ILK Studies and Implications
| Empirical Study Sample | Validity-Related Results | Implications for Future ILK Revision |
|---|---|---|
| 120 competency examinees, setting unspecified8 | ILK correlations with Rey FIT = 0.67, RDS = 0.63, and TOMM-Trial 2 = 0.60 (all P < .01) | Significant convergence with other measures of effort or cognition support use of the ILK as a measure of response style |
| 70 jail detainees9 | ILK cutoff score of ≤47; PPP = 0.60 at 15% base rate, PPP = 0.74 at 25% base rate | Positive preliminary findings using well-defined and revised cutoff scores, as well as the R-ILK-90 and R-ILK-95 sub-measures |
| 195 college students5 | ILK/M-FAST correlation = –0.51 (P < .01); ILK score ≤ 47 had PPP = 0.92, NPP = 0.60, sensitivity = .35, specificity = 0.97 | Revised ILK cutoff score ≤ 40, 42 may balance concerns about false positives and negatives |
| 130 state hospital patients in competency proceedings5 | ILK/M-FAST correlation = –0.54 (P < .01); ILK score ≤ 47 had PPP = 0.80, NPP = 0.67, sensitivity = 0.57, specificity = 0.86 | ILK cutoff score from ≤ 35 to ≤ 42 may better balance concerns about false positives and negatives |
| 203 state hospital patients in competency proceedings1 | ILK/M-FAST correlation = –0.49 (P < .01); higher cutoff score (ILK ≤ 35, M-FAST ≥ 16), correlation (χ2 = 14.72, P < .001) | Feigning individuals often use more than one strategy, particularly in cases of very low ILK scores or very high M-FAST scores |
| 100 college students10 | ILK/M-FAST correlation = –0.55 (P < .001, Cohen's d = 1.48); ILK and CAST-MR correlations ranged from 0.80 to 0.90 | Feigning individuals often use more than one strategy; balance of sensitivity and specificity not highly improved at ILK cutoff score ≤ 37, or ≤ 42 |
| 65 state hospital patients in competency proceedings, 65 college students11 | ILK/M-FAST correlation = –0.52, (P < .01); Shipley-2 intelligence/ILK score correlation = 0.39 (P = .002) | Affirms manual cautions on ILK scores and IDD; further study needed on item discrimination and severely impaired groups |
| 32 adults with IDD, most had no active legal cases3 | ILK/M-FAST correlation = –0.34 (moderate effect size); ILK and IQ (KBIT-2) correlation was significant (P < .05) | Affirms manual cautions regarding ILK scores and IDD; nonsignificant convergent validity with TOMM |
| Dissertations: Each had a sample of 40–732,4,7 | Manual-comparable rates of sensitivity, specificity, and reported false positives | As before, affirmed manual cautions about ILK scores and < fifth-grade reading level, acculturation, effort, and lower cognition or IQ |
ILK, Inventory of Legal Knowledge; FIT, Rey 15-Item Test; RDS, Reliable Digit Span; TOMM, Test of Memory Malingering–Trial 2; M-FAST, Miller Forensic Assessment of Symptoms Test; PPP, positive predictive power; NPP, negative predictive power; CAST-MR, Competence Assessment for Standing Trial for Defendants with Mental Retardation; IDD, intellectual and developmental disabilities.