Tipultech logo

Results

Guy Doron

Analysis was performed using SPSS 12.0. Preliminary screening led to the discovery of 3 cases of missing data, in which the participants did not complete all relevant questionnaires. For these cases, relevant questionnaire item scores were replaced by the mean item score. No outliers with excessive influence were detected using Cook's Distances. Although residual plots showed moderate non-normality to which regression is robust, examination of residual plots did not suggest violations of homoscedasticity in total PI-R scores (Tabachnik & Fidell, 1996). For the individual PI-R subscales, square-root transformations were undertaken to increase normality. In order to decrease multicollinearity, all relevant world assumption scale (WAS) sub-scales were centered (i.e., put in deviation score form so that their means are zero) and interaction terms formed by multiplying together the two relevant centered predictors (Aiken & West, 1991).

The means and standard deviations (sd) for the scales are shown in Table 1. The mean PI-R total score in this sample was comparable to those observed in other Australian samples (Einstein & Menzies, 2004a; Scarrabelotti, Duck, & Dickerson, 1995). All scales, except randomness, exhibited adequate internal consistency in this sample (see Table 1). Thus, the randomness scale was excluded from further analysis.

Table 1. Means, Standard Deviations (SD) and Internal Consistency Alpha Statistics

Name of Scale Mean (SD) Cronbach's
    Alpha
PI-R 21.28 (16.14) .92
BDI 10.94 (8.07) .89
OBQT 146.18 (35.50) .92
WAS Benevolence World 2.94 (0.79) .82
WAS Justice 3.87 (1.05) .77
WAS Controllability 3.58 (.87) .74
WAS Randomness 3.24 (.95) .63
WAS Self Deservingness 2.72 (1.11) .84
WAS Self Controllability 2.72 (.80) .75
WAS Luck 2.61 (.97) .80

Note. PI-R-Padua Inventory Revised; OBQT - Obsessional Beliefs Questionnaire; BDI-Beck depression Inventory total; WAS - World Assumption Scale. In order to test our first hypothesis (i.e., whether a multi dimensional model of world

assumptions would predict overall OC severity scores), a sequential multivariate linear regression was undertaken with the mean PI-R total as the dependent variable. The regression was conducted according to the following sequence of entering blocks: First, the following world assumption scale (WAS) sub-scales were entered: benevolence of world (BW), justice (J), controllability(C), self-deservingness (SD) and self-controllability (SC) and luck (SL). Next, the two interaction terms (i.e., justice*self-deservingness; controllability*self-controllability) were entered.

Table 3 presents the summary of the regression statistics for this analysis. Results are presented with the Beta values of the final step. The overall model explained 33% of the variance (F (8, 159) = 9.67, p <0.001). The addition of the two interaction terms, namely, justice*self-deservingness and controllability*self-controllability, explained an additional 9% of the variance (Fchange (8, 159)= 10.54, p<0.001) over and above the main effects. The four statistically significant main effects were: benevolence of the world, justice, self-deservingness and self controllability. The two interactions terms were also statistically significant.

Table 2. Summary of Sequential Regression Analysis for
World Assumption Scale Variables Predicting Obsessive and Compulsive Symptoms.

Predictors PI-R B (SE ß) ß
Step 1
WAS BW -.14 (.04) -.27**
WAS J .08 (.03) .19*
WAS C .05 (.04) .10
WAS SD -.09 (.03) -.24**
WAS SC .09 (.04) .17*
WAS SL .01 (.03) .04
R² change .24  
F change 8.39**  
Step 2
J*SD .08 (.02) .24**
C*SC - .12(.04) -.23**
R² change .09  
F change 10.54**  
R² total .33  
F total 9.67**  

Note. PI-R = Padua Inventory-Revised; WAS = World Assumption Scale; BW= Benevolence of World; J=justice; C=controllability; SD=Self-deservingness; SC=Self controllability, SL=Luck. Negative beta signifies negative relationship between variables.
* p<.05 ; ** p<.01

As illustrated in Figure 2, the interactions patterns conformed to the expected pattern. Simple slope analysis (see Aiken & West, 1991) revealed that a decrease in self-deservingness scores led to a statistically significant increase (ß = .18, t(159)=4.63, p < 0.001) on OC severity scores only for participants scoring high (analyzed at 1 SD above the mean) on the justice scale. For participants scoring low on the justice scale (analyzed at 1 SD below the mean) the effect of self deservingness on OC severity scores was not significant (ß =0.01, t (159)=0.260, p=n.s.).

Interaction between justice and self-deservingness on PI-R scores

Figure 3 shows a similar pattern of results for the self-controllability and controllability interactions. As expected, increased self-controllability scores led to a significant increase (ß =-.19, t (159)= -3.62, p<0.01) in OC severity scores for participants scoring high on controllability, but not for participants scoring low on controllability (ß =-0.02, t (159)= 0.34, p=n.s.).

Interaction between controllability and self-controllability on OC severity scores

In order to ascertain that world assumptions are a related but not redundant construct with regards to OC related cognitions, a zero-order correlation analysis was conducted. Table 3 presents the results of this analysis.

OBQ
WAS subscales Total Threat/Resp Perfect/uncert Imp.Th
BW -.21* -.19* -.14 -.20**
Just .30** .32** .23** .18*
Contr .31** .28** .28** .17*
SD -.27** -.16* -.27** -.23**
SC -.11 -.15* -.10 .01
SL .12 .06 .07 .22**

Note: ** P<.01; * P<.05; WAS=World Assumption Scale; BW= benevolence of World; Just=Justice; Contr=Controllability; Rand=Random; SD=Self Deservingness; SC=Self Controllability; SL=SelfLuck; PI-R= Padua Inventory Revised; OBQ = Obsessive Beliefs Questionnaire; Threat/resp= overinflated threat/responsibility subscale; Perfect/uncert=perfectionism/uncertainty subscale; Imp Th=importance of/need to control thoughts subscale.

As can be seen in table 3, overall, zero-order correlations analysis suggests small to medium sizes in the expected directions. That is, negative perceptions of the world (i.e., social and physical) world and positive assumptions regarding the predictability of the world (i.e., justice and controllability) were associated with higher conviction levels in OC cognitions. Negative perceptions of self deservingness were also associated with such higher conviction levels. Unexpectedly, self controllability and benevolence of the physical world did not show significant relationships with most OC symptoms and cognitions.

In order to test whether a multidimensional model of world view would add to the prediction of OC symptoms over and above OC-related beliefs and depression symptoms, a sequential regression was undertaken with the mean PI-R total as the dependent variable. The regression was conducted according to the following sequence of entering blocks: First, the total score of the OBQ and the BDI total scores were entered. This was done in order to covary out OC-related beliefs and depression symptom scores when using a regression format (Tabachnik & Fidell, 1996). Next, the WAS variables were entered followed by the interaction terms.

Table 4 presents the summary of the regression statistics of this analysis. Results are presented based on the values of the final step. The overall model explained 54% of the variance (F(10, 157)= 18.11, p <0.001). The addition of world-view assumptions and interaction terms added 11% of the explained variance over and above OC related cognitions and depression scores. As can be seen in Table 4, benevolence of the world and the two interactions terms were statistically significant predictors of OC severity. The OBQ and BDI were also statistically significant predictors of OC severity.

Table 4. Summary of Sequential Regression Analysis for World Assumption Scale Variables controlling for OCD Relevant Beliefs and Depression Symptoms and Predicting Obsessive and Compulsive Symptoms.

Predictors PI-R B (SE ß) ß
Step 1
BDI .19 (.08) .18*
OBQ .23 (.03) .46**
R² change .43  
F change 62.53**  
Step 2
WAS BW -.09 (.03) .18**
WAS J .04 (.03) .11
WAS C -.03(.04) -.06
WAS SD -.01 (.03) -.02
WAS SC .04 (.04) .07
WAS SL .01 (.03) .02
R² change .04  
F change 2.01  
Step 3
J*SD .08 (.02) .22**
C*SC -.10 (.03) -.19**
R² change . 07  
F change 11.80**  
R² total .54  
F total 16.67**  

Note. PI-R = Padua Inventory-Revised total score; WAS = World Assumption Scale; BDIT = Beck Depression Inventory; OBQ = Obsessive Beliefs Questionnaire; BW= Benevolence of World; J=justice; C=controllability; SD=Self-deservingness; SC=Self controllability; SL=luck. Negative beta signifies negative relationship between variables.
** p<.01; * p<.05.

A simple slope analysis revealed the same patterns of interaction as in the previous model. That is, a decrease in self-deservingness scores led to a statistically significant increase (ß = .10, t(157)= 2.41, p<0.05) on OC severity scores only for participants scoring high on the justice scale, but not for participants scoring low on the justice scale (ß = -.07, t(157)= -1.96, p=n.s.). Similarly, increased self-controllability scores led to a significant increase (ß = -.12, t (157)= -2.72, p<0.01) in OC severity scores for participants scoring high on the controllability, but not for participants scoring low on controllability (ß = .06, t(157)= 1.23, p=n.s.). Finally, in order to explore whether different OC symptoms are predicted by different world assumptions when controlling for OC relevant cognitions and depression symptoms, four sequential regressions with the four PI-R subscales (i.e., contamination, checking, thoughts of harm and impulses of harm) were undertaken. The regressions were conducted according to the following sequence of entering blocks: first, the mean total OBQ and the mean BDI scores were entered. Next, the following world assumption scale (WAS) sub-scales were entered: benevolence of world (BW), justice (J), controllability(C), self-deservingness (SD), self-controllability (SC) and luck (SL). Finally, the two interaction terms (i.e., justice*self-deservingness; controllability*self-controllability) were entered. Table 5 presents a summary of the regression statistics of these analyses. Results are presented based on the values of the final step. All the regression models were significant at ß<0.01, from this table, addition of the world view model explained statistically significant additional variance in the prediction of all OC dimensions (except impulses of harm) over and above OC-related cognitions and depression scores . As predicted, the controllability and self-controllability interactions predicted both overt symptom dimensions. The justice and self-deservingness interaction predicted contamination and obsessions of harm.

More than 60% of the participants scored zero on the PI-R Dressing Rituals subscale. This subscale was excluded from further analysis. Zero order correlations between the WAS luck subscale and PI-R impulses of harm subscale were not significant. However, when entering the WAS luck subscale in step two, a significant relationship between this subscale and the PI-R impulses of harm was revealed (i.e., suppression effect). Thus, the WAS luck scores were omitted from this regression analysis.

Table 5. Summary of Sequential Regression Analysis for World Assumption Scale Variables controlling for OCD Relevant Beliefs and Depression Symptoms and Predicting Obsessive and Compulsive Symptom dimensions.

  Contamination Checking Obsessions of Harm Impulses of Harm
Predictors B (SE ß) ß B (SE ß) ß B (SE ß) ß B (SE ß) ß
Step 1
BDI .07 (.09) .07 .11 (.08) .12* .14 (.09) .14 .34(.09) .38**
OBQ .18 (.03) .41** .21 (.03) .47** .22 (.04) .46** .01(.03) .02
R² change .27   .34   .33   .23  
F change 29.87**   42.23**   41.38**   24.48**  
Step 2
WAS BW -.06 (.03) -.13 -.02 (.03) -.05 -.07(.04) -.15* -.03 (.03) -.13
WAS J .04 (.03) .11 .03 (.03) .08 .01 (.03) .05 .02 (.03) .05
WAS C .01(.04) .03 -.02 (.04) -.04 -.00 (.02) -.01 -.06 (.04) .15
WAS SD -.01(.03) -.03 -.01 (.03) -.03 -.01 (.03) -.04 .01 (.03) .02
WAS SC .04 (.04) .08 -.01 (.04) -.03 .02 (.04) .04 .03 (.03) .14
WAS SL .01 (.03) .11 -.02 (.03) -.04 .02 (.03) .05    
R² change .03   .02   .01   .04  
F change 1.14   0.70   0.54   1.64  
Step 3
J*SD .05 (.02) .16* . 04(.02) .12 .06 (.02) .19** .03 (.02) .09
C*SC -.08 (.03) -.18* -.06(.03) -.13* -.05 (.03) - .10 -.00 (.03) -.01
R² change .05   .03   0.04   .01  
F change 5.63**   3.25*   4.91**   0.92  
R² total .34   .38   0.39   .28  
F total 8.20**   10.74**   9.87**   6.66**  

Note. PI-R = Padua Inventory-Revised total score; WAS = World Assumption Scale; BDIT = Beck Depression Inventory; OBQ = Obsessive Beliefs Questionnaire; BW= Benevolence of World; BP= Benevolence of people; J=justice; C=controllability; SD=Self-deservingness; SC=Self controllability; SL=luck. Negative beta signifies negative relationship between variables.
** p<.01; * p<.05.

<< >>