| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
RESEARCH ARTICLE |
a Department of Psychology, Kent State University, Ohio
b Department of Psychiatry, University of Pittsburgh, Pennsylvania
c Mandel School of Applied Social Sciences, Case Western Reserve University, Cleveland, Ohio
Tina R. Norton, Department of Psychology, Kent State University, P.O. Box 5190, Kent, Ohio 44242-0001.
Decision Editor: Fredric D. Wolinsky, PhD
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
Method. Interviews were conducted at 2 time points approximately 1 year apart. Participants were asked to rate the personal importance of each role on a scale of 1 to 10. Stress and rewards in each role were also assessed.
Results. Analyses revealed considerable change in role centrality, especially in the parent-care and employee roles. Increases in wife and employee stress were associated with decreases in the centrality of these roles, whereas increases in rewards in each of the four roles were related to increases in the centrality of the respective roles. When considered simultaneously, role rewards were stronger predictors of change in centrality than role stress.
Discussion. These findings suggest that the centrality of a social role can change over time in response to stressful and rewarding role experiences.
WITH the aging of the population, many midlife women occupy the role of caregiver to an impaired parent or parent-in-law in addition to other social roles such as mother, wife, and employee (Doress-Worters 1994
; Moen, Robison, and Fields 1994). According to census data, 70% of American women who are married and have a child under age 18 living at home are also in the labor force (U.S. Bureau of the Census 2000). Other research has estimated that 14% of women ages 40 to 64 provide at least 3 hr of assistance per week to a parent, and 4% provide that much assistance to a parent-in-law (Spitze and Logan 1990). Furthermore, the likelihood of being a family caregiver has been shown to increase both with age and across birth cohorts (Moen et al. 1994), suggesting that many more women will occupy the parent care role in the future.
An accumulating literature suggests that individuals attach greater importance or centrality to some roles than to others (e.g., McCall and Simmons 1966; Rosenberg 1979; Stryker 1987). Consequently, some roles are more important to a given woman than other roles, and the role that is most central to one woman may not be the most central to another. Some theorists have suggested that the centrality of a given role not only varies across individuals, but may also vary over time for each individual (e.g., Thoits 1995). The purpose of the present study was to investigate the extent of change occurring in social roles over 1 year and to examine increases in role-specific stress and increases in role-specific rewards as predictors of change in the centrality of that same role.
Role centrality has been defined as the degree to which a role is a person's source of identity or self-definition (e.g., Gurin, Veroff, and Feld 1960; Stryker and Serpe 1994). This construct has also been referred to as role or identity salience (e.g., Krause 1994; Luchetta 1995; Simon 1992; Thoits 1992), role commitment (e.g., Brown, Bifulco, and Harris 1987), role or identity importance (Hoelter 1983), and psychological or personal involvement (e.g., Frone, Russell, and Cooper 1995; Gurin et al. 1960; Pleck 1985). The concept of role salience or centrality dates back to James 1961, who proposed that individuals possess multiple roles or identities on which different importance or value is placed.
It has been assumed that because individuals gain meaning, purpose, and behavioral guidance from their roles or identities, roles that are highly central should be more important sources of psychological well-being than roles that are less central (Gurin et al. 1960; Thoits 1992). Consequently, it is argued that greater role centrality directly benefits a person's psychological well-being. This proposition has been supported by findings that higher levels of centrality in the parent-care, wife, or employee role have positive effects on the mental health of the role occupant (Martire, Stephens, and Townsend 2000; Pleck 1985).
Social roles can be characterized by the quality of the experiences they encompass. Theorists have conceptualized role quality in terms of negative and positive experiences within a role (Barnett and Baruch 1985). This perspective argues that two people may evaluate the same roles differently depending on the quality of their role experiences. Research has documented that both negative and positive experiences in multiple roles (e.g., role stress and role rewards) are simultaneously related to the psychological well-being of the women who occupy them (e.g., Baruch and Barnett 1986; Stephens, Franks, and Townsend 1994).
It has been proposed that the quality of experiences within a role can also alter the extent to which the role occupant perceives the role to be central (Thoits 1995). There are competing views, however, regarding the manner in which centrality is affected by role experiences. Some theorists suggest that people maintain role commitment or centrality when faced with adversity in that role (e.g., Lydon and Zanna 1990). We refer to this perspective as a "maintenance model" of role centrality. In contrast, other theories have proposed a "change model" of centrality in which individuals may change the centrality of a role in response to the quality of their role experiences (Thoits 1992, Thoits 1995).
The maintenance model of role centrality suggests that individuals maintain their role commitment or centrality in the presence of negative experiences. From this perspective, adversity is assumed to be a necessary condition for commitment. Therefore, a threat or challenge to an endeavor is needed before an individual can recognize it as a commitment (Brickman 1987).
Some research has supported this maintenance model of role centrality. When high levels of stress and difficulty were experienced in completing a task, individuals who judged the task as important at the outset (in that it was consistent with their values) felt more committed to this task later than those who did not consider it important (Lydon and Zanna 1990). Thus, commitment to the task was maintained rather than diminished by adversity. It has also been found that older adults who experienced undesirable events in the social role they ranked as most important (relative to other roles) were less likely to devalue the relative ranking of that role over a 4-year period (Krause 1999). Thus, individuals who experienced stressful events in their most important social role continued to value that role highly in the face of adversity.
Another perspective on the manner in which role experiences affect role centrality, the change model of centrality, posits that role centrality may change as a result of role experiences. Theories based on this model most often focus on the de-emphasis or devaluation of a highly central role as a means of coping with a threat to this role (Hobfoll 1989; Thoits 1991). It is argued that individuals may cope with a stressor by devaluing or changing their perceptions rather than confronting or changing the stressful situation itself (Hobfoll 1989; Pearlin and Skaff 1996). This restructuring in values or priorities in response to a crisis or stressor has been referred to as a cognitive redefinition coping strategy (Moos and Schaefer 1986).
Identity theory (e.g., Burke 1991) and affect control theory (e.g., Mackinnon 1994; Smith-Lovin and Heise 1988) posit that whether or not an individual changes the centrality of a role depends on both the quality of experiences in that role and the extent to which that role is initially regarded as central. According to these theories, change in role centrality would be most likely to occur when experiences in a highly valued role disconfirm this role identity. Thus, to the extent that stressful role experiences continually strain a valued role identity, these theories argue that the centrality of that role has the potential to be diminished.
Some studies have shown that, in fact, role centrality does change in response to role experiences. In a laboratory experiment, college students who were randomly told their performance on a task was inferior when compared with that of others decreased the importance of that domain of performance (Major, Sciacchitano, and Crocker 1993). It has also been found that individuals who were economically stressed were more likely to devalue the importance of monetary success, presumably to help protect themselves from the stress of limited financial resources (Pearlin and Schooler 1978).
Although theories based on the change model of centrality typically emphasize devaluation of the importance of a stressful or problematic domain, it could also be argued that individuals may increase the value or importance of a rewarding or positive domain as a means of self-enhancement. Positive experiences regarding one's performance in a role have been associated with higher centrality or salience of that role (Hoelter 1983). Furthermore, college students who were randomly told their performance was superior when compared with that of others (a presumably rewarding situation) increased the importance of that domain or aspect of performance (Major et al. 1993).
These studies suggest that change in importance or centrality occurred as a result of negative (or positive) experiences in the same domain. Their findings, however, are limited in a number of ways. The study by Major and colleagues 1993 manipulated negative feedback (or positive feedback) in a laboratory setting, and thus it is unknown to what extent naturally occurring experiences would have altered the perceived importance of a domain. Although the study by Pearlin and Schooler 1978 and the one by Hoelter 1983 focused on naturally occurring experiences, both were cross-sectional in design and were therefore unable to assess changes in centrality over time. Additionally, research in this area is limited by the paucity of longitudinal studies and by the common practice of examining the separate effects of negative and positive role experiences on centrality. Role quality theorists have emphasized the value of simultaneously considering negative and positive role experiences to understand the nature of a particular role (e.g., Barnett and Baruch 1985).
The primary purpose of the present study was to investigate change in centrality over 1 year as a function of stress and rewards within women's roles of parent care provider, mother, wife, and employee. This study had three objectives. The first objective was to determine the extent to which the centrality of women's social roles changes over time. Although some theorists have posited that the centrality of social roles may vary over time, few empirical studies have examined this assertion. The second objective was to examine increases in stress or increases in rewards in a given role as predictors of change in centrality 1 year later. As suggested by the change model of centrality, it was predicted that an increase in role-specific stress would be related to decreases in the centrality of that same role. It was further predicted that an increase in role-specific rewards would be related to increases in centrality.
The third objective was to determine whether increases in stress or increases in rewards in a role would be a stronger predictor of change in centrality when stress and rewards were considered simultaneously. It was predicted that role-specific stress would be more strongly associated with change in centrality than would role-specific rewards. This prediction was based on the stronger support for the effects of stress on centrality as compared with the effects of rewards (e.g., Hobfoll 1989; Pearlin and Skaff 1996; Thoits 1991).
The design of the present study provided the opportunity to assess change in the quality of role experiences and role centrality in the context of a stable role constellation. All participants in this study simultaneously occupied the same four roles (parent care provider, mother, wife, and employee) at both times of assessment. This strategy increases confidence that change in role centrality is in part accounted for by the quality of role experiences rather than by the different types of roles occupied or by the loss of a role.
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
). The sample of women in the present study was composed of 182 women who simultaneously occupied the roles of primary care provider to an impaired parent or parent-in-law (hereafter referred to as "parent"), mother to at least one child living at home, wife, and employee. Participants in this study occupied these four roles at Time 1 (T1) and 1 year later at Time 2 (T2).
Primary care provider 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, the parents were required to be living in the community at the first wave of data collection rather than in a nursing home or with the participants. This parent population was chosen because research has indicated that adult-child caregivers who live with their impaired parent differ in well-being from those who live in separate households (e.g., Deimling, Bass, Townsend, and Noelker 1989). To qualify for the mother role in this study, participants were required to have at least one child 25 years of age or younger (and no child older than 25) living at home on initial interview. This upper age limit of 25 years for children was selected to account for an increasing tendency of young adults to remain at home for longer periods of time (U.S. Bureau of the Census 1992c). In addition, each participant was required to be married and currently living with the same husband at both waves of data collection. Full-time or part-time employment at both time waves was also required of participants.
Participants were recruited from northeastern Ohio and two counties in Pennsylvania. Several means of recruitment were used, including newspaper and radio announcements and notices in the newsletters of businesses and social organizations. Special efforts were made to recruit African American women in proportion to their regional representation (13%; U.S. Bureau of the Census 1992b). A total of 898 women were screened for eligibility and 296 (33%) were eligible and willing to participate. The most frequent reasons for ineligibility among the 602 nonparticipants were not providing care to a parent (22%), having children over the age of 25 living at home (16%), and not being the primary care provider (15%). Another 5% of women who contacted the project office declined to participate.
At T2, 22 women who had participated at T1 were not interviewed because they could not be contacted (54%), were too busy (32%), or were ill (14%). Compared with the 274 women interviewed at both T1 and T2, the 22 women who were interviewed only at T1 were younger (p 
.01), were less educated (p .05), had been married fewer years (p .05), had lower household incomes (p .001), had younger parents (p .05), had children with more functional needs (p .05), had younger husbands (p .01), and were more likely to be African American (p .05). There were no significant differences in initial levels of role stress, role rewards, or role centrality between women who were interviewed only at T1 and women who were interviewed at both T1 and T2.
Of the 274 women interviewed at T1 and T2, 92 women were excluded from the present study because of the loss of one or more of the four roles at T2. This strategy allowed the role constellation of the participants to be consistent across time. On the basis of T1 demographic data, these 92 women had been married longer (p .01), worked fewer hours (p .05), had lower personal incomes (p .05), provided more hours of care to their parent (p .05), had parents requiring more help with personal and instrumental tasks (p .001), and had fewer (p .05), yet older (p .05), children at home who had fewer functional needs (p .01) than the 182 women who occupied all four roles at both points of assessment. Again, there were no significant differences in initial levels of stress, rewards, or centrality between women who occupied all four roles at both time waves and women who had lost one or more roles by T2.
Additional attrition analyses were conducted to determine whether initial levels of stress, rewards, or centrality in a given role were associated with the exit of that role between T1 and T2 (defined as losing that role for reasons other than death of the role partner). The 182 women who maintained all four roles between T1 and T2 were compared with women who had exited each role by T2. Women who exited the parent care role (n = 36) or the wife role (n = 4) did not differ significantly from women who maintained all four roles at both time waves on initial levels of stress, rewards, or centrality in those roles (p > .05). Although women who exited the mother role (n = 21) did not differ significantly from women who maintained all four roles on initial levels of stress or centrality in that role, they did have lower initial levels of mother rewards (p .05). There were no significant differences between women who exited the employee role (n = 9) and those who occupied all four roles at both time waves on initial levels of work stress or rewards. However, women who exited the employee role had lower initial levels of employee centrality than did women who occupied all four roles at both time waves.
Of the 182 women composing the sample for the present study, 88% were Caucasian and 12% were African American. At T1, the average age of the participants was 43.7 years (SD = 5.61; range = 28–57). They had completed an average of 14.8 years of education (SD = 1.88; range = 10–17) and had been married for an average of 17.8 years (SD = 8.24; range = 1–34). On average, participants reported having two children living at home (SD = 1.1; range = 1–9) with an average age of 13.6 years (SD = 5.88; range = 1–25).
In terms of the participants' relationship to the parents, 71% of the sample was providing care to a mother, 15% to a father, 10% to a mother-in-law, and 3% to a father-in-law at T1. Participants had been providing care to the parent for an average of 6.7 years (SD = 6.12, range = 0–36). The average age of the parents was 75.9 (SD = 7.29, range = 50–94). Participants spent an average of 2.5 hr assisting with caregiving tasks on a typical weekday (SD = 2.36; range = 0–12) and 3.3 hr on a typical weekend day (SD = 2.71; range = 0–24). Parents' health was rated by 59% of participants as being either "fair" or "poor." Most (96%) of the participants reported that their parents needed help with three or more instrumental activities of daily living (e.g., housework), but only 22% reported that their parents needed help with three or more personal activities of daily living (e.g., grooming).
Other demographic information collected at T1 indicated that participants worked an average of 37.9 hr per week (SD = 11.87, range = 9–75) and an average of 49.7 weeks per year (SD = 6.28; range = 12–52). Most of the participants were employed in managerial or professional occupations (49%) or in technical, sales, and administrative-support occupations (44%); occupational classifications were taken from U.S. Bureau of the Census, 1992a). The average household income in this sample was $40,000–$59,000. This is comparable to the median family income ($51,000) of married couples or families in which the wife is in the paid labor force (U.S. Bureau of the Census 1995).
Measures
Role stress
Stress was assessed separately for each of the four roles. Items were adapted from other research on the parent care role (Albert 1991; Kinney and Stephens 1989; Vitaliano, Russo, Young, Becker, and Maiuro 1991; Zarit and Zarit 1983), mother role (Baruch and Barnett 1986; Crnic and Greenberg 1990; Pearlin and Schooler 1978; Veroff, Douvan, and Kulka 1981), wife role (Baruch and Barnett 1986; Veroff et al. 1981), and employee role (Baruch and Barnett 1986; Quinn and Staines 1979; Veroff et al. 1981). For each role, participants were asked to indicate how stressful particular experiences had been during the past 2 months using a scale ranging from 1 (not at all) to 4 (very much). Participants could indicate that items did not apply to their situation.
Role-stress scores were calculated for each respondent. These scores represent the average amount of stress experienced in each role. They were calculated by summing the ratings across items and dividing by the number of items endorsed as being applicable to the participant's situation, the same procedure as used in earlier research (Stephens et al. 1994). Role-stress scores were calculated in this way to create measures that excluded events that the participant had not experienced in the past 2 months, that she had never experienced, or that were irrelevant to her situation. The focus of these measures was on the appraisal of events that had occurred in the participants' lives.
Parent-care role (T1 
= .86; T2 = .78). A total of 15 items (e.g., "Dealing with parent's memory or cognitive problems," "Doing housework for your parent") made up the measure of parent care stress. These items assessed the stress associated with providing care to the parent as well as the interpersonal relationship with the parent. The average parent care stress score was 2.23 at T1 (SD = 0.55; range = 1.10–3.47) and 2.16 at T2 (SD = 0.56; range = 1.00–3.42).
Mother role (T1 = .80; T2 = .85). Stress in the mother role was assessed using 12 items (e.g., "Problems in communicating with your children," "Disciplining or correcting the children"). These items assessed stress associated with parental tasks as well as the interpersonal relationship with the child. Women with more than one child living at home could respond to these items in regard to any or all of their children. The average mother stress score was 2.21 at T1 (SD = 0.61; range = 1.00–4.00) and 2.17 at T2 (SD = .57; range = 1.00–3.75).
Wife role (T1 = .79; T2 = .80). Wife stress was assessed with 12 items (e.g., "Problems in communicating with husband," "Doing housework"). These items included stress associated with household responsibilities as well as interactions with the husband. Inclusion of household responsibilities was based on prior research that suggests that even in two-income households women retain primary responsibility for housework (e.g., Biernat and Wortman 1991). The average wife stress score was 2.22 at T1 (SD = 0.57; range = 1.00–3.56) and 2.17 at T2 (SD = 0.57; range = 1.00–3.75).
Employee role (T1 = .71; T2 = .79). A total of nine items (e.g., "Having too much work to do," "Receiving inadequate pay or employee benefits") composed the employee role stress measure. These items assessed stress associated with the nature of the work performed as well as the work environment. The average employee stress score was 2.36 at T1 (SD = 0.69; range = 1.00–4.00) and 2.35 at T2 (SD = 0.69; range = 1.00–4.00).
Role rewards.
Rewards were assessed in all four roles. For each role, participants were asked to indicate how rewarding particular experiences had been during the past 2 months using a scale ranging from 1 (not at all) to 4 (very much). Rewards scores for each role were calculated in the same way as role stress scores, and the potential range for all of the role rewards measures was 1 to 4. The items in the parent care–rewards scale were adapted from previous work on parent care (e.g., Lawton, Kleban, Moss, Rovine, and Glicksman 1989). All of the items in the mother and wife rewards scales were derived from earlier work on family roles (e.g., Baruch and Barnett 1986; Veroff et al. 1981), whereas the employee rewards items were derived from the same sources as the employee-stress items.
Parent-care role (T1 = .82; T2 = .86). A total of six items (e.g., "Doing things to help your parent," "Feeling needed by your parent") composed the parent care–rewards scale. These items assessed rewards associated with providing care to a parent. The average parent care rewards score was 3.10 at T1 (SD = 0.74; range = 1.00–4.00) and 3.08 at T2 (SD = 0.70; 1.00–4.00).
Mother role (T1 = .86; T2 = .83). A total of seven items (e.g., "Sharing in your children's accomplishments," "Watching your children develop as individuals") composed the mother rewards scale. These items assessed rewards associated with parenting. The average mother rewards score was 3.67 at T1 (SD = 0.47; range = 1.20–4.00) and 3.64 at T2 (SD = 0.45; range = 1.75–4.00).
Wife role (T1 = .87; T2 = .90). A total of six items (e.g., "Seeing your marriage mature and grow," "Doing things with your husband") assessed rewards in the wife role. These items assessed rewards associated with interactions with the husband. The average wife rewards score was 3.31 at T1 (SD = 0.66; range = 1.00–4.00) and 3.25 at T2 (SD = 0.68; range = 1.00–4.00).
Employee role (T1 = .80; T2 = .86). A total of nine items assessed employee rewards (e.g., "Utilizing your skills and abilities," "Seeing the results of your work"). These items assessed rewards associated with the nature of work or the work environment. The average employee rewards score was 3.36 at T1 (SD = 0.48; range = 1.78–4.00) and 3.28 at T2 (SD = 0.57; range = 1.22–4.00).
Role centrality.
Centrality of each of the four roles was assessed with one item ("How important to you is: being a caregiver to your parent/being a mother/being a wife/doing the work that you do?"). Participants were asked to rate the personal importance of each role on a scale from 1 (not at all important) to 10 (very important). This operationalization of role centrality is consistent with that used in other studies (e.g., Hoelter 1983; Simon 1992) and allows for the possibility that two given roles may be of equal or differing importance to the individual. Such a strategy was deemed appropriate given that family and work roles have been shown to be highly important to women (Veroff et al. 1981). The average parent care–centrality score was 8.41 at T1 (SD = 1.90; range = 1.00–10.00) and 8.05 at T2 (SD = 2.07; range = 1.00–10.00). The average mother centrality score was 9.56 at T1 (SD = 1.06; range = 5.00–10.00) and 9.39 at T2 (SD = 1.26; range = 3.00–10.00). The average wife centrality score was 8.97 at T1 (SD = 1.54; range = 3.00–10.00) and 8.89 at T2 (SD = 1.50; range = 3.00–10.00). The average employee centrality score was 8.01 at T1 (SD = 1.72; range = 3.00–10.00) and 7.56 at T2 (SD = 2.07; range = 1.00–10.00).
Procedure
Women who met eligibility requirements were contacted by a trained interviewer. Structured, in-person interviews were conducted in the participants' homes or another place of their choosing. These interviews were conducted using computer-assisted interviewing software, and typically took 1.5 hr to administer. Interviews were conducted approximately 1 year apart. Participants received $20 for each interview.
Analysis Plan
In the Results section, zero-order correlations among the study variables and descriptive data on change in role stress and role rewards are first presented. To address the first objective of the study, we report descriptive data regarding change in centrality over time at both the group and individual levels. Because role centrality may change in different directions (increase or decrease) across individuals, change in centrality may be masked at the group level but revealed at the individual level. Paired samples t tests were performed to examine differences between the means of T1 and T2 centrality scores within each role. In addition, individual variability in role centrality between T1 and T2 was examined using difference scores. Difference scores for each role were calculated by subtracting each participant's T1 centrality score from her T2 centrality score. Thus, a positive difference score indicates an increase in centrality between T1 and T2, and a negative score indicates a decrease. Centrality difference scores are categorized as no change (<0.5 SDs from 0), small change (between 0.5 and 1.5 SDs from 0), and moderate to large change (>1.5 SDs from 0). Defining change as more than half a standard deviation is consistent with the criterion for a medium effect size (Cohen 1988) and has been used by other investigators (e.g., Li, Seltzer, and Greenberg 1999).
Hierarchical multiple regression analyses were conducted to test the hypotheses concerning change in stress and rewards as predictors of T2 role centrality. To determine whether any variables in addition to T1 role centrality should be included as control variables in the regression analyses, we computed partial correlation coefficients between 24 background variables (e.g., age, race, education, income) and T2 role centrality for each role. In each partial correlation, the relationship between T1 role centrality and T2 role centrality was controlled. Any demographic or background variable that was significantly correlated (p .05) with T2 centrality for a given role (adjusted for T1 centrality) was included as a control variable in the regression analyses regarding that role. No variable met this criterion for the parent care or employee roles. For the mother role, two variables met this criterion (husbands' employment status and children's supervision needs), and for the wife role, two variables met this criterion (participants' education and parents' level of impairment).
Three sets of regression analyses were conducted for each of the four roles. In the first set, we entered the difference score for stress in a given role as the predictor of T2 centrality in that same role while controlling for T1 levels of both centrality and stress in that role. Difference scores for stress were calculated by subtracting each participant's T1 stress score from her T2 stress score. Both a stability component (T1 stress) and a change component (difference score) were used in each analysis to account for the possibility that the magnitude of change in stress may be constrained by the T1 level (Finkel 1995; Kessler and Greenberg 1981). In addition, T1 stress was controlled to ensure that change in stress, not the T1 level, was a predictor of T2 centrality.
In the second set of analyses, we entered a difference score for rewards in a given role as the predictor of T2 centrality in that same role while controlling for T1 levels of both centrality and rewards. Difference scores for rewards were calculated in the same way as the stress difference scores described earlier (T2–T1). In the third set of analyses, we entered difference scores for both stress and rewards simultaneously as predictors of T2 centrality in that same role, while controlling for T1 levels of centrality, stress, and rewards. In each analysis conducted, a statistically significant coefficient (p .05) for change in stress or change in rewards was judged as indicative of a main effect of that predictor.
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
.01), with shared variance ranging from 23% to 49%.
|
Descriptive analyses were conducted to examine change in stress and change in rewards over time at both the group and individual levels. Paired t tests were used to examine differences in the means of T1 and T2 role stress scores and differences in the means of T1 and T2 role reward scores. Significant differences were found between the means of T1 and T2 parent care stress (t = 2.11, p .05) and T1 and T2 employee rewards (t = 2.04, p .05). The average level of parent care stress and the average level of employee rewards decreased significantly between T1 and T2. No significant differences were found between the means of T1 and T2 stress for the mother, wife, or employee roles or between the means of T1 and T2 rewards for the parent care, mother, or wife roles.
Change in stress and change in rewards were also examined at the individual level. Difference scores for stress and rewards were categorized in the same manner as the centrality difference scores (e.g., no change is < 0.5 SDs from 0). In the parent care role, 60% of participants reported change (either an increase or decrease) in stress between T1 and T2, and 60% reported change in rewards. In the mother role, 58% of participants reported change in stress between T1 and T2, and 38% reported change in rewards. In the wife role, 58% of participants reported change in stress between T1 and T2, and 53% reported change in rewards. In the employee role, the largest proportion of women (52%) indicated change in stress between T1 and T2, and 51% reported change in rewards.
Change in Centrality Over Time
Table 2 shows the means and standard deviations of parent care, mother, wife, and employee role centrality at T1 and T2. Paired t tests were used to examine change in centrality over time at the group level. Significant differences were found between the means of T1 and T2 parent care centrality (t = 2.58, p .01) and T1 and T2 employee centrality (t = 3.11, p .01). The average level of parent care centrality and the average level of employee centrality decreased significantly between T1 and T2. No significant differences were found between the means of T1 and T2 centrality for either the mother role or the wife role.
|
|
.001, and the employee role, ß = -.25, F(1,178) = 12.28, p .001. In the wife and employee roles, the sign of the significant regression coefficients indicated that an increase in stress between T1 and T2 was significantly related to a decrease in centrality at T2.
|
.001, the mother role, ß = .21, F(1,175) = 8.44, p .01, the wife role, ß = .43, F(1,176) = 77.32, p .001, and the employee role, ß = .43, F(1,178) = 48.18, p .001. In each of the four roles, the sign of the regression coefficient indicated that an increase in rewards between T1 and T2 was significantly related to an increase in centrality, and these findings were particularly strong for the parent care, wife, and employee roles (Table 5 ).
|
.001, but an increase in parent care stress was not. In the mother role, an increase in rewards was significantly related to change in centrality, ß = .22, F(2,174) = 9.06, p .01, but an increase in stress failed to significantly predict change in mother centrality. Similar effects were found for the wife role, with only an increase in wife rewards predicting change in wife centrality when increases in stress and rewards were considered simultaneously, ß = .41, F(2,174) = 61.75, p .001. In the employee role, both an increase in employee stress, ß = -.20, F(2,176) = 9.36, p .01, and an increase in employee rewards, ß = .43, F(2,176) = 49.36, p .001, were significantly related to change in employee centrality, with an increase in employee rewards emerging as the stronger predictor, t(176) = 7.40, p .001. As indicated by the magnitude of the regression coefficients, an increase in rewards was more predictive of change in centrality than was an increase in stress in each of the four roles (Table 6 ).
|
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Although some women in our study did not alter the perceptions of the centrality of their roles over the 1-year period, many others did alter their perceptions. The greatest degree of change in centrality was evident in the parent-care and employee roles. Thus, our study provides empirical support for the theoretical proposition that centrality can, in fact, change over time (e.g., Thoits 1995).
The greater degree of change in centrality observed in some roles than in others may be due to the extent to which a given role was initially viewed as being central. Centrality ratings that were at or near the upper limit of the scale at baseline were restricted in their potential for increase over time. Large portions of women gave the mother (88%) and wife (76%) roles the two highest centrality ratings, and these roles demonstrated the least amount of change in centrality over the year. In contrast, the parent-care and employee roles were given the two highest centrality ratings by only 56% and 45% of women, respectively. Thus, the mother and wife roles may have been rated so highly central initially that any enhancement that occurred could not be detected by our measure of centrality.
Another possible explanation for the differential change in centrality across various roles may be the amount of time these roles had been occupied by the women. In our study, women had been in the wife role for an average of 18 years and in the mother role for an average of 14 years. In contrast, the parent-care role, in which more change in centrality was observed than in the wife and mother roles, had been occupied for an average of only 7 years. The importance of roles that have been occupied for a greater period of time may be more stable (and thus undergo less change) than roles of lesser duration because women have had more time to integrate these long-standing roles into their self-concepts. We were unable to examine this assertion for the employee role because women were not asked to indicate the number of years they had been in the labor force.
Our findings provide evidence for the theoretical assumption that increases in stress and increases in rewards would be associated with change in role centrality. When stress was considered independently from rewards, women who experienced an increase in stress in the wife role or in the employee role perceived that same role to be less central 1 year later. No association was found between an increase in stress and a decrease in centrality in the parent care or mother roles. These findings suggest that women devalued the importance of the wife and employee roles over the course of the year, possibly as a means of coping with the increased levels of stress occurring in these two roles (e.g., Hobfoll 1989). Additionally, reduced centrality may reflect an attempt to restore the balance between one's perception of a role as important and the disconfirming evidence presented by stressful experiences (e.g., Burke 1991; Mackinnon 1994). The different pattern of findings for the parent care and mother roles versus the wife and employee roles may have been due, in part, to the nature of the responsibilities associated with these roles. Given that the role partners in the parent care and mother roles (ill or disabled parents and children) were dependent on the women in our study, these women may have expected the roles of parent care provider and mother to be highly stressful (e.g., Veroff et al. 1981). Thus, the stressful experiences of parent care or motherhood may have been viewed as normative, and as a result, may not have threatened the women's identity in these roles. In contrast to the varied effects of stress on centrality, women who experienced an increase in rewards in each of the four roles perceived those same roles to be more central 1 year later. It is possible that women who experienced increased rewards in each of the four roles increased the importance of these roles as a means of enhancing their self-images in important life domains.
Our findings regarding the relationship between role stress and role centrality are both consistent with and different from findings reported by Krause 1999. Like our study, Krause also found that substantial change took place in the centrality or salience of roles that older adults regarded as most important at baseline. Over one half of these older participants devalued the centrality of the role they initially ranked as most important (Krause 1999). Unlike our study, however, devaluation in role centrality was not predicted by stress or adversity occurring in the most important role. Instead, older adults who experienced undesirable events in the role that was initially most central were less likely to devalue the importance of that role over a 4-year period (Krause 1999).
The differences between our findings and those of the Krause 1999 study regarding the linkages in role stress and role centrality may be due in part to the different life stages of the participants studied. In contrast to the midlife women in our study, the older adults in the Krause study may have had fewer roles, roles of longer duration, and roles that were more strongly integrated into their self-conceptions. Hence, when an important role was threatened by stress or adversity, the older adults may have had few desirable alternatives, and therefore remained committed to the role. Another possible explanation for the different findings in these two studies is the way centrality was conceptualized and measured. In the Krause study, older adults were asked to rank their most important role relative to other roles they occupied, and change in centrality was operationalized as assigning a role a different relative ranking at follow-up from the one assigned at baseline. In contrast, our study assessed role centrality by asking women to rate the importance of a given role (without regard for the relative importance of their other roles). Had our study used the relative rankings of the women's four roles, we may have detected less devaluation of the most important role in response to adversity within that role. Taken together, these two studies suggest that people may well devalue the centrality of a stressful role relative to its centrality on a prior occasion, without necessarily devaluing the centrality of that role relative to other roles.
Contrary to our hypothesis, increases in rewards in each of the four roles were more strongly related to change in centrality than were increases in stress. One explanation for this pattern of findings may be women's desire to confirm or strengthen their self-identity or self-concept (e.g., Sedikides 1993; Swann 1997). Research has shown that family and work roles are keenly important to women's self-concepts (e.g., Veroff et al. 1981). Thus, the women in our study may have given more weight to positive experiences in these roles than to negative experiences because these positive experiences confirm their role commitments and validate the time and effort they invest in the roles. It is also possible that women's reports of role experiences were influenced by a need to characterize these valued roles in socially desirable ways. Consequently, the greater weight given to positive experiences may reflect women's biased recall or biased reporting of role experiences such that positive experiences were overestimated and negative experiences were downplayed. Another explanation for these findings may be that the women in our study simply experienced more rewards than stress in these roles.
One potential limitation of this study is the generalizability of the findings. Because participants were self-selected, questions could be raised about the representativeness of the sample. It could be argued that these women were extremely stressed by their multiple roles and chose to participate and remain in this study to gain insight into their problems. It could also be argued that these women experienced less stress and therefore had the time and energy to participate in this research. A variety of recruitment sources were used to reduce the potential of either bias (Barer and Johnson 1990). Given that, on average, stress scores in each of the four roles were in the middle range at both time periods, it appears that our recruitment strategy was highly successful.
The generalizability of the present findings may also be limited by the ethnic composition of the sample. The participants were primarily Caucasian (88%) and, to a lesser extent, African American (12%). Special efforts to recruit African American women in proportion to their representation in the recruiting area were successful. It is not known, however, to what extent the present findings would apply to women with ethnic backgrounds that differ from those of women who were included in this study.
Biases in sample attrition may also limit the generalizability of the findings. Compared with women in the longitudinal panel, those women who were interviewed only at baseline tended to be younger and of lower socioeconomic status and were more likely to be African American. Thus, despite the initial success in recruiting African American women into the study, the longitudinal findings reported here may be less likely to generalize to this group of women because they left the study at a higher rate.
The requirement that all participants initially occupy the same role constellation and maintain this constellation over a year may pose another limitation to the generalizability of the findings. Women were included in the present study only if they simultaneously occupied the roles of parent care provider, mother, wife, and employee over a 1-year period. Because we did not consider any roles other than these four, it is unclear to what extent the present findings would apply to women who initially had fewer or a different set of roles. Research suggests that women who occupy fewer social roles tend to have fewer financial and psychosocial resources than do women with more roles (e.g., Baruch and Barnett 1986). Furthermore, it is not known how the current findings would apply to women whose role constellation changed as a result of role loss. In our study, women who lost one or more roles during the year differed from those who maintained all four roles in that they provided more care to their parents at baseline and had lower household incomes. Additionally, compared with women who maintained all four roles, those who exited the mother role because their children left home had lower initial levels of mother rewards, and women who either quit or were laid off from their jobs had lower initial levels of employee centrality. It is also possible that perceptions of the centrality of one role could be altered by the loss or exit of another role.
Several strengths of the present study should be noted. Our strategy of examining change in the centrality of social roles in the context of a stable role constellation strengthens the inferences that can be made regarding the relationships between perceptions of centrality of a given role and the quality of experiences that took place in that same role. The longitudinal design of our study also allowed for an examination of change over time, thus providing a dynamic portrayal of the relationship between change in perceptions of role centrality as a function of role experiences rather than the static view offered by cross-sectional studies.
Another strength of our study is the use of a sample not previously studied in research on role centrality. Previous work in this area has investigated college students (e.g., Hoelter 1983; Lydon and Zanna 1990; Major et al. 1993) and older adults (e.g., Krause 1999). Because perceptions of role centrality are derived from a person's sense of identity, these perceptions may be more stable at some points in the life span than at others. College students, for example, may still be forming their personal identities, and older adults may have fewer family and work roles at this stage of their lives (Krause 1999). In contrast, women at midlife are likely to have a better formulated sense of identity than most younger adults, and they may have more competing roles and responsibilities than most older adults. Thus, the four-role constellation of the midlife women investigated in our study provides a rich context for testing hypotheses regarding change in role centrality over time.
Given the growing number of women simultaneously occupying the roles of parent care provider, mother, wife, and employee (e.g., Moen et al. 1994), the importance of studying this population is clear. The issue of change in centrality is particularly important for the parent care role, as midlife women provide an invaluable service to society by shouldering most of the responsibility for parent care. Understanding the relationship between a caregiver's experiences and the perceived importance of this role may have direct implications for improving role performance and decreasing the likelihood of role exit.
The present findings extend the research on centrality and multiple roles in three important ways. First, our findings demonstrate that even over a 1-year period, considerable change can occur in the centrality of family and work roles. Second, unlike previous research that has examined either stress or rewards as antecedents to change in perceptions of centrality, the present study examined role stress and role rewards simultaneously. This strategy is consistent with the role quality perspective, which emphasizes the importance of simultaneously examining negative and positive experiences to understand the nature of a particular role. Finally, contrary to the emphasis on stress that has characterized the literature on role centrality, the present study revealed that centrality is more likely to be bolstered in the presence of rewards than to be devalued in the presence of stress.
|
Acknowledgments |
|---|
Received for publication January 5, 2001. Accepted for publication June 29, 2001.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  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]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
