Examining Trajectories of Change on Risk and Protective Factors among White and Black Men Offenders on Community Supervision in Iowa

Abstract

Relying on the results of validated risk assessments is vital to evidence-based decision making in corrections. Advancements in the approach to risk assessment has seen an emphasis on measuring dynamic risk factors alongside protective factors, as both are expected to be useful for identifying treatment targets and measuring changes in risk over time. Despite these advancements, empirical evidence of change on dynamic risk and protective factors remains limited. Utilizing a sample of 3,976 White and Black men offenders on community supervision with a minimum of three waves of assessment, this thesis tested a three-step analytic approach to establish whether: (a) the Dynamic Risk Assessment for Offender Re-entry (DRAOR) measured the constructs of dynamic risk and protective factors consistently over time, (b) there was evidence of trajectories of within-offender change throughout community supervision, and (c) individual trajectories of change predicted recidivism above and beyond baseline DRAOR score and static risk. Results indicated that the DRAOR was measuring the same constructs in the same way over repeated assessments. Change was observed across each of the DRAOR domains, indicating that on average, offenders were expected to decrease in stable and acute risk over time, and increase in their protect score. Variations in the rates of change were unrelated to static risk, age, and race, but change on each DRAOR domain was partially explained by scores on the remaining domains. Change trajectories were significant predictors of revocations of community supervision after considering the effects of initial score, static risk, and age, but change trajectories were not related to new convictions. The findings indicate that the DRAOR is sensitive to change and First, I want to express my gratitude to my supervisor, Dr. Ralph Serin. I am truly grateful for the years of mentorship that you provided, which were filled with support, advice, understanding, and opportunities. I am appreciative of the constructive feedback that I received from my committee members, Dr. Shelley Brown and Dr. Julie Blais, throughout the development of this research. I also wish to thank Dr. Ryan Labrecque, my external examiner, and Dr. Katharine Kelly, my internal examiner, for contributing their knowledge to strengthen the overall quality of this document. This project would not have been possible without the support of Iowa Department of Corrections. In particular, I want to thank the members of the Research Department at the Central Office for their role in preparing and updating the datasets and always being willing to answer my questions. Finally, the unwavering support, patience, and encouragement from my family and friends made accomplishing this milestone possible. Thank you for being a constant source of inspiration.

Psychometric Properties of a Risk Tool Across Indigenous Māori and European Samples in Aotearoa New Zealand: Measurement Invariance, Discrimination, and Calibration for Predicting Criminal Recidivism

“…Following previous studies (Chadwick, 2020;Davies et al, 2022;Lloyd, Hanson et al, 2020), we used a threefactor structure for DRAOR within the ESEM analyses. For ordinal categorical data with less than five categories, weighted least squares estimation methods with robust corrections are more appropriate than maximum likelihood methods (see Sellbom & Tellegen, 2019).…”

“…Corrections agencies in other countries use DRAOR (e.g., see Chadwick, 2020;K. Singh & Samion, 2016), but most current DRAOR research originated from Aotearoa NZ and, thus, these samples contain substantive M aori participants (e.g., see Lloyd, Hanson et al, 2020;Polaschek & Yesberg, 2017;Scanlan et al, 2020;.…”

Investigating the Dynamic Risk Assessment for Offender Re-entry’s (DRAOR) Ability to Operate as a Weighted Risk Prediction Instrument

“…12 For example, the DRAOR Total score has been strongly correlated with the Stable (r = .87 to .92), Acute (r = .81 to .87), and Protective subscales (r = -.82 to -.89; Averill, 2016;Ferguson, 2015;Muirhead, 2016;Perley-Robertson, 2018;Smeth, 2013). Additionally, the DRAOR Stable and Acute subscales have been consistently found to be moderately to strongly positively correlated (r = .46 to .73), indicating that the two subscales measure similar constructs (i.e., criminogenic risk; Averill, 2016;Chadwick, 2014;Chadwick, 2020;Ferguson, 2015;Muirhead, 2016;Perley-Robertson, 2018;Smeth, 2013;Tamatea & Wilson, 2009;Wardrop, 2020).…”

“…Research has shown that DRAOR scores are normally distributed and demonstrate change over the course of supervision (Chadwick, 2020;Lloyd et al, 2020a;Tamatea & Wilson, 2009). However, previous investigations into the DRAOR's internal consistency, a common indicator of scale reliability that measures whether a series of items have been appropriately grouped based on their interrelatedness, have yielded consistently mixed results.…”

“…Conversely, researchers have consistently reported negative associations between the Protective subscale and the Stable (r = -.40 to -.73) and Acute subscales (r = -.27 to -.63), suggesting that the Protective subscale is composed of factors that operate differently than dynamic risk factors do (Averill, 2016;Chadwick, 2014;Chadwick, 2020;Ferguson, 2015;Muirhead, 2016;Perley-Robertson, 2018;Smeth, 2013;Tamatea & Wilson, 2009;Wardrop, 2020). In an initial validation study using a sample of 59 New Zealand probationers, Tamatea and Wilson (2009) compared the DRAOR to the Risk of re-Conviction X Risk of re-Imprisonment model (ROC*ROI; Bakker et al, 1999), a validated measure of static risk that has been implemented in New Zealand.…”