Dana L. Haynie

Dana L. Haynie is a Professor of Sociology at The Ohio State University whose research applies criminological and social network methods and theories to better understand processes related to crime and delinquency. Examples include work examining the diverse ways in which peer relationships and social networks affect adolescent involvement in risky behavior, race and gender differences in social networks and crime, and the role of neighborhood and school networks for understanding criminal behavior. Current projects focus on mass incarceration and re-entry, illegal drug trafficking occurring on the Darkweb, and the opioid crisis impacting Ohio. She is a founding member of the Prison Incarcerated Persons Network Studies (PINS), an ongoing, multi-site study of the social networks among incarcerated men and women and the implications of network position on re-entry. Her research has been funded by the National Science Foundation, the National Institutes of Health, and the National Institute of Justice and published in the leading sociological and criminological journals with more than 6,000 citations to her published work. She received her Ph.D. in Sociology from the Pennsylvania State University in 1999.