Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


36419.pdf (3.31 MB)

ETD Abstract Container

Abstract Header

Confirmatory Factor Analyses of the Level of Service Inventory-Revised

Arnold, Thomas K
http://orcid.org/0000-0003-3537-3774
http://rave.ohiolink.edu/etdc/view?acc_num=ucin1573224799095916

Abstract Details

2019, PhD, University of Cincinnati, Education, Criminal Justice, and Human Services: Criminal Justice.
The purpose of this dissertation is to provide a set of confirmatory factor analyses of the Level of Service Inventory-Revised (LSI-R; Andrews & Bonta, 1995). The LSI-R is a widely used offender risk/needs assessment with 54 dichotomous items broken down into 10 subscales. The LSI-R is a “dual purpose” “risk/needs” offender assessment instrument that is designed to assess 1) recidivism risk and 2) treatment needs. The first purpose of the LSI-R, assessment of recidivism risk, is achieved by examining the total LSI-R score, which is the sum of the 54 dichotomous item scores. Higher scores are correlated with higher recidivism levels. The second purpose of the LSI-R, assessment of treatment needs, is achieved by the examination of the ten LSI-R subscale scores. A high score on a subscale indicates that the offender may need treatment related to that particular domain (employment, attitude, drug use, etc.). Previous analyses suggest that the subscale structure of the LSI-R is not an accurate representation of how the items are grouping into factors. The results from three previous item level exploratory factor analyses of the LSI-R suggest that the factor structure of the LSI-R does not match the subscale structure. This dissertation provides a set of confirmatory factor analyses of the LSI-R using 3,493 LSI-R assessments collected from male offenders while they were on probation in a Midwestern county from 2002 to 2006. The initial confirmatory factor analysis of the LSI-R using the subscales as factors produced a Comparative Fit Index (CFI) of only .760, which suggests that the subscale structure does not provide a good fit to the item covariance structure of the LSI-R. After 29 modifications, a confirmatory factor analysis model was produced with a CFI of .950, which indicates a good fitting model. The 29th model had 19 factors and 11 items that loaded onto multiple factors. These results suggest that further analyses and discussions of the LSI-R item covariance structure are needed.
Edward Latessa, Ph.D. (Committee Chair)
Paula Smith, Ph.D. (Committee Member)
Keith Warren, Ph.D. (Committee Member)
John Wooldredge, Ph.D. (Committee Member)
272 p.
Psychological Tests
Offender; Risk Needs; Confirmatory Factor Analysis; LSI R; Psychometrics; Scale Development

Recommended Citations

Citations

  • Arnold, T. K. (2019). Confirmatory Factor Analyses of the Level of Service Inventory-Revised [Doctoral dissertation, University of Cincinnati]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1573224799095916

    APA Style (7th edition)

  • Arnold, Thomas. Confirmatory Factor Analyses of the Level of Service Inventory-Revised. 2019. University of Cincinnati, Doctoral dissertation. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=ucin1573224799095916.

    MLA Style (8th edition)

  • Arnold, Thomas. "Confirmatory Factor Analyses of the Level of Service Inventory-Revised." Doctoral dissertation, University of Cincinnati, 2019. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1573224799095916

    Chicago Manual of Style (17th edition)

ucin1573224799095916
526
© 2019, all rights reserved.
This open access ETD is published by University of Cincinnati and OhioLINK.