15581-m-nanninga

CHAPTER 3 52 Internal consistency The internal consistencies of the total scale and subscales of the parent and adolescent BTPS-exp were good, with the lowest Cronbach’s α of 0.85 found for the parent-subscale “perceived irrelevance of treatment” (Table 2). Examination of the consistencies without item 28, accidentally missing in the questionnaire at first, did not yield different findings, except for a slightly lower α for the adolescent-subscale “demands and issues” (α=0.90). Table 2 Internal consistency (Cronbach’s Alpha) of the total and subscale scores on the BTPS-exp parent and adolescent versions BTPS-exp scales Parent version Adolescent version Total barriers 0.95 0.96 Stressors and obstacles 0.92 0.93 Demands and issues 0.89 0.92 Perceived irrelevance 0.85 0.88 Problematic relationship with therapist 0.92 0.91 Scale structure Results on the model with the four subscales for the parent BTPS-exp showed a medium fit (CFI=0.94, TLI=0.94, RMSEA=0.07 [CI=0.07-0.07]). Two items had loadings <0.60 (item 1=0.55, item 36=0.47). Correlations coefficients between the four factors ranged from 0.52-0.79. A second-order one-factor solution showed a similar fit (CFI= 0.93, TLI= 0.93, RMSEA= 0.07 [CI= 0.07-0.08]). Results for the model with four subscales for the adolescent BTPS-exp also showed a medium fit (CFI=0.90, TLI=0.90, RMSEA=0.08 [0.07-0.08]), but slightly poorer compared to the parent BTPS-exp. One item had a loading <0.60 (item 1=0.55). Correlation coefficients between the four factors ranged from 0.57-0.77. Again, a second-order one- factor solution gave exactly the same results. Parent-adolescent agreement The parent and adolescent BTPS-exp correlated significantly (Table 3). Correlation coefficients were low, ranging from 0.16 (Pearson’s r ) for the subscale “stressors and obstacles” to 0.29 for “perceived irrelevance.”