TEXAS INMATE STUDY (PAGES 41-78) NAIL SCREW HAMMER SAW BUBBLE LEVELER   
      
   TEXAS INMATE STUDY (PAGES 41-78) KENNETH MARTIN DOLAN IS DIFFERENT THAN ALL >   
   64 TEXAS INMATES AND NEVER LEGITIMATELY THE SAME AS ALL > 64 TEXAS INMATES   
   THEREFORE KENNETH MARTIN DOLAN CAN NEVER BE LEGITIMATELY AN INMATE AS STATED   
   SINCE AT LEAST 2009    
      
   (PAGES 41 THROUGH 78)    
      
   = 0.22, p = .001. Follow-up univariate analyses, with a Bonferroni corrected   
   alpha level    
   of .01, indicated that the standard instructions group scored significantly   
   higher than the    
   simulating group on standard validity scales K (p < .001) and S (p < .001),   
   which    
   registered moderate effect sizes (ds = 1.00 and 0.96, respectively) in the   
   differentiation of    
   the groups. The remaining standard validity scales produced small to   
   negligible effect    
   sizes. See Table 2 for detailed results of these analyses.    
   For the second comparison involving the simulating and clinical comparison    
   groups, age and sentence length were held constant (covaried) and scores on   
   the standard    
   validity scales were entered as dependent variables. Wilks’ Λ was employed   
   as the    
   statistical criterion since homogeneity of variance was met, F(15, 34331) =   
   1.16, Box’s    
   M = 18.46, p > .05 (Tabachnick and Fiddell, 2001). Results indicated a   
   significant group    
   difference on the standard validity scales, F(5, 97) = 3.25, Wilks’ Λ =   
   0.86, p = .009.    
   Follow-up univariate analyses, with an alpha level of .01 using a Bonferroni   
   correction,    
   indicated that the clinical comparison group scored significantly higher than   
   the    
   simulating group on standard validity scale S (p = .001). A possible trend was   
   evidenced    
   as validity scale K approached significance, with the clinical comparison   
   group scoring    
   higher than the simulating group (p = .04). Nonetheless, these differences   
   were of small    
   to negligible magnitude, as indicated by their effect sizes. It should be   
   noted that the    
   relative lack of significant and clinically meaningful differences are not   
   unexpected,    
   given that the MMPI-2 protocols were screened and excluded based on extreme    
   elevations on validity scales, thereby imposing an artificial ceiling on   
   participants’    
   scores. See Table 2 for more detailed information about these analyses.    
   42    
   Group differences on the validity indicators    
   The first hypothesis predicted that the simulating group would produce    
   significantly higher mean scores on the validity indicators than the standard   
   instructions    
   group and the clinical comparison group. To evaluate the first hypothesis,   
   MANCOVA    
   procedures examined possible group differences between (1) the simulating   
   group and    
   standard instructions group and (2) the simulating group and clinical   
   comparison group    
   on the validity indicators (F, Fb, Fp, and Ds). Since these groups differed   
   with respect on    
   demographic variables, separate MANCOVA analyses were conducted for each    
   comparison. Cohen’s d (Cohen, 1992) effect sizes were computed using the   
   output from    
   univariate tests with the covariates controlled in the analyses. The effect   
   sizes were    
   interpreted according to Rogers and colleagues’ recommendation for   
   malingering studies    
   (Rogers et al., 2003) (d: > .75 = “moderate,” > 1.25 = “large,” and >   
   1.75 = “very large”).    
   For the first comparison, index offense and sentence length were held constant   
   (covaried)    
   and scores on the validity indicators were entered as dependent variables.   
   Pillai’s Trace    
   was employed as the statistical criterion since homogeneity of variance was   
   not met,    
   F(10, 31324) = 2.28, Box’s M = 24.08, p < .05 (Tabachnick & Fiddell, 2001).   
   Results    
   indicated that there was an omnibus group difference on the validity   
   indicators, F(4, 76)    
   = 10.69, Pillai’s Trace = 0.36, p < .001. Follow-up univariate analyses,   
   with a Bonferroni    
   corrected alpha level of .013, indicated that the simulating group scored   
   significantly    
   higher than the standard instructions group on validity indicators Fp (p <   
   .001), Ds (p <    
   .001), F (p < .001), and Fb (p < .001). Ds and F evidenced large effect sizes   
   (ds = 1.43    
   and 1.31, respectively) and Fp and Fb produced moderate effect sizes (ds =   
   1.18 and 1.02,    
   43    
   respectively) in the differentiation of the groups. See Table 3 for more   
   detailed results of    
   these analyses.    
   For the second comparison involving the simulating and clinical comparison    
   groups, age and sentence length were held constant (covariates) and scores on   
   the validity    
   indicators were entered as dependent variables. Pillai’s Trace was employed   
   as the    
   statistical criterion since homogeneity of variance was not met, F(10, 38355)   
   = 2.19,    
   Box’s M = 22.94, p < .05 (Tabachnick & Fiddell, 2001). There was an overall   
   group    
   difference on the validity indicators, F(4, 97) = 6.54, Pillai’s Trace =   
   0.21, p < .001.    
   Follow-up univariate analyses, with a Bonferroni corrected alpha level of   
   .013, indicated    
   that the simulating group scored significantly higher than the clinical   
   comparison group    
   on validity indicators Fp (p < .001), Ds (p < .001), and F (p = .002). Effect   
   sizes showed    
   that Fp and Ds evidenced moderate discrimination of malingered and genuine    
   psychopathology (ds = 0.83 and 0.81, respectively), whereas F and Fb produced   
   less than    
   moderate differentiation (ds = 0.63 and 0.24, respectively). Refer to Table 3   
   for more    
   information on these results. To summarize these results, the validity   
   indicators were    
   more effective in differentiating inmate simulators from healthy inmates   
   answering    
   honestly than in distinguishing inmate simulators from psychiatric inmates   
   when    
   demographic differences between the groups were held constant as covariates.    
   Nonetheless, Fp and Ds were particularly effective in the latter   
   differentiation of    
   malingered from genuine psychopathology. Overall, the results supported the   
   first    
   hypothesis, with the exception of the Fb scale, for the simulating and   
   clinical comparison    
   groups.    
   44    
   Group differences on the Basic Clinical Scales    
   MANCOVA procedures examined possible group differences between (1) the    
   simulating group and standard instructions group and (2) the simulating group   
   and    
   clinical comparison group on the Basic Clinical scales. Since the groups   
   differed with    
   respect todemographic variables, separate MANCOVA analyses were conducted for   
   each    
   comparison. Cohen’s d (Cohen, 1992) effect sizes were computed using the   
   output from    
   univariate tests with the covariates controlled in the analyses. The effect   
   sizes were    
      
   [continued in next message]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)