Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Giving and receiving help in aging families : The independent and joint effects of gender and class Bednarski, Valerie Ann 1995

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1995-0202.pdf [ 5.83MB ]
Metadata
JSON: 831-1.0086889.json
JSON-LD: 831-1.0086889-ld.json
RDF/XML (Pretty): 831-1.0086889-rdf.xml
RDF/JSON: 831-1.0086889-rdf.json
Turtle: 831-1.0086889-turtle.txt
N-Triples: 831-1.0086889-rdf-ntriples.txt
Original Record: 831-1.0086889-source.json
Full Text
831-1.0086889-fulltext.txt
Citation
831-1.0086889.ris

Full Text

Giving and Receiving Help in Aging Families: The Independent and Joint Effects of Gender and Class b y Valerie Ann Bednarski B.A., The University of British Columbia, 1991 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS in THE FACULTY OF GRADUATE STUDIES Department of Anthropology and Sociology We accept this thesis as conforming to th^ r^uljred standard THE UNIVERSITY OF BRITISH COLUMBIA April 1995 © Valerie Ann Bednarski, 1995 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at. the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. The University of British Columbia Vancouver, Canada Department Date &4 DE-6 (2/88) ABSTRACT This thesis examines the independent and joint effects of gender and class on patterns of giving and receiving help in aging families. Using data from the 1990 General Social Survey, this study looks at adult children who help their parents who live in another household. This study also examines an unrelated sample of elderly parents who receive help from their children not living within their household. The findings show that more adult daughters than sons helped their parents with a majority of tasks. The results also show that "competing demands" such as work, marital status, and health status do not pull either sons or daughters away from providing assistance. Sons and daughters who live within an hour of their parents are more likely to help with most tasks compared to those children whose parents live farther away. Among elderly parents, more mothers than fathers received filial assistance on a couple of tasks. Marital status did not appear to influence likelihood of receiving help. Mothers and fathers who had at least one child living within an hour's distance were more likely to receive help than those whose children lived farther away, but only for some tasks. There was a relationship between class, as measured by one's level of education, and helping. However, results varied by task and were not always linear or in the anticipated direction. Class was also related to receipt of help but only for housework, transportation, and help with at least one task. In most cases the results were in the expected direction: working class elderly were more likely to receive help compared to the elderly from the upper and middle class. Expectations that class and gender interact to influence helping received mixed i i support. Contrary to the literature, middle class women were the most likely to give help with a couple of task categories such as housework. For the elderly, interaction effects only existed for those who received help with maintenance and with at least one task. For most tasks, gender is a stronger predictor than class for either giving or receiving help. Results from this thesis demonstrate that future studies on class, gender and caring should examine tasks separately as key differences emerge according to the task involved. i i i TABLE OF CONTENTS Abstract ii Table of Contents iv List of Tables vi Acknowledgement ix CHAPTER ONE INTRODUCTION 1 CHAPTER TWO GENDER AND CARING 5 Literature Review: Caregivers 5 Elderly Care Recipients 11 Theoretical Issues 13 Psychological/Individual Perspectives 13 Domestic Labour Theories 15 Feminist Perspectives 19 CHAPTER THREE CLASS AND CARING 23 CHAPTER FOUR GENDER, CLASS AND CARING 30 Central Propositions and Research Questions 32 CHAPTER FIVE METHODOLOGY 36 Measurement of Independent Variables 39 Limitations of the Study 49 iv CHAPTER SIX FINDINGS-ADULT CHILDREN WHO PROVIDE HELP TO THEIR PARENT(S) 50 Gender Differences in Helping Behaviour 50 Characteristics of Adult Children Who Help Their Parents 54 Education Differences in Helping Behaviour Towards Parents 61 Gender and Class: Interaction Effects 69 Closer Look at Education Differences Among Adult Children Helpers 74 Logistic Regression Results 80 CHAPTER SEVEN FINDINGS-ELDERLY PARENTS WHO RECEIVE HELP FROM THEIR CHILDREN Gender Differences in the Receipt of Help 87 Characteristics of Elderly Parents Who Receive Help 91 Education Differences and Interaction Effects in the Receipt of Help 95 Closer Look at Education Effects Among Elderly Parents 100 Logistic Regression Results 103 CHAPTER EIGHT CONCLUSIONS AND AREAS FOR FURTHER RESEARCH 109 Children Who Help Their Parents 109 Elderly Parents Who Receive Help 115 Contributions To The Literature 118 Directions for Further Research 119 Closing Comments 122 BIBLIOGRAPHY 124 v LIST OF TABLES Table 1 Frequencies and Coding for Independent Variables Used in Logistic Regression Analyses For Adult Children Who Help Their Parents 43 Table 2 Frequencies and Coding For Independent Variables Used in Logistic Regression Analyses: Parents Who Receive Help From Their Children 45 Table 3 Gender and Level of Education Frequencies For Adult Children Sample 51 Table 4 Adult Children Who Help Their Parents By Gender of Adult Child 52 Table 5 Daughters Who Help Their Parents 56 Table 6 Sons Who Help Their Parents 59 Table 7 Adult Children Who Help Their Parents By Children's Level of Education 62 Table 8 Adult Children Who Help Their Parents By Children's Level of Education Controlling For Gender of Helper 70 Table 9 Daughters Who Help Their Parents: Education and Work Status 76 Table 10 Daughters Who Help Their Parents: Education and Distance From Parents 76 Table 11 Sons Who Help Their Parents: Education and Work Status 79 Table 12 Sons Who Help Their Parents: Education and Distance From Parents 79 v i Table 13 Logistic Regression For Adult Children Who Help Their Parents With Housework 81 Table 14 Logistic Regression For Adult Children Who Help Their Parents With Personal Care 81 Table 15 Logistic Regression For Adult Children Who Help Their Parents With Transportation 83 Table 16 Logistic Regression For Adult Children Who Help Their Parents With Financial Support 83 Table 17 Logistic Regression For Adult Children Who Help Their Parents With Housework: Multiple Variables 85 Table 18 Gender and Level of Education Frequencies For Elderly Parents Sample 88 Table 19 Elderly Parents Who Receive Help From Their Children By Gender of Parent 89 Table 20 Elderly Mothers Who Receive Help From Their Children 92 Table 21 Elderly Fathers Who Receive Help From Their Children 92 Table 22 Elderly Parents Who Receive Help From Their Children By Parent's Level of Education 96 Table 23 Elderly Parents Who Receive Help From Their Children By Parent's Level of Education Controlling For Gender of Parent 98 Table 24 Elderly Mothers Who Receive Help According to Mother's Level of Education and Distance From Children 102 v i i Table 25 Elderly Fathers Who Receive Help According to Father's Level of Education and Distance From Children Table 26 Logistic Regression For Elderly Parents Who Receive Help With Housework Table 27 Logistic Regression For Elderly Parents Who Receive Help With Transportation Table 28 Logistic Regression For Elderly Parents Who Receive Help With Maintenance Table 29 Logistic Regression For Elderly Parents Who Receive Help With Housework From Their Children: Multiple Variables v i i i ACKNOWLEDGEMENT There are several people whose assistance I would like to acknowledge. First, thank you to Dr. Fiona Kay for her kindness and encouragement. I would also like to express my sincerest appreciation in being able to work with a wonderful committee. I thank Drs. Nancy Waxier-Morrison and Andrew Wister for their time, very helpful comments and suggestions, and continuous support. A tremendous heartfelt thank you goes to Dr. Neil Guppy who was everything a student could ever dream of in a supervisor. I have truly benefitted from, and will always appreciate the generous gifts of his time, advice, patience, good humour and never-ending support. Finally, thank you to my parents without whose love and encouragement all this never would have been possible. This thesis is for them. ix CHAPTER ONE INTRODUCTION As the number of elderly Canadians grows, so do the concerns of policy makers over the financial impact of this trend on our health care system (Aronson, 1990b:235-236). In an era of shrinking public resources and increasing worries about the rising costs of health care, the elderly are often portrayed as an economic burden (Aronson, 1986:47; 1990:236; Chappell, 1993:42; Kaden and McDaniel, 1990:4-5; McDaniel, 1992:62; McDaniel and Gee, 1992:63). As a result, provincial governments are developing policies that support "community care," while reducing institutional and acute care services (Aronson, 1986:47-48; Chappell, 1993:46). The 1991 British Columbia Royal Commission on Health Care and Costs stated that "a clear message received by the commission is that, whenever possible, care provided in the home...is preferable to institutionalization. ... There are many costs involved in keeping people in institutions, and from a quality of life as well as an economic perspective, we must encourage home and community care" (Province of British Columbia, 1991:C-154). In response to the Royal Commission, the British Columbian government stated in its recent health initiative that "future health policy for seniors will be based upon a commitment to helping the elderly maintain their independence through care at home or in the most home-like setting as possible" (Ministry of Health and Ministry Responsible for Seniors, 1993a: 11). It was also stated that "The health care system could not function without the valuable contribution of informal care providers. Strengthened support for care providers is a vital 1 aspect of bringing health closer to home" (original emphasis, Ministry of Health and Ministry Responsible for Seniors, 1993b: 17). The concept of community care is not unique to the British Columbia Royal Commission. Policy makers have heralded the idea in other countries, such as the United Kingdom and the United States, where aged populations are also increasing in numbers (Arber and Ginn, 1991:261-263; Aronson, 1985:116; Hooyman, 1992:185; Walker, 1983:120). In Britain, for example, a community care policy for elderly people with disabilities has been in effect for over twenty-five years (Walker, 1983:120). In 1958, the British Minister of Health was quoted as saying "[the] underlying principle of our services for the old should be this: that the best place for old people is in their own homes, with help from the home services if need be" (Townsend, 1962:196 in Walker, 1983:120). This statement is not unlike the one recently expressed by the B.C. government's new health care policy. Some proponents of the current community care policy argue that it is a financially viable alternative to a costly institutional care system, and it would accommodate seniors' desires to remain at home (Chappeli, 1993:46). Others, however, have expressed concerns (Aronson, 1985:122; Chappeli, 1993:46; McDaniel and Gee, 1993:67; Walker, 1991:105). For example, some suggest that community care is really a glorified term for care that is given mostly by women: "terms like 'community care' and 'family care' tend to mask the fact that care in the private arena is provided largely by female family members" (Aronson, 1985:122). Hooyman also notes that "in practice, community care is based on women's unpaid domestic labour" (1992:185). These comments reflect research findings showing that 2 women are the predominant carers of the elderly (Dwyer and Coward, 1992a:5; Horowitz, 1985a:616; Horowitz, 1985b:216; Kaden and McDaniel, 1990:17; Lee, Dwyer and Coward, 1993:S13-S14; Rosenthal, 1987:326; Rosenthal and Gladstone, 1993:18; Stoller, 1983:854; Stone, Cafferata and Sangl, 1987:625; Stone, 1988:64,66,68). As the provinces begin implementing community care policies, questions arise about how they will be implemented, who will be affected, and what, if any, role these people will have in the decision making process (Chappeli, 1993:49). These issues will directly affect both seniors and their caregivers, for as Chappeli states, "discussions taking place about the health care system and resultant policy changes not only draw on knowledge about seniors and caregivers to seniors but also have implications for both groups" (1993:40). A step towards answering these questions comes in understanding how caregiving of the elderly is socially organized. Using data from the 1990 General Social Survey (Statistics Canada, 1991) this research investigates the factors that are associated with adult children providing assistance to their community-dwelling parents. The focus of this thesis will be on gender and class, and whether there is an interaction between the two. That is, besides examining the independent effects of class and gender on helping behaviour, I also ask whether the class effect on helping is the same for men and women. In addition to examining the issue from the perspective of giving help, I also explore the issue from the perspective of receiving help. Is the receipt of assistance organized by gender and class? Following the introduction, Chapter Two reviews the recent literature on gender and caregiving/care-receiving. Specific theories on the gender/caring relationship are also discussed. Chapters Three investigates the findings on class and caring while Chapter Four 3 reviews research on the interaction between gender, class and caring. Chapter Five outlines the study's methodology, while Chapters Six and Seven will present and discuss the findings. Chapter Eight is the concluding chapter. 4 CHAPTER TWO GENDER AND CARING Literature Review: Caregivers It is widely documented in the literature that daughters are more likely than sons to be the primary caregivers to their elderly parents (Chappell, 1992:35; Horowitz, 1985a:614; Montgomery, 1992:69; Montgomery and Datwyler, 1990:36; Stoller and Earl, 1983:69; Stone et al., 1987:620). Beyond differences in amount of caregiving, adult sons and daughters also have different caregiving roles, activities and experiences (Montgomery, 1992:67; Montgomery and Datwyler, 1990:36). For example, daughters are more likely than sons to give assistance with tasks such as meal preparation, household chores, and personal care (Chappell, 1992:36; Dwyer and Coward, 1991:S266; Horowitz, 1985a:614; Kaden and McDaniel, 1990:17; Montgomery and Kamo, 1989:215; Stoller, 1983:854; Stone et al., 1987:622). These tasks require what Horowitz describes as "hands on" (1985a:614) routine kinds of care, and are generally stereotyped as female duties. In addition to assisting with these specific tasks, daughters are also more likely to give overall assistance and more hours of care (Stoller, 1983:854; Stoller, 1990:234). Many of these studies also found that daughters provide more help with shopping and transportation. However, Stone et al. (1987:622) report little difference between sons and daughters on assistance with these tasks. Sons, on the other hand, tend to provide fewer hours of helping over a shorter timeframe, even when they are the primary caregivers (Horowitz, 1985a:615; Montgomery and Datwyler, 1990:36; Montgomery and Kamo, 1989:220; Stoller, 1983:854). The tasks 5 that sons engage in typically involve those that are male-stereotyped and can be performed on a sporadic basis. These tasks include home repairs, maintenance, and yard work, including snow shovelling (Chappell, 1992:36; Kaden and McDaniel, 1990:19). Further, when the demands of the care recipient escalate, sons still help with the intermittent tasks, but it is daughters who help with the routine assistance that is needed (Matthews and Rosner, 1988:190; Montgomery and Kamo, 1989:220; Stoller, 1990:234). Therefore, as Montgomery states, "sons, in contrast, more often assume supportive roles that require commitments over shorter periods of time and tend to be peripheral helpers within a caregiving network rather than the central actors" (1992:69). Despite these differences, most research has found that sons and daughters do not differ significantly with regard to who gives the most help with financial matters such money management (Horowitz, 1985a:614; Montgomery and Kamo, 1989:216; Stoller, 1983:854; Stoller, 1990:231). However, one study did find that daughters were more likely to provide help with handling finances (Dwyer and Coward, 1991:S266). Meanwhile, McPherson (1990:348), Rosenthal (1987:326) and McDaniel (1992:62) state that sons tend to provide their parents financial assistance. There are various caregiving roles that adult children can adopt when looking after their elderly parents. For example, they can be care managers where they coordinate services provided to their parents, or they can be the actual care providers (Archbold, 1983:41). Some research has found that there is no difference between sons and daughters as to which roles they assume (Finley, 1989:84). Other research has suggested that sons are more likely to be the care managers (Montgomery and Kamo, 1989:224). Matthews and Rosner have further refined possible caregiving roles by reporting that 6 daughters provide "routine" caregiving more often than sons, while their sisters assume "backup" roles (1988:188-189). Alternatively, sons are more likely to be described as providing "circumscribed" assistance. They can be relied upon to help, but only to a certain extent, or only on a particular task. Sons are also more likely to be described as providing help sporadically, or else are disassociated from caregiving altogether. Sons and daughters also differ in their caregiving experiences as they relate to employment. It has been suggested that working in paid employment acts as a "competing demand" on the caregiver's time (Lang and Brody, 1983:194; Stoller, 1983:853). The majority of the findings show that employment does not significantly affect the amount of time daughters spend caregiving (Brody and Schnoover, 1986:378; Matthews and Rosner, 1988:192; Montgomery and Kamo, 1989:222; Seecombe, 1992:171; Stoller, 1983:856). Some research has found that employed and non-employed daughters provide similar levels of help with IADL 1 tasks such as shopping, household tasks and transportation. Employed daughters, however, give less help with A D L 2 tasks such as personal care and cooking (Brody and Schnoover, 1986:378, Dwyer and Coward, 199LS267; Merrill, 1993:85). Brody and Schnoover (1986:378) report that employed daughters make up the difference by enlisting paid help for these tasks. Lang and Brody (1983:199) found some evidence that working daughters contributed less help than non-working daughters, but report that the difference was 1 IADL (instrumental activity of daily living) is a common term in the gerontological and social support literature. Chappeli defines IADL as referring "to any activity that is not considered essential for survival, but nevertheless is important for independence. Examples include help with housework, preparing meals, household maintenance, transportation, shopping and banking." (1992:80). J ADL refers to activities of daily living. These activities include being able to walk, move around physically, eat and bathe on one's own, and use the toilet. These tasks differ from the instrumental activities of daily living in that they are considered necessary to survival (Chappeli, 1992:79). 7 minimal (see also Merrill, 1993:85). The effect of employment on caregiving by sons is more substantial. Stoller (1983:856) found that working outside the home reduced the average number of hours of help given by sons. Employment status had no effect on the amount of care provided by daughters, however. On the separate question on how caregiving affects employment, Stone et al. (1987:620) report that daughters were more likely than sons to quit their jobs, cut back on their work hours, rearrange their schedules or take time off without pay to become caregivers (see also Alford-Cooper, 1993:48). Additional competing demands that may influence caregivers' time and energy are their marital status, and the number of children in their household (Lang and Brody, 1983:194; Stoller, 1983:855; Stone et al., 1987:620). Stoller (1983:855) found that both married sons and daughters contributed significantly less care to their parents compared to adult children who were not married (see also Dwyer and Coward, 1991:S265). Lang and Brody (1983:197) similarly found that married daughters gave three times less help than unmarried daughters. Stoller (1983:855) also reports that the number of dependent children of adult daughters does not influence the number of hours of assistance they provide to their parents. However, among adult sons there is a positive relationship between the number of caregiving hours and how many of their children are under the age of six. Stoller (1983:856) suggests that perhaps early child care responsibilities pulls daughters and daughters-in-law away from parental caregiving duties. In the meantime, adult sons may assume more caregiving responsibilities during this stage in the family life-cycle. Other variables that may influence the caregiving relationship for adult sons and 8 daughters include: age, configuration of the sibling network, parental living arrangement, proximity to parents, and health status (Dwyer and Coward, 1991; Lang and Brody, 1983; McPherson, 1990; Rosenthal, 1987). For example, research has found that older adult children, and older adult daughters in particular, give more parental assistance (Dwyer and Coward, 199LS265; Lang and Brody, 1983:197; Merrill, 1993:85). This likely reflects the correlation between age of child and parent. As parents grow older, their functional abilities usually decline. They are therefore more likely to require assistance from their children (Lang and Brody, 1983:197). It may also be that this apparent aging effect is masking a birth order effect. The eldest child may be expected to assume different parent-care responsibilities that younger siblings do not share (see Dwyer and Coward, 199LS261; Matthews and Rosner, 1988:190). The composition of the sibling network has also been shown to have a relationship with gender differences in parent care. Dwyer and Coward (1990:175) found that there were fewer differences between sons' and daughters' caregiving participation levels when the caregivers were only children, or were from single-gender sibling families. This contrasts with daughters from mixed gender networks who spent the most time caring for their parents compared to sons. The experiences of caregivers also varies according to their parents living arrangement. Adult children who live with their parents often provide more assistance (Arber and Ginn, 1992:630; Lang and Brody, 1983:197). Proximity to parents is also another factor that may influence the likelihood of providing assistance (Dwyer and Coward, 199LS265; McPherson, 1990:347; Rosenthal, 1987:318). Dwyer and Coward (199LS265) found that 9 the less proximate children were from their parents, the less likely they were to provide both ADL and IADL assistance. However, women tend to live closer to their parents so this may be another reason why daughters provide more help (Horowitz, 1985a:612,614; McPherson, 1990:348). With respect to health status, Horowitz states that "References to the health status of the caregiver in the empirical literature are scarce, except by implication on the basis of the caregiver's age" (1985b:219). Some studies have included health status as a variable. For example, Stone et al. (1987:620) report that one-third of the caregivers had a self-rated health status of either fair or poor. However, they did not explore health related differences in helping behaviour. Merrill (1993:85) found that daughters and daughters-in-law in poor health who were caregivers gave more hours of care and helped with more IADLs but not with ADLs. She is unclear as to why there are differences between help given with IADLs and ADLs. It would be interesting to know if health status has the same impact for both sons and daughters who provide care. While it should be noted that the studies reported here used different samples and research techniques3, the general conclusion is that daughters are the predominant caregivers of their elderly parents. However, the research has also shown how various differential characteristics between and among sons and daughters can influence their caregiving duties, roles and experiences. For example, some studies included adult children who were caring for either their parents or their in-laws. Other research, however, did not include in-laws. Some studies were based on face-to-face interviews while others used secondary data sources. 10 Elderly Care Recipients The bulk of the social support literature concentrates on caregivers, while comparatively less attention has been devoted to the experiences of those who receive assistance (Aronson, 1991:143; Aronson, 1990b:235). Research has found that elderly care recipients are most likely to be women (Aronson, 1991:138; 1986:48; Dwyer and Coward, 1992a:5; Kaden and McDaniel, 1990:5; Lee, Dwyer and Coward, 1993:S15; Lee, 1992:129; Rosenthal, 1987:326; Stone et al., 1987:619). One reason for this is the longer life span of women. Older men who need assistance are often married so it is their wives who tend to their functional health needs. Since married women usually outlive their husbands, they must often turn to their children for assistance (Chappell, 1992:34-35; Gee and Kimball, 1987:88; Kaden and McDaniel, 1990:17). Research has shown that marital status is an important factor influencing whether an elderly parent receives assistance from children (Dwyer and Coward, 1991.S267; Stoller, 1983:854; StoUer and Earl, 1983:67). Stoller (1983:854) reports that if an elderly parent was married, daughters provided less help than if the parent was not married. However, parents' marital status made no difference in whether they received help from sons. Stoller and Earl (1983:67) found that spouses were the main source of support for the married elderly while a greater number of daughters or family members not including sons gave help to unmarried elders. Other research has shown that non-married women such as widows or divorcees are more likely to name children as their main sources of support than are non-married men (Chappell, 1992:35). 11 There may be a relationship between gender of the care recipient and whether he/she receives help from a son or daughter, although the evidence is mixed. (Lee et al., 1993; Stoller, 1990). Stoller (1990) found that when elders were asked to name a helper other than a spouse, fathers were more likely to name sons, while mothers named daughters. Lee et al. report that "Although daughters are more likely than sons to be the primary caregivers (among adult children) for elderly parents of both genders, their predominance is much greater among the primary caregivers of mothers than of fathers" (1993:S14). In contrast, Horowitz found that "sons were no more or less likely than daughters to care for fathers as opposed to mothers" (1985a:614). There are gender differences in the types of assistance that elderly parents receive from their children. Penning (1990:225) found that the elderly who were male, older, had more education and had a higher level of functional disability were more likely to receive help with housekeeping or grocery shopping. Meanwhile, other research has found that more elderly women than men receive help from their children with household chores, repairs, personal care, financial support and emotional support (Kaden and McDaniel, 1990:19; Rosenthal, 1987:326). Women also receive the most overall care from children (Kaden and McDaniel, 1990:18). The fact that elderly women receive the most care is alarming for some researchers. The elderly do not wish to be dependent upon their children, yet it is elderly women who are more likely to require their children's assistance (Kaden and McDaniel, 1990:19; Rosenthal, 1987:335). The concern is that this negatively portrays elderly women who do not receive help from their family as being the most reliant on the state and public resources for help (Kaden and McDaniel, 1990:19). 12 Theoretical Issues Several explanations have been offered as to why women are the predominant caregivers of the elderly. What follows are brief discussions of the main perspectives that stem from the disciplines of psychology, family science, and feminist theory. Psychological/Individual Perspectives One of the most common explanations for the predominance of female caregivers comes from what Walker (1992) refers to as the psychological/individual perspectives. The psychological component owes much to the work of Nancy Chodorow (1978). This perspective views caring as an intrinsic part of a woman's identity, constituting a major aspect of her life (Chodorow, 1978; Graham, 1983:18; Walker, 1992:36). Caregiving becomes identified with women because infants tend to be cared for primarily by women. During the early mother-child relationship infants develop a special bond or attachment to their mothers, which leads them to associate caregiving and nurturing qualities, with women: Women's early mothering, then, creates specific conscious and unconscious attitudes or expectations in children. Girls and boys expect and assume women's unique capacities for sacrifice, caring, and mothering and associate women with their own fears of regression and powerlessness (Chodorow, 1978:83). Mothers experience their sons and daughters differently in that "Mothers tend to experience their daughters as more like, and continuous with, themselves. ...By contrast, mothers experience their sons as a male opposite" (Chodorow, 1978:166,109). Through the relationship with their mothers, daughters develop a sense of empathy and sensitivity to the 13 needs and feelings of others that is not necessarily adopted by sons (Ibid:\61 ,\16). Consequently, men and women assume their respective feminine and masculine roles. This enables the sexes to differentiate themselves from each other (Ibid:ll6; see also Graham, 1983:18). Chodorow states, "Masculine identification processes stress differentiation from others, the denial of affective relation... . Feminine identification processes are relational, whereas masculine identification processes tend to deny relationship" (1978:176; Chodorow, 1978 in Graham, 1983:18). Women's self concept, therefore, stems from doing things for others, such as providing care, while men achieve self worth by accomplishing things for themselves and are not involved in caregiving activities. Although it is not clear what Walker (1992:36) means by the "individual" aspect of this perspective, she briefly discusses Gilligan's contention that women often engage in a struggle between fulfilling their own needs and the needs of others because they have a "sensitivity to the needs of to others" and assume "responsibility for taking care" of others (Gilligan, 1982:16). Women are motivated to care for others to whom they feel emotionally attached. Men may also feel attached to their care recipient, but they express their feelings by offering more instrumental assistance, such as financial support (Walker, 1992:36). Therefore, Walker concludes: psychological/individual perspectives suggest that women are naturally, nurturers and caregivers. In particular, both their identity and their attachment to loved ones conspire to create nurturing beings. Men are seen as having a fundamentally different identity and as responding to attachment through labour force participation, that is, by supporting family members economically (1992:37). There have been several critiques of the psychological explanations of caregiving (see Graham, 1983; Walker, 1992). First, it is not always the case that women prefer to provide 14 care to children or the elderly. Some men are very keen to perform these tasks (Walker, 1992:37). Studies show that sons and daughters have similar feelings of obligation to provide care to their parents (Finley, 1989:84; Montgomery and Kamo, 1989:222; Walker, 1992:38). Second, Graham contends that "the psychological perspective sees caring as the mechanism through which the consciousness of women and men is recreated generation by generation: but as a result, it has tended to ignore the economic and political forces which determine that consciousness in the first place" (1983:17). In other words, by concentrating on the emotional aspects of caring, the tendency is to accept the dual roles of men as achievers and women as carers as being normative. Overlooked are the ways in which these gender roles serve to legitimate the subordination of women in the wider society (Graham, 1983:21; Walker, 1992:38). Domestic Labour Theories Finley's (1989) study on adult children caregivers tests four exploratory models used to explain gender differences in the levels of domestic labour. The objective was to determine their explanatory power when applied to gender differences in caregiving to elderly mothers. The four models are the time-available hypothesis, the socialization/ideology hypothesis, the specialization-of-tasks hypothesis and the external resources hypothesis (see also Condran and Bode, 1982; Ross, 1987). The time-available hypothesis posits that women have more time available than men to care for the elderly because they have fewer competing demands, such as employment 15 (Finley, 1989:79; see also Horowitz, 1985a:612). According to this proposition, employed women and/or women with other role conflicts/competing demands (e.g., marriage, young children) are expected to spend less time in caregiving activities (Finley, 1989:79). However, as discussed earlier, most research has found that daughters' employment status does not significantly influence the amount of time spent in caregiving (Brody and Schnoover, 1986:378; Seccombe, 1992:171; Stoller, 1983:856). Instead, employment status is more likely to affect the amount of care sons provide (Stoller, 1983:856). Research on other "role conflicts" found that married adult children (both sons and daughters) provide less help to their parents compared to unmarried children (Lang and Brody, 1983:197; Stoller, 1983:856). Finley tested this hypothesis using the concept of role conflict rather than total time available. She developed a scale that measured whether respondents felt that caring for a parent interfered with their other roles such as work or meeting family responsibilities. Finley found that, even when role conflict was statistically controlled, adult daughters were more involved in caregiving than were sons. The socialization/ideological hypothesis proposes that the social division of labour stems from gender role beliefs learned through socialization processes (Condran and Bode, 1982:422; Finley, 1989:80; Montgomery and Datwyler, 1990:37; Walker, 1992:39). This hypothesis bears some similarity to what Walker (1992:38) refers to as the "sociological perspectives." Part of the socialization process involves men and women internalizing the ideology of various gender role dichotomies such as the "instrumental" and "expressive" dichotomy. The social expectation is that women are best suited for work inside the home 16 as it enables them to demonstrate their expressive qualities. Meanwhile, it is appropriate for men to adopt instrumental responsibilities such as providing financial support. This enables men to more freely delegate caregiving tasks to women who are expected to perform them (Montgomery and Datwyler, 1990:38; Walker, 1992:39-40). This proposition therefore suggests that, because of socialization, daughters feel greater obligation to tend to family concerns such as caring for parents than do sons. Finley tested this idea by creating a series of indicators that measured filial obligation about responsibilities towards parents. She found no gender differences for this variable. Montgomery and Datwyler (1990:37) suggest that, despite studies that show that both sons and daughters feel obligated to care for their parents (Finley, 1989:84; Montgomery and Kamo, 1989:226), there may be differences in how sons and daughters translate that feeling of obligation into action. This, in itself, may be a product of socialization. When feelings of obligation were taken into account in Finley's study, daughters were more likely to provide care. The third hypothesis is the specialization-of-tasks hypothesis. This perspective suggests that to maximize the overall well-being of the family, men and women are delegated different tasks. For example, men are the breadwinners because they tend to earn more, while women look after the household chores. As these specialized tasks complement each other, a kind of equilibrium or parity is achieved (Condran and Bode, 1982:81; Finley, 1989:81). Finley suggests that task specialization may come about through men being care managers, delegating caring chores to others, while women are the care providers (see Archbold, 1983). There is some support for this task-specialization hypothesis. As discussed earlier, 17 research has found that women provide help with the most tasks, especially those involving household chores and personal assistance. Some research has also found that sons specialize in tasks such as home repairs and heavy home chores such as snow shovelling (Horowitz, 1985a:614; Kaden and McDaniel, 1990:17; Montgomery, 1992:74; Montgomery and Kamo, 1989:218). Finley hypothesized that if sons were more likely to be care managers, or if they were more likely to perform certain types of tasks compared to daughters, the specialization-of-tasks hypothesis would be supported. She found that there was no category where sons helped significantly more than daughters. Instead, daughters gave significantly more help with activities of daily living and cognitive assistance. Finley also found that sons were no more likely than daughters to be care managers. The final hypothesis Finley tests is the external resources hypothesis. It suggests that external resources emanating from access to income and education help to decide power dynamics and consequentiy the division of labour within the family (Condran and Bode, 1982:422; Finley, 1989:80; Ross, 1987:819). Possession of these resources enables some family members to delegate caregiving tasks to others (Montgomery, 1992:72). This hypothesis implies that having greater amounts of employment income results in the individual providing less care. Research findings have not been consistent in supporting this hypothesis. There is evidence that employed daughters and daughters with more income are able to purchase assistance (Archbold, 1983:43; Brody and Schnoover, 1986:378); however, research also shows that employment does not dramatically affect the amount of care daughters give (Brody and Schnoover, 1986:378; Stoller, 1983:856). Finley found that although there were gender differences in employment status and education levels, when these 18 were statistically controlled, daughters still provided more care than sons. Feminist Perspectives4 Recent theoretical work by feminists has provided yet another set of views in explaining gender differences in caregiving. Feminist perspectives, in particular those that adopt a political economy approach, share the contention that female caregiving is a socially constructed role that serves to legitimate the division of gender roles in society (Aronson, 1985:116; Graham, 1983:25; Hooyman, 1992:182; Walker, 1992:41). Supporters of this view argue that "caring has been associated with women and seen in a psychological focus, rather than as an expression of political and social relations" (Aronson, 1985:116). The gendered pattern of caregiving has less to do with women identifying this role as part of their identity, and more to do with the needs of the wider society and the labour that women are expected to perform: an understanding of caring requires a fundamental assessment of both the institutions of caring (the family, the community, the state) and the conditions to which they give rise (dependency, poverty, powerlessness). In this reassessment, caring emerges not so much as an expression of women's natural feelings of compassion and connectedness, as the psychological analyses suggest, but as an expression of women's position within a particular kind of society in which the twin forces of capitalism and patriarchy are at work. Caring, it appears, describes more than the universal feelings women have: it describes the specific kind of labour they perform in our society (Graham, 1983:25). The state, it is pointed out, has a vested interest in maintaining the status quo where the family, and women in particular, are the main source of support for the elderly 4For the sake of brevity, I summarize the general perspectives of feminists. I do not provide details on the debates within feminism. 19 (Hooyman, 1990:223; Walker, 1983:120; Walker, 1991:102). First, it is argued that, as the state is patriarchal, there is little incentive to alter the traditional gendered relations of caregiving (Walker, 1983:122; Walker, 1991:102). Second, proponents of the status quo argue that state intervention would serve to weaken family ties, and in many cases substitute for the family (Hooyman, 1992:184; Walker, 1991:102). However, critics contend that although family (female) care is idealized as being superior to assistance provided by the formal sector, the state is more concerned with the money it saves in institutional and health services costs. The quality of care given appears to be a secondary concern (Hooyman, 1992:185). Many feminists have long argued that unpaid domestic labour such as caregiving is essential to the economy, yet because it has low status in relation to the labour market, it goes unrecognized and unrewarded {lbid.22%; Aronson, 1985:118; Graham, 1983:27; Walker, 1983:123; Walker, 1992:43). It is estimated that family care saves the state millions of dollars in health care service expenditures (Abel, 1986:481-482; Hooyman, 1990:223; Walker, 1991:102). Aronson states that compared to paid service for elderly parents, "family care is portrayed as a cheap and attractive solution" (1985:116). However, what the state saves economically is often borne by female caregivers physically, emotionally, and financially (Arber and Ginn 1992:622; Aronson, 1985:118; Hooyman, 1992:187; National Advisory Council on Aging, 1990b: 14-15; Walker, 1992:42). It has also been suggested that the burden particularly affects economically disadvantaged women since they do not have the financial resources to help alleviate their caregiving duties (Archbold, 1983; Glazer, 1990:488; Walker, 1983:124). 20 The policy of community care, as argued by some feminists, serves as an example of the state perpetuating the ideology of familism, that is, maintaining that the family provides the primary care of the elderly (Aronson, 1991:142; Glazer, 1990:486; Walker, 1991:104). As women are perceived as naturally being carers and nurturers, family care really refers to 'female' care: "Today social policies, such as community care, reflect and reinforce the ideology of familism by assuming that the family is necessarily the right location for the care of older relatives and that, within it, female kin are the most appropriate carers" (Walker, 1991:105). The ideology is so pervasive that women who decide to seek outside assistance often have feelings of guilt and inadequacy (Aronson, 1991:162; Hooyman, 1990:230-231). Women who choose to fulfil career aspirations and thus provide less care must often deal with being labelled selfish and uncaring (Aronson, 1991:162; Aronson, 1986:53; Hooyman, 1990:230-231). The latter part of this chapter has reviewed the psychological, domestic labour and feminist theories of why more women than men care for the elderly. To summarize, the psychological/individual perspectives suggest that women have an inherent propensity to care for others, and this constitutes a large part of their self-identity as women. The domestic labour theories constitute four propositions which suggest that women provide more care because they either have fewer competing demands and hence more available time; they are socialized into adopting and demonstrating expressive, caring qualities; they are better able to specialize in providing caregiving duties, especially those that require hands-on kinds of care; or they have less access to resources such as high employment income when compared to men. Finally, a feminist/political economy approach argues that female caregiving is a 21 socially constructed role which benefits men and the state in particular as the care women provide is often unpaid, unnoticed and taken for granted. These theories have been reviewed to set the context of why there are gender differences in caregiving. Of particular interest to this thesis are the ideas from the feminist/political economy perspective that economically disadvantaged women bear more of the caregiving burden. As will be fiirther explained shortly, these concepts will be used to as a rationale for testing whether there are interaction effects between class and gender in helping behaviour. In addition, ideas stemming from the time-available, or competing demands hypothesis will be examined to guide the upcoming analyses that explore the characteristics of those sons and daughters who do provide care. The gender Inequality in caregiving has been well documented, yet Abel states, "In a society riven by divisions of gender, class and race, the experience of caregiving will differ dramatically among different groups" (1990:71). In response to this, this study will examine how caregiving and care-receiving are organized along both gender and class lines. This chapter has focused on gender. Chapter Three takes up the issue of class and caring. As this thesis seeks primarily to determine if class differences in caregiving are the same for both men and women, it is important to understand both gender and class influences and why they may exist in order to determine if and why there is an interaction between the two. 22 CHAPTER THREE CLASS AND CARING Recently, researchers have lamented the fact that not enough attention has been paid to class differences in caregiving/care-receiving (Abel, 1990:73; Arber and Ginn, 1992:619; Horowitz, 1992:139; Rosenthal, 1987:326). Arber and Ginn state that "It is remarkable that within this extensive [caregiving] literature there is little mention of class divisions. ...the caring literature.. .has emphasised gender inequalities while neglecting class differences in the location and social relations of caring" (1992:619). Research on the relationships between class and receipt of social support shows that there are class variations with respect to availability and proximity of support networks, and types of support provided by children (Glazer, 1990; McPherson, 1990; Rosenthal, 1987; Taylor and Ford, 1980). For example, working class elderly parents tend to live closer to their children compared to middle and upper class parents (Glazer, 1990:487-488; McPherson, 1990:344; Rosenthal, 1987:318). Research has shown that proximity between adult child and parent influences the likelihood of caregiving/care receiving (Dwyer and Coward, 1991:S265; Glazer, 1990:487; Rosenthal, 1987:318) as well as the style of contact (Dewit, Wister and Burch, 1988:75). Similarly, there are class differences in the availability of different kinds of social support. Taylor and Ford (1983:192) found in their study that elderly Britons from a working class background are more likely to have either children or siblings living locally than the middle class elderly. These findings may reflect the tendency for working class families to 23 have a greater number of offspring. However, Taylor and Ford also found that elderly from a middle class background have a greater number of close friends living close-by than those from the working class. When the variables of age, sex and class were included in a multivariate analysis, young-old working class females were shown to have the most children living close-by, while the old-old middle class females had the fewest. This latter group tends to replace family social support with support from close friends. Taylor and Ford state, "thus their disadvantage in available family support is offset by their advantage in having close friends. In effect, there is a form of substitution, friends for family members" (1983:195). It is recognized that "available support" does not necessarily translate into receipt of care, however, it can be assumed that there is a relationship between these two variables. A few researchers have explored the link between class and the use of paid help. In another British study, Caldock found that although paid help is not used often, "the most important and significant correlation in determining the likelihood of employing private help is social class..." (1992:105). Use of paid help was found to predominate among the elderly from the upper or professional classes {Ibid: 105; Abel, 1990:74; Hooyman, 1992:186; Victor and Evandrou, 1987:264). In a Canadian study, Aronson notes that only elderly women possessing economic resources have the option of purchasing private support services (1991:161). Rosenthal (1987:326), and Rosenthal and Gladstone (1993:127) report that the elderly from the working class tend to receive more assistance from their offspring than the middle class elderly. They also note that class appears to be associated with different kinds of assistance. For example, working class families support each other by providing services 24 while those from the middle class offer financial assistance or else provide help through gifts (Rosenthal and Gladstone, 1993:127). Most of the aforementioned studies focus upon the care recipient. Few researchers have examined the class background of the caregiver and how this affects his/her caregiving experiences. One exception is Arber and Ginn's (1992) study of older Britons. They suggest that whether a frail or disabled person needs informal assistance is in part determined by his/her class position. How one provides the care is also conditioned by one's background. Arber and Ginn coin the term "leverage" (1992:621) to describe either the caregiver's or recipient's possession of financial, material and cultural resources. This leverage can influence the ability of either party to purchase extra assistance, aids, adaptations, and/or modified housing. This affects the extent to which the elderly person becomes dependent on the caregiver. Meanwhile, those caregivers who have more leverage may not only be able to purchase assistance, but may also have sufficient space to allow an elderly or infirm relative to move in with them. Therefore, "carers with financial resources have more opportunity to manage caring in a way which suits both themselves and their relatives" (Ibid:623). Possession of these resources also enables the caregiver to negotiate with health and social welfare agents for the most appropriate support services (see also Archbold, 1983). Studies previous to Arber and Ginn's also report that carers from higher socio-economic backgrounds are more able to purchase various market services to help them in their caregiving tasks (Abel, 1990:74; Archbold, 1983:41; Kinnear and Graycar, 1984:20). Arber and Ginn's notion of leverage is similar in some ways to the external resources hypothesis utilized in Finley's research (1989). However, Arber and Ginn provide a more 25 useful conceptualization of how possession of the three kinds of resources influence the caring process. For example, they emphasize the effect of cultural resources on being able to utilize social services more skilfully. This is in addition to their more detailed discussion of the impact of financial and material resources. Further, they discuss how leverage can affect both the giver and receiver of care. Using their idea of leverage, Arber and Ginn examined how informal care varies by class of the caregiver. In addition, they test a separate but related hypothesis that, given the relationship between class and health, working class caregivers will need to provide informal assistance to their parents at an earlier age than those from a middle class background. Arber and Ginn report that class differences emerge when care is provided on a co-resident basis versus care provided outside of the home. Working class caregivers were slightly more likely to provide co-resident care, which is more time-consuming and demanding, compared to middle class carers. Care provided to those outside the home was more often provided by middle class men and women compared to their working class counterparts. Arber and Ginn conclude that leverage may influence caregiving since the middle class carers probably had more resources to provide care "at a distance" (1992:168) than did working class caregivers. They also found very weak evidence that working class carers who provide extra-resident care tend to provide it at an earlier age than middle class carers. Social class can also influence the type of caregiving role which one assumes. Although she did not use the term class or leverage, Archbold's (1983) study on women who were caring for their parents found that socio-economic status influences whether one 26 becomes a care provider or a care manager. The former performs the tasks necessary, whereas the latter manages the provision of caregiving services performed by others. Archbold found that care managers came from higher socio-economic backgrounds and were employed full time in professional positions. Not only were care managers better able to afford outside services but, An indirect benefit of the higher socioeconomic status of the care managers was contact with a broad range of social supports. Such supports included physicians, lawyers, social workers and nurses. As a result, knowledgeable professional advice was available to many managers "unofficially" and at no cost. Providers, on the other hand, had significantly fewer social supports in general. None had unofficial connections with the health or legal systems. Providers made contact with these systems without the assistance of an "insider". Thus, ease of access and knowledge of options were limited (Archbold, 1983:41). This relates to Arber and Ginn's notion of leverage, and the associated cultural resources that higher classes may possess, to ease the caregiving process. While all of these studies explore different ways in which class or socioeconomic status influences the giving or receiving of care, few provide a theoretical definition of class. A precise definition of class is contested in the literature as it is dependent on the researcher's theoretical perspective (Dale, Gilbert and Arber, 1985:384; Grabb, 1990:6; Victor and Evandrou, 1987:253). Victor and Evandrou state that "any discussion of social class, because of the varying perspectives, can soon become complex, confined to the consideration of technical issues and therefore fail to reach a consensus" (1987:254). A detailed description of the history and meaning of "class" will not be covered here, however, a brief outline of the major approaches to the definition of class will be highlighted. The stratification position groups people according to some objective signifier of economic rank such as income, education, or occupation, and places them along a continuum 27 (Grabb, 1990:108; see also Dale etal., 1985:384; Victor and Evandrou, 1987:254). Classes, in this view are not necessarily groups of real people, but instead are statistical aggregates, often predetermined by the researcher (Grabb, 1990:109). Marxian theory views classes as defined in relation to the mode of production. Those who own property, the bourgeoisie, are able to exercise power and control over the proletariat, those without property (Grabb, 1990:22; Marx and Engels, [1846] 1978:179; [1846] 1978:473). Weber's theory, alternatively, focuses more on the distribution of valued resources, where some groups have greater access to these than others (Grabb, 1990:54). Classes are primarily economic entities, determined by one's position in relation to the labour market (Weber, 1922:927,928; see also Dale et al., 1985:384; Victor and Evandrou, 1987:254). Individuals within the same economic position form aggregates and share similar economic interests and life chances within this relational continuum (Weber, 1922:927; see also Grabb, 1990:54). Weber further separates the term class, from the term "social class." The latter refers to an economic class that develops group consciousness and a sense of a shared economic position (Grabb, 1990:55; 1988:3). For this study, I will opt for a Weberian conceptualization of class, whereby groups of people in similar economic situations in relation to the market have possession of, or access to, similar resources. This results in different life chances for the different groups (Grabb, 1988:4; Dale et al., 1985:387). This relates to Arber and Ginn's notion of leverage. According to these authors, different classes have access to different financial, material and cultural resources. These can influence the caregiving relationship by facilitating the process of support for those who possess such resources. Similarly, the elderly who have leverage 28 do not need to be as dependent on their children. As will be described in Chapter Five, class will be operationalized according to one's level of education. 29 CHAPTER FOUR GENDER, CLASS AND CARING Arber and Ginn (1992:625) contend that we know little about the interrelationships between gender and class and how they intersect with caring. This seems to be an underresearched area for there are few studies that focus specifically on how class and gender interact. Arber and Ginn's study suggests that among people providing co-resident care, the class gradient is stronger for men than for women as more working class men gave help. However, their evidence is rather weak. Several researchers, including some proponents of the feminist/political economy perspective discussed earlier, suggest that working class female caregivers are more disadvantaged than men and women from other classes (Abel, 1986:483; Archbold, 1983:41; Glazer, 1990:487; Walker, 1983:124). For example, Walker states that "it is the daughters of working-class elderly people who are bearing the brunt of informal care in the community" (1983:124). Glazer (1990:487) also reports that there are class differences in how women cope with caregiving. Upper class women are not as likely to provide continuous care to their parents compared to working class women. Upper class women are also able to hire necessary outside help such as attendants (see also Archbold, 1983:44; Kinnear and Graycar, 1984:20). From the perspective of those who are likely to receive care, Aronson reports that elderly women from higher socio-economic backgrounds are more able to purchase assistance (1991:149,161). Taylor and Ford (1983:195) also found that, in terms of available family 30 social support, older middle class women are more disadvantaged when compared to middle class men, and men and women from a working class background. Other researchers have implied that sons have more options available to either purchase services, or delegate caregiving tasks to their sisters and/or wives, thereby providing less care themselves (Montgomery and Kamo, 1989:222,224). Since men are more likely to be perceived as being the primary source of family income, this helps fortify the social norms of "men as provider, women as carer" roles. Therefore, upper class men may provide more financial support than other gender/class groups. By exploring the links between gender, class and caring, I hope to determine whether there are class differences among men and women who help their parents. The literature shows that women do the bulk of the caregiving, but there may be important differences among women in different economic situations. The literature also suggests that working class caregivers do not have the resources to facilitate their caregiving chores. This is particularly emphasized in Arber and Ginn's notion of leverage. Therefore, women who are poor may face the greatest caregiving burdens compared to men, and women of different classes. Arber and Ginn state that "current policies which shift a greater responsibility for the care of disabled and elderly people on to relatives are likely to have a disproportionate effect on women and on the working class, compounding existing inequalities in informal care" (1992:631). Women from disadvantaged backgrounds may face the double jeopardy of lacking the leverage to displace some of their caregiving duties. The same questions can be applied to those who receive care. Taylor and Ford claim that "...in most discussions of the elderly, the class dimension is ignored and all elderly 31 women are held to be disadvantaged" (1983:196). The literature implies that poor elderly women, and poor widows in particular, may be the most likely group to rely on support from their children; they do not have the material and/or cultural resources to help alleviate their condition due to their functional limitations (Hooyman, 1992:189). Central Propositions and Research Questions Consequently, based on the review of the literature, I anticipate finding the following with respect to adult children giving help to their parents and elderly parents receiving help from their children. 1) There are gender differences in helping behaviour. The literature suggests that daughters are more likely than sons to help their parents with tasks such as housework, personal care and transportation. Sons, meanwhile, tend to help more with home maintenance as well as give their parents financial support. This study will test these propositions to determine if the findings support previous research outcomes. 2) There are class differences in helping behaviour. Adult children from a working class background will be more likely to help their parents than adult children from a middle or upper class background, as suggested by the literature. In contrast to the literature on gender and caregiving, there is less emphasis on the relationship between class and help with specific tasks. Therefore, I propose that for tasks 32 such as housework, maintenance and personal care, working class adult children are the most likely to give help compared to middle and upper class adult children. However, on other tasks such as financial support and transportation, there is a greater probability that upper class adult children will provide these kinds of help. 3) There are interaction effects between gender and class in helping behaviour. I suspect these interaction effects will differ according to the different tasks involved. To elaborate, there will be an interaction effect between gender and class for help with housework and personal care as working class daughters will provide such help more that any other class/gender group. However, for assistance with transportation, the interaction is expected to occur with daughters from the upper class as more women tend to help with transportation than men, and the upper classes tend to have more resources so they are best able to afford a vehicle. This combined will lead to an interaction effect. Working class sons are expected to be the most likely to provide help with home maintenance, while upper class sons will be the most likely to give financial support. 4) Receipt of help will vary according to gender. Research has found that elderly mothers tend to receive more help from their children than do fathers. Using this as a foundation, I expect to find that elderly mothers will be more likely than fathers to receive help with the following tasks: housework, transportation, maintenance, personal care and financial support. 33 5) Receipt of help will vary according to class status. Elderly parents from a working class background will be more likely to receive help from their children. There has been little research on the kinds of assistance the elderly from different classes receive. Therefore, I anticipate that more working class elderly will receive help with all tasks compared to the elderly from the middle and upper class. 6) There are interaction effects between gender and class among elderly parents who receive help. Elderly women are one of the poorest groups in Canadian society. I therefore expect to find an interaction effect where elderly working class mothers are the most likely to receive help from their children across the majority of tasks. These are the main research propositions this study will investigate. On a lesser scale, this research will also consist of exploratory analyses of what are some of the characteristics of those men and women who do help, or receive help. For example, the time-available hypothesis examined in Finley's research (1989), as well as other studies in the literature suggests that competing demands such as marriage and employment may limit the probability or amount of help that sons and daughters provide. Poor health may also negatively affect one's ability to give help. Thus it is expected that those adult children who are either married, working, or have health problems are less likely to provide assistance. Another influencing factor may be distance. The expectation is that children who live closer to their parents are more likely to provide help. In addition, separate analyses will be conducted for 34 adult sons and daughters of different classes who help their parents to explore the effects of work status and distance between caregiver and the care recipient. For example, are sons and daughters who work as likely to give help regardless of their class background? Are sons and daughters from different classes more likely to help their parents if they live close-by than if they live farther away? With regards to elderly parents, I anticipate that elderly parents will be more likely to say they received help from their children if they have at least one child living close-by compared to those who have more distant offspring. In addition, elderly parents who are married will be less likely to receive help from their children than those who are not married (i.e., widows/widowers). Further analyses will also be done to examine what are the effects of distance from children on receipt of help according to the recipient's class background. 35 CHAPTER FIVE METHODOLOGY This study uses data from cycle five of the 1990 General Social Survey (Statistics Canada, 1991). The annual General Social Survey, implemented by Statistics Canada in 1985, has two main objectives: to collect data on Canadian social trends and gather information on specific policy issues. Every year the survey focuses on a key theme or issue of interest (Ibid.I). For example, the first survey in 1985 focused on health and social support. The following year time use, language and the social mobility of Canadians were the key themes. In 1990, cycle five focused on the issue of family and friends. The target population of the 1990 survey included all Canadians aged fifteen years or older, with the exception of residents from the Yukon and Northwest Territories, and those living in institutions. A total of 18,325 households were contacted using Random Digit Dialling sampling techniques. This group also included a sample of people aged sixty-five and over drawn from households used in the Labour Force Survey. Seventeen and a half percent (3,206) of the 18,325 households were non-responding households5. From the remaining 15,119 responding households, one person was randomly selected from eligible household members and an interview was attempted. A final sample size of 13,495 was obtained resulting in a response rate of 75.8%. Al l interviews were conducted by telephone by enumerators trained by Statistics Canada. The interviews were completed between January 5This includes households that refused to participate, households where because of language difficulties or other problems a response could not be obtained, and households for whom no contact was made throughout the survey timeframe (Statistics Canada, 1991:26). 36 and March 1990. The data underwent numerous weighting procedures by Statistics Canada to adjust for non-response rates, multiple telephone households, personal weight calculations, external totals and sampling procedures that varied for different provinces. The 1990 General Social Survey asked respondents a series of questions on whether they gave unpaid help to, or received unpaid help from someone outside their household within twelve months of the interview. The survey focused on five tasks: help with housework such as cooking, sewing or cleaning; house maintenance or outside work such as repairs, painting, carpentry, lawn mowing or shovelling snow; transportation; personal care such as bathing or dressing; and financial support. No questions were asked with regards to help given to someone within the household. Therefore, this thesis concentrates on help given to and received from someone outside the household. Since this study examines those who gave help, as well as received help, two different, unrelated samples from the survey are examined. One sample consists of people aged forty-five to sixty-nine who have at least one parent alive, who neither co-resides with the respondent nor lives in an institution. This sample of 1,311 respondents will be used to determine whether there are gender and class differences in helping behaviour with each of the aforementioned tasks. Because the 1990 General Social Survey examines help given outside the household, people who live with a parent have been excluded from the analysis. Those whose parent(s) live in an institution have also been excluded. Much of the caregiving literature investigates caregiving to non-institutionalized parents (see Horowitz, 1985a; Kaden and McDaniel, 1990; Montgomery and Kamo, 1989; Stoller, 1983). One of the primary dichotomous dependent variables in this research will be whether 37 the respondent helped their parent(s) with each of the specified tasks. The General Social Survey did not ask if the parent was either the mother or father, nor is it clear if this includes parents-in-law6. Consequently, the findings of this study will be limited to those who indicated they have helped at least one parent. The second sample consists of people aged sixty-five and over who have children7 and answered yes to the question "Because of a long-term physical condition, mental condition or health problem, are you limited in the kind or amount of activity you can do at home, at work, at school or in other activities such as transportation or leisure?" (Statistics Canada, 1991:16). Once again, since the survey examines help received from outside the household, this sample will only include elderly parents who do not reside with any of their children. The other main dichotomous dependent variable in this study will be whether or not the respondent in this second sample indicated that he/she received assistance with each task from either their son or daughter. The son and daughter categories will be combined to allow for analyses based on help given by children. The sample size for this group is 760 elderly parents. 6Studies on caregiving do not always examine help given to in-laws. Studies which do include in-laws rarely analyze caregiving patterns for help given to parents versus help, given to in-laws. One exception is Merrill's work (1993). 7To determine whether the respondent has children, a variable was constructed from three questions asking respondents if they had ever raised either a natural child, adopted child or stepchild. 38 Measurement of Independent Variables The first key independent variable is gender, a dichotomous variable coded one when the helper was female, and zero when the helper was male. For this study, education level is selected as a measure of class, the second key independent variable. In this thesis, I have defined class in a Weberian perspective. I have defined class in relation to different people's access to valued goods and financial, material and cultural resources as suggested by Arber and Ginn's (1992) notion of leverage. It would seem appropriate to operationalize class according to occupation. However, this is problematic in this study for a few reasons. As I am examining both the givers and receivers of help, it is preferable to use the same measure of class for both groups. The General Social Survey's questions on occupation may not be applicable for those over the age of sixty-five as they are likely retired. For those few who are still working, the General Social Survey only asks what work the person was doing in the last twelve months. It is not possible to obtain information on the person's main occupation during their majority of their working years (see Strain and Payne, 1992:38). The respondent's level of income (personal or household) is also not a good measure of class for this study. First, for both groups the percentage of missing values is high. Second, for many elderly, their level of income may not reflect their level of wealth (e.g., property ownership, savings, stocks and bonds). One of the advantages of using education level to measure social class is that it is a relatively stable measure. Strieb notes that education and training are "measurable attributes that cannot be taken away from a person, and can be used as a precise indicator of rank. ... 39 the person with higher educational attainments retains certain objective advantages that can never be removed" (1985:345). Strain and Payne (1992) also use social class as one of their independent variables. They utilize the 1985 General Social Survey to study social networks and patterns of social interaction among ever-single and separated/divorced elderly Canadians aged fifty-five and over. They comment that "education is the 'best' measure of social class. Major occupation throughout life is not available for the majority of the sample (75 %) as only individuals who worked in the last five years were asked about their work histories and income data are missing for 25 per cent of the sample" (1992:38). There is a known association between education, occupation and income. Therefore, taking all these considerations into account, class will be operationalized according to one's level of education. Education level is coded as an ordinal variable with three categories. Low education level represents those who possess some secondary schooling or less; middle education level is operationalized as those who have completed secondary school and have some post secondary education. Finally, those who have a post-secondary degree or certificate are categorized as having a high level of education. For the sample of adult children, gender and education are each crosstabulated with whether one gave help to his/her parent(s). For the sample of elderly parents, gender and education are also separately crosstabulated with whether one received help with each task. A dichotomous variable was also created combining all the tasks to measure whether one gave/received help on at least one of the tasks. A key research question is whether there are interaction effects between gender and 40 education for those who help their parents, or those parents who receive help from their children. To test this, education was once again crosstabulated for each task while statistically controlling for gender. Further crosstabular analyses are conducted among subgroups, such as sons and daughters who help their parents and elderly mothers and fathers who receive help. Analyses are also done for sons and daughters, and mothers and fathers from the three different education groups. The objective is to examine what are some of the characteristics of these women and men. For adult helpers the additional variables examined are marital status, employment status, health status, and distance between parent and child. For the recipients of help, only the effects of distance are examined. A subsequent analysis involves logistic regression of helping/receipt of help to further investigate the effects of the key independent variables when other possible predictors are statistically controlled. Logistic regression is a multivariate technique that permits an analysis that is similar to multiple regression (Walsh, 1987:179). This method is used to predict a dichotomous dependent variable from a set of independent variables (Wister and Gee, 1994:115). Walsh states, "No other technique will allow the researcher to analyze the effects of a set of independent variables on a dichotomous dependent variable (or a qualitative polytomous one) with such minimal statistical bias and loss of information" (1987:178). The relevant tables will show the logistic coefficient, its standard error and odds ratio. The logistic coefficient is the change in the log odds of either helping, or receiving help, associated with a one-unit change in the independent variable (Norusis, 1990b:B-43). The odds of an event occurring can be understood as being the ratio of the probability that it will 41 occur to the probability that it will not occur (Jbid:B-43). If a coefficient is positive, the odds of helping, or receiving help, are greater than one. If the coefficient is negative, then the odds of helping or receiving help are less than one but never reach zero (Wister and Gee, 1994:117). Goodness-of-fit chi square is provided for all the models, in addition to a statistical significance test. Large significance levels for chi square suggest that a model is not significantly different from a "perfect" model (Norusis, 1990b:B-46; Wister and Gee, 1994:117). The frequencies for the independent variables, as well an indicator of the reference categories for the logistic regression, are listed in Table 1 and Table 2. Measurement of variables used in the logistic regression analysis merit some discussion. Most of these variables are categorical, the majority being dichotomous. Respondent's age and age of parent are continuous variables coded in years. Education was coded as an indicator variable, with the reference category varying according to the task being examined. This study will also explore the possible effects of a series of competing demands on helping behaviour. These include marital status and work status. The literature implies that either working or being married may pull adult children away from caregiving responsibilities (Lang and Brody, 1983; Seccombe, 1992; Stoller, 1983). The respondent's work status was determined by the question "during the past 12 months what best describes your main activity?" (Statistics Canada, 1991:24). The category of interest here is "working at a job or business"8. 8 For explanation purposes, analyses in reference to the work status variable will use the terminology of working versus not working. It is recognized that the original Statistics Canada question did not explicitly ask this question, so the question on respondent's main activity in the last 12 months is the best measure for employment status. 42 Table 1 Frequencies and Coding for Independent Variables Dsed in Logistic Regression Analyses For Adult Children Who Help Their Parents Variable Name Description Percent [n] Gender Respondent's Gender 100.0 1311 0 Hales 48.7 639 1 Fenales 51.3 673 Education Respondent's Level of Educ. 100.0 1311 1 Low Educ. 33.8 443 2 Hiddle Educ. 29.9 392 3 High Educ. 34.4 451 9 Hissing 2.0 26 EducH Dummy Var/High Education 100.0 1311 0 Else 63.6 834 1 High Educ. 36.4 477 EducH Dummy Var/Middle Education 100.0 1311 0 Else 70.2 920 1 Hiddle Educ. 29.9 392 EducL Dummy Var/Low Education 100.0 1311 0 Else 66.2 868 1 Low Educ. 33.8 443 Work Respondent's Work Status 100.0 1311 0 Not Working 26.5 348 1 Working 71.9 943 9 Hissing 1.5 20 Marital Respondent's Marital Status 100.0 1311 0 Not Married 22.0 289 1 Harried 77.3 1013 9 Missing 0.7 9 Income Respondent's Personal Income Mean ($$) 27852.45 Standard Deviation ($$) 22045.31 43 Table 1 continued Variable Name Description Percent [n] Income Respondent's Personal Income 100.0 1311 0 $0to $28,000 66.2 868 1 $28,001 or more 33.8 443 9 Missing 23.0 299 RHealth Respondent's Health Status 100.0 1311 0 No Health Limitation 88.0 1154 1 Has Health Limitation 11.5 152 9 Missing 0.3 4 Age Respondent's Age Mean 51 Years Standard Deviation 5.3 Years Distance Distance From Parent(s) 100.0 1311 0 Lives >1 Hr.From Parent 43.7 573 1 Lives <1 Hr.From Parent 51.4 675 9 Missing 4.8 63 Agepar Age of Parent(s) Mean 78 Years Standard Deviation 6.3 Years ParLivArr Parental Living Arrangement 100.0 1311 1* Lives Alone/others only 48.2 632 2 Lives w. spouse/sp and others 31.0 407 3 Lives w. children/children 20.8 273 & others/children & spouse 9 Missing 0.2 2 WSpouse Dummy Var/Lives W. Spouse 100.0 1311 0 Else 69.0 905 1 Lives With Spouse 31.0 407 WChild Dummy Var/ Lives W. Children 100.0 1311 0 Else 78.9 1035 1 Lives With Children 20.8 273 * Reference Category 44 Table 2 Frequencies and Coding For Independent Variables Used in Logistic Regression Analyses: Parents Who Receive Help From Their Children Variable Name Description Percent [n] Gender Respondent's Gender 100.0 760 0 Hales 39.5 301 1 Females 60.5 460 Education Respondent's Level of Educ. 100.0 760 1 Low Educ. 58.8 447 2 Middle Educ. 23.1 176 3 High Educ. 15.1 115 9 Missing 2.9 22 EducH Dummy Var/High Education 100.0 760 0 Else 84.9 646 1 High Educ. 15.1 115 EducH Dummy Var/Middle Education 100.0 760 0 Else 76.8 584 1 Middle Educ. 23.2 176 EducL Dummy Var/Low Education 100.0 760 0 Else 38.3 291 1 Low Educ. 61.2 469 Marital Respondent's Marital Status 100.0 760 0 Not Married 38.9 296 1 Married 59.4 452 9 Missing 1.7 13 Age Respondent's Age Mean 73 Years Standard Deviation 5.1 Years Income Respondent's Personal Income Mean ($$) 13646.32 Standard Deviation ($$) 11620.29 Income Respondent's Personal Income 100.0 760 0 $0 to $13,800 (Mean) 82.3 626 1 Over $13,800 17.7 134 9 Missing 43.0 327 45 Table 2 continued Variable Name Description Percent [n] NumChild Respondent's I of Children 100.0 760 0 Only 1 Child Outside House 19.1 145 1 1+ Child Outside House 67.2 511 9 Hissing 1.9 15 Child/dist Distance From Children 100.0 760 0 Children Live > 1 Hr. Away 17.4 133 1 Children Live < 1 Hr. Away 79.1 601 9 Hissing 3.6 27 46 Health status of the adult child is examined to determine if it has an influence on the helping relationship, specifically as it relates to another kind of competing demand. In this study, health status was measured according to whether people said they were limited in the amount or kind of activity they could perform due to a physical or mental health condition. Parental living arrangement is another indicator variable included in the model for helping behaviour. The adult child's parent living alone is the reference category9. This is contrasted with two other categories. The first is the adult child's parent lives with a spouse, or spouse and others. The second is the parent lives with his/her children, not including the respondent10. Although a continuous variable, income was coded into a dichotomous variable where the reference category was between no income and the mean for the entire distribution. The reason for this is the income variable was highly skewed and contained a high percentage of missing data. Computing the distance variable for the adult children sample proved to be a more challenging task. Due to the nature of the question in the survey, as well as not being able to ascertain whether the adult child was helping the mother, father or both, it was Impossible to account for distance between adult child and elderly parents who were both alive, yet lived apart. As this frequency was small in relation to the other possibilities (e.g., parents live together, only one parent is alive) this is unlikely to make a large difference in the results. Distance was coded into a dichotomous variable, coded one if the adult child lived one hour 9Due to the small numbers involved, those who live with others only are included in this category. l0Included in this group are those whose parents live with their spouse and children. 47 or less away from his/her parent(s), and zero if he/she lived more than an hour away. There are a few additional differences for the analyses of those receiving help. For example, number of children was dichotomized and refers to whether the respondent had one child or more than one child living outside the household. The coding procedure for married elderly is the same as for the caregivers, however, the premise is different. For the elderly, being married may entail the individual is more apt to rely on support from the spouse than from children. Finally, distance is measured according to whether the respondent had at least one child over the age of fourteen living within 100 kilometres or not. Most of the variables had missing data. Logistic regression only permits the listwise11 deletion of missing data (Wister and Gee, 1994:117). To account for this, missing data are coded into the modal category of categorical variables. Where necessary, these variables are then coded into a binary variable format. Missing values for continuous variables were coded into their mean value. Where variables had to be constructed (e.g., distance), the desired categories were first developed, then the missing values were coded into the modal category. These are methodologically acceptable methods of recoding missing data. Al l analyses reported here were conducted using SPSS/PC Version 4.0. Most of the analyses use weighted and scaled data. The General Social Survey utilizes a disproportionate stratified sampling scheme. It is necessary therefore to weight the data to enable the sample to be representative of the population from where it came (Babbie, 1987:G8; Statistics "Listwise deletion of missing data means that a case is eliminated if any variable in the list has a missing value. That is, the researcher, for analysis purposes, is able to keep the variables of interest but exclude those cases with missing values (Norusis, 1990a:B-19). 48 Canada, 1991:28; Vogt, 1993:245). The data here have been scaled back from the population count to reflect the sample size12. The exception is for analyses utilizing logistic regression. Analyses here involve unweighted data. Limitations of the Study There are a few limitations in this study that should be pointed out. First, Statistics Canada does not include a category of "parent-in-law or son/daughter-in-law" among its categories of helpers/receivers of help. This is lamentable since research has suggested that daughters-in-law also provide assistance to their in-laws (Merrill, 1993). Second, the General Social Survey incorporates a cross-sectional methodological design. This makes it impossible to analyze patterns of helping/receipt of help over time (McDaniel and McKinnon, 1993:83). Third, the survey does not include those who are living in institutions. Thus, the sample of elderly parents will include those whose health has not declined to the point of needing institutionalization (Strain and Payne, 1992:39). Some of these people may have a strong enough social support network that precludes them from entering an institution (Ibid:39). Despite these shortcomings, this study can make several contributions to the social support/caring literature. 1 2As suggested by Statistics Canada, the weights on these data have been rescaled so that the average weight is one. This was done by dividing each weight by the overall average weight (Statistics Canada, 1991:29). 49 CHAPTER SIX FINDINGS -- ADULT CHILDREN WHO PROVIDE HELP TO THEIR PARENT(S) Gender Differences in Helping Behaviour Of the 1,311 adult children whose parent(s) neither lived in an institution, nor co-resided with the respondent, 49% were males, and 51% were females (see Table 3). Table 4 shows13 that more daughters than sons helped their parent(s) with most of the tasks surveyed. For example, 10.4% of daughters14 helped their parent(s) with household chores such as cooking, sewing or cleaning. This is almost five times greater than the 2.3% of sons who also provided these kinds of assistance (x2=34.30, p< .001). A larger proportion of daughters than sons also helped their parent(s) with personal care, 6.3% versus 1.6% (x2= 17.28, p<.001). These findings are not surprising; they support earlier studies that found that daughters are more likely than sons to assist with those tasks that require "hands on", routine assistance (Horowitz, 1985a:614; Stoller, 1983:856; Kaden and McDaniel, 1990:28; Stoller, 1990:234; Stone et al., 1987:622). More daughters (16.7%) than sons (9.1%) also provided their parent(s) transportation (x2= 16.04, p<.001). This finding confirms Horowitz (1985a), and Montgomery and Kamo's (1989) research, but is not consistent with the outcomes of Stone et al. (1987). The latter found there was no difference in the percentage of sons and daughters who helped with this task. 1 3 The distributions in this section pertain to those when no other variables are statistically controlled. 1 4 Unless otherwise specified, statistical significance for all crosstabular analyses involving two by two tables use the significance level of the continuity corrected chi-square. 50 Table 3 Gender and Level of Education Frequencies For Adult Children Sample Variable Gender Males Females Total Level of Education Low Education Middle Education High Education Missing Total Percent [n] 48.7 639 51.3 673 100.0 1311 33.8 443 29.9 392 34.4 451 2.0 26 100.0 1311 51 Table 4 Adult Children Who Help Their Parents By Gender of Adult Child Task Housework Personal care Maintenance Financial Support Transportation At Least 1 Task Male Percent I W 2.3 [15] 1.6 [10] 5.1 [32] 2.9 [18] 9.1 [58] 15.6 [100] n=639 Feraale Percent \ [n] 10.4 [70] 6.3 [42] 4.4 [30] 2.2 [15] 16.7 [113] 28.6 [193] n=673 34.30 *** 17.28 *** 0.16 n.s 0.43 n.s 16.04 *** 31.11 *** * p<.05; ** p<.01; *** p<.001 52 Table 4 also shows that there was no difference between sons and daughters in help given to parent(s) with house maintenance or outside work such as repairs, carpentry or lawn mowing (x 2=. 16, n.s). Previous research (Chappeli, 1992; Kaden and McDaniel, 1990) has found that these are the kinds of tasks that sons typically help with more than daughters. It was therefore expected that more sons would give these kinds of assistance. Sons and daughters also did not significantly differ in providing their parent(s) with financial support (x2=.43, n.s.). This is contrary to Rosenthal's earlier findings (1987:326). Montgomery and Kamo (1989:222) suggest that sons' feelings of obligation towards their parents do not usually materialize in their performing a personal service. Instead, sons convey their sense of obligation through acting as a care manager where they purchase services, or else delegate chores to others, most notably their wives and sisters. The rationale that more sons than daughters would provide financial support, is not, however, supported here. To determine whether there are gender differences in support of all types, a dichotomous variable was created that measured whether someone gave help on any of the five tasks mentioned (see Table 4). Considering the above results that daughters were more likely than sons to provide assistance across most chores, it is not surprising that more daughters (28.6%) than sons (15.6%) gave their parent(s) at least one type of help (x2=31.11, p<.001). What all of these findings demonstrate is that daughters are more likely to give help to their parent(s) across most tasks. There were no duties in which more sons gave assistance. In her research, Finley (1989:81) proposed that if sons were more likely than daughters to perform certain types of tasks, this would support the specialization-of-tasks hypothesis. Her 53 findings were similar to the ones reported here: sons do not specialize in providing certain types of assistance. Therefore, her hypothesis is also not supported here. Characteristics of Adult Children Who Help Their Parents The predominance of daughters being the most likely to help their parent(s) prompts one to ask what are some of the characteristics of these women? Once gender was statistically controlled, several variables were crosstabulated across the different kinds of helping tasks. These included: employment status, marital status, health status and distance between respondent and parent(s). Given the relatively small frequencies of adult children helping their parent(s), and that statistically controlling for women further decreases the sample size, crosstabular analysis partitions the cell size of tables to small numbers making statistical significance testing problematic15. Where appropriate, the results are meant to be interpreted as indicators of possible trends and their substantive significance is discussed. It is hoped that these discussions will lead to the development of further research questions for future analyses. As discussed earlier, it has been speculated in the caregiving literature that employment outside the home may serve as a competing demand on the adult child's available time and energy (Finley, 1989:80; Lang and Brody, 1983:194; Stoller, 1983:853). 15Chi-square tests are not appropriate when over 20% of the cells have an expected frequency of less than five (Norusis, 1991:270). In such instances where tables also contain two rows and two columns, Fisher's exact test is used. Norusis states: "Fisher's exact test evaluates the same hypothesis as the chi-square test, and its suitable for tables having two rows and two columns with small expected frequencies" (7M/:270-271). 54 The time-available hypothesis discussed in Finley's research (1989:80) suggests that employed women may not give as much help, if they give any at all, compared to men and non-employed women. Previous research found little difference exists between the amount of help employed daughters give versus unemployed daughters (Brody and Schnoover, 1986:378; Stoller, 1983:856). The results in Table 5 show that there are no statistically significant differences between adult daughters who work and those who do not on whether they give help on any of the tasks. It has also been suggested that marriage may act as another competing demand that impacts on an adult daughter's available caregiving time (Finley, 1989:80; Lang and Brody, 1983:194; Stoller, 1983:855). Thus, parents often turn to their unmarried children first for assistance. The findings in Table 5 do not support this association except for personal care. A statistically significant difference was found between married and unmarried daughters, where 7.9% of married daughters helped with this task compared to 1.5% of unmarried daughters (x2=7.85, p< Ol) 1 6 . This finding partially contradicts Lang and Brody's (1983:194) findings that marriage acts to pull women away from helping their parents. Perhaps unmarried women are more likely to work outside the home compared to married women. Married women may be home during the day and can provide this kind of help. Little research has examined the consequences of health status of caregivers. Table 5 indicates that there are no significant differences between daughters with a health problem who help their parent(s) compared to healthier daughters. It was expected that being in poor 16Caution must be exercised in interpreting this one significant finding. In one of every twenty tests a nonsignificant chance finding may be significant. 55 Table 5 Daughters Who Help Their Parents Work Status Marital Status Not Is Not Is Working Working X 2 Harried Married I1 Percent \ Percent 1 Percent 1 Percent % Task [n] [n] In] •tn] Housework 12.1 9.3 1.12 n.s. 9.7 10.7 .05 n.s. [35] [35] . [16] [54] Personal Care 7.8 5.2 1.39 n.s. 1.5 7.9 7.85 ** [23] [20] [2] [40] Transportation 16.0 17.5 .17 n.s. 18.2 16.4 .17 n.s. [47] [99] [34] [82] Maintenance 2.8 5.8 2.65 n.s. 5.8 4.0 .55 n.s. [8] [22] [10] [20] Financial Support 1.6 2.6 .41 n.s. 2.5 2.1 .00 n.s. [5] [10] [4] . [10] At Least 1 Task 29.8 28.2 .13 n.s 27.8 29.1 .04 n.s. [86] [106] [47] [146] n=290 n=377 n=168 n=501 Table 5 continued Health Limitation Distance From Parents No health Has health Lives >lHr. Lives <lHr. Limitations Limitations X 2 From Parents From Parents X2 Percent % Percent 1 Percent 1 Percent I Task • W [n] [n] tn] Housework 11.2 5.5 2.22 n.s. 3.9 16.3 24.61 *** [65] [5] [11] [58] Personal Care 6.8 3.3 1.08 n.s. 3.4 8.9 7.34 ** [39] [3] [10] [43] Transportation 17.1 15.0 .12 n.s. 3.6 27.1 62.35 *** [99] [14] [11] [96] Maintenance 4.8 2.1 .78 n.s. 2.9 5.8 2.44 n.s. [28] [2] [9] [21] Financial Support 1.9 4.1 .99 n.s. 3.7 0.4 7.91 ** [11] [4] [11] [1] At Least 1 Task 29.6 23.0 1.38 n.s. 12.1 41.5 66.77 *** [172] [21] [36] [148] n=579 n=91 n=293 n=356 * p<.05; ** p<.01; *** p<.001 56 health would restrict one's ability to provide assistance as poor health would act as another competing demand. Another variable that may influence the probability of helping is physical distance between the place of residence of caregivers and care-receivers. The literature suggests that distance between adult children and their parents influences the likelihood of providing help (Dwyer and Coward, 1991:S265; McPherson, 1990:347) as well as the style of contact (DeWit et al., 1988:75). Focusing just on women, the results in Table 5 show a statistically significant relationship between distance and help given to parent(s) on various tasks. Daughters whose parent(s) lived within an hour's drive were approximately four times more likely to help with housework compared to those whose parent(s) were farther away (16.3% and 3.9% respectively; x 2 = 24.61, p<001). They were also nearly eight times more likely to provide transportation (27.1% and 3.6% respectively; x 2 = 62.35, p< .001), and nearly three times more likely to help with personal care (8.9% and 3.4% respectively; x 2 = 7.34, p < .01). Forty-one percent of daughters who lived close to their parent(s) helped with at least one of the five tasks, compared to 12.1% of those whose parent(s) lived farther away (x2=66.77, p< .001). In contrast, 3.7% of daughters whose parent(s) lived further than an hour's distance provided financial support compared to less than one percent (0.4%) of daughters who lived closer (x2=7.91, p< .01). This latter finding can be explained by suggesting that daughters who are not in close proximity to their parents may feel that giving financial support is the most convenient method of providing tangible assistance. Distance was not significantly related to help given with maintenance. Although fewer sons gave assistance with the various tasks, it is important still to 57 know the characteristics of those who did help. Table 6 shows that, as was the case with daughters who provide help, there are no significant differences between sons who help and whether they work and or not17. Comparing married sons with those who are not married shows that there is also no difference between the two groups in who gives help on any of the tasks. As was discovered when statistically controlling for women, the results in Table 6 show that sons with a health hmitation are no more likely to help their parent(s) compared to sons without such limitations. One possibility for these findings is that there exists a dual kind of competing demands situation. Healthy sons and daughters may have more time and opportunity to participate in other activities that do not necessarily include helping their parents. These activities may include work or family responsibilities. Adult children with a health limitation may not be as able to help because of their health restrictions. Alternatively, healthy adult children may help their parents, but those with a health limitation may put the needs of their parents first, regardless of their own limitations (Horowitz, 1985b:219). Future research may want to explore to what extent men and women with health problems overcome them to provide help to their parents. This is an especially relevant issue given that the National Advisory Council on Aging (1990b: 9) estimates that anywhere between fifteen to twenty-five percent of adult children over the age of sixty-five have a parent who may require assistance (see also Chappeli, 1992:36). These adult children may themselves be starting to encounter some of their own health problems. Due to the low frequencies of sons who help their parent(s) with personal care, this task was not examined. 58 Table 6 Sons Who Help Their Parents Not is Not Is Working Working X2 Married Harried X2 Percent 1 Percent % Percent 1 Percent \ [n] [ii] [n] [n] Task Housework 2.8 2.3 .00 n.s. 3.1 2.2 .06 n.s. [2] [13] [4] [11] Transportation 6.9 9.6 .19 n.s. 5.9 10.0 1.51 n.s. [4] [54] [7] [58] Maintenance 3.1 5.4 .21 n.s. 3.4 5.5 .53 n.s. [2] [31] [4] [28] Financial Support 0 3.1 .92 n.s. 3.3 2.8 .00 n.s. [0] [18] [4] [15] At Least 1 Task 10.9 16.4 .83 n.s. 13.8 16.2 .27 n.s. [6] [93] [17] [83] n=58 n=566 •11=121 n=512 Table 6 continued Health Limitation Distance From Parents No health Has health Lives >lHr. Lives <lHr. Limitations Limitations From Parents From Parents X2 Task Housework 2.5 1.0 .08 n.s. 1.2 3.0 1.38 n.s. [14] [1] [3] [10] • Transportation 9.4 6.9 .17 n.s. 3.5 13.6 17.53 *** [54] [4] [10] [43] Maintenance 5.1 4.9 .00 n.s. 1.2 9.1 16.82 *** [29] [3] [3] [29] Financial Support 2.9 3.0 .00 n.s. 2.8 2.4 .00 n.s. [17] [2] [8] [8] At Least 1 Task 16.1 12.1 .39 n.s. 7.3 22.7 25.59 *** [93] [7] [21] [72] n=61 n=575 n=280 n=320 * p<.05; **p<.01; *** p<.001 59 Finally, 9.1 % of sons who lived within an hour's drive of their parent(s) helped with maintenance compared to 1.2% who lived further away (x2= 16.82, p<.001). Thirteen percent of sons who lived closer provided transportation compared to 3.5 % who resided more than an hour away (x2= 17.53, p< .001). Sons who lived closer were also more likely to help with at least one task, 22.7% versus 7.3% (x2=25.59, p< .001). However, they were not more likely to help with housework or financial support. For the latter task, it is interesting that for daughters there was a relationship between distance and provision of financial support, but no such relationship existed for sons. To summarize thus far, these results demonstrate that, similar to other research findings, more adult daughters than sons gave help to their parent(s) with housework, personal care and transportation. However, both daughters and sons helped with maintenance and financial support. When examining the characteristics of these sons and daughters who do help, distance from parent(s) emerges as an important variable. This supports McPherson's (1990:347) contention that distance is an important variable that helps determine the likelihood of helping one's parent. Daughters were more likely to help with tasks such as housework, personal care and transportation if they lived within an hour from their parent(s). They were more likely to give financial support, however, if they lived over an hour away. Sons who lived close-by were more likely to provide transportation, help with maintenance and help with at least one task. It was also found that consistent with some earlier research, employment did not serve to pull women away from helping their parents. However, it also did not seem to pull sons away either. Marital status and health status also had no significant effect on helping. The 60 lone exception was the finding that married women were more likely to help with personal care than non-married women. Although the small cell sizes for the various tables may have influenced tests of significance, these findings imply that these various characteristics do not necessarily act as "competing demands" for either gender, contrary to what is suggested by some researchers. Further work is needed to determine why this is the case. Education Differences in Helping Behaviour Towards Parents As suggested earlier, much of the focus of the caregiving research has concentrated on the role of gender. Dwyer and Coward state: ...to date, research has identified a large number of factors in addition to gender that influence who provides care, what kind of care is provided, and how much care is given. These other factors include characteristics of the caregiver, characteristics of the care recipient, and circumstances surrounding the provision of care. Many of these other influences, however, also covary with gender (1992:159). The other caregiver characteristic that is of interest to this thesis is social class, as measured by level of education. Within the sample, 34% had an education level of some secondary school or less, 30% had either completed secondary school or had some post-secondary education, while the remaining 34 % had received either a post-secondary degree or certificate (see Table 3). As expected, a statistically significant relationship was found between level of education and help given to parent(s) with household tasks (see Table 7; x 2 = 9.98, p< .01). More adult children with a "middle" level of education provided help with housework, 9.8%, while those with the least education were the least likely to provide help, 4.5 %. Interestingly, 6 1 Table 7 Adult Children Who Help Their Parents By Children's Level of Education Low Educ. Middle Educ. High Educ. V Percent 1 Percent 1 Percent % [n] [n] [n] Task Housework 4.5 9.8 6.0 9.98 ** [20] [38] [27] Personal Care 3.2 4.8 4.4 1.62 n.s [14] [19] [20] Transportation 9.1 15.2 15.4 9.86 ** [40] [60] [69] Maintenance 4.5 4.9 5.0 0.15 n.s [20] [19] [23] Financial Support 2.0 0.8 4.0 9.81 ** [9] [3] [18] At Least 1 Task 18.2 24.1 24.9 6.71 * [81] [94] [112] n=443 n=392 n=451 * p<.05; ** p<.01; *** p<.001 62 almost as many people with a high level of education gave help (6%) relative to this latter group. This rinding does not support the earlier proposition that those with the lowest level of education would be the most likely to help with housework. Instead, the results at first glance seem to support what Arber and Ginn (1992) found in their study when they examined people aged forty-five to sixty-four who gave help to someone outside of the home. They found some limited evidence that those from the middle and upper classes provided more extra-residential help compared to the working class. As mentioned earlier, they suggest that those who have access to more financial, material and cultural resources are more likely to provide extra-residential help because those with few of these resources are more apt to live with the care recipient. Their findings, however, should be interpreted with caution. Their sample consisted of those who may have given help to a sick, handicapped or elderly relative, friend or neighbour. Therefore, their sample was not as focused as the one used here, where the emphasis is on adult children with community-dwelling parents who do not co-reside with them. Including friends among the care recipients may have also confounded their findings for as Taylor and Ford (1983) found, the middle class tend to have more friends available to potentially provide social support. Further, "help" in Arber and Ginn's study refers to any kind of help instead of help with a specific task. One explanation why fewer highly educated adult children helped their parent(s) with housework relative to those with a middle level of education may be consistent with Arber and Ginn's notion of leverage: those with the most financial and material resources may be able to purchase services to assist their parents with household duties. They may be what Archbold (1983) refers to as care managers, directing the provision of help to their parents, 63 rather than actually providing the service themselves. These people's parents may also have access to financial resources that enable them to purchase various services. Because of the cultural resources that the upper classes possess, perhaps the highly educated are more aware or are better able to access appropriate community and professional services (Arber and Ginn, 1992:614; Archbold, 1983;44). The concept of leverage may thus help explain why fewer highly educated people gave help compared to those with a middle level of education. However, the results show that those with the least education were the least likely to help. It was expected that most of the helpers would come from this group. Perhaps fewer low educated adult children helped their parent(s) with housework because it is economically necessary for them to be employed, often in more than one job. These occupations may not have flexible time schedules that would facilitate helping with this task (Abel, 1990:73; Archbold, 1983:43). The time these people spend at work may reduce the time they have available to help their parents with household tasks, especially if the adult children have other family responsibilities at home. It may also not be economically feasible for them to take time off to help their parents, assuming their employer gives them this option (Abel, 1990:73). Research has also demonstrated that people from a working class background tend to have larger family networks and live in closer proximity to their parents (Glazer, 1990:487-488; Rosenthal, 1987:316,318). It may be that those with the least education have larger sibling networks living close-by that may assist them in giving help to their parents with housework. Matthews and Rosner (1988:192) report that geographical proximity among siblings was an important consideration in determining which siblings would assist in 64 caregiving duties. Therefore, perhaps low educated adult children have more siblings upon which to rely upon as a back-up (IbidASS). Why then, are those with a middle level of education the most likely to provide assistance with housework? Perhaps these people, while not possessing the financial leverage to purchase outside help, also do not have the sibling network available to help them with this task. Research has shown that as socio-economic status increases, so does geographic mobility between parents and children (see Glazer, 1990:488). Middle class helpers may have fewer siblings living close-by to share in the housework duties for their parents. To further investigate the question of why more middle class adult children gave help with housework, level of education was crosstabulated with people whose community dwelling parent(s) live with at least one of the respondent's siblings, but not with the respondent. Interestingly, it was found that fewer people with a middle educational background had parents who lived with their children compared to the other two education groups (x2=20.74, p<.01). Those with the least education were the most likely to have parents in this situation (26.6%). Those with the most education also had a higher likelihood of parents living with children (19%) compared to the middle education group (14%). These findings may help explain why more people from the middle educated group provided assistance; It may be that as Arber and Ginn imply, those with little leverage must often live with their frail parents, while those "carers who are well-off may have sufficient space to accommodate an elderly relative in their own home..." (1992:623). While this does not necessarily apply to the respondents themselves, it seems to apply more to their siblings. Arber and Ginn state, "The resources associated with being middle class can be used 65 to facilitate or alter the tasks of caring. We will consider one way in which resources can be used to advantage—through car ownership, which is more likely for those in higher social classes" (1992:629). They found that adults under age sixty-five who did not own a car were more likely to be caring for someone in the same household compared to those who owned more than one car. Those who did own a car were more likely to give help to someone outside their home. While the 1990 General Social Survey did not ask about car ownership, it did ask whether the respondent provided transportation to his/her parent(s). Not surprisingly, those with the higher levels of education were more likely to provide transportation. Approximately 15% of those with a middle level of education or higher gave this kind of assistance compared to 9% of those with a low level of education (x2=9.86, p<.01). The influence of education on help with transportation seems to manifest itself in a kind of threshold effect. That is, level of education has its greatest influence between those with a low level of education and those with a middle level or higher. Partitioning the contingency tables reveals that there is no statistically significant difference between those who have a middle level of education compared to those with a high level. However, when these two higher education levels are combined, the results show that 15.3% gave help with transportation compared to the 9% of those with the least education (x2=9.32, p< .01). This reflects that those with more education are probably more able to afford a car, or more than one car, thus making it easier to provide the transportation. Crosstabulations between education level and help given with financial support (see Table 7) show that, as expected, those with the most education were the most likely to give 66 this kind of help (4.0%; x2=9.81, p<.01). It was also expected that there would be a monotonic pattern across the education categories, with those with the least education being the least likely to give this kind of assistance. However, there was little difference between those who gave help and were the least educated (2.0%), and those with a middle level of education (0.80%). Interestingly, because slightly more adult children with the least education contributed financial support than those from the middle group, this prevented a monotonic pattern from emerging. This may be an indicator that people with the least education have parents who are in the most need of financial assistance. These parents may also come from low educational backgrounds. Their children therefore feel more pressure to try to meet their parents' financial needs. Meanwhile, the highly educated adult children may have more financial resources enabling them to provide financial support to their parents. Despite the differences across the different education groups, very small percentages of adult children within each group give their parent(s) this kind of assistance. A statistically significant difference was not found for the association between education and help with personal care (x2=1.62, n.s.) or maintenance (x2=0.15, n.s). Dwyer and Coward upon finding that there was no relationship between parent care with ADL tasks such as personal care, and income of the elder state, "It may be, when the more basic personal care needs associated with ADL impairment are present, that resources other than adult children (both formal and informal) respond to elders' needs, regardless of their economic status.. .in ways not observed for IADL impairments" (1991 :S267). Although their research examined the economic status of the elder and not the child(ren), it Is still not clear why their findings pertain to ADLs such as personal care and not other IADLs. 67 With respect to maintenance, it was expected that those with a higher level of education would have enough financial leverage to be able to pay for the provision of necessary maintenance services. It may be that many of the respondents' parent(s) live in an apartment/condominium or congregated/social housing where most of the maintenance tasks are already looked after (see McDaniel, 1992:63). Help given with at least one task versus no help was also examined (see Table 7). Results show that education level has a statistically significant effect on support. Of those with a low level of education, 18.2% gave help compared to 24.1% of the middle educated group, and 24.9% of those with a high level pf education (x2=6.71, p< .05). This finding more closely supports Arber and Ginn's (1992) findings that the middle and upper classes were more likely to provide help. While this is an interesting, and unexpected finding, it illustrates the point that it is important to explore helping across different kinds of tasks because amalgamating them may cover key differences among different groups. With respect to the earlier propositions, therefore, it appears that education level as a proxy measure for class is associated with adult children helping their parent(s) on a majority of tasks which include housework, transportation, financial support, or else help with at least one task. However, contrary to expectations, it was the middle educated children and not those with the least education who were the most likely to help with housework. This partially contradicts the earlier proposition that the working class, those with the least leverage, are the most disadvantaged in terms of providing support. With respect to provision of transportation, as expected, the most highly educated were the most likely to provide this service. Although there was almost no difference between the highly educated group and 68 those with a middle level of education, this is not an entirely unexpected finding. The education effect occurred between the least educated and those with more education. The results also showed some support for the earlier proposition that more highly educated adult children would provide financial support to their parents. However, unexpectedly, the least educated were a little more likely to provide this support compared to the middle educated group. Contrary to the earlier propositions, there was no difference between education level and likelihood of giving help with either maintenance or personal care. Gender and Class: Interaction Effects One of the central research questions in this thesis is to explore whether class and gender interact to influence helping. The findings in Table 8 show that there may be an interaction effect for those daughters who assist their parent(s) with housework. Daughters with the least education were the least likely to help with this task (6.4%) while those with a middle level of education were the most likely to provide help (16%). Highly educated daughters fell somewhere in the middle as almost ten percent of them provided assistance x*=11.19, p<.01). In contrast, due to the low frequencies of sons helping their parent(s) with housework, analyses of these results must be done cautiously. The results seem to suggest that education level does not make a difference in whether sons help their parent(s) with housework. The findings show that very few sons (approximately two percent) assisted their parent(s) with this task regardless of their educational background. 69 Table 8 Adult Children Who Help Their Parents By Children's Level of Education Controling For Gender of Helper Low Educ. Middle Educ. High Educ. X* Percent I Percent \ Percent \ Hales [ii] [n] [n] Task Housework 2.30 2.10 2.60 n/c [5] [4] [6] Personal Care 0.80 1.70 2.40 n/c [2] [3] [6] Transportation 7.30 7.40 11.90 3.70 n.s. [151 [13] [28] Maintenance 6.2 3.5 5.5 1.59 n.s. [13] [6] [13] Financial Support 3.30 0.30 4.40 6.37 * . [7] [1] [14] At Least 1 Task 16.2 10.5 19.00 5.51 n.s. [34] [19] [45] n=210 n=176 n=236 Low Educ. Middle Educ. High Educ. XJ Percent % Percent \ Percent 1 Females [n] [n] [n;] Task Housework 6.4 16.0 9.7 11.19 ** [15] [34] [21] Personal Care 5.3 7.4 6.5 0.80 n.s. [12] [16] [14:] Transportation 10.6 21.6 19.2 10.72 ** [25] [47] [41] Maintenance 3.0 6.2 4.5 2.63 n.s. [7] [13] [10] Financial Support 0.8 1.3 3.6 n/c [21 [3] [8] At Least 1 Task 20.1 35.1 31.5 13.66 ** [47] [76] [68] n=233 n=216 n=214 * p<.05; ** p<.01; *** p<.001; n/c not calculated 70 These findings on help with housework do not support the earlier hypothesis that working class women are doubly disadvantaged. It was thought that as women they would be expected to help with this task more than men, as is suggested in the literature. In addition, since they would have less financial, material and cultural leverage than better educated/upper class women, they would not have the resources to displace some of their helping duties (Glazer, 1990; Walker, 1983). That more middle educated women helped is certainly an unexpected finding and is contrary to the dominant view in the literature. One reason why women with a mid-level educational background were the most likely to help compared to the least and highly educated women is perhaps they come from relatively small families where there is little proximity among siblings. Alternatively, it may be these women's parents who are the least likely to co-reside with their other children. It may also be that compared to middle class women, working class women are in poorer health so they are less able to manage heavier household tasks, thus they are less likely to help. Statistically controlling for gender reveals evidence of an interaction effect for assistance with transportation (see Table 8). For sons, education level has no significant influence on help given with this task. In contrast, for daughters, the least educated were the least likely to assist their parent(s) (10.6%). The pattern that emerges, however, is not monotonia. Slightly more middle educated daughters provided transportation compared to highly educated daughters, 21.6% and 19.2% respectively (x 2-10.72, p< .01). These results show that class effects on provision of transportation seen earlier in Table 7 vary depending on whether the helper is a son or a daughter. Sons are perhaps able to delegate this task to their sisters so sons do not help regardless of education level. 71 Meanwhile, the education effect for daughters who provide their parent(s) with transportation occurs between the least educated and those with a middle level of education or greater. There is no statistically significant difference between the middle and highly educated women who give this kind of support. However, when these two categories are combined, 20.4% give help compared to the 10.6% of women with the least education (x2=9.59, p<.01). Low educated women may not own a car so this may help explain why those with more education are almost twice as likely to help. There is a statistically significant difference between sons' level of education and provision of financial support to parentis). More sons who were either highly educated or had a low level of education gave this support, 4.4% and 3.3% respectively. This compares to less than one percent of sons with a middle level of education (x2=6.37, p< .05). Tests of significance are not appropriate for analyses of daughters who provide this type of support as over 20% of cells have an expected frequency of less than five. The results must therefore be interpreted carefully. The percentage distributions show that the direction is monotonic as fewer lower educated daughters gave financial support (0.8%) while 3.6% of the most educated also helped their parents financially. It appears that daughters provide little financial support, and instead give assistance with other tasks such as personal care, housework and transportation. As suggested earlier, sons may feel more comfortable in giving financial support in the place of these other kinds of assistance. It would follow that the most educated would be the most able to provide such support. Concurrently, sons with the least education may have parents who are in more need of the assistance. The results in Table 8 reveal that there was no interaction effect for assistance with 72 maintenance or personal care. The earlier propositions that low educated sons would be the most likely to help with house maintenance, and that low educated daughters would help with personal care were not supported. The education effect for help with at least one task differs for sons and daughters. Focusing first on sons, we find that the relationship is not statistically significant When examining the pattern for daughters, the findings show that those with the lowest education levels were the least likely to give help (x2=13.67, p< .01). Notice that on the task variable working class men and women do not differ much, 16.2% and 20.1% respectively. Perhaps both sons and daughters help. Among middle class families, the burden falls on women: 35.1 % of daughters helped compared to 10.5% of sons. This latter pattern also holds for the upper class except that here the gender difference is smaller, mainly because upper class sons give the highest levels of help on transportation and financial support. To summarize, the propositions on interaction effects for helping behaviour received little support. The findings hint that there is an interaction effect for help with housework, however, small cell size warrants cautious interpretations. It was interesting to find that middle educated daughters were most likely to help with this task instead of the least educated women as was anticipated. There was clear evidence of interaction effect for daughters who help with transportation. It was the middle and highly educated daughters who were most likely to help relative to the other gender/education groups. It was thought that highly educated women would provide this help but a monotonic pattern did not emerge. With respect to financial support, it was expected that highly educated sons would be the most likely to provide this service. The results do seem to support this, but it should be noted 73 that the difference between this group and low educated sons is minimal. Relative to this, middle educated sons were the least likely to give this support. For daughters, a monotonic pattern did emerge in the expected direction, but analyses must be done cautiously due to small cell size. There was no evidence of an interaction effect for help with either personal care or maintenance contrary to what was expected. Closer Look at Education Differences Among Adult Children Helpers Little research has been done that examines how certain variables thought to influence caregiving varies by class of the caregiver. This section explores the effects of work status and physical distance between parent(s) and child on helping behaviour according to the adult child's level of education. As was the case earlier in the subanalyses among sons and daughters, the frequencies are small. The objective is to explore the direction of possible relationships and highlight particular trends. Due to small frequencies, analyses of daughters will revolve around help given with housework, personal care, transportation and the combined task variable. For sons, only maintenance, transportation and the combined task variable will be examined. It was suggested earlier that perhaps one reason why those with a lower level of education may be less likely to help with housework is because they find it economically necessary to work in more than one job. They also may be working in occupations that do not have flexible hours. Yet at the same time, it was found that among sons and daughters, there was little difference in the percentage of helpers who work versus those who do not. 74 Table 9 shows that for daughters with the least education, there is no relationship between work status and help with housework. Moving up the educational scale, we observe a larger difference such that more women who were not working give help. The results, however, are still not statistically significant. The difference becomes even larger, and approaches statistical significance, for those women with the highest level of education. Sixteen percent of those not working provided assistance compared to 6.8% of those women who were working (x2=3.58, p=06). These results may suggest that work is perhaps more of a competing demand for highly educated daughters. Archbold found that for female caregivers from high socio-economic backgrounds "Career commitment provides a salient competing role to caregiving" (1983:41). Perhaps these well educated women who work can afford to purchase assistance. Meanwhile, low educated women who work may bear a greater burden than their low educated, non-working counterparts because they both work and help their parent(s). Although not statistically significant, the data hint that competing demands might also be a factor for highly educated daughters who help with personal care and transportation. The results for low educated daughters who help with personal care were significant. Seven percent of those not working helped compared to 1.8% of those not working (x2=2.00, p< .05). Further research is needed to examine whether the competing demands hypothesis, with respect to employment, operates differently for women from different classes. Table 10 shows that regardless of level of education, daughters who live an hour or less away from their parent(s) were significantly more likely to provide assistance to their parent(s) with housework, transportation and help with at least one task. Close to 12% of low 75 Table 9 Daughters Who Help Their Parents: Education and Work Status Low Education diddle Education High Education Daughters Daughters Daughters Not Is Not Is Not Is Working Working Working Working Working Working Percent % Percent \ Percent % Percent % Percent 1 Percent ! Task [n] [n] [n] [n] [n] [n] Housework 6.7 5.7 n.s 20.4 13.9 n.s. 16.3 6.8 *+ [10] [5] [14] [20] [10] [10] Personal Care 7.1 1.8 ++ 7.0 7.6 n.s. 10.5 4.8 n.s. [11] [1] <[5] [11] [7] [7] Transportation 10.6 10.7 n.s 21.4 21.8 n.s. 23.8 17.2 n.s. [16] [9] [15] [32] [15] [26] At Least 1 Task 21.1 18.2 n.s. 41.8 32.0 n.s. 34.9 30.1 n.s. [32] [14] [29] [46] [22] [45] n=154 n=80 n=71 n=145 n=64 n=151 *+ p=.06; * p<.05;** p<.01; *** p<.001 ++ p value statistically significant at .05 level using Fisher's Exact Test Table 10 Daughters Who Help Their Parents: Education and Distance From Parents Low Education Middle Education High Education Daughters Daughters Daughters Lives >lHr Lives < lib- Lives >lHr Lives < lHr Lives >lHr Lives < lHr Fron Parent FroB Parent Froi Parent Fron Parent Fron Parent Froi Parent Percent ! Percent % Percent % Percent 1 Percent \ Percent % Task [n] [n] [n] [n] [n] [n] Housework 0.0 11.7 ** 8.0 23.8 ** 4.0 14.6 * [0] [15] [7] .[27] [4] [16] Personal Care 0.0 9.7 ** 6.0 9.1 n.s. 4.3 8.2 n.s [0] [12] [6] [10] [4] [9] Transportation 0.7 18,9 *** 8.5 33.9 *** 2.0 30.7 *** [1] [24] [8] [38] [2] [34] At Least 1 Task 3.5 34.0 *** 18.9 50.5 *** 14.6 42.8 *** [3] [43] [18] [57] [14] [47] n=100 n=127 n=94 n=113 n=98 n=110 * p<.05; ** p<.01; *** p<.001 76 educated daughters who lived within an hour of their parent(s) helped with housework compared to no daughters who lived farther away (x2=10.68, p<.01). Almost 24% of middle educated daughters who lived close-by gave assistance with this task compared to eight percent of more distant daughters (x2=8.22, p< .01). A similar pattern emerges for highly educated daughters. Close to fifteen percent of those in close proximity to their parent(s) helped with housework compared to four percent who lived over an hour away (x2=5.58, p<.05). Distance was significantly related to help given with personal care by low educated daughters, but made no difference for women with higher levels of education. Help with transportation, as well as with at least one task, show once again that similar to the findings for help with housework, daughters who lived closer were significantly more likely to help. Interestingly, the difference between daughters who live close-by who help with either housework, transportation or with at least one task, and those who live over an hour away is greatest for the least educated. For example, close to nineteen percent of near-by daughters assisted with transportation. This is twenty-seven times greater than the 0.7% of daughters who lived farther away (x2= 17.40, p< .001). The difference is close to four times for the middle educated (x2= 17.63, p<.001), while it is fifteen times for the highly educated, 30.7% and 2.0% respectively (x2=28.00, p<.001). Low educated daughters who lived within an hour of their parent(s) were nearly twelve times more likely to help with housework compared to similarly educated daughters who lived farther away. Daughters with a middle level of education or higher who also lived closer to their parent(s) were about three times more likely to help compared to more distant daughters. For help with the combined 77 task variable, the results once again show the same pattern: daughters who lived closer were more likely to help, with the greatest difference occurring among the least educated. One possible reason for these results is that for those daughters whose parents lived farther away, they may not have had the means available to travel to their parent's home. Table 11 examines the effects of work status for each of the education levels for sons who help with either maintenance or transportation. Since the expected frequencies for most of the cells presented are less than five, Fisher's exact test is used for analyses pertaining to this table. Al l of the results in Table 11 show that work status is not significantly related to helping regardless of sons' level of education. The findings for the effects of sons proximity to parents are found in Table 12. Sons who lived close to their parents were more likely to help with maintenance if they had either a high (x2=9.91, p< .01) or low (x2=4.12, p< .05) level of education. Interestingly, there was no significant relationship for those with a middle level of education. It is unclear why this is so. Highly educated sons who lived less than one hour away were significantly more likely to provide transportation (19%) compared to the 4% of those who lived farther away (x 2= 11.28, p<.001). The results for low educated sons only approach significance (x2=3.49, p=.06), but are not significant for sons with a middle level of education. In examining these data more closely, distance from parents once again emerges as an important variable, for daughters in particular. Those sons and daughters who live within an hour of their parents are the most likely to help with most tasks, in many cases regardless of their level of education. However, for daughters who help, the difference between those whose parents live close-by and those whose parents live farther away is the greatest for the 78 Table 11 Sons Who Help Their Parents: Education and Work Status Task Maintenance Transportation At Least 1 Task * p<.05; '** p<.01; Low Education Sons Not Working Percent % [n] Middle Education Sons High Education Sons Is Working Percent \ [n] 0.0 [0] 2.9 [1] 5.3 [2] n=30 *** p<.001 7.3 [13] 8.1 [15] 18.0 [32] n=180 n.s. n.s. n.s. Not Working Percent \ [n] 10.1 [2] 15.1 [2] 15.1 [2] n=15 Is Working Percent I [n] 2.9 [5] 6.7 [11] 10.2 [16] n=160 n.s. n.s. n.s. Not Working Percent % [n] 2.0 [0] 6.5 [1] 15.8 [3] n=14 Is Working Percent \ [n] 5.8 [13] 12.3 [27] 19.1 [42] n=222 n.s. n.s. n.s. Table 12 Sons Who Help Their Parents: Education and Distance Froi Parents Low Education Sons Middle Education Sons High Education Sons Lives >lHr Lives < lHr Froa Parent From Parent Lives >lHr Lives < lHr Froi Parent From Parent Lives >lHr Lives < lHr Frou Parent Froi Parent Percent % Percent 3 Percent % Percent % Percent \ Percent % Task [n] [n] [n] [n] [n] [n] Maintenance 1.8 10.2 * 1.0 5.5 n.s 1.0 11.9 [1] [12] [1] [5] [1] [12] Transportation 2.7 10.9 *+ 3.9 10.4 n.s. 4.0 19.0 [2] [12] [3] [10] [5] [19] At Least 1 Task 9.5 20.8 *+ 6.6 14.2 n.s. 6.2 33.1 [8] [24] [5] [14] m [33] n=81 n=114 n=73 n=97 n=120 n=loo ** *** *** *+ p=.06; * p<.05;** p<.01; *** p<.00l 79 least educated. Work status, on the other hand, makes no difference whether one helps or not regardless of one's level of education. There were some signs, however, that the effect of employment as a competing demand for women of different classes warrants further research. Logistic Regression Results To further explore the effects of gender and education, both independently and jointly, logistic regression analyses were conducted. The logistic regression for help given with housework is presented in Table 13. Gender remains the strongest predictor of help with this task regardless of the inclusion of other variables. Daughters are approximately five times more likely than sons to give help with housework (B=1.59, p< .001, odds ratio=4.92). Consistent with earlier crosstabular results, Model 2 shows that even when gender is held constant, individuals with a middle level of education are twice as likely to provide help to their parent(s) with this particular task compared to those with a low level of education (B=.83, p< .01, odds ratio-2.30). Low education level is the reference category since it was expected that those with the least education would have the least leverage to displace some of duties associated with housework. Gender, however, is still a stronger predictor variable (B=1.58, p<.001, odds ratio=4.86). Model 3 incorporates the interaction term for women with low education into the model. The literature suggests that working class women are more likely to bear the burden of caregiving tasks, as well as are more likely to be care providers rather than care managers (Archbold, 1983; Glazer, 1990; Walker, 1983). Thus it was hypothesized that low educated women would be the most likely to provide assistance 80 Table 13 Logistic Regression For Adult Children Who Help Their Parents With Housework Model 1 Model 2 Model 3 Odds Odds Odds Variable 6 S.E. Ratio 6 S.E. Ratio B S.E. Ratio Gender 1.59 *** 0.29 4.92 1.58 *** 0.29 4.86 1.77 *** 0.35 5.86 EducH 0.33 0.31 -0.23 0.58 EducH 0.83 ** 0.29 2.30 0.27 0.58 IntLED -0.70 0.64 Constant -3.74 *** 0.26 4.16 *** 0.34 -3.75 *** 0.46 Goodnes  Of Fit XM311.00 df=1309 n.s. x2=1367.17 df=1307 n.s. x2=1336.48 df=1306 n.s. IntLED= interaction term for low educated daughters * p<.05 ** p<.01 *** p<.001 Table 14 Logistic Regression For Adult Children Who Help Their Parents With Personal Care Model 1 Model 2 Model 3 Odds Odds Odds Variable B S.E. Ratio B S.E. Ratio 8 S.E. Ratio Gender 1.40 *** 0.35 4.06 1.40 *** 0.35 4.08 1.25 ** 0.39 3.47 EducH 0.35 0.36 0.96 0.87 EducH 0.42 0.36 1.04 0.88 IntLED 0.74 0.93 Constant -4.10 *** 0.31 -4.37 *** 0.39 -4.87 *** 0.79 Goodnes  Of Fit x2=1311.34 df=1309 n.s. x2=1294.17 df=1307 n.s. x2=1312.94 df=1306 n.s. IntLED= interaction ten for low educated daughters * p<.05? ** p<.01; *** p<.001 8 1 with housework. The results do not support the interaction term. Earlier results hinted that there was an interaction between middle educated daughters and helping behaviour. It would be expected that the logistic coefficient would likely be statistically significant for an interaction term representing middle educated women. For assistance with personal care (see Table 14), daughters are four times as likely to provide this kind of help compared to sons as shown in Model 1 (B = 1.40, p < .001, odds ratio=4.06). Education, however, is not a good predictor as the log odds compared to those who are poorly educated (the reference category) are not statistically significant. The interaction term for low educated daughters is also not statistically significant. Table 15 shows the results for help given with transportation. As was the case for housework, gender is again the strongest predictor of this kind of assistance relative to level of education. Model 1 shows that daughters are twice as likely to give this kind of support compared to sons (B=.69, p< .001, odds ratio=2.00). Model 2 of this table reflects earlier findings that those with a low level of education are less likely to provide transportation (B= -.61, p< .01, odds ratio=.54) when compared to the reference category of those with high education. This reference category was selected as it was expected, as Arber and Ginn suggest, that those with more leverage are more likely to own a vehicle. When the interaction term of highly educated women is added to the model, there is no interaction effect. This may be because, as seen earlier in the contingency tables, almost as many women with a middle level of education provided their parent(s) with transportation compared to highly educated women. Table 16 examines children who provide their parents with financial support. Gender, 82 Table 15 Logistic Regression For Adult Children Who Help Their Parents With Transportation Variable Gender 0 EducH EducL IntHED Constant -2 Goodnes  Of Fit x2=1311.34 IntHED= interaction ten for highly educated daughters * p<.05; ** p<.01; *** p<.001 Table 16 Logistic Regression For Adult Children Who Help Their Parents With Financial Support Model 1 Model 2 Model 3 Odds Odds Odds S.E. Ratio B S.E. Ratio B S.E. Ratio *** 0.17 2.00 0.71 *** 0.17 2.03 0.86 *** 0.23 2.36 -0.03 0.19 -0.25 0.29 -0.61 ** 0.21 0.54 -0.82 ** 0.30 0.44 -0.34 0.35 *** 0.14 -2.12 *** 0.17 -2.01 *** 0.19 df=1309 n.s. x2=1319.54 df-1307 n.s. x2=1327.11 dM306 n.s. Model 1 + Model 2 Model 3 Odds Odds Variable B S.E. Ratio B S.E. Ratio B S.E. Gender -0.29 0.35 0.75 -0.21 0.36 -0.61 0.60 EducM -1.71 ** 0.60 0.18 -2.06 ** 0.74 EducL -0.83 * 0.41 0.44 -1.18 * 0.60 IntHES -0.63 0.75 Constant -3.51 *** 0.24 -2.98 *** 0.27 -2.45 *** 0.68 Goodnes  Of Fit x2=1310.75 df=1309 n.s. xJ=1320.61 df=1307 n.s. x2=1373.37 df=1306 Odds 0.13 0.31 IntHES= interaction term for highly educated sons Model Chi-Square Not Statistically Significant at .05 level * p<.05; ** p<.01; *** p<.001 83 for this task, is not a significant predictor. In contrast, education is a good predictor variable. The odds of providing this support are lower for those with a low (B=-.83, p< .05, odds ratio =. 44) or medium level of education (B=-1.71, p <. 01, odds ratio=.18) when compared against the reference category of highly educated individuals18. This is not surprising. It was expected that those with the most resources would be the most likely to provide this type of support. It was also expected that an interaction term for highly educated men would be statistically significant. The logistic regression results do not support this. This finding goes against the rationale that, as implied in the literature, men are the higher income earners and may feel more comfortable in either purchasing services for their parents, or providing their parents with financial support (see Montgomery and Kamo, 1989:224). The results for help with maintenance are straightforward (not presented in a table). Neither gender nor education emerge as significant predictors: the overall model chi square was not statistically significant. This indicates that it diverges significantly from the "perfect" model. Therefore, the analyses are not discussed here. In the preceding literature review, it was shown that various factors may influence whether sons and daughters help their parents. Using help provided with housework as an illustrative example, some of the more salient variables were added to the model already consisting of gender and education. The results in Table 17 show that even controlling for all these additional variables, gender is the strongest predictor of helping parents with housework (B= 1.52, p< .001., odds ratio=4.57). As suggested by the earlier contingency 18The standard error is relatively high for the middle education term. One likely reason for this is the small number of cases who provided financial support (n=33). Therefore, this result should be interpreted cautiously. 84 Table 17 Logistic Regression For Adult Children Who Help Their Parents With Housework: Multiple Variables Variable 8 S.E. Exp (B) Gender 1.52 *** 0.35 4.57 EducH 0.33 0.33 EducH 0.82 ** 0.31 2.27 Work -0.52 *+ 0.29 0.60 Harital -0.03 0.29 RHealth 0.78 0.47 Age -0.03 0.03 Distance 1.40 *** 0.30 4.05 Agepar 0.04 0.02 Income 0.09 0.35 WSpouse -0.26 0.27 WChild -0.95 * 0.41 0.39 Constant -7.51 *** 2.07 Goodnes  of Fit x2=1320.11 df=1298 n.s. *+ p=.07; * p<.05; ** p<.01; *** p<.001 85 table analyses, distance from parents also emerges as a strong predictor: the closer one is to one's parents, the higher the likelihood of helping (B=1.40, p< .001, odds ratio=4.05). Those with a middle level of education are more likely to help compared to those with a low level of education (B=.82, p<.01, odds ratio= 2.27). This findings provides more contradictory evidence to the expectation that the least educated would be the most likely to help. People whose parent(s) live with their children other than the respondent, are 39% less likely to help their parent(s) with housework compared to those whose parent(s) live alone ( B - -.95, p < .05, odds ratio= 0.39). Lastly, it appears that those who work were less likely to provide housework compared to those who do not work, however, the results only approach significance (B=-.52, p=.07, odds ratio=.60). 86 CHAPTER SEVEN FINDINGS - ELDERLY PARENTS WHO RECEIVE HELP FROM THEIR CHILDREN Gender Differences in the Receipt of Help Another important research question in this study is whether elderly parents who do and do not receive assistance from their children differ by gender and education, both independently and jointly. Of the 760 people aged sixty-five and over who report a health limitation and have children yet do not live with them 40% were men, while 60% were women (see Table 18). The findings in Table 19 show that 15.5% of elderly mothers received help with home maintenance compared to 7.5% of elderly fathers (x2=9.92, p< .01). The findings support similar findings in the literature (Kaden and McDaniel, 1990:19; Rosenthal, 1987:326). One explanation for why fewer men receive assistance with maintenance is that despite their poorer health, they still may be able to perform lighter maintenance duties. Elderly mothers may be more likely to receive help from their children with these tasks because many of them may be widows living alone. They rely on their children's assistance, especially if they do not live in a housing environment where some of the maintenance is looked after (McDaniel, 1992:63). Alternatively, perhaps some of these women never had to perform maintenance chores around the house and either do not know how or are unable to manage them. Over a quarter of elderly mothers (28.7%) reported receiving help with transportation compared to nearly fourteen percent of elderly fathers (x2=22.36, p< .001). Once again, 87 Table 18 Gender and Level of Education Frequencies For Elderly Parents Sample Variable Gender Hales Females Total Level of Education Low Education Hiddle Education High Education Hissing Total Percent [n] 39.5 301 60.5 460 100.0 760 58.8 447 23.1 176 15.1 115 2.9 22 100.0 760 88 Table 19 Elderly Parents Who Receive Help From Their Children By Gender of Parent Task Housework Personal care Maintenance Financial Support Transportation At Least 1 Task Male Percent % mi 9.4 [28] 1.2 . [4] 7.5 [23] 1.4 [4] 13.7 [41] 27.7 [83] n=301 Female Percent I [n] 13.0 [60] 3.2 [15] 15.5 [21] 1.8 [8) 28.7 [152] 43.5 [200] n=460 1.97 n.s. 2.26 n.s. 9.92 ** 0.89 n.s. 22.36 *** 18.70 *** p<.05; ** p<.01; *** p<.001 89 elderly men are more likely to have a spouse available compared to elderly women, so it is their wives who may be providing them with transportation, if they are not able to provide it themselves. As more women are likely than men to be widowed, they probably rely on their children for help with this task. Being male or female did not influence whether one received help with housework from children (x2= 1.97, n.s.). One possibility for this finding is that older mothers and fathers may both receive help with housework from other formal sources such as home support groups. Alternatively, it may be that those elderly who are married are looking after their spouses so they may need help with household chores. Widows and widowers may also similarly require assistance from their children, especially if they have a health limitation. There was no significant difference between mothers and fathers who received assistance with personal care or financial support, thus not totally supporting the findings of Rosenthal (1987) and Kaden and McDaniel (1990). For both of these tasks the frequencies were very small suggesting that elderly parents do not receive these kinds of support. It may be the case that the elderly do not need these kinds of help. Branch and Jette (1983:52) found that in their sample of elders over the age of 70, over 80% of them were self-sufficient in performing basic ADLs such as bathing, dressing, grooming and eating. Interestingly, over 80% of their sample used informal networks for help with I ADLs. The low frequencies in Table 19, however, may also be underestimating the number of elderly who do receive assistance with personal care and financial support. Elderly respondents may not have wanted to admit that they receive help with these tasks because of the stigma involved in being dependent on their children for such assistance (Aronson, 1986). Alternatively, these parents 90 may be receiving personal care from formal sources instead of from their children. As was the case in the analysis of adult children who helped their parent(s), a variable was created that measured whether people received help on at least one of these tasks versus not receiving any help. Forty-three percent of elderly mothers reported receiving help with at least one task compared to 27% of elderly fathers (x2= 18.70, p<.001). The greater predominance of women receiving help with either transportation or maintenance is a likely factor for the magnitude of the difference of this result. Characteristics of Elderly Parents Who Receive Help ' To investigate some of the characteristics of elderly mothers who receive help from their children, the role of two additional variables were examined: marital status and distance from children. Research has shown that elderly parents tend to rely first on their spouses for assistance before turning to children (Dwyer and Coward, 1991:S260; Stoller and Earl, 1983:67) so the expectation is that non-married (e.g., widowed, divorced, single) parents would be more likely to receive filial assistance. Meanwhile, it would be expected that elderly parents who have at least one child over the age of fourteen living within an hour's drive would be more likely to receive help than those parents whose children live farther away. Table 20 shows that there are no statistically significant differences between married and unmarried elderly mothers in receipt of help from their children with either housework (x2=1.19, n.s.), maintenance (x2=2.03, n.s.) or with more than one task (x2=2.31, n.s). 91 Table 20 Elderly Mothers Who Receive Help From Their Children Marital Status Distance From Children Not Is No Child Lives > 1 Child Lives Married Harried X 2 Within 1 Hour Within 1 Hour X2 Percent \ Percent \ Percent 1 Percent 1 [n] [n] [n] [n] Task Housework 14.9 11.0 1.19 n.s. 6.5 14.4 2.29 n.s. [35] [24] [4] [55] Maintenance 13.1 18.5 2.03 n.s. 8.6 17.3 2.44 n.s. [31] [40] [5] [66] Transportation 35.4 21.7 9.66 ** 9.4 32.2 12.63 ** [83] [47] [6] [122] At Least 1 Task 47.0 39.5 2.31 n.s. 20.8 47.9 15.19 *** [110] [86] [13] [182] n=233 n=218 n=64 n=380 * p<.05; ** p<.01; *** p<.001 Table 21 Elderly Fathers Who Receive Help From Their Children Marital Status Distance From Children Not Is No Child Lives > 1 Child Lives Married Harried X2 Within 1 Hour Within 1 Hour X2 Percent 1 Percent \ Percent \ Percent % [n] [n] [n] [n] Task Housework 6.8 9.1 .11 n.s 7.3 10.5 .29 n.s [4] [21] [5] [23] Maintenance 5.3 8.2 .27 n.s. 1.5 9.8 3.95 * [3] [19] [1] [22] Transportation 15.5 13.0 .10 n.s 6.2 16.2 3.65 *+ [10] [30] [4] [36] At Least 1 Task 23.8 27.5 .19 n.s. 13.5 33.0 8.88 ** [15] [65] [9] [73] n=62 n=234 n=69 n=221 *+ p=.06; * p<.05; ** p<.01; *** p<.001 92 With respect to maintenance, perhaps married women's spouses are still able to do some of these chores. Mothers who are no longer married may have moved into accommodations where much of the maintenance is already looked after (see McDaniel, 1992:63). One explanation for the non-significant findings for help with housework is that these married women, since they are in poor health, are likely being looked after by their husbands who may be doing the housework. Widows, however, may be receiving help with these chores from other sources such as friends, neighbours or community agencies. It is this support that perhaps enables these women to still live in their homes. More non-married women received help with transportation (35.4%) than the 21.7% of married women (x2=9.66, p< .01). Not having a spouse probably results in having to rely more upon children for this task. Results in Table 21 reveal that for elderly fathers marital status makes no difference on whether one receives help from children with either housework (x 2 =. l l , n.s.), maintenance (x2=.27, n.s.), transportation (x2=.10, n.s.) or help with at least one task (x2=.19, n.s.). A similar explanation as the one above may be offered: married men may be receiving support from their spouses, while widowers receive assistance from community sources, friends or neighbours. The fact that marital status did not make a difference with receipt of help with most of these tasks for both mothers and fathers is an interesting finding. Previous research shows that non-married parents are more likely to receive support from their children compared to married parents (Dwyer and Coward, 1991:S265) Indicators on the effects of distance between parent and child show that for elderly mothers, having at least one child living within an hour's distance is associated with an 93 increased likelihood of receiving help with either transportation or at least one task (see Table 20) . Approximately 32% of mothers with at least one child over age fourteen living within an hour's distance received transportation compared to 9.4% of those whose children lived farther away (x2= 12.63, p<.01). Meanwhile, 47.9% of mothers with proximate children received support with at least one task compared to 20.8% of those with more distant children (x2=15.19, p<.001). It is somewhat surprising that distance did not make a difference for receipt of help with either housework or maintenance. Unfortunately, the sample size of those who had children living farther away is small so this may have influenced tests of significance to some degree. The effect of distance from children is slightly different for elderly fathers (see Table 21) . Almost ten percent of those who had at least one child over age fourteen living within an hour's distance received help with maintenance compared to 1.5 % of those whose children lived farther away (x2=3.95, p< .05). The results for transportation show a similar pattern where 16.2% of those with at least one proximate child received help compared to 6.2% of those who did not have such children living within an hour's distance. However, the results only approach statistical significance (x2=3.67, p= .06). Thirty-three percent of fathers with near-by children received assistance with at least one task compared to 13.5% of fathers with children who lived farther away (x2=8.88, p<.01). Similar to the Fmdings for elderly mothers, distance made no difference in receipt of housework (x2=.29, n.s.). To summarize, more elderly mothers than fathers received assistance from children with tasks such as maintenance and transportation, in addition to receiving help with at least one task. There was no task in which fathers were the most likely to receive help. Marital 94 status only made a difference for help received by mothers with transportation as more non-married mothers received this support. Contrary to expectations, marital status was not associated with receipt of assistance with any of the other tasks examined. Alternatively, having at least one child living close-by only made a difference in receipt of help with transportation for mothers, maintenance for fathers, and help with at least one task for both parents. It was expected that distance would have a greater effect for all of the tasks involved. Education Differences and Interaction Effects in the Receipt of Help In this sample of elderly people, 59% had an education level of some secondary or less, 23% had completed secondary school or had some post-secondary education, while the remaining 15% had a post-secondary degree or diploma (see Table 18). When receipt of help with housework is crosstabulated with education level of the respondent, a monotonic pattern emerges (see Table 22) as more people with a low level of education received help compared to the other two education groups. That is, 14.5% of elderly parents with a low level of education received assistance with housework compared to 8.9% of those with a middle education and 3% of those who were highly educated (x2= 13.56, p< .01). There is considerable difference between the two extreme education categories. One explanation for this finding is that those who have the most education may be able to purchase assistance or else have it purchased for them by their offspring. This would lend support for the leverage hypothesis. It may also be that socioeconomic status 95 Table 22 Elderly Parents Who Receive Help Fron Their Children By Parent's Level of Education Low Hiddle High Education Education Education Percent 3 Percent % Percent \ W [n] [n] Task Housework 14.5 8.9 3.0 13.56 ** [65] [16] [3] Transportation 24.9 22.9 14.6 5.45 *+ [111] [40] [17] Maintenance 13.0 11.9 11.9 0.20 n.s [58] [21] [14] Financial Support 1.9 1.2 0.4 n/c [9] [2] [0] Personal care 2.7 3.4 0.6 n/c [12] [6] [1] At Least 1 Task 40.1 38.8 21.8 13.48 ** [179] [68] [25] n=447 n=176 n=115 *+ p=.07; * p<.05; ** p<.01; *** p<.001; n/c not calculated 96 tends to be related to low geographic mobility. The elderly with low educational backgrounds may be more likely to have children who live close-by thus facilitating the receipt of help. In addition, while the elderly in this sample were said to have a health limitation, it may be that those with the least education may have limitations that are more severe. This would reflect the relationship between low socio-economic status and quality of health. When the table for this task is partitioned, the results show that there is a statistically significant difference between the least educated and the middle educated. There is also a relationship between the middle educated and the highly educated. This suggests that the effect of education is distributed somewhat evenly throughout the education levels. In other words, no threshold effect exists. There was no statistically significant relationship between education level and receipt of help with maintenance (x2=.20, n.s.) and transportation. The latter did approach significance (x2=5.45, p= .07) and the relationship was in the expected direction: those with the most education received considerably less help than those with the least education, 14.6% and 24.9% respectively. With respect to help with personal care and financial support, unfortunately, the cell sizes are small so chi-square tests are not appropriate. Based on the distributions, it would appear that there is no association with level of education. In contrast, help with at least one task was significantly associated with one's level of education as 40.1 % of the least educated elderly received such assistance compared to 38.8% of the middle educated group. Twenty-two percent of those with the most education reported receiving help with at least one task (x2= 13.48, p< .01). Table 23 shows that there are no interaction effects between gender and education on 97 Table 23 Elderly Parents Who Receive Help Froi Their Children By Parent's Level of Education Controling For Gender of Parent Low Middle High Education Education Education Hales Percent \ Percent \ Percent % [n] [n] [n] Task Housework 11.1 9.2 0.0 6.09 * [19] [6] [0] Transportation 15.2 14.2 3.3 5.03 n.s. [26] [9] [2] Maintenance 10.3 3.9 4.0 4.02 n.s. [18] [3] [2] Personal Care 2.1 0.0 0.0 n/c [4] [0] 10] Financial Support 1.2 3.1 0.0 n/c [2] [2] [0] At Least 1 Task 30.7 30.4 6.3 12.65 ** [54] [20] [3] n=175 n=67 11=51 Low Middle High Education Education Education X2 Females Percent \ Percent \ Percent \ [n] [n] [n] Task Housework 16.8 8.6 5.4 8.45 ** [46] ' [9] [3] Transportation 31.1 28.2 23.5 1.53 n.s. [85] [31] [15] Haintenance 14.7- 16.7 18.1 0.56 n.s. [46] [15] [12] Personal Care 3.0 5.4 1.0 n/c [8] [6] [1] Financial Support 2.4 0.0 0.0 n/c [6] [0] [0] At Least 1 Task 46.1 44.0 33.9 3.13 n.s. [126] [48] [22] n=273 n=110 n=64 * p<.05; ** p<.01; *** p<.001; n/c not calculated 98 help received with housework. Regardless of whether one is an elderly mother or father, those with the least education were the most likely to receive this assistance. These results, therefore, do not support the earlier preposition that disadvantaged elderly women would be the most likely to receive this kind of help. This suggests that having a low level of education transcends gender when it comes to elderly parents receiving household help from children. Interestingly, there were no highly educated elderly fathers who reported receiving help from their children with this task compared to 9% of those with a middle education and 11.1% of those with a low education level (x2= 6.09, p<.05). Low educated mothers were three times as likely to receive help from their children with this task compared to highly educated mothers, 16.8% and 5.4% respectively (x2=8.45, p<.01). These findings may be explained using Arber and Ginn's idea of leverage. The working class elderly do not have the financial and cultural resources to obtain paid external help with these chores, and therefore need to rely on their children for support. Further, this happens for both elderly men and women. It may also be that both working class elderly fathers and mothers are the most likely to have children living close-by compared to the other class groups. There was also no interaction effect for help received with transportation or maintenance. Results were not statistically significant for either men or women. This is contrary to the expectation that low educated elderly women would be the most likely to receive help with these tasks. On the other hand, there was an interaction effect when all of the tasks were combined. Approximately thirty percent of fathers with either a low or middle level of 99 education received help with at least one task. This is nearly five times more than the 6.3% of highly educated fathers (x2= 12.65, p< .01). One possible explanation for these findings is that these men have a lesser likelihood of being single than elderly mothers so their spouses are probably providing support. However, highly educated men have the extra advantage of possessing greater leverage, therefore decreasing their need to rely on children for support. For mothers, on the other hand, there was no statistically significant association between help received and level of education. Elderly mothers with a health limitation appear to be much more reliant on their children for help with at least one task regardless of their level of education. The proposition that poor elderly women would be most likely to receive assistance is not supported. The results in this section, therefore, show that previous propositions on the effect of education/class were not supported for receipt of assistance with maintenance or transportation. Unfortunately, frequencies for financial support and personal care are too low to comment on. However, the hypothesis that the least educated would be the most likely to receive help with housework was strongly supported. Surprisingly, education had an effect regardless of gender. This lends some support to Arber and Ginn's leverage hypothesis. Perhaps the least educated have the least material and cultural resources available so they have to rely on their children for support. Closer Look at Education Effects Among Elderly Parents The contingency tables were further partitioned to explore the effects of geographic 100 distance for elderly men and women of different educational backgrounds. Due to the small frequencies for men and women who receive help, differences across all three education levels are only presented for women who receive help with transportation, maintenance and at least one task. For women Who receive help with housework, and men who receive help with housework, transportation, maintenance, and at least one task, only those with the least education will be examined. The results in Table 24 show that 33.9% of elderly mothers who have at least one child over age fourteen living within an hour's drive received transportation compared to 14.2% of those whose children lived farther away (x2=4.77, p<.05). Similarly, more middle educated mothers with children living close-by also received transportation compared to those whose children were more distant (31.2% and 6.4% respectively, x2=2.60, p<.05). The same pattern emerges for highly educated mothers (30.9% versus 0% respectively; x2=3.60, p<.05). What is interesting to note is that the magnitude of the difference between the two distance values is the largest for highly educated women, and is the lowest for women with the least education. Unfortunately, these results cannot be compared with those for elderly fathers (as their frequencies are too small) to determine whether this is the case for mothers only or if it extends to fathers as well. Distance did not have a statistically significant effect on receipt of help with maintenance. However, for the combined task variable, those who had either a low or middle level of education and who had children close-by were more likely to receive help. This time, however, the greatest difference between the two categories occurs for the middle educated, 49.1 % of those who had children close by received such support compared to 10.1 % of those 101 Table 24 Elderly Mothers Who Receive Help According to Mother's Level of Education and Distance From Children Low Education , Hiddle Education High Education Mothers Mothers Mothers No Child > 1 Child No Child > l Child No Child > 1 Child Lives Within Lives Within Lives Within Lives Within Lives Within Lives Within 1 Hour 1 Hour 1 Hour 1 Hour 1 Hour 1 Hour Percent Percent Percent Percent Percent Percent Task [n] [n] [n] [n] [n] Housework 10.0 18.0 n.s. - _ _ [4] [41] Transportation 14.2 33.9 * 6.4 31.2 ++ 0.0 30.9 ++ [5] [77] [1] [29] [0] [15] Haintenance 8.2 16.4 n.s. 0.0 19.6 n.s 20.2 18.5 n.s. [3-] [37] [0] [18] [3] [9] At Least 1 Task 25.7 49.5 * 10.1 49.1 * 20.2 39.3 n.s. [9] [112] [1] [46] [3] [19] n=36 n=227 n=14 n=113 n=13 n=49 * p<.05; ** p<.01; *** p<.001 ++ p value statistically significant at .05 level using Fisher's Exact Test Table 25 Elderly Fathers Who Receive Help According to Father's Level of Education and Distance From Children Low Education Middle Education High Education Fathers Fathers Fathers No Child > 1 Child No Child > 1 Child No Child > 1 Child Lives Within Lives Within Lives Within Lives Within Lives Within Lives Within 1 Hour 1 Hour 1 Hour 1 Hour 1 Hour 1 Hour Percent Percent Percent Percent Percent Percent Task [n] fn] [n] [n] [n] [n] Housework 15.2 10.8 n.s. - -[5] . [14] Transportation 3.5 19.1 * - -[1] [25] Haintenance 3.1 12.8 n.s. - -[1] [17] At Least 1 Task 18.7 35.7 n.s. 16.0 34.7 n.s. [6] [47] [3] [16] n=33 n=133 n=18 n=47 * p<.05 102 who did not (x2=6.05, p< .05). The findings for low educated fathers (see Table 25) reveal that 19.1 % of those whose children were close-by received help with transportation compared to 3.5% of those whose children were at a greater distance (x2=3.74, p< .05). Distance did not make a difference for receipt of help with housework, maintenance or at least one task. Logistic Regression Results Due to the small frequencies for help received with personal care and financial support, logistic regression analyses were only conducted with the following three tasks: housework, transportation and maintenance. Model 1 on Table 26 shows that, similar to the crosstabular results, gender is not a good predictor of help received with housework as the model chi-square is not statistically significant. Meanwhile, Model 2 shows that education level is a strong predictor. Controlling for gender, the results show that the middle educated elderly are about half as likely to receive such help as the low educated elderly (B=-.58, p< .05, odds ratio=.56). Highly educated elderly parents, meanwhile, are about one-fifth less likely to receive this help relative to the least educated (B=-1.71, p<.01, odds ratio =.18). As the literature suggests that working class elderly women would receive the most help, this interaction term was incorporated into Model 3. The interaction term is not statistically significant, probably because both low educated men and women are the most likely to receive help. In Table 27, Model 1 shows that elderly mothers are two and a half times as likely 103 Table 26 Logistic Regression For Elderly Parents Who Receive Help With Housework Model 1 + Model 2 Model 3 Odds Odds Odds Variable 6 S.E. Ratio B S.E. Ratio B S.E. Ratio Gender 0.37 0.24 0.35 0.24 0.32 0.51 EducH -1.71 ** 0.56 0.18 -1.68 * 0.67 0.19 EducH -0.58 * 0.30 0.56 -0.56 0.49 IntLEH 0.04 0.58 Constant -2.27 *** 0.20 -1.98 *** 0.21 n.s. -1.99 *** 0.23 n.s. Goodnes  Of Fit x2=760.50 df=758 n.s. x2=750.88 df=756 n.s. x2=751.37 df=756 n.s. IntLEH= interaction terra for low educated elderly mothers +Hodel Chi-Square Not Statistically Significant at .05 level * p<.05; ** p<.01; *** p<.001 Table 27 Logistic Regression For Elderly Parents Who Receive Help With Transportation Model 1 Model 2 Model 3 Odds Odds Odds Variable B S.E. Ratio B S.E. Ratio B S.E. Ratio Gender 0.93 *** 0.20 2.53 0.92 *** 0.20 2.51 1.21 *** 0.36 3.35 EducH -0.62 * 0.29 0.54 -0.93 * 0.43 0.65 EducH -0.11 0.21 -0.43 0.39 IntLEH -0.42 0.43 Constant -1.84 *** 0.17 -1.73 *** 0.18 -1.75 0.20 Goodnes  Of Fit x2=760.34 df=758 n.s. x2=753.01 df=756 n.s. x2=753.73 df=755 n.s. IntLEM= interaction term for low educated elderly mothers * p<.05; ** p<.01; *** p<.001 104 to receive transportation assistance than fathers (B = .93, p< .001, odds ratio=2.53). Model 2 shows that when gender is statistically controlled, those with a high level of education are about half as likely to receive transportation compared to the least educated (B=-.62, p < .05, odds ratio =.54). However, gender is still a stronger predictor (B=,92, p<.001, odds ratio=2.51). The interaction term for low educated women was not statistically significant, therefore, the proposition that low educated elderly women would be the most likely to receive transportation is not supported. Results for help with maintenance (Table 28) show that similar to transportation, mothers are twice as likely to receive assistance than fathers (B=.81, p<.01, odds ratio=2.25). However, when gender is statistically controlled, education level is not a significant predictor. Interestingly, Model 3 shows that there is a statistically significant effect for the interaction term of low educated mothers; however, it shows that they are less likely to receive help when compared to other gender and education groups (B=-1.19, p. <05, odds ratio=.30). It was expected that this group would be the most likely to receive assistance. It should be noted, however, that the standard error value is rather high (.60) so caution should be used in interpreting this finding. Once again, for illustrative purposes, help with housework is examined when several potential influential variables are statistically controlled. The results in Table 29 show that the highly educated elderly are much less likely to receive help from their children compared to those who with a low level of education (B=-l .69, p < .01, odds ratio=. 18). The results for those with a medium level of education show the same pattern yet they approach statistical significance (B= .56, p= .06, odds ratio = .57). This reiterates the earlier findings 105 Table 28 Logistic Regression For Elderly Parents Who Receive Help With Haintenance Model 1 Model 2 Model 3 Odds Odds Odds Variable 8 S.E. Ratio 6 S.E. Ratio 8 S.E. Ratio Gender 0.81 ** 0.25 2.25 0.81 ** 0.25 2.25 1.63 ** 0.52 5.86 EducH -0.03 0.32 -0.93 0.58 EducH -0.08 0.28 -1.01 0.56 IntLEH -1.19 * 0.60 0.30 Constant -2.51 *** 0.22 -2.49 *** 0.24 -2.03 *** 0.25 Goodnes  Of Fit x2=760.47 df=758 n.s. x2=758.90 df=756 n.s. XJ=760.06 df=755 n.s. IntLEH= interaction ten for low educated elderly mothers * p<.05; ** p<.01; *** p<.001 106 Table 29 Logistic Regression For Elderly Parents Who Receive Help With Housework Froi Their Children: Multiple Variables Variable •6 S.E. Exp (B) Gender 0.31 0.27 EducH -1.69 ** 0.57 0.18 EducH -0.56 *+ 0.30 0.57 Marital -0.06 0.25 Age 0.05 ** 0.02 1.06 Income 0.00 0.38 NumChild 0.37 0.33 Child/dist 0.49 0.28 Constant -6.78 *** 1.90 Goodnes  of Fit x2=749.92 df=751 n.s. *+ p=.06; * p<.05; ** p<.01; *** p<.001 107 from the crosstabular analyses. Another statistically significant predictor variable here is age (B=.05, p< .01, odds ratio=1.06). This is not surprising. The older one gets, the greater one's health care needs are likely to be. No other variables emerged as significant predictors. 108 CHAPTER EIGHT CONCLUSIONS AND AREAS FOR FURTHER RESEARCH The recent British Columbian government health initiative "New Directions for a Healthy British Columbia" (Ministry of Health and Ministry Responsible for Seniors, 1991b) is premised on the idea of moving health care services closer to people's homes. Within this framework, it is acknowledged by both this provincial government as well as those throughout the country that caregivers to the elderly and disabled need strengthened support based at the community level (Chappeli, 1993:48). As these "closer to home" policies slowly come into effect, a key question that emerges is how caregiving in Canada socially organized? This study was designed to examine both the independent and joint effects of gender and class on intergenerational support. In so doing, the intent was to explore some of the characteristics of parental caregivers, and who they are likely to be helping. Children Who Help Their Parents The literature has shown that the caregiving burden is borne disproportionately by women (Finley, 1989; Horowitz, 1985a; Kaden and McDaniel, 1990; Stoller, 1983). Consistent with previous research findings, this thesis showed that many more daughters than sons provided their parents with help across most tasks. There were no tasks in which sons were more likely to give assistance. What emerges is the striking difference between the likelihood of daughters and sons providing help. These findings are consistent with the 109 feminist viewpoint that gender is key in determining who are the caregivers of the elderly: women provide the care. Taking a step further, the next question was who are these sons and daughters, and what are their characteristics? Utilizing key ideas from the time-available or competing demands hypothesis (Finley, 1989; Lang and Brody, 1983; Stoller, 1983) the results showed that for both sons and daughters, demands such as work, marriage and poor health status did not pull them away from their helping duties. It should be reiterated, however, that in performing these subanalyses, cell sizes decreased substantially thus possibly influencing tests of significance. On the other hand, what emerged from this study was the importance of proximity to parents and its influence on helping behaviour. Both sons and daughters who lived within an hour of their parent(s) were more likely to provide various kinds of help, most notably housework, transportation and maintenance. This suggests that the effects of distance between caregiver and the recipient should continue to be examined in future caregiving studies, especially those which utilize a multivariate analytic framework. Interestingly, while provincial governments are discussing bringing health care services closer to home, the results show that, conveniently, the sons and daughters who provide this help are also closer to the recipient's home. The concern, however, is for those elderly parents who do not have children living close-by. How will the services these people receive differ from those whose children are more proximate? While the caregiving literature has focused much of its attention on the effects of gender, comparatively little work has been done on the effects of class. This is particularly 110 true for Canadian research. The results of this study show that as suggested in the literature, in particular in the work of Arber and Ginn (1992) who discuss their notion of leverage, there are class differences in helping behaviour. Class, as measured by one's level of education, was related to adult children helping their parents with tasks such as housework, transportation, financial support, as well as help with at least one task. Key differences emerged, however, according to the task involved. In addition, the results were not always linear or in the anticipated direction. That different tasks were examined in this study is an important point. Arber and Ginn (1992), along with many studies in the class and caring literature (see Abel, 1990; Glazer, 1990; Kinnear and Graycar, 1984), did not examine the effects of class according to specific tasks. Instead, they discuss helping behaviour in a generic sense. Arber and Ginn (1992:630) do at least acknowledge that there may be class differences in nature of tasks performed. Meanwhile, Rosenthal and Gladstone (1993:127) comment that middle class support consists of providing money or gifts, while the working class tend to give more direct services. These researchers do not, however, state what those services are. Interestingly, Horowitz comments that "the social class of the caregiver tends not to affect the amount of help given as much as it affects the type of assistance given" (1985b:219). However, she too gives very few examples. The results from this thesis show that it is important to examine individual tasks as there are class differences in helping behaviour. For example, the findings here show that, as expected, fewer upper class children assisted with housework than those from the middle class. However, the middle class were more likely to give help than those from the working 111 class. This was a surprising finding for the literature suggests that the caregiving burden is felt more by the working class than people from other class backgrounds (Abel, 1990; Arber and Ginn, 1992, Kinnear and Graycar, 1984). The results for transportation revealed that adult children from the upper and middle classes were the most likely to provide this kind of assistance. The education effect occurred between the least educated and those with a middle level of education or higher. This in itself is not a surprising result. This finding can be explained using Arber and Ginn's notion of leverage. The upper classes have the most financial, material and cultural resources and are thus able to afford a car compared to the working class. Other results showed that there was no support for the association between class and help with tasks such as maintenance and personal care. Therefore, all these findings reveal that the earlier proposition that class is related to helping is only partially supported, depending on which task and which classes are being compared. Future studies on class and caregiving ought to examine the various tasks people help their parents with. Failure to do so potentially masks important class differences. In addition to examining the independent effects of class and gender on helping behaviour, this thesis sought primarily to determine whether gender and class interact to influence helping. Some researchers utilizing a feminist/political economy perspective suggest that working class women bear a greater burden with respect to caregiving (Glazer, 1990; Walker, 1983). Arber and Ginn's notion of leverage (1992) supplements this premise. Women are socially expected to help their parents more than men, but women with the least leverage, that is, having the least financial, material and cultural resources, have the least 112 ability to displace their caring tasks. Therefore, it was expected that working class, or low educated, daughters would be the most likely to provide help with tasks such as housework and personal care, while upper class men would provide financial support. The data revealed that there may indeed be an interaction effect for help with housework. Unfortunately, the frequencies of sons helping were low so interpretations must be done cautiously. However, it would appear that level of education for sons did not make a difference in whether they give help: few sons provided assistance regardless of educational background. Alternatively, for daughters, there was an association between education level and help, but not in the expected direction. In a similar pattern seen earlier, middle class daughters were the most likely to give this assistance, while the least educated were the least likely to help. This was not expected as it is contrary to the dominant view in the literature. Further contradictory evidence to this view emerges from the results for help with at least one task. The findings here also show that it is middle class women who are the key helpers relative to other education/gender groups. Perhaps middle class women, while not possessing the leverage that the upper class have, have fewer siblings living close-by to assist them in their helping duties relative to working class women. It was also suggested earlier that middle class women may be the most likely to have parents who live with their other children compared to the upper and working class. There was a clear interaction effect for daughters providing transportation. More women from, once again, the middle educated group provided this support. However, this was only slightly more than the highly educated. The education effect, however, occurs between the least educated and those with a middle level of education or higher. Level of 113 education did not make a difference for sons who give this type of assistance. Therefore, similar to the situation discussed above, the expectation that gender and class interact to influence helping was only partially supported, depending once again on the tasks involved. Further, some unanticipated findings emerged, such as that which shows middle class women being the most likely to help with housework. Future research may want to examine more carefully why this is the case. In addition, future caregiving studies on interaction effects would be wise to examine help given with different tasks. Little research has examined how the effect of variables thought to influence likelihood of helping varies by class. By statistically controlling for education level, the findings hinted that for upper class women who help with housework, work may be a competing demand. This was not the case for either the other women or for men. For these particular analyses the sample size decreased considerably so cell sizes became very small. This may have influenced tests of significance so the findings should be treated carefully. However, the results suggest that researchers who examine work as a competing demand perhaps should explore whether this demand is similar for people from different class backgrounds. The effect of distance was also examined, and this variable has an important influence, in particular for daughters. More women who lived closer to their parent(s) gave help compared to those who lived farther away, but the magnitude of the difference was strongest for the least educated. This suggests that working class daughters who live a considerable distance from their parents may not have the resources to travel to their parent's home, therefore, they are unable to provide help. Once again, these results intimate that 114 geographic distance is an important variable to consider in future studies on class and helping behaviour. Finally, with respect to helpers, the logistic regression results show that although class differences do emerge in helping behaviour with some tasks, gender is the much stronger predictor variable. Being a woman results in being much more likely to help one's parent than does belonging to a certain class group. The only task for which this was an exception was the provision of financial support where belonging to the upper classes yielded a greater likelihood of giving this type of assistance. Elderly Parents Who Receive Help While most of the caregiving literature examines helping behaviour from either the givers or receiver's point of view, this study examined a sample of helpers as well as an unrelated sample of elderly parents with a health hmitation who receive help. In reference to this latter group, the preceding literature review found that elderly mothers tend to be the care-receivers compared to elderly fathers. The findings in this study reveal that more mothers than fathers receive support with maintenance, transportation and at least one task, but are as likely as fathers to receive help with housework, financial support and personal care. It is thought that perhaps both mothers and fathers are also receiving household assistance from community home support groups, so this may explain why there is no gender difference with help received with housework. The results of the effects of education level on receipt of help reveal that the working 115 class elderly are the most likely to receive help from their children with housework, while the upper class are the least likely. The same pattern emerges for mothers and fathers, so while there is no support for an interaction effect, it is nevertheless interesting that an education effect exists for both men and women who have a health limitation. One explanation for this finding is that the working class have the least leverage so they are unable to purchase assistance to undertake modifications to their homes that would facilitate performing these tasks. For help received with housework, logistic regression results show that level of education is a stronger predictor of receiving help than is gender. Meanwhile, gender has a stronger effect for tasks such as transportation and maintenance. The only other class effect that emerged was for help received with transportation. The logistic regression showed that highly educated parents are about half as likely to receive this assistance compared to the least educated. This is also consistent with the leverage idea. However, the middle educated were not significantly less likely to receive help compared to the least educated, as seen earlier in the contingency tables. No interaction effects emerged with help received on any of the singular tasks. This was not anticipated. However, in contrast, there was an interaction effect for help received with at least one task. The results show that fewer fathers with a high level of education received such help relative to any other gender/education group. This is a possible indicator that upper class elderly fathers turn to their wives for support. If their wives cannot provide it, or they are widowers, then these men probably have the leverage to enlist outside help for the task they need assistance with. They may also be more able to purchase medical equipment, or else pay for necessary modifications to their homes. While there was no 116 evidence to support that working class elderly women are the most disadvantaged, some support does appear showing that upper class men are the most likely not to have to rely on their children for help. It is interesting to note that for a task such as housework, or help with at least one task, most of the helpers came from the middle class. Meanwhile, elderly from a working class background were the most likely to receive such filial assistance: As these two samples are unrelated, little information is known about who the helpers are helping, and from whom the receivers are receiving help. However, these findings hint that perhaps adult children who emerge from working class backgrounds and experience some social mobility are helping their parents who lack the leverage to purchase outside help. In other words, working class elderly may be in the most need of their children's assistance: they may not be able to pay for outside help, specialized equipment or home modifications, nor may they be aware of where to go to for community support. They may be receiving help from their children who are in the same class background as themselves, but some of these parents may also be receiving help from children who have experienced greater economic advances than themselves. Closer examination of the data on recipients of help reveals that marital status did not influence whether fathers or mothers receive help from their children across most tasks. The exception being that more unmarried women received assistance with transportation compared to married women. These results were surprising for the literature shows that marital status is an important variable in determining likelihood of receiving assistance from children (Dwyer and Coward, 1991; Stoller and Earl, 1983). 117 Another unanticipated finding was that distance between parent and child only influenced receipt of help on only a few tasks. For women this was with transportation, while for men, the influence was on help with maintenance. It may be that the small sample sizes may have influenced significance tests for these results as well as those for marital status. Contributions to the Literature While in the course of conducting this research it was acknowledged there were some limitations. These were discussed earlier in Chapter Five. However, despite these limits, this research has also made some contributions to the literature. As suggested earlier, a review of the literature found that there is very little Canadian research on class and caring, and even less on how class and gender interact to influence caring. Consequently, this study's findings contribute to the gaps in this literature. In addition, while there has been a fair amount of research on gender differences in intergenerational support (see Chappeli, 1992; Kaden and McDaniel, 1990; Penning, 1990), few studies utilize a nationally representative sample such as the one analyzed in this thesis. The majority of studies concentrate on small, local samples (see Aronson, 1991; Kaden and McDaniel, 1990; Rosenthal, 1987). As suggested earlier, most caring research concentrates on either the giver or the receiver. In this study, both groups were examined. Although the two samples were unrelated, the same key independent variables were used. For example, the existence of interaction effects between class and gender were examined separately for both the helpers 118 and the receivers of help. As discussed previously, interesting differences emerged. Another contribution that was commented upon earlier was that helping/receipt of help was examined according to various tasks as well as on a combined task basis. This gives a richer understanding of how gender and class influence the likelihood of either giving or receiving help. From a statistical analysis perspective, this study also tested several of the key propositions using contingency tables as well as the more sophisticated logistic regression technique that statistically controls for all the variables in the model. Directions for Further Research While this study has examined the independent and joint effects of gender and class on intergenerational support, there are still many areas open for further research. For example, this study treated helping and receipt of help as dichotomous variables coded as one when either help was given by someone in the adult children sample, or else when help was received by someone in the elderly parents sample. This method is not unlike that used in other research (see Arber and Ginn, 1992; Dwyer and Coward, 1991; Rosenthal, 1987). However, as Arber and Ginn (1992:630) note, there may also be class and gender differences in the amount of time devoted to helping, as well as how much help is received. This thesis found that more middle class women help their parents with tasks such as housework compared to other class groups. The proposition that working class women are the most disadvantaged compared to men, and middle and upper class women may receive stronger 119 support if help is examined according to quantity of help given. For example, working class women may actually give the most help on tasks such as housework, personal care, or help with at least one task. Meanwhile, upper class men may provide the least amount of help with these tasks, but may give more financial assistance. Further, perhaps elderly mothers from the working class receive more help with transportation, financial support, as well as housework while the upper class elderly receive the least. Measuring quantity of help would also be another way of examining the competing demands, or time available hypothesis. Many studies which test this hypothesis measured help in terms of number of hours of help given (see Lang and Brody, 1983; Stoller, 1983). Perhaps key differences among men and women who either work, are married or are in poor health emerge when helping is measured in this manner. In addition, it would be interesting to know for how long the help has been given. Elderly people from the working class may be in a poorer state of health, and begin encountering health difficulties earlier (Arber and Ginn, 1992:620). Therefore, their caregivers may be forced to help them over a longer timeframe. If findings show that they do give the most help for the longest time, while the upper class give the least help for the least amount of time, this would lend more support the idea that the working class are the most disadvantaged with respect to caregiving. With respect to gender differences in caregiving, it would be expected that, similar to Montgomery and Kamo's (1989) findings, daughters give help for longer periods of time for tasks such as housework, personal care and transportation. In addition, given that as Gee and Kimball state, "women get sick but men die" (1987:31), it may be the case that elderly 120 mothers may also receive help for a longer period of time. This thesis concentrated on unpaid assistance between children and their parents. Future research may wish to more closely test for interaction effects between class and gender by examining the use of paid help. Using Arber and Ginn's idea of leverage, one would expect the upper classes (either helpers or receivers) to be more likely to pay for extra assistance, medical equipment, or necessary home modifications. Meanwhile, the working class would utilize these services the least. Other questions worth investigating include whether it is the case that men rely more on paid help than women. How does use of paid help vary according to specific tasks? There have been many calls in the gerontological literature for more longitudinal studies (Dwyer and Coward, 1992b; Horowitz, 1992). Given the richness of information and the unique insights this particular methodology offers, a longitudinal study oh gender, class and caring would be very useful. Questions to be explored may include at what point do the the different class groups turn to formal assistance? When does work begin to act as a competing demand and does this vary by class and gender? If, as the literature suggests, that sons tend to help for shorter periods of time (Montgomery, 1992), do results from a longitudinal study support this? If so, why does this occur? This study has used education as a proxy measure for class. Future studies may wish to tap into other measures of class, and measure it for both giver and receiver. It would be interesting to see if the same kinds of results emerge. Examples for other measures include asking about the person's major occupation throughout the employment years. In addition, future studies may want to examine for interaction effects between class, gender, and 121 ethnicity. One other consideration for future research would be to examine gender and class differences in helping/receipt of help using two samples where the giver is related to the receiver. For example, what are the gender and class backgrounds of the helper's parents? If it is the case that middle class women are the most likely to help with housework, as found here, then are their parents from the same class background? Or, is there a pattern where children from working class backgrounds experience some social mobility and end up belonging to the middle class, yet given their parents' lack of leverage, these children provide the needed assistance? Finally, throughout this thesis, I was very consistent in using the term "help" instead of caregiving/reeeipt of care because this was the wording used in Statistics Canada survey. However, Lapierre (1993), who also uses the same survey, explicitly uses the term "caregiving" to describe the survey's results. Barer and Johnson (1990:26) comment that throughout the literature there are multiple definitions of caregivers which influence how these people are identified and studied. While developing standardized terms to define caregiving is far beyond the scope of this thesis, it is important to point out this issue that exists in the literature, with the hope that some work can be done to create some commonality in terminology. Closing Comments As health care for seniors moves "closer to home", it is evident that the burden of 122 informal caregiving will fall disproportionately on women. Interestingly, the results from this thesis show that middle class women in particular are at risk for increased caregiving burdens. It would appear that government programs that are designed to "support" caregivers consist of respite care programs and support groups (Aronson, 1991:143). Other kinds of programs such as home care will be of little help to caregivers if they are targeted primarily at those who do not have an informal support network (Chappeli, 1993:49). Programs such as these do not necessarily lighten the caregiving load, instead they merely sustain it (Wid:U3; Hooyman, 1992; McDaniel and Gee, 1993:68). Perhaps it is time that policymakers begin to address the larger issues of why there are gender and class imbalances in the caregiving process and how they can be overcome. 123 BIBLIOGRAPHY Abel, Emily K. 1990 "Family Care and the Frail Elderly." in Emily K. Abel and Margaret K. Nelson (Eds.) Circles of Care: Work and Identity in Women's Lives. Albany, New York: State University of New York Press, 65-91. Abel, Emily K. 1986 "Adult Daughters and Care For the Elderly." Feminist Studies. 12(3), 479-497. Alford-Cooper, Finnegan 1993 "Women as Family Caregivers: An American Social Problem." Journal of Women and Aging. 5(1), 43-57. Arber, Sara and Jay Ginn 1992 "Class and Caring: A Forgotten Dimension." Sociology. 26(4), 619-634. Arber, Sara and Jay Ginn 1991 "The Invisibility of Age: Gender and Class in Later Life." Sociological Review. 9(2), 260-291. Arber, Sara and Jay Ginn 1990 "The Meaning of Informal Care: Gender and the Contribution of Elderly People." Ageing and Society. 10(4), 429-454. Archbold, Patricia G. 1983 "Impact of Parent-Caring on Women." Family Relations. 32(1), 39-45. Aronson, Jane 1991 "Dutiful Daughters and Undemanding Mothers: Constraining Images of Giving and Receiving Care in Middle and Later Life." in Carol Baines, Patricia Evans and Sheila Neysmith (Eds.) Women's Caring: Feminist Perspectives on Social Welfare. Toronto: McLelland & Stewart, 138-168. Aronson, Jane 1990a "Women's Perspectives on Informal Care of the Elderly: Public Ideology and Personal Experience of Giving and Receiving Care." Ageing and Society. 10(1), 61-84. Aronson, Jane 1990b "Old Women s^ Experiences of Needing Care: Choice or Compulsion?" Canadian Journal on Aging. 9(3), 234-247. 124 Aronson, Jane 1986 "Care of the Frail Elderly: Whose Crisis? Whose Responsibility?" Canadian Social Work Review. 45-58. Aronson, Jane 1985 "Family Care of the Elderly: Underlying Assumptions and Their Consequences." Canadian Journal on Aging. 4(3), 115-125. Babbie, Earl 1989 The Practice of Social Research. (5th ed). California: Wadsworth. Barer, Barbara M . and Colleen L. Johnson 1990 "A Critique of the Caregiving Literature." The Gerontologist. 30(1), 26-29. Blalock, Herbert M . Jr. 1972 Social Statistics. (2nd ed). New York: McGraw-Hill Company. Branch, Laurence G. and Alan M Jette 1983 "Elders' Use of Informal Long-Term Care Assistance." The Gerontologist 23(1), 51-56. Brody, Elaine M . , Morton H. Kleban, Pauline T. Johnsen, Christine Hoffman and Claire B. Schoonover 1987 "Work Status and Parent Care: A Comparison of Four Groups of Women." The Gerontologist. 27(2), 201-208. Brody, Elaine M . and Claire B. Schoonover 1986 "Patterns of Parent-Care When Adult Daughters Work and When They Do Not." The Gerontologist. 26(4), 372-381. Caldock, Kerry 1992 "Domiciliary Services and Dependency: A Meaningful Relationship?" in Frank Lacko and Christina R. Victor (Eds.) Social Policy and Elderly People. Hants, England: Avebury Ashgate Publishing Ltd., 99-108. Chappeli, Neena L . 1993 "Implications of Shifting Health Care Policy for Caregiving in Canada." Journal of Aging and Social Policy. 5(1/2), 39-55. Chappeli, Neena L . 1992 Social Support and Aging. Toronto: Butterworths Canada Ltd. 125 Chappell, Neena L . 1985 "Social Support and the Receipt of Home Care Services." The Gerontologist. 25(1), 47-54. Chodorow, Nancy 1978 The Reproduction of Mothering: Psychoanalysis and the Sociology of Gender. Los Angeles, California: University of California Press. b Condran, John G. and Jerry G. Bode 1982 "Rashomon, Working Wives, and Family Division of Labor: Middletown, 1980." Journal of Marriage and the Family, 44(2), 421-426. Dale, Angela G., Nigel Gilbert and Sara Arber 1985 "Integrating Women Into Class Theory." Sociology. 19(3), 384-408. Dewit, David J., Andrew V. Wister and Thomas K. Burch 1988 "Physical Distance and Social Contact Between Elders and Their Adult Children." Research on Aging. 10(1), 56-80. Dwyer, Jeffrey W. and Raymond T. Coward 1992a "Gender, Family, and Long-Term Care of the Elderly." in Jeffrey W. Dwyer and Raymond T. Coward (Eds.) Gender, Families, and Elder Care. Newbury Park, California: Sage Publications, Inc., 3-17. Dwyer, Jeffrey W. and Raymond T. Coward 1992b "Gender and Family Care of the Elderly: Research Gaps and Opportunities." in Jeffrey W. Dwyer and Raymond T. Coward (Eds.) Gender, Families, and Elder Care. Newbury Park, California: Sage Publications, Inc., 151-162. Dwyer, Jeffrey W. and Raymond T. Coward 1991 "A Multivariate Comparison of the Involvement of Adult Sons Versus Daughters in the Care of Impaired Parents." Journal of Gerontology. 46(5), S259-S269. Dwyer, Jeffrey W. and Raymond T. Coward 1990 "The Association of Gender, Sibling Network Composition, and Patterns of Parent Care By Adult Children." Research on Aging. 12(2), 158-181. Finley, Nancy J. 1989 "Theories of Family Labour as Applied to Gender Differences in Caregiving for Elderly Parents." Journal of Marriage and the Family. 51(1), 79-86. Gee, Ellen M . and Meredith M . Kimball 1987 Women and Aging. Toronto: Butterworths Canada. 126 Gilligan, Carol 1982 In A Different Voice: Psychological Theory and Women's Development. Massachusetts: Harvard University Press. Glazer, Nancy 1990 "The Home As Workshop: Women as Amateur Nurses and Medical Care Providers." Gender and Society. 4(4), 479-499. Grabb, Edward G. 1990 Theories of Social Inequality: Classical and Contemporary Perspectives. (2nd Ed.) Holt, Rinehart and Winston of Canada. Grabb, Edward G. 1988 "Conceptual Issues in the Study of Social Inequality" in James Curtis, Edward Grabb Neil Guppy and Sid Gilbert (Eds.) Social Inequality in Canada: Patterns, Problems, Policies. Scarborough, Ontario: Prentice-Hall Canada Inc., 1-19. Graham, Hilary 1983 "Caring: A Labour of Love." in Janet Finch and Dulcie Groves (Eds.) A Labour of Love: Women, Work and Caring. London: Routledge and Kegan Paul, 13-30. Horowitz, Amy 1992 "Methodological Issues in the Study of Gender Within Family Caregiving Relationships." in Jeffrey W. Dwyer and Raymond T. Coward (Eds.) Gender, Families and Elder Care. Newbury Park, California: Sage Publications Inc. 132-150. Horowitz, Amy 1985a "Sons and Daughters as Caregivers to Older Parents: Differences in Role Performance and Consequences." The Gerontologist. 25(6), 612-617. Horowitz, Amy 1985b "Family Caregiving to the Frail Elderly." in C. Eisdorfer, M . Lawton and G. Maddox (Eds.) Annual Review of Gerontology and Geriatrics. 5, 194-246. Hooyman, Nancy R. 1992 "Social Policy and Gender Inequities in Caregiving." in Jeffrey W. Dwyer and Raymond T. Coward (Eds.) Gender, Families, and Elder Care. Newbury Park, California: Sage Publications, Inc., 181-201. Hooyman, Nancy R. 1990 "Women as Caregivers of the Elderly: Implications for Social Welfare Policy and Practice." in David E. Biegel and Arthur Blum (Eds.) Aging and Caregiving: Theory, Research and Policy. Newbury Park, California: Sage Publications, Inc., 221-241. 127 Kaden, Joan and Susan A. McDaniel 1990 "Caregiving and Care-Receiving: A Double Bind for Women in Canada's Aging Society." Journal of Women and Aging. 2(3), 3-26. Kinnear, David and Adam Graycar 1984 "Ageing and Family Dependency.n Australian Journal of Social Issues. 19(1), 13-26. Lang, Abigail M . and Elaine M . Brody 1983 "Characteristics of Middle-aged Daughters and Help to Their Elderly Mothers." Journal of Marriage and the Family. 45(1), 193-202. Lapierre, Louise 1993 "Measures of Outside Care Given and Received by Seniors." Health Reports. 4(4) Statistics Canada, Catalogue Number 82-003. 355-366. Lee, Gary R. 1992 "Gender Differences in Family Caregiving: A Fact in Search of a Theory" in Jeffrey W. Dwyer and Raymond T. Coward (Eds.) Gender, Families, and Elder Care. Newbury Park, California: Sage Publications, Inc., 120-131. Lee, Gary R., Jeffrey W. Dwyer and Raymond T. Coward 1993 "Gender Differences in Parent Care: Demographic Factors and Same-Gender Preferences." Journal of Gerontology. 48(1), S9-S16. Marx, Karl and Frederich Engels [1845-46] 1978 The German Ideology in Robert C. Tucker (Ed.) The Marx-Engels Reader. New York: W.W. Norton & Company, Inc., 146-200. Marx, Karl and Frederich Engels [1848] 1978 Manifesto of the Communist Party in Robert C. Tucker (Ed.) The Marx-Engels Reader. New York: W.W. Norton & Company, Inc., 469-500. Matthews, Sarah H. and Tena Tarler Rosner 1988 "Shared Filial Responsibility: The Family as the Primary Caregiver." Journal of Marriage and the Family. 50 (1), 185-195. McDaniel, Susan 1992 "Women and Family in the Later Years: Findings from the 1990 General Social Survey." Canadian Woman Studies. 12(2), 62-64. 128 McDaniel, Susan A. and Ellen M . Gee 1993 "Social Policies Regarding Caregiving to Elders: Canadian Contradictions." Journal of Aging and Social Policy. 5(1/2), 57-72. McDaniel, Susan A. and Allison L. McKinnon 1993 "Gender Differences in Informal Support and Coping Among Elders: Findings from Canada's 1985 and 1990 General Social Surveys." Journal of Women and Aging. 5(2), 79-98. McPherson, Barry D. 1990 Aging As A Social Process. (2nd ed). Toronto: Butterworths. Merrill, Deborah M . 1993 "Daughters-in-Law as Caregivers to the Elderly." Research on Aging. 15(1), 70-91. Ministry Responsible for Health and Ministry Responsible for Seniors 1993a New Directions: Meeting the Health Needs of Today's Families. Victoria: Ministry of Health Ministry Responsible for Health and Ministry Responsible for Seniors 1993b New Directions for a Health British Columbia: Meeting the Challenge, Action For A Healthy Society. Victoria: Ministry Responsible for Health and Ministry Responsible for Seniors Montgomery, Rhonda J. V. 1992 "Gender Differences in Patterns of Child-Parent Caregiving Relationships." in Jeffrey W. Dwyer and Raymond T. Coward (Eds.) Gender, Families, and Elder Care. Newbury Park, California: Sage Publications, Inc., 65-83. Montgomery, Rhonda J. V. and Mary McGlinn Datwyler 1990 "Women and Men in the Caregiving Role." Generations. 14(3), 34-38. Montgomery, Rhonda J. V . and Yoshinoro Kamo 1989 "Parent Care by Sons and Daughters." in Jay A. Mancini (Ed.) Aging Parents and Adult Children. Lexington, Massachusetts: Lexington Books, 213-230. National Advisory Council on Aging 1990a "The NACA Position on Community Services in Health Care for Seniors." National Advisory Council on Aging, No. 8. Minister of Supply and Services Canada. National Advisory Council on Aging 1990b "The NACA Position on Informal Caregiving: Support and Enhancement." National Advisory Council on Aging, No. 9. Minister of Supply and Services Canada. 129 Norusis, Marija, J 1991 The SPSS Guide to Data Analysis for SPSS/PC+. Chicago, Illinois: SPSS Inc. Norusis, Marija J. 1990a SPSS/PC+ Statistics 4.0 for the IBM PC/XT/AT and PS/2. Chicago Illinois: SPSS Inc. Norusis, Marija J. 1990b SPSS/PC+ Advanced Statistics 4.0 for the IBM PC/XT/AT and PS/2. Chicago, Illinois: SPSS Inc. Penning, Margaret J. 1990 "Receipt of Assistance by Elderly People: Hierarchical Selection and Task Specificity." The Gerontologist. 30(2), 220-227. Province of British Columbia 1991 Closer to Home: The Report of the British Columbia Royal Commission on Health Care and Costs. Victoria: Crown Publications Inc. Rosenthal, Carolyn J. 1987 "Aging and Intergenerational Relations in Canada." in Victor W. Marshall Aging in Canada: Social Perspectives. Markham, Ontario: Fitzhenry and Whiteside, 311-342. Rosenthal, Carolyn J. and James Gladstone 1993 "Family Relationships and Support in Later Life." Journal of Canadian Studies. 28 (1), 122-138. Ross, Catherine E. 1987 "The Division of Labour At Home." Social Forces. 65(3), 816-833. Seccombe, Karen 1992 "Employment, the Family, and Employer-Based Policies." in Jeffrey W. Dwyer and Raymond T. Coward (Eds.) Gender, Families, and Elder Care. Newbury Park, California: Sage Publications, Inc., 165-180. Speare, Alden Jr. and Roger Avery 1993 "Who Helps Whom in Older Parent-Child Families" Journal of Gerontology. 48(2), S64-S73. Statistics Canada 1991 The 1990 General Social Survey — Cycle 5, Family and Friends: Public Use Microdata File Documentation and User's Guide. Ottawa: Minister of Supply and Services. 130 Stoller, Eleanor Palo 1990 "Males as Helpers: The Role of Sons, Relatives, and Friends." The Gerontologist. 30(2), 228-235. Stoller, Eleanor Palo 1983 " Parental Caregiving by Adult Children." Journal of Marriage and the Family. 45 (4), 851-858. Stoller, Eleanor Palo and Lorna L. Earl 1983 "Help with Activities of Everyday Life: Sources of Support for the Noninstitutionalized Elderly." The Gerontologist. 23(1), 64-70. Stone, Leroy O. 1988 Family and Friendship Ties Among Canada's Seniors: An Introductory Report of Findings from the General Social Survey. Catalogue 89-508. Ottawa: Minister of Supply and Services. Stone, Robyn, Gail Lee Cafferata and Judith Sangl 1987 "Caregivers of the Frail Elderly: A National Profile." The Gerontologist. 27(5), 616-626. Strain, Laurel A. and Barbara J. Payne 1992 "Social Networks and Patterns of Social Interaction Among Ever-Single and Separated/Divorced Elderly Canadians." Canadian Journal on Aging. 11(1), 31-53. Strieb, Gordon 1985 "Social Stratification and Aging." in Robert H. Binstock and Ethel Shanas (Eds.) Handbook of Aging and the Social Sciences. 2nd edition. New York: Van Nostrand Reinhold Company Inc., 339-368. Taylor, Rex and Graeme Ford 1983 "Inequalities in Old Age: An Examination of Age, Sex and Class Differences in Sample of Community Elderly. "Ageing and Society. 3(2), 183-208. Victor, Christina R. and Maria Evandrou 1987 "Does Social Class Matter, in Later Life?" in Silvana di Gregorio (Ed.) Social Gerontology: New Directions. London: Croom Helm, 252-267. Vogt, W. Paul 1993 Dictionary of Statistics and Methodology: A Nontechnical Guide for the Social Sciences. California: Sage Publications. 131 Walker, Alan 1991 "The Relationship Between the Family and the State in the Care of Older People." Canadian Journal on Aging. 10(2), 94-112. Walker, Alan 1983 "Care For Elderly People: A Conflict Between Women and the State." in Janet Finch and Dulcie Groves (Eds.) A Labour of Love: Women, Work and Caring. London: Routledge and Kegan Paul, 106-128. Walker, Alexis J. 1992 "Conceptual Perspectives on Gender and Family Caregiving." in Jeffrey W. Dwyer and Raymond T. Coward (Eds.) Gender, Families, and Elder Care. Newbury Park, California: Sage Publications, Inc., 34-46. Walsh, Anthony 1990 Statistics for the Social Sciences With Computer Applications. New York: Harper & Row Publishers, Inc. Walsh, Anthony 1987 "Teaching Understanding and Interpretation of Logit Regression." Teaching Sociology. 15(1), 178-183. Weber, Max [1922] 1968 Economy and Society: An Outline of Interpretive Sociology Volume 2. Guenther Roth and Claus Wittch (Eds.). Bedminster Press Incorporated. New York. Wister, Andrew 1990 "Living Arrangements and Informal Social Support Among the Elderly." Journal of Housing for the Elderly. 33-43. Wister, Andrew V. and Ellen M . Gee 1994 "Age at Death Due to Ischemic Heart Disease: Gender Differences." Social Biology. 41 (1-2), 110-126. 132 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0086889/manifest

Comment

Related Items