CHAPTER FIVE (part 4)

Table 21. Pearson Correlation Coefficients: Demographic variables and perpetrated partner abuse comparing Wave 1 and Wave 2 data based on the same sample of female respondents

 Note:  * p < .05, ** p < .01, *** p < .001

+ Z Scores were derived from the prevalence data and demographic measures measured in Wave 1 and Wave 2, respectively.

Asymptotic Standard Errors are provided in brackets.

Partner abuse was based on the full measure prior to any transformations being conducted

Life Stress Events and Perpetrated Partner Abuse

Pearson's Correlational analyses were conducted on Wave 2 data for males and females using weighted and unweighted stress scales.
     Overall, correlations between stress and perpetrated partner abuse were positive and were low to moderate in strength.  The strongest correlations were demonstrated by males in both weighted (r=.25) and unweighted (r=.21) scale formats.  By comparison, correlations for females were weaker in both weighted (r=.12) and unweighted formats (r=.10).  Table 22 summarizes the results of these analyses

Table 22. Pearson Correlation Coefficients: Stress experienced during the past two years and current perpetrated partner abuse among males and females based on Wave 2 data

Note: * p < .05, ** p < .01, *** p < .001

Partner abuse was based on the full measure prior to any transformations being conducted.

Exposure to Violence in the Family of Origin and Partner Abuse

Table 23 provides the results of Pearson's Correlations conducted on Wave 2 prevalence and incidence data on perpetrated partner abuse for males and females.  Each of the three exposure to violence in the family of origin variables were coded as follows: 1=observed violence and 0=did not observe violence.  When testing the relationship between the prevalence of perpetrated partner abuse and violence in the family of origin, results provided low to moderate positive relationships for all violence in the family of origin measures for both males and females.  The strongest relationships were found among males who observed father hitting mother (r=.18, p < .001) and parents' mutual violence (r=.17, p < .001).   For females, significant relationships were also found between the prevalence of perpetrated partner abuse and observing mother hitting father (r=.11, p < .05), father hitting mother (r=.12, p < .05), and parents' mutual violence (r=.14, p < .01).

The relationship between the incidence of perpetrated partner abuse and violence within the family provided moderate correlations for males, but weak correlations for females.   The strongest relationships were found among males who observed their fathers hitting their mothers (r=.27, p <.001) and who observed their parent's mutual violence (r=.27, p <.001).

Table 23. Pearson Correlation Coefficients: Violence in the family of origin and perpetrated partner abuse by male and female respondents

Note: * p < .05, ** p < .01, *** p < .001

Partner abuse was based on the full measure prior to any transformations being conducted.

Alcohol Consumption and Dependence and Perpetrated Partner Abuse

Tables 24 and 25 provide correlation coefficients for alcohol measures and perpetrated partner abuse comparing Wave 1 and Wave 2 data for same samples of males and females.  Alcohol measures assessed during Wave 1 were correlated with the prevalence of partner abuse reported in Wave 1 and Wave 2 as well as with the incidence of partner abuse reported in Wave 2.  Alcohol measures assessed during Wave 2 were correlated with the prevalence and incidence of partner abuse reported in Wave 2.  Overall, correlations for males were weak to moderate in strength.

Z scores were computed to assess the significance in proportions between correlations at Wave 1 and Wave 2.  The following are the alcohol variables found to differ with respect to the perpetration of partner abuse by males based on prevalence of abuse data:

1. The association between MAST scores and perpetrated partner abuse was significantly stronger in Wave 1.

2. The association between alcohol consumption and perpetrated partner abuse was significantly stronger in Wave 1.

3. The association between the Alcohol Dependence Index (ADI) and perpetrated partner abuse was significantly stronger in Wave 1.

Overall, Kendal's Tau-b values assessing the relationships between alcohol measures and the incidence of perpetrated partner abuse differed from those obtained by Pearson's Correlation coefficients.
     This suggests that a more conservative approach to evaluating the relationship between these measures among males may be needed.
     Bonferroni T tests (p=.05) conducted on the Wave 1 alcohol measures with respect to the incidence of perpetrated partner abuse by males indicated that ALC3R, MAST and SADD showed significant differences between groups.  No significant differences were found on the Wave 2 alcohol measures.  These findings suggest an increased likelihood of a Type 1 error in the association between partner abuse and alcohol consumption variables as measured in Wave 1.

Table 24. Pearson Correlation Coefficients: Alcohol consumption and dependence and perpetrated partner abuse comparing Wave 1 and Wave 2 data based on the same sample of male respondents

Note: * p < .05, ** p < .01, *** p < .001

+ Z Scores were derived from the prevalence data and alcohol measures for Wave 1 and Wave 2, respectively.

Partner abuse was based on the full measure prior to any transformations being conducted.

Compared to the male data, correlations between Wave 1 and Wave 2 alcohol measures and the prevalence and incidence of partner abuse reported in Wave 1 and Wave 2 were found to be weaker for females.
     However, for some of the measures, the strength of the correlations improved in Wave 2.  Z scores were computed to assess the significance in proportions between correlations at Wave 1 and Wave 2.  The following are the alcohol variables found to differ with respect to the perpetration of partner abuse by females based on prevalence of abuse data:

1. The association between alcohol consumption and perpetrated partner abuse was significantly stronger in Wave 2.

2. The association between SADD scores and perpetrated partner abuse was significantly stronger in Wave 2.

3. The association between ALC3R (lifetime diagnosis for alcoholism) scores and perpetrated partner abuse was stronger in Wave 2.

The Kendal's Tau-b values assessing the relationship between alcohol measures and the incidence of perpetrated partner abuse approximated those obtained by Pearson's Correlations coefficients.
     Bonferroni T tests (p=.05) conducted on Wave 1 and Wave 2 alcohol measures with respect to the incidence of perpetrated partner abuse by females did not show any significant differences in groups suggesting an increased likelihood of a Type 1 error in the associations between partner abuse and the alcohol measures.

Table 25. Pearson Correlation Coefficients: Alcohol consumption and dependence and perpetrated partner abuse comparing Wave 1 and Wave 2 data based on the same sample of female respondents

Note: * p < .05, ** p < .01, *** p < .001

+ Z Scores were derived from the prevalence of abuse data and alcohol measures from Wave 1 and Wave 2, respectively.

Partner abuse was based on full measure prior to any transformations being conducted.

Personality Measures and Perpetrated Partner Abuse

As in previous analyses, personality measures assessed during Wave 1 and Wave 2 were correlated with the prevalence of partner abuse reported in Wave 1 and Wave 2 as well as with the incidence of partner abuse reported in Wave 2. As in the previous sets of analyses, correlations were likewise found to be low to moderate in strength.
     Z scores testing differences in proportions between Wave 1 and Wave 2 correlations revealed a number of inconsistent relationships between the prevalence of perpetrated partner abuse and personality measures.  For males, they are as follows:

1. The relationship between EPQP scores and the perpetration of partner abuse was significantly stronger in Wave 2.

2. The relationship between MacAndrew Scale scores and the perpetration of partner abuse was significantly stronger in Wave 2.

3. The relationship between EPQL scores and the perpetration of partner abuse were significantly stronger in Wave 1.

4. The relationship between EPQN scores and the perpetration of partner abuse was significantly stronger in Wave 1.

With the exception of the relationships between EQPQ, EPQE and EPQL and the perpetration of partner abuse, the values obtained by Kendal's Tau-b (assessing the relationship between personality measures and the incidence of perpetrated partner abuse) and Pearson's Correlation coefficients were found to differ.  This suggests a more conservative approach to evaluating these personality measures among males.  Bonferroni T tests (p=.05) conducted on Wave 1 personality measures with respect to the incidence of perpetrated partner abuse by males indicated significant differences between groups for EPQL and EPQN.  However, analyses conducted on Wave 2 personality measures did not find any significant differences between groups (high v. low scores on measures).  These findings suggest an increased likelihood for a Type 1 error in the associations between partner abuse and EPQP (Wave 2), MacAndrew (Wave 2), Ego-strength (Wave 1 and Wave 2), Self Esteem (Wave 2), Trait Anxiety (Wave 1 and Wave 2) and EPQN (Wave 2). A summary of the correlation coefficients for personality and partner abuse by males respondents are reported in Table 26.

Table 26. Pearson Correlation Coefficients: Personality measures and perpetrated partner abuse comparing Wave 1 and Wave 2 data based on the same sample of male respondents

Note:  * p < .05, ** p < .01, *** p < .001

+ Z Scores were derived from the prevalence of partner abuse data and personality measures from Wave 1 and Wave 2, respectively.

Asymptotic Standard Errors are presented in brackets.

Partner abuse was based on the full measure prior to any transformations being conducted.

The following are the relationships between personality and partner abuse from Wave 1 and Wave 2 that were found to differ significantly based on Z scores computed for females:

1. The relationship between EPQP scores and the perpetration of partner abuse was significantly stronger in Wave 1.

2. The relationship between EPQL scores and the perpetration of partner abuse was significantly stronger in Wave 1.

3. The relationship between EPQE scores and the perpetration of partner abuse was significantly stronger in Wave 1.

4. The relationship between MacAndrew Scale scores and the perpetration of partner abuse was significantly stronger in Wave 1.

5. The relationship between Ego-strength Scale scores and the perpetration of partner abuse was significantly stronger in Wave 1.

6. The relationship between Self Esteem Scale scores and the perpetration of partner abuse was significantly stronger in Wave 1.

7. The relationship between Trait Anxiety Scale scores and the perpetration of partner abuse was significantly stronger in Wave 1.

8. The relationship between Neuroticism Index scores and the perpetration of partner abuse was significantly stronger in Wave 1.

The values obtained by Kendal's Tau-b assessing the relationship between personality measures and the incidence of perpetrated partner abuse approximated those obtained by Pearson's Correlation coefficients with the exception of the relationships between Trait Anxiety, Ego-strength and EPQP and the perpetration of partner abuse.
     Bonferroni T tests (p=.05) conducted on Wave 1 personality measures with respect to the incidence of perpetrated partner abuse by females indicated significant differences between groups for EPQN (high v. low scores).  Analyses conducted on Wave 2 personality measures likewise found significant differences between groups for EPQN as well as for EPQP and Ego-strength. These findings suggest an increased likelihood for Type 1 error in the associations between partner abuse and EPQP (Wave 1) and Trait Anxiety (Wave 1 and Wave 2).  A summary correlation coefficients for personality and partner abuse by female respondents are reported in Table 27.

Next: Chapter 5 Part 5