Caregiver Mental Health and Potentially Harmful Caregiving Behavior: The Central Role of Caregiver Anger

Gordon MacNeil, PhD; Jordan I. Kosberg, PhD; Daniel W. Durkin, MSW; W. Keith Dooley, PhD; Jamie DeCoster, PhD; Gail M. Williamson, PhD

Disclosures

Gerontologist. 2010;50(1):76-86. 

In This Article

Results

Correlational Analyses

Table 2 presents the correlations among anger, anxiety, depression, resentment, and PHB. All the predictor variables were positively correlated with one another, indicating the expected relationships between these emotional states. As anticipated, all predictor variables were significantly related to PHB.

Hypothesis 1: Mediation Analyses.—Using the guidelines specified by Baron and Kenny (1986), we tested whether the significant relations of anxiety, depression, and resentment with PHB were mediated by feelings of anger. Mediation can be tested by determining whether the relation between a predictor and the outcome is significantly reduced after controlling for the mediator. Sobel (1982) provides a test of whether the relation between the predictor and the outcome is significantly reduced after controlling for the mediator. If the relation between the predictor and the outcome is not significant after controlling for the mediator, it is referred to as full mediation. If there is still a relation between the predictor and the outcome after controlling for the mediator but the relation is significantly reduced, it is referred to as partial mediation. As seen in Table 3 , anger fully mediated the relation between anxiety and PHB (Hypothesis 1a), partially mediated the relation between depression and PHB (Hypothesis 1b) and partially mediated the relation between resentment and PHB (Hypothesis 1c).

Hypothesis 2: Moderation Analyses.—We next sought to determine whether anger moderated the relations of anxiety, depression, or resentment with PHB. We were interested in the extent to which the strength of the relation of anxiety, depression, and resentment with PHB vary across different levels of caregiver anger. To test for moderation, we used multiple regression to examine whether interactions between each mental health predictor variable and anger explained a significant proportion of variance in PHB scores. In order to avoid multicollinearity, we zero-centered anger and all predictor variables before computing interaction terms. More importantly for our interests, zero centering the predictor variables also changes the interpretation of the tests so that the noninteraction coefficients represent the main effects of the independent variables.

The results of moderation analyses are presented in Table 4 . Anger moderated the relation of depression (Hypothesis 2b) and resentment (Hypothesis 2c) with PHB. Our results indicate that anger did not moderate the relation of anxiety with PHB; thus, Hypothesis 2a was not supported. Our results indicate that anger did not moderate the relation of anxiety with PHB. The presence of anger intensifies the likelihood that caregivers with either high levels of depression or high levels of resentment will exhibit PHB. There was also a main effect of anger such that caregivers reporting more anger scored higher on PHB.

Hypothesis 3: Moderated Mediation Analysis.—The prior analyses demonstrated that anger mediates the relations of anxiety, depression, and resentment with PHB and that the relation between anger and PHB is moderated by depression and resentment. These findings indicate the presence of moderated mediation of anger by both depression and resentment (Preacher et al., 2007). As a reminder, moderated mediation means that the ability of a mediator to explain the relation between predictor and outcome depends on the level of the predictor (i.e., the moderator). There is no need for additional overall tests to determine the presence of moderated mediation because it is always present when the relation between a significant mediator and the outcome is moderated by another variable. Our focus therefore turns to understanding how the strength of the mediated effect changes. Preacher and colleagues provide a formula to assess the strength of a mediated effect across different levels of a moderating variable. We used this formula to explore how the ability of anger to mediate the relations between mental health predictors and PHB depends on the levels of those mental health predictors.

Figure 2 illustrates the conditional mediated effect of anger on PHB for different levels of anxiety. Although anxiety did not moderate the relation between anger and PHB, we wanted to illustrate the absence of conditional mediated effects for anxiety to act as a contrast for the results obtained for depression and resentment. Figure 2 shows that anger has a similar mediating effect across all values of anxiety, as noted by the horizontal nature of "anxiety"; thus, Hypothesis 3a was not supported. Figure 3 illustrates the conditional mediated effect of depression through anger for different levels of depression; Figure 4 illustrates the conditional mediated effect of resentment through anger for different levels of resentment. As shown in these figures, the mediating effect of anger substantially increases with higher scores on depression and resentment, respectively, thus supporting Hypotheses 3b and 3c.

Figure 2.

Conditional mediated effect for anxiety.

Figure 3.

Conditional mediated effect for depression.

Figure 4.

Conditional mediated effect for resentment.

Comments

3090D553-9492-4563-8681-AD288FA52ACE
Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as:

processing....