Skip to main content

Table 1 Comparison of Program Fidelity Tools in Criminal Justice

From: Assessing the reliability and validity of the Risk-Need-Responsivity (RNR) program tool

 

CPAI

CPC

SPEP

Domain

Reliability

Predicted Validity

Reliability

Predicted Validity

Reliability

Predicted Validity

Program Implementation

.49d

.56 + d

unavailable

.41

n/a

n/a

Client Preservice/Offender Assessment

.67d

.42 + d

unavailable

.42

n/a

n/a

Characteristics of Program

.43d

.52 + d

unavailable

.38

n/a

n/a

Characteristics/Practice of Staff

-.30d

.27 + d

unavailable

.55

n/a

n/a

Evaluation/Quality Assurance

.41d

.41 + d

unavailable

.16

n/a

n/a

Miscellaneous

-.01d

.16d

n/a

n/a

n/a

n/a

Primary/Supplement Service Typeb

n/a

n/a

n/a

n/a

unavailable

−.178

Amount of Serviceb

n/a

n/a

n/a

n/a

unavailable

−.186

Risk Level of Youthb

n/a

n/a

n/a

n/a

unavailable

−.42***

Quality of Serviceb

n/a

n/a

n/a

n/a

unavailable

–

Overall

.74

.60 + e

unavailable

.72**f

unavailable

−.36***g

Number of Itemsc

66

–

77/73

–

26

–

  1. Note: Cronbach’s alpha is used for reliability estimates. Pearson’s r is used for predictive validity estimates. CPAI estimates (Lowenkamp, 2004), CPC estimates (Latessa et al., 2010), and SPEP estimates (Redpath & Brandner, 2010)
  2. +p < .10, *p < .05, **p < .01, ***p < .001
  3. aCPC’s Leadership and Development domain is the combined Program Implementation and Miscellaneous domains of CPAI
  4. bSPEP-specific domains that are similar to CPAI and CPC but use vastly different measures to assess the domains
  5. cNumber of scoring items only
  6. dEstimate for significant items only (see Lowenkamp, 2004)
  7. ePredicting return to Ohio Correctional Facility for any reasons (technical violation or new arrest)
  8. fPredicting any new misdemeanor or felony conviction
  9. gPredicting whether a new complaint was recorded for either delinquency or status offenses