In order to test our hypotheses, individuals were classified as sensitive or not sensitive in each domain. ?Sensitive? was operationalized as reporting on average that the particular domain was important in defining oneself (i.e. greater than the midpoint on the importance items of the subscale), while not feeling competent in this domain (less or equal to the midpoint of the competence items of the subscale) . This grouping was performed separately for the four domains of interest, with the group sizes for sensitivity in the various domains as follows: Morality (N=29), Job Competence (N=68), Scholastic Competence (N=87) and Social Acceptability (N=49). Following this, ANOVA analyses were performed with the group membership as the independent variable, and the OC-related belief or OC symptom severity score as dependent variables. To test the hypothesis that individuals who were sensitive in the relevant domain would experience greater severity scores even after accounting for differences in global self-worth, ANCOVAs were performed with global self-worth entered as a covariate .
Analysis was performed using SPSS 12.0. To increase normality, square-root transformations of the Padua-Inventory total and subscales were undertaken. Preliminary screening led to the discovery of 5 cases of missing data, in which the participants did not complete all relevant questions. For these cases relevant questionnaire item scores were replaced by the mean item score. Homogeneity of regression was acceptable for all analyses except the relationship between global self-worth and impulses of harm, and therefore this ANCOVA was not performed.
Firstly, ANOVAs were run on total levels of OC-related beliefs, as a function of sensitive self-domains (see Table 6). The results indicated that sensitivity in all four self-domains was related to higher levels of OC-related beliefs. Furthermore, ANCOVAs indicated that morality and job-competence sensitivity were related to higher OC-related beliefs, over and above general self-worth (see Table 6).
Table 6. Summary of ANOVA/ANCOVA indicating means (SE),
and F-statistics for differences in OCD-related total cognitions,
as a function of whether the individual was sensitive/not sensitive in the four domains.
| Means (OBQ) | ||||
|---|---|---|---|---|
| Self-domain | Covariate | Not sensitive | Sensitive | F |
| Morality | 142.08 (2.93) | 164.76 (6.42) | 10.316* | |
| Self Worth | 143.36 (2.91)ª | 158.62 (6.66)ª | 4.243* | |
| Job-Competence | 138.85 (3.46) | 156.50 (4.19) | 10.552* | |
| Self Worth | 141.07 (3.48)ª | 153.24 (4.27)ª | 4.613* | |
| School-Competence | 137.98 (3.65) | 153.46 (3.73) | 8.305** | |
| Self Worth | 141.04 (3.92)ª | 150.62 (3.78)ª | 2.89 | |
| Social-Acceptability | 141.50 (3.21) | 156.07 (5.01) | 5.712* | |
| Self Worth | 144.13 (3.23)ª | 150.54 (5.22)ª | 1.015 | |
Note: OBQ = Revised Obsessive Beliefs Questionnaire.
a Means estimated at global self-worth=13.03.
* p<.05; ** p<.01; *** p<.001
ANOVA analyses were then conducted on total OC-severity scores, as a function of sensitivity in self-domains (see Table 7). The results indicated that sensitivity in the morality, job-competence and social acceptability domains was related to higher levels of OC-related symptoms. Furthermore, sensitivity regarding morality and job-competence was related to higher OC-related symptoms when controlling for general self-worth.
Table 7. Summary of ANOVA/ANCOVA indicating means (SE), and F-statistics
for differences
in OC severity scores as a function of whether the individual was sensitive/not sensitive in the four domains.
| Means (PI-R) | ||||
|---|---|---|---|---|
| Self-domain | Covariate | Not sensitive | Sensitive | F |
| Morality | 4.11 (.131) | 5.13 (.286) | 10.494** | |
| Self Worth | 4.16 (.131)ª | 4.90 (.299)ª | 4.979* | |
| Job-Competence | 3.97 (.144) | 4.74 (.168) | 10.280** | |
| Self Worth | 4.57 (.156)ª | 4.63 (.192)ª | 5.009* | |
| School-Competence | 4.04 (.175) | 4.51 (.168) | 3.751 | |
| Self Worth | 4.18 (.178)ª | 4.39 (.171)ª | 0.705 | |
| Social-Acceptability | 4.12 (.144) | 4.68 (.224) | 4.414* | |
| Self Worth | 4.22 (.145)ª | 4.46 (.235)ª | .751 | |
Note. PI-R = Padua Inventory Revised. Means of square root of Padua total.
a Means estimated at global self-worth=13.03.
* p<.05; ** p<.01; *** p<.001
To test the hypothesis that specific OC-related beliefs would differentially relate to sensitivity in different self-domains, ANOVAs were run with the three OC-related beliefs as dependent variables (see Table 8). Morality sensitivity was related to higher levels of all three OC-related beliefs. After controlling for global self-worth, sensitivity in the Morality domain was significantly related to responsibility/threat, and the importance/control of thoughts. Job and scholastic-competence was related to higher levels of responsibility/threat and to the importance of thoughts. After controlling for self-worth, job competence remained related to these two OC related beliefs, but scholastic competence remained related only to importance/control of thoughts. Social acceptability related to perfectionism/intolerance of ambiguity, but only prior to controlling for self-worth.
Table 8. Summary of ANOVA/ANCOVA indicating means (SE), and F-statistics for differences in OCD-related beliefs, as a function of whether the individual was sensitive/not sensitive in the four domains.
| Responsibility/Overestimation of Threat | Perfectionism/Intolerance of Ambiguity | Importance of/Control of Thoughts | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Self-domain | Covariate | not sensitive | sensitive | F | not sensitive | sensitive | F | not sensitive | sensitive | F |
| Morality | 52.30 (1.29) | 60.69 (2.84) | 7.221** | 58.49 (1.42) | 66.24 (3.10) | 5.173* | 31.29 (0.88) | 37.30 (1.92) | 9.563** | |
| Self Worth | 52.83 (1.30)ª | 5933 (2.10)ª | 4.096* | 59.19 (1.40)ª | 62.92 (3.20)a | 1.105 | 31.59 (0.88)ª | 36.37 (2.01)ª | 4.557* | |
| Job Competence |
50.90 (1.52) | 57.83 (1.85) | 8.635** | 57.83 (1.68) | 62.77 (2.03) | 3.512 | 30.12 (1.03) | 35.79 (1.25) | 12.329** | |
| Self Worth | 51.37 (1.56)ª | 57.25 (1.91)ª | 5.345* | 59.09 (1.67)ª | 60.93 (2.05)ª | .458 | 30.62 (1.04)ª | 35.06 (1.28)ª | 6.819* | |
| Scholastic Competence |
50.99 (1.71) | 56.32 (1.65) | 5.033* | 57.60 (1.87) | 61.91 (1.80) | 2.753 | 29.40 (1.14) | 35.23 (1.10) | 13.651**** | |
| Self Worth | 51.71 (1.77)ª | 55.64 (1.70)ª | 2.403 | 59.30 (1.88)ª | 60.32 (1.81)ª | .141 | 30.02 (1.17)ª | 34.65 (1.13)ª | 7.614** | |
| Social Acceptability |
52.47 (1.42) | 56.85 (2.21) | 2.774 | 57.75 (1.52) | 64.89 (2.38) | 6.403* | 31.63 (.97) | 34.33 (1.51) | 2.250 | |
| Self Worth | 53.05 (1.45)ª | 55.45 (2.35)ª | .707 | 58.82 (1.53)ª | 62.28 (2.48)ª | 1.316 | 32.26 (.98)ª | 32.80 (1.58)ª | .077 | |
a Means estimated at global self-worth=13.03.
* p<.05; ** p<.01; *** p<.001
Finally, separate ANOVAs were run with the four OC-symptom dimensions as dependent variables (see Table 4) . Morality was related to higher levels of all symptom dimensions. After controlling for self-worth, morality was related to checking compulsions, contamination obsessions/washing compulsions and obsessions of harm. Job-competence sensitivity was significantly related to higher levels of all symptom dimensions, except obsessional impulses of harm, even after adjusting for self-worth. Social acceptability related to both obsessional thought and obsessional impulses. This relationship did not remain when controlling for self worth. Scholastic-competence sensitivity was related to higher levels of obsessional impulses of harm. However, this relationship was not significant over and above self-worth.
Table 9. Summary of ANOVA/ANCOVA indicating means (SE), and F-statistics for differences
in OC-symptom dimensions, as a function of whether the individual was sensitive/not sensitive in the four domains.
| Checking Compulsions | Contamination/Washing | Obsessional Thoughts of Harm | Obsessional Impulses of Harm | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Self-domain | Covariate | NS | SENS | F statistic | NS | SENS | F statistic | NS | SENS | F statistic | NS | SENS | F statistic |
| Morality | 2.31 (.09) | 2.97 (.20) | 8.872** | 2.38 (.09) | 2.86 (.20) | 4.719* | 1.48 (.08) | 2.19 (.18) | 12.351** | .89 ( .08) | 1.56 (.18) | 10.757** | |
| Self Worth | 2.33 (.09)¹ | 2.83 (.21)¹ | 4.463* | 2.40 (.09)¹ | 2.79 (.21)¹ | 2.777* | 1.50 (.08)¹ | 2.07 (..19)¹ | 7.005** | .94 (.08)¹ | 1.34 (.19)¹ | 3.678 | |
| Job Competence |
2.18 (.11) | 2.78 (.13) | 12.389** | 2.37 (.11) | 2.61 (.13) | 1.988* | 1.36 (.10) | 1.95 (.12) | 14.603*** | .90 (.10) | 1.16 (.12) | 2.539 | |
| Self Worth | 2.22 (.11)¹ | 2.71 (.14)¹ | 12.389** | 2.40 (.11)¹ | 2.56 (.14)¹ | 1.988* | 1.40 (.10)¹ | 1.89 (.12)¹ | 14.603*** | 1.00 (.10)¹ | 1.02 (.12)¹ | 2.539 | |
| School Competence |
2.25 (.12) | 2.58 (.12) | 3.582 | 2.34 (.12) | 2.58 (.12) | 2.090 | 1.40 (.11) | 1.78 (.11) | 5.901* | .85 (.11) | 1.15 (.11) | 3.568 | |
| Self Worth | 2.33 (.13)¹ | 2.50 (.12)¹ | 0.910 | 2.38 (.13)¹ | 2.54 (.12)¹ | 0.794 | 1.47 (.12)¹ | 1.72 (.11)¹ | 2.240 | .97 (.11)¹ | 1.04 (.11)¹ | 0. .165 | |
| Social Acceptability |
2.32 (.10) | 2.66 (.16) | 3.247 | 2.43 (.10) | 2.57 (.16) | .573 | 1.48 (.09) | 1.89 (.14) | 5.700* | 0.90 (.09) | 1.26 (.14) | 4.299* | |
| Self Worth | 2.38 (.10)¹ | 2.53 (.17)¹ | 0. 531 | 2.46 (.10)¹ | 2.48 (.17)¹ | 0. .005 | 1.53 (.090)¹ | 1.78 (.15)¹ | 1.771 | - ² - ² |
- ² - ² |
- ² | |
Note: NS=Not Sensitive in Domain; SENS=Sensitive in Domain. Means based on square root of subscales.
¹ Means estimated at global self-worth=13.03.
² ANCOVA not performed due to violation of homogeneity of regression.
* p<.05; ** p<.01; *** p<.001