Hypotheses Testing. The first hypothesis is that co-morbid internalizing behavior problems will predict worse treatment outcomes. This hypothesis was tested using multiple regression analyses with CBCL externalizing behavior scores and self reported delinquency as the dependent measures. These analyses revealed a significant positive association between internalizing scores at pre-treatment and delinquency scores at post-treatment (β =.16, p=.04), controlling for age, gender, and delinquency scores at pre-treatment. It was found that internalizing scores at pre-treatment accounted for 2.3% of the variance in delinquency scores at post-treatment. No significant association was found between internalizing scores at pre-treatment and CBCL externalizing scores at post-treatment (β =.05, p=.50). The second hypothesis was that co-morbid internalizing behavior problems would be associated with greater stress during treatment. Regression analyses found a significant association between internalizing scores at pre-treatment and total caregiver stress (β =.24, p=.01), controlling for age and gender. The regression showed that internalizing scores at pre- treatment accounted for 5.1% of the variance in total caregiver stress. Regression analyses also found a significant association between internalizing scores at pre-treatment and youth-related caregiver stress (β =.21, p=.01), controlling for age and gender. Internalizing scores at pre- treatment accounted for 4.0% of the variance in youth-related caregiver stress.
Hypotheses Testing. In chapter three, we proposed three hypotheses. These were tested using statistical results from regression estimates. The hypotheses were: H1: The effect of relational duration on relational norms is dependent / contingent on the level of logistics performance. H2: The effect of relationship duration is negative on the relational norms when the logistic performance is low / modest. H3: The effect of relationship duration is positive on the relational norms when the logistic performance is high. The hypothesis is supported by the results of our regression model. The model is significant at F= 10.65 at p<0.01, however as we look in the equation given below. δRN ⁄ δLogRD = 5.524 – 1.670 (logRD) + 0.305 (LSP) = -1.670 + 0.305 (LSP) The value of coefficient of interaction effect is 0.305 which is obviously non zero, and the t- value = 3.475 at p<.01 which shows that the value of interaction effect is highly significant and depending upon the level of logistic service of the seller. Therefore, we can safely say that the effect of relationship duration on relational norms is contingent on the level of logistic performance which is our main hypothesis H1. Similarly, from the above graph of Figure 6.1, the value of interaction effect of relationship duration and relational norms is -1.670 (which is a negative value). So, when the logistic performance level is zero or less than 5.47, the value of interaction effect remains in the negative region of the graph. Thus, we can interpret from this result that at a low or modest level of logistic service performance the effect of relationship duration on relational norms is negative which proves our second hypothesis (H2) which states that the effect of relationship duration is negative on the relational norms when the logistic performance is low / modest. The same can be discussed by elaborating the results from the model / graph in the following matrix. Table 6-3: Extended Relationship Matrix When the relationship duration is short and the level of logistic performance is low / modest, we are in cell 3. This shows the modestly negative relationship between relationship duration and relational norms. We can land into cell 3 from cell 1 as well. In this situation it may be a case of adverse selection that the performance level of the seller deteriorated sharply but the buyer is still in relationship with this seller. This situation may be attributed to various reasons, for example, the buyer is new entrant into this marke...
Hypotheses Testing. The first hypothesis in this study was that higher levels of stress would be correlated with increased externalizing behaviors prior to treatment. A partial correlation controlling for ethnicity and SES did not yield significant correlations between the different types of stressors and externalizing behavior (see Table 3). Regression analyses testing for a curvilinear relationship between cortisol and externalizing behaviors at pretreatment also yielded nonsignificant results (F Change=.14). The second hypothesis was that dysregulation of the HPA axis, indicated by higher or lower levels of cortisol, would be correlated with externalizing behaviors before treatment. This hypothesis was not supported. A partial correlation controlling for ethnicity, SES, and time of cortisol collection did not yield significant results (see Table 3). Multiple regression analyses were used to test the third hypothesis that that HPA axis dysregulation would moderate the relationship between stress and externalizing behaviors at pretreatment. Both independent variables (cortisol and the stress measures) were centered and cortisol X stress interaction terms were computed. Ethnicity, SES, and time of cortisol collection were entered into Block 1, the centered cortisol and stress variables were entered into Block 2, and the interaction terms were entered into Block 3. Results from the analysis did not support this hypothesis (see Table 4). The fourth hypothesis was that stress and HPA axis dysregulation would predict treatment response, such that more self-reported stressors and HPA axis dysregulation at pretreatment would be related to less of a reduction in externalizing behavior following MST treatment. This hypothesis was supported for daily hassles and higher levels of cortisol. A partial correlation controlling for SES and pretreatment levels of externalizing behaviors revealed a significant positive relationship between daily hassles and externalizing behavior at post treatment (r=.30, p<.005; see Table 3). A second partial correlation controlling for SES, time of cortisol collection, and externalizing behavior at pretreatment revealed a positive relationship between cortisol levels and externalizing behaviors after treatment (r=.22, p<.05). Multiple regression analyses were used to test for the moderating role of cortisol on the relationship between stress and externalizing behaviors at post treatment. Both independent variables (cortisol and the stress measures) were cen...
