| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
RESEARCH ARTICLE |
a Department of Psychology, Kent State University, Ohio
b Mandel School of Applied Social Sciences, Case Western Reserve University, Cleveland, Ohio
c Department of Psychiatry, University of Pittsburgh, Pennsylvania
d Department of Psychology, Cleveland State University, Ohio
Mary Ann Parris Stephens, Department of Psychology, 118 Kent Hall, Kent State University, Kent, OH 44242 E-mail: mstephen{at}kent.edu.
Toni C. Antonucci, PhD
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
PROVIDING care to a chronically ill or disabled older adult can have a variety of negative psychosocial consequences for family caregivers. Caregivers frequently experience higher levels of depressive symptoms when compared with noncaregivers and with population norms (Haley, Levine, Brown, Berry, and Hughes 1987
; Schulz, Visintainer, and Williamson 1990). Caregivers have also been shown to be more restricted than noncaregivers in their ability to participate in leisure and social activities (Gwyther and George 1986; Haley, Levine, Brown, Berry, et al. 1987). These psychosocial consequences often result from the stress engendered by the instrumental tasks of caregiving (e.g., providing help with daily activities) and by the problematic behaviors of the impaired relative (e.g., criticizing and complaining). Caregivers who appraise these caregiving demands as more stressful tend to have higher levels of depressive symptomatology (Haley, Levine, Brown, and Bartolucci 1987; Kinney, Stephens, Franks, and Norris 1995; Schulz and Williamson 1991) and greater restrictions in their leisure activities (Kinney et al. 1995; Miller and Montgomery 1990).
Caregivers often experience conflict arising from their attempts to balance the demands of the caregiver role and the demands of other social roles (e.g., Aneshensel, Pearlin, Mullan, Zarit, and Whitlach 1995; Barling, MacEwen, Kelloway, and Higginbottom 1994; Brody 1990). Interrole conflict may be especially relevant for adult daughters who assume the role of caregiver to an impaired parent or parent-in-law. Daughters who take on primary responsibility for parent care are often referred to as women in the middle (Brody 1981). This term alludes to the multiple roles of these women; they frequently occupy several major roles, such as mother, wife, and employee, each of which can place additional demands on them (Stone, Cafferata, and Sangl 1987). In the present study we examined conflict occurring between parent care and these three other roles frequently occupied by adult daughter caregivers.
The competing demands perspective evolved to explain the negative impact that caregiving often has on caregivers' psychological well-being (Stephens and Franks 1999). According to this perspective, multiple role responsibilities create demands on these women and compete for the women's time and energy. The competing demands hypothesis rests on the assumption that individuals have limited personal resources and that roles and role partners demand all of these resources (Goode 1960). Thus, an individual's total role obligations are thought to be overly demanding, making interrole conflict the norm (Goode 1960).
Interrole conflict has been conceptualized as the extent to which pressures within one role are incompatible with the pressures that arise within another role (Kahn, Wolfe, Quinn, Snoek, and Rosenthal 1964). Such conflict may result when the requirements and responsibilities of two roles compete for limited resources of the role occupant (Goode 1960; Kahn et al. 1964; Kopelman, Greenhaus, and Connolly 1983). In addition, interrole conflict may occur when experiences in one or more roles are especially stressful (Kopelman et al. 1983). Consistent with this perspective, Pearlin and colleagues have argued that stress experienced in the caregiver role can proliferate (spill over) into other life domains and social roles (Aneshensel et al. 1995; Pearlin, Mullan, Semple, and Skaff 1990). As with interrole conflict, it is assumed that this proliferation process occurs because the stress encountered in caregiving limits the amount of time and energy that is available for other roles and role partners.
In the literature on later life caregiving, most research on interrole conflict has investigated conflict between caregiving and two other roles: employee or spouse. Because many caregivers who provide care to their aging parents (hereafter referred to as adult-child caregivers) occupy one or both of these other roles (Stone et al. 1987), research on interrole conflict has often focused on this group of family caregivers. Some studies have shown that the demands of the caregiver role and the employee role frequently interfere with each other (Aneshensel et al. 1995; Barling et al. 1994; Gignac, Kelloway, and Gottlieb 1996; Gottlieb, Kelloway, and Fraboni 1994; Neal, Chapman, Ingersoll-Dayton, and Emlen 1993; Scharlach 1994; Stephens, Franks, and Atienza 1997). Other research has shown that conflict can also occur between the caregiver and spouse roles (Barling et al. 1994; Kleban, Brody, Schoonover, and Hoffman 1989; Stephens and Franks 1995; Walker, Pratt, and Wood 1993). Prior research has also demonstrated that adult daughters who feel greater tensions between the demands of their parent care role and the demands of either their employee or marital role experience higher levels of depressive symptoms (Stephens and Franks 1995; Stephens et al. 1997). This existing research on later life caregiving has been based on the assumption that a woman's role as caregiver will necessarily conflict with her other roles, but this assumption has received little empirical attention. It is possible, therefore, that some women may experience no conflict between the responsibilities of parent care and the responsibilities of their other roles.
Empirical research on the intersection of work and family roles has demonstrated that one way stress in a given role (e.g., employment or family roles) exerts its effects on psychosocial well-being is indirectly through the incompatible pressures of two roles, most often conceptualized as the extent to which one role limits the time and energy available for the other role (Frone, Russell, and Cooper 1992; Kopelman et al. 1983). These findings suggest that stressful experiences in one role are likely to tax the individual's time and energy and thereby create conflict between that role and another role. This resulting conflict between roles can predispose an individual to psychological distress. These studies of interrole conflict, however, did not include individuals who occupied the role of caregiver to an impaired older family member.
Some studies of later life caregiving have examined interrole conflict as an intervening mechanism in the caregiving stress process. One study demonstrated that conflict occurring between adult daughters' roles as caregiver and employee functioned as a mediator in the relationship between parent care stress and depression (Stephens et al. 1997). Another study found that, for employed caregivers, caregiving stress was indirectly related to work-relevant outcomes (e.g., absenteeism) through the conflict occurring simultaneously between the caregiver role and several other roles, including those of parent, spouse, and employee (Barling et al. 1994).
These prior studies of interrole conflict as an intervening mechanism in the caregiving stress process are limited in several ways. By focusing exclusively on conflict occurring between parent care and the employee role, Stephens and colleagues 1997 did not address the possibility that some women may also have experienced conflict between parent care and their other roles (e.g., mother or wife). Moreover, that study offers no insight about whether conflict between parent care and these other roles also may have mediated the caregiving stress and well-being relationship. Barling and colleagues 1994 assessed the interrole conflict occurring simultaneously between the caregiver role and an aggregate of several other roles, but they did not examine psychosocial well-being as an outcome of caregiving stress or interrole conflict.
Our overall objective in the present study was to investigate conflict occurring between the parent care role and three other roles of women who were providing care to an impaired parent or parent-in-law and who were also mothers to children living at home, wives, and employees. As part of the first two aims of our study, we investigated patterns of conflicting roles that involve the parent care role. In contrast to prior research, we allowed for the possibility that some adult daughter caregivers may experience no conflict between the parent care role and their other roles as mother, wife, and employee. We also considered the possibility that for some women, the parent care role might conflict more with one of their roles than with the other two.
Our first aim in the study was to identify factors that differentiated women whose pattern of conflicting roles either did or did not include the parent care role. Specifically, we addressed the question of whether these two groups of women differed on sociodemographic characteristics; on characteristics of their impaired parent, children, and husband; or on characteristics of their employment. Our second aim was to identify factors that differentiated women whose pattern of conflicting roles included the parent care role and the mother, wife, or employee role. Comparisons were made among three groups of women (those whose parent care role conflicted most with their mother role, their wife role, or their employee role). This comparison addressed the question of whether these three groups of women differed on sociodemographic characteristics; on characteristics of their impaired parent, children, and husband; or on characteristics of their employment.
Our third aim in the study was to examine a theoretical model of the caregiving stress process (Fig. 1) for those women who reported experiencing the most conflict between their parent care role and one of their other roles. In this model, interrole conflict was conceptualized as incompatible pressures between the requirements and responsibilities of the parent care role and those of one other role. Specifically, it reflected the extent to which the demands of these two roles taxed a woman's time and energy.
|
; Kinney et al. 1995; Miller and Montgomery 1990; Schulz and Williamson 1991). Depressive symptomatology and leisure activity restriction were chosen as indicators of psychosocial well-being because they represent common consequences of caregiving stress (Gwyther and George 1986; Haley, Levine, Brown, Berry et al. 1987; Schulz et al. 1990). We predicted that higher levels of interrole conflict would serve as a mediating mechanism in the relationship between both dimensions of parent care stress (instrumental and behavioral) and both indicators of psychosocial well-being (depressive symptomatology and leisure activity restriction). |
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
). At the first wave of this study, all respondents occupied four roles: primary caregiver to an impaired parent or parent-in-law (both hereafter referred to as parent), wife, mother to at least one child living at home, and employee. Primary caregiver refers to the family member who spends the most time caring for the parent. To qualify as a caregiver, a woman had to assist the parent with at least one personal or instrumental activity of daily living (e.g., bathing, meal preparation) or provide supervision to the parent. In addition, respondents were required to be caring for a parent who lived in the community but in a separate household from the respondent. The decision to select respondents who initially did not share a residence with the parent was based on research indicating that adult-child caregivers who live with the impaired parent differ in well-being from those who live in separate households (e.g., Deimling, Bass, Townsend, and Noelker 1989).
Additional requirements for eligibility were that the respondents be currently married and living with their husband and have at least one child aged 25 years or younger living at home (and no children aged more than 25 in the household). We selected this upper age limit of 25 years for children to account for an increasing tendency for young adults to remain at home for longer periods of time (U.S. Bureau of the Census 1992b). Finally, respondents had to be employed either full time or part time. Respondents had to have occupied each of these four roles for at least 2 months prior to the interview.
Respondents were recruited from northeastern Ohio and two counties in Pennsylvania bordering Ohio. A variety of sources was used for recruiting respondents, including newspaper articles, newspaper and radio advertisements, and notices in newsletters published by businesses and social organizations (see Stephens and Townsend 1997, for details on recruitment). A total of 898 women were screened for eligibility, and 296 (33%) were eligible. The three most common reasons for ineligibility included not providing care to a parent or parent-in-law (22% of all ineligible women), not being the primary caregiver (15%), and having children aged more than 25 living at home (16%). Although 296 women comprised the final sample for the larger study, 11 (4%) women reported that they did not experience any conflict among their 4 roles. An additional four respondents (1%) were excluded because they did not report household income (which was used as a control variable), and 3 other respondents (1%) were excluded because they indicated that none of the behavioral stressors (a key construct in our theoretical model) were applicable to their situation. Eliminating these 18 women from all analyses reported in this article resulted in a total sample of 278 women.
Structured, in-person interviews were conducted in the respondents' homes or other places of their choosing. The interviews typically took 1.5 h to administer. Respondents received a nominal fee for their participation in the study.
Most (89%) of the sample was Caucasian; the remainder (11%) was African American, which approximates the proportion of African American women in the region where the study was conducted (13%; U.S. Bureau of the Census 1992a). The average age of the respondents was 43.6 years (SD = 6.1; range = 25–60). They had an average of 14.6 years of education (SD = 1.9; range = 10–17) and had been married for an average of 18.2 years (SD = 8.4; range = 1–35). On average, respondents reported having two children living at home (SD = 1.1; range = 1–9). The average age of all children living at home was 14.1 years (SD = 6.2; range = 1–25).
In terms of their relationship to the care recipient, 68% of respondents were providing care to a mother, 19% to a father, 10% to a mother-in-law, and 3% to a father-in-law. Respondents had been providing care to the parent for an average of 6.3 years (SD = 6.1; range = 2 months to 37.0 years). The average age of the impaired parent was 76.1 years (SD = 7.4; range = 50–94). More than half (60%) of the respondents rated the physical health of the parent as either "fair" or "poor."
Of seven personal activities of daily living, the average number with which parents required help was 1.9 (SD = 2.0, range = 0–7). Of eight instrumental activities of daily living, the average number with which parents required help was 6.1 (SD = 1.7, range = 1–8). Respondents reported that of 11 possible memory and behavior problems, their parent experienced an average of 5.1 (SD = 2.7, range = 1–11). More than three fourths of the parents (77%) reportedly needed supervision at least some of the time. The respondents spent an average of 2.8 h assisting with caregiving tasks on a typical weekday (SD = 2.9; range = 0–20) and 3.4 h on a typical weekend day (SD = 3.1; range = 0–24).
Slightly less than half of the women were in professional or managerial occupations (46%); a similar proportion were in technical, sales, or administrative support occupations (44%); and the remainder were in precision production or service occupations (10%). The average number of hours that respondents worked per week was 36.7 (SD = 12.1; range = 8–72), and the average number of weeks respondents worked during the previous year was 48.9 (SD = 7.8; range = 8–52). On average, respondents reported that their annual household income was between $40,000 and $59,999. Respondents earned, on average, between $15,000 and $24,999 from their own employment. The average household income in our sample was comparable to the median family income ($51,000) of married couples with the wife in the paid labor force (U.S. Bureau of the Census 1995).
Instruments
Tests of the hypothesis that interrole conflict involving parent care would mediate the relationship between parent care stress and psychosocial well-being included only those women who indicated that their parent care role conflicted most with one of their other roles. Therefore, the descriptive information on all measures presented in the following section is based on this subsample of women who were included in the tests of this hypothesis (n = 171). The internal consistency coefficients (Cronbach's alpha) for each of the measures based on this subsample of women did not differ appreciably from those based on the total sample of women who experienced role conflict of any kind (N = 278).
Parent care stress.—Two dimensions of stress in the parent care role were assessed: the stress of providing assistance with the parent's instrumental activities of daily living (instrumental stress) and the stress of dealing with the parent's behavior problems (behavioral stress). Because many parents (36%) did not require assistance with any personal care activities, this dimension of caregiving stress was not included in the model. The parent care stress items were adapted from other research on the caregiver role (Albert 1991; Kinney and Stephens 1989; Vitaliano, Russo, Young, Becker, and Maiuro 1991; Zarit and Zarit 1983). For both stress measures (available from the first author), respondents indicated how stressful experiences within the parent care role had been during the past 2 months, using a scale ranging from "not at all" to "very much." Higher scores indicate more stress.
The parent care stress scores represent the average amount of stress experienced. We calculated both scores by summing the ratings across items and dividing by the number of items endorsed as being applicable to the respondent's situation (Stephens and Townsend 1997). We calculated these scores in this way to create measures that excluded events that a woman had not experienced in the past 2 months, that she had never experienced, or that were irrelevant to her situation. We were concerned only with appraisals of events that had occurred. The potential range for both stress measures was 1 (low stress) to 4 (high stress).
= .79). Respondents were asked how stressful it had been to help their parent with each of nine instrumental activities of daily living (transportation, supervision, personal finances, health care decisions, shopping, housework, meal preparation, laundry, and errands) in the past 2 months. The average number of activities with which respondents reported assisting their parent was 7.4 (SD =1.5; range = 3–9). The average level of instrumental stress was 2.2 (SD = .60; range = 1–3.8). = .74). Respondents were asked how stressful it had been to deal with each of six behavior problems of the parent (parents' emotional problems, memory problems, cognitive problems, communication problems, endangering self or others, and agitation) in the past 2 months. The average number of problems experienced was 4.6 (SD = 1.4; range = 2–6). The average level of behavioral stress was 2.6 (SD = .63; range = 1–4).
Patterns of conflicting roles
We used two items to assess a respondent's pattern of conflicting roles. Respondents were asked whether any of their four roles had conflicted during the past 2 months. Those respondents who indicated that they had experienced conflict were then asked to identify the two roles that had conflicted with one another the most.
Interrole conflict ( = .74)
Respondents experiencing conflict between any combination of roles were presented with seven items describing ways in which these two roles might have conflicted with one another. We asked respondents to indicate the extent to which they agreed (or disagreed) with each item on a 5-point scale (1 = "strongly disagree" to 5 = "strongly agree"). Items were derived from theory and empirical research on interrole conflict (Frone et al. 1992; Goode 1960; Kopelman et al. 1983; Small and Riley 1990). Data on these seven items were submitted to a principal components factor analysis with varimax rotation. Although two factors emerged (the first factor accounted for 42% of the variance, and the second accounted for an additional 16%), only the first factor was retained for further analysis because of its stronger psychometric properties.
The four items that had factor loadings greater than .65 on the first factor and less than .30 on the second factor were included in the measure of interrole conflict. Items that made up this factor reflected the extent to which the two roles taxed the respondents' time and energy. The four items were (a) not enough time to do everything, (b) not enough energy to do everything, (c) these two roles are emotionally draining, and (d) both of these roles pose considerable demands and responsibilities. The potential range of scores was from 4 to 20, with higher scores indicating greater interrole conflict. Interrole conflict scores ranged from 8 to 20, with a mean level of 17.5 (SD = 2.2).
Well-being
We used two measures of well-being.
= .90). We used the Center for Epidemiologic Studies-Depression Scale (CES-D; Radloff 1977) to assess depressive symptoms. The CES-D is a 20-item scale that asks respondents to indicate how frequently they experienced certain symptoms or feelings during the past week. Scores can range from 0 to 60, with higher scores reflecting greater symptomatology. A score of 16 or higher on the CES-D is indicative of a risk for clinical depression (Radloff 1977). Approximately 33% of the present sample scored in this range, as compared with 23.5% of women from the general population (Clark, Aneshensel, Frerichs, and Morgan 1981). The average depressive symptoms score was 13.2 (SD = 10.2; range = 0–54), as compared with a mean of 10.4 for women in the general population (Clark et al. 1981). = .86). Respondents were asked to indicate the extent to which they agreed (or disagreed) with eight items that stated that they had less time than they would have liked in the past year to engage in selected leisure activities. The activities were (a) socializing with friends, relatives, or neighbors; (b) participating in group or organized activities; (c) going out to dinner, the theater, a concert, or a show; (d) traveling for pleasure; (e) participating in crafts or hobbies; (f) watching television, listening to music, or reading; (g) exercising or participating in sports (including going for walks); and (h) just relaxing or doing nothing. Items were adapted from prior research on the social impact of caregiving (Poulshock and Deimling 1984). Respondents reported how strongly they agreed or disagreed with each statement using a 5-point scale (1 = "strongly disagree" to 5 = "strongly agree"). The potential range of scores was from 8 to 40, and higher scores indicate greater restriction in leisure activities. The average level of activity restriction was 30.4 (SD = 6.4, range = 15–40).
Analysis Plan
To address the first aim of the study, we compared the 171 respondents whose pattern of conflicting roles included the parent care role (62% of the sample) to the 107 respondents whose pattern of conflicting roles did not include the parent care role. We used a series of analyses of variance (ANOVAs) to compare these two groups (parent care conflict and no parent care conflict). These groups were compared on the sociodemographic characteristics of the respondents (age, education, race), their impaired parent (age, functional impairment, memory and behavior problems), children (age, functional needs, number of children living at home) and husband (age, physical health, number of years married) and on characteristics of their employment (professional versus nonprofessional occupation, hours worked per week, weeks worked per year, household and personal income).
All remaining analyses are based on only the 171 respondents whose pattern of conflicting roles included the parent care role. To address the second aim of the study, we used another series of ANOVAs to compare three groups. Group 1 was composed of respondents who reported that the parent care role conflicted most with the mother role (n = 58, 34% of the subsample); Group 2 included respondents whose parent care role conflicted most with the wife role (n = 48, 28%); and Group 3 included respondents whose parent care role conflicted most with the employee role (n = 65, 38%). These three groups were compared on sociodemographic characteristics of the respondents; characteristics of their impaired parent, children, and husband; and characteristics of their employment.
To determine whether data from these three groups could be aggregated to test the path models specified in the third aim of the study, we used ANOVAs to compare these groups on all study measures (i.e., mean levels of parent care stress, interrole conflict, and psychosocial well-being). Because the groups did not differ significantly on any of these variables (with p 
.05 as the criterion), the groups were aggregated.
To address the third aim of the study, we used ordinary least squares hierarchical regression to test the mediating effects of interrole conflict in the association between two dimensions of parent care stress (instrumental and behavioral) and two indicators of psychosocial well-being (depressive symptomatology or leisure activity restriction). A separate path analysis was conducted for each index of well-being. In testing for mediation, the procedures set forth by Baron and Kenny 1986 were followed.
To determine whether any control variables should be used in the path analyses, we examined bivariate correlation coefficients between 24 demographic and caregiving-related variables (e.g., age, race, income, number of children at home, parent's age) and the two indicators of well-being. Because of the size of the sample, many of these coefficients were significant (p .05). Therefore, a criterion of r 
.20 was used to select a variable for inclusion as a control variable. For depressive symptoms, respondent's age (r = -.20) and household income (r = -.26) met this criterion, and for leisure activity restriction, respondent's age (r = -.25) met this criterion. For consistency, age and household income were used as control variables in both path analyses. Table 1 presents the bivariate correlations between all study and control variables for the 171 respondents who reported conflict between parent care and another role.
|
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
.05, df = [1,277]), had lower household (F = 11.88, p .001, df = [1,277]) and personal income (F = 18.96, p .001, df = [1,277]), were less likely to be in professional or managerial occupations (F = 8.62, p .01, df = [1,277]), had children who were older (F = 3.90, p .05, df = [1,277]), and had been married longer (F = 4.31, p .05, df = [1,277]) than their counterparts with no parent care conflict. In addition, the parents of the respondents with parent care conflict were more functionally impaired (i.e., needed more assistance with activities of daily living; F = 9.83, p < .01, df = [1,277]) and exhibited more memory and behavior problems (F = 6.49, p .05, df = [1,277]). These groups did not differ significantly on respondents' age or race, parents' age, children's functional needs, number of children at home, husbands' age or physical health, hours worked per week, or weeks worked per year.
|
.001, df = [2,170]) and education (F = 4.43, p .05, df = [2,170]), their children's age (F = 10.58, p .001, df = [2,170]) and functional needs (i.e., the extent to which the children needed assistance with daily activities such as dressing and transportation; F = 9.76, p .001, df = [2,170]), their husbands' age (F = 6.04, p .01, df = [2,170]), and the number of years they had been married (F = 6.36, p .001, df = [2,170]). Post hoc analyses (Tukey's honestly significant difference test) revealed significant differences between respondents whose parent care role conflicted most with the mother role and those whose parent care role conflicted most with the employee role. Respondents reporting conflict between the parent care role and their mother role, on average, were younger, had less education, had younger children with greater needs for assistance, and had husbands who were younger. Moreover, those respondents whose parent care role conflicted most with either of their other family roles (mother or wife) had been married for fewer years than respondents who reported that the parent care role conflicted most with employment. The three groups did not differ significantly on respondents' race; parents' age, functional impairment, or memory and behavior problems; number of children at home; husbands' physical health; or on any characteristics of employment.
|
). Results of the mediation analyses are presented in Table 4 and Table 5 and in Fig. 1.
|
|
After controlling for respondents' age and household income, we found that higher levels of behavioral stress were significantly related to more depressive symptoms (ß = .19, p .05) but that instrumental stress was not. An opposite pattern of findings emerged for leisure activity restriction. After age and income had been considered, higher levels of instrumental stress (ß = .28, p .001), but not behavioral stress, were related to more restriction in leisure activities.
Table 5 displays the beta coefficients, standard errors, and explained variance for the two path analyses examining interrole conflict as a potential mediator. To test the second requirement of mediation, we regressed interrole conflict on the two indicators of parent care stress. These findings are presented in the top portion of Table 5 . Both behavioral stress (ß = .16, p .05) and instrumental stress (ß = .24, p .001) were significantly related to interrole conflict, after age and income had been controlled.
Findings from the path analysis examining depressive symptoms are presented in the middle portion of Table 5 . To test the third and fourth requirements of mediation, we regressed depressive symptoms on interrole conflict and the two indicators of parent care stress, controlling for age and income. Results indicated that greater interrole conflict was significantly related to greater depressive symptoms (ß = .22, p .01).
The resulting path coefficient between behavioral stress and depression was reduced slightly in magnitude but not to nonsignificance (from ß = .19 in Table 4 to ß = .16 in Table 5 ). Moreover, the indirect effect of behavioral stress on depression via interrole conflict was significant. These findings indicate that interrole conflict partially mediated the relationship between behavioral stress and depressive symptoms. Because the first requirement of mediation was not met for instrumental stress in the analysis of depressive symptoms (Table 4 ), results for this predictor in the path analysis of depression are not reviewed further. Adding interrole conflict to the model significantly increased the explained variance in depressive symptoms (from R2 = .16 in Table 4 to R2 = .20 in Table 5 , p < .01).
Findings pertaining to the path analysis of leisure activity restriction are presented in the lower portion of Table 5 . Interrole conflict was strongly related to leisure activity restriction, after we controlled for age, income, and the two indicators of parent care stress (ß = .41, p .001). The magnitude of the path coefficient between instrumental stress and leisure activity restriction was reduced substantially but not to nonsignificance when interrole conflict, age, and income were controlled (from ß = .28 in Table 4 to ß = .18 in Table 5 ). The indirect effect of instrumental stress on leisure activity restriction via interrole conflict was significant. These findings indicate that interrole conflict partially mediated the relationship between instrumental stress and leisure activity restriction. Because the first requirement of mediation was not met for behavioral stress in the analysis of leisure activity restriction (Table 4 ), the results for this predictor in the path analysis of leisure activity restriction are not reviewed further. Adding interrole conflict to the model significantly increased the explained variance in leisure activity restriction (from R2 = .19 in Table 4 to R2 = .33 in Table 5 , p < .001).
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Our results revealed that more than three fifths (62%) of the women reported patterns of conflict that included their parent care role. Although these women represented a majority in this sample, a sizable proportion of the women (38%) did not indicate that parent care was a major source of interference with their other roles. These latter findings are especially important because they call into question a widely held (albeit largely implicit) assumption of the competing demands perspective: that caring for an impaired parent necessarily interferes with adult daughters' perceived ability to meet the demands of other family and work roles (e.g., Brody 1990).
Regarding the first aim of the study, several demographic and caregiving characteristics differentiated those women who identified parent care as one of their conflicting roles from those women who did not. Compared with their counterparts, women experiencing parent care conflict tended to have parents who were more functionally and cognitively impaired. These women also were likely to have less education, to have lower personal and household income, and to be employed in lower status occupations. Our findings on socioeconomic status are consistent with prior work on later life caregiving showing that caregivers with fewer economic resources are more likely to be providing care that interferes with the enactment of other social roles (Aneshensel et al. 1995). In addition, other research has demonstrated that employed caregivers in nonprofessional occupations who provide care to an elderly relative have more difficulty combining their work and family responsibilities (Neal et al. 1993).
The women in our study who reported that the most conflict occurred between their parent care role and one other role also tended to have older children and marriages of longer duration, when compared with women who did not report such conflict. It is possible that these findings can be explained by the fact that women with older children and those who had been married longer had parents with greater impairment. Additional analyses revealed, however, that the age of children and the duration of marriages of the women in our study were unrelated to the functional and cognitive impairment of their parent, but these two variables were strongly related to each other (r = .62, p < .001). These children were typically adolescents, with nearly half between the ages of 12 and 18. The adolescent years can be a time of heightened strains and conflict in families (Montemayor 1983; Montemayor 1986; Silverberg and Steinberg 1987). Those women in our study with older children at home may have been dealing with problems associated with having teenaged children, and such family strains could have increased the likelihood that their parent care responsibilities would conflict with other roles. Taken together, our findings suggest that adult daughters' parent care responsibilities are more likely to conflict with their other roles when the demands of caregiving are high, children are older, and socioeconomic resources are low.
Although most prior research examining conflict between caregiving and other roles has focused on employment and marriage (Aneshensel et al. 1995; Barling et al. 1994; Gignac et al. 1996; Gottlieb et al. 1994; Kleban et al. 1989; Neal et al. 1993; Scharlach 1994; Stephens and Franks 1995; Stephens et al. 1997; Walker et al. 1993), our findings show that for some women, parent care can also conflict with motherhood. One third (34%) of the women whose pattern of role conflict included the parent care role reported that the parent care role most often conflicted with the mother role. Conflicts between the parent care role and the employee role were identified by a slightly larger proportion of the women (38%), and conflicts with marriage were identified by a slightly smaller proportion (28%). Compared with women reporting conflict between their parent care and employee roles, those whose parent care role conflicted most with their mother role tended to be younger, to have husbands who were younger, to be in marriages of shorter duration, and to have younger children who were less independent in their daily activities. The likelihood that conflict will occur between the roles of caregiver to an impaired parent and caregiver to one's children, therefore, seems to decrease as children age and become more self-sufficient and as marriages mature. In contrast, women who reported that their parent care role conflicted most with employment had been married longer than women reporting conflict between parent care and their mother or wife roles. Taken together, our findings suggest that the point in the life course at which a woman provides care to an impaired parent may help to determine the configuration of conflict she experiences between her parent care role, and her mother, wife, or employee roles.
Findings from our study also provide some support for our theoretical model, which links parent care stress to psychosocial well-being through interrole conflict between parent care and other roles. Specifically, our results demonstrate that the relationship between the stress of the parents' behavior problems and women's depressive symptoms, and the relationship between the stress of providing assistance to the parents with instrumental activities and restrictions in leisure activities, were partially mediated by interrole conflict. These findings suggest that one way parent care stress exerts its deleterious effects on the well-being of adult daughter caregivers is through the incompatible pressures of parent care and another role on women's time and energy (Aneshensel et al. 1995; Goode 1960; Kahn et al. 1964; Kopelman et al. 1983; Pearlin et al. 1990).
The findings that interrole conflict only partially explained the relationships between parent care stress and well-being suggest that other mediational forces were operating. Given the complexity of the stress process, it is unlikely that a single mediator, such as interrole conflict, could fully account for the stress and well-being relationship. Another explanation for our findings is that the conflict occurring between parent care and other roles threatened the self-identity of these women in two important life domains (Burke 1991; Schlenker 1987) and thereby increased their susceptibility to poor psychosocial functioning (Frone et al. 1992). Theoretical and empirical work have suggested that the erosion of self-worth and a sense of control over one's life are other mediating mechanisms that help to explain how stress manifests itself in psychological distress (Krause 1990; Pearlin, Lieberman, Menaghan, and Mullan 1981). In support of this perspective, lower levels of self-esteem and mastery as a caregiver have been shown to mediate the relationship between greater caregiving stress and poorer psychological well-being among middle generation women caregivers, most of whom were adult daughters and daughters-in-law (Franks and Stephens 1992). Other research that has taken an interpersonal perspective has shown that poor relationship quality between the caregiver and the care recipient mediated the linkages between the stress of the care recipient's behavior problems and caregiver's depressive symptoms (Lawrence, Tennstedt, and Assmann 1998). It seems likely, therefore, that in addition to the demands on time and energy resulting from conflicting role responsibilities, these other stress mechanisms might have been operating among the women in our study.
Regarding our findings for the direct effects of parent care stress on well-being, when the two types of parent care stress were considered simultaneously, only the stress of the parents' behavior problems was significantly related to depressive symptoms, whereas only the stress of instrumental assistance to the parent was significantly related to leisure activity restriction. These findings are consistent with prior research that has shown that behavioral stressors are more strongly associated with negative emotional states than are instrumental stressors (Haley, Levine, Brown, and Bartolucci 1987; Kinney et al. 1995; Schulz and Williamson 1991) and that instrumental stressors are more strongly associated with leisure activity restrictions than are behavioral stressors (Kinney et al. 1995). This pattern of findings strongly suggests that care recipients' problematic, and often unpredictable, behaviors can be emotionally upsetting to caregivers and may dispose them to psychological distress, whereas the more routine activities of providing tangible assistance can limit the amount of time that caregivers have for social and leisure pursuits. Because more than one third of the women in our study did not provide personal care assistance to their parent, we did not include these forms of assistance as potential instrumental stressors. Given the frequency with which personal care tasks must be performed, and the routines that are often established around them, it is possible that the links between instrumental stress and leisure activity restriction would have been even stronger had the stress of providing personal care assistance been included.
Our approach to investigating interrole conflict extends prior caregiving research in several ways. Unlike studies that have examined conflict occurring between parent care and one other role at a time, typically either employment or marriage (Stephens et al. 1997; Stephens and Franks 1995), or those that have examined conflicts between caregiving and an aggregate of several "other" roles (Barling et al. 1994), we asked women to identify which two of their four roles conflicted the most. This approach allowed for the possibility that, for some women, parent care might not conflict with other roles at all or it might conflict with one role more than others. Although our method permitted women to choose which roles conflicted the most, it assessed conflict between two roles only. It is possible that a woman might experience equivalent amounts of conflict between parent care and two (or three) other roles; that she might experience lower, but still meaningful, levels of conflict between parent care and the roles not selected; or that she might experience conflict with roles we did not include. Given the nature of our measurement strategy, we are unable to offer insight into the larger context of conflict among multiple roles, in particular, the cumulative effects of conflicts among three or more roles on well-being.
Our approach to investigating interrole conflict relied solely on women's perceptions of role incompatibility. As such, our findings do not address ways in which the objective pressures of multiple roles function in the caregiving stress process (Kopelman et al. 1983). It can be argued, however, that women's appraisal of the incompatible pressures among their multiple roles may be more useful for understanding their subjective well-being than the objective features of these role demands.
Because the design of our study was cross-sectional, the direction of influence among our study constructs remains ambiguous. Rather than functioning as an antecedent condition to well-being, as our theoretical model posits, parent care stress and interrole conflict may be consequences of poor psychological and social well-being. Prior work in the literature on the intersection of employment and family life (Frone et al. 1992; Kopelman et al. 1983) and later life caregiving (Barling et al. 1994; Stephens et al. 1997) has also investigated interrole conflict as an intervening mechanism using cross-sectional findings. These studies, therefore, cannot contribute substantively to the causal interpretation of the findings from the present study. Some longitudinal studies have shown that role experiences (such as stress) are antecedents to psychological well-being (e.g., Aneshensel et al. 1995; Pearlin et al. 1981), but other longitudinal studies have shown the opposite causal direction among these constructs (Link, Mesagno, Lubner, and Dohrenwend 1990). Thus, it appears that the influences among parent care stress, interrole conflict, and well-being are likely complex and reciprocal.
Because all women in our study occupied the same four roles, the extent to which our findings would generalize to women with different numbers or types of roles is not known. It is also not known how well our findings would apply to women with different socioeconomic and racial characteristics from those of the women in our study. To ensure that our sample was as representative as possible of the adult daughter caregivers in our region, we used a broad range of strategies to recruit participants. Given that the average household income of our respondents was equivalent to that for dual-earner households in the United States and the proportion of African American women in the sample was similar to that in the region where our study was conducted, we believe that our strategy for recruiting a demographically diverse sample of adult daughter caregivers was largely successful.
Our study of interrole conflict contributes to the literature on later life caregiving in several important ways. First, in contrast to widely held beliefs, our findings provide evidence that caring for an impaired parent or parent-in-law does not necessarily interfere with women's perceived ability to carry out the requirements and responsibilities of their other roles as mothers, wives, and employees. Second, we identified several demographic and background characteristics of women, their family members, and their employment that were associated with different patterns of interrole conflict. Finally, our findings contribute to a growing literature on the caregiving stress process by demonstrating that the conflicts that occur between parent care and other family and work roles can help to explain how the stress of parent care manifests itself in poor psychosocial well-being.
|
Acknowledgments |
|---|
Received for publication June 28, 1999. Accepted for publication March 30, 2000.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J. Savla, D. M. Almeida, A. Davey, and S. H. Zarit Routine Assistance to Parents: Effects on Daily Mood and Other Stressors J. Gerontol. B. Psychol. Sci. Soc. Sci., May 1, 2008; 63(3): S154 - S161. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. E. Scharlach, K. Gustavson, and T. S. Dal Santo Assistance Received by Employed Caregivers and Their Care Recipients: Who Helps Care Recipients When Caregivers Work Full Time? Gerontologist, December 1, 2007; 47(6): 752 - 762. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Read and J. Wuest Daughters Caring for Dying Parents: A Process of Relinquishing Qual Health Res, September 1, 2007; 17(7): 932 - 944. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
