UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Investigating the significance of single and married mothers’ child care arrangements on work/family… Madaisky, Dolores P. 1993

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

Item Metadata


831-ubc_1994-0015.pdf [ 1.77MB ]
JSON: 831-1.0087303.json
JSON-LD: 831-1.0087303-ld.json
RDF/XML (Pretty): 831-1.0087303-rdf.xml
RDF/JSON: 831-1.0087303-rdf.json
Turtle: 831-1.0087303-turtle.txt
N-Triples: 831-1.0087303-rdf-ntriples.txt
Original Record: 831-1.0087303-source.json
Full Text

Full Text

INVESTIGATING THE SIGNIFICANCE OF SINGLE AND MARRIED MOTHERS’CHILD CARE ARRANGEMENTS ON WORK/FAMILY ROLE BALANCEByDOLORES P. MADAISKY -B.A., The University of British Columbia, 1990A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF ARTSinTHE FACULTY OF GRADUATE STUDIES(Department of Family Studies)We accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAOctober, 1993©Dolores P. Madaisky, 1993In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.Department of ( ‘flLLj chndu..z’The University of British ColumbiaVancouver, CanadaDate 29iDE.6 (2/88)1]AbstractThe relationship between a mother’s child care arrangements and herwork/family role balance has been investigated. Differencesbetween married and single mothers’ child care arrangements andtheir work/family role balance has also been investigated. Sevenhypotheses that phrase these relationships empirically have beenpostulated. From the Statistics Canada General Social Survey Cycle5 version (1990) a sub-sample of 504 mothers with children underthe age of 5 years was selected. Multiple regression, chi squareand ANCOVA procedures were applied. The results indicated that 2of the 7 hypotheses were significantly supported. In particular,the significant findings suggest that family care, day care andrelative care vary negatively on satisfaction and positively onnumber of hours the respondent works per week (two of the fourdependent measures that represented the work/family role balanceconcept). Also, the results suggested that married mothers sufferfrom less work/family role imbalance than single mothers. Inaddition, single and married mothers differed in their use of daycare and relative care. The findings were explained by currentwork/family literature. Work/family role balance was partiallyexplained by the methodological representation. Rather, thistheory was used as a guide to further the search for a theoreticalframework that incorporates both enhancement and conflict. Inaddition, an implication for future research is to locateadditional variables that exhaustively explain this concept so thatwork/family role balance may become a theory that can stand up tothe most vigorous empirical testing.11).Table of ContentsPageAbstractTable of ContentsListofTablesChapterIntroduction . . . .Review of the Literature .Empirical Literature: Research onSingle Mothers . . . . . .Empirical Literature: Research onChild Care Arrangements . .P.iroses .HypothesesMethod .Sample and Data CollectionMeasuresI.II.III.11. ii).. vi1. 59• . . • • . • . 142020. 22. 22. 24Dependent Variables: Work/Family RoleBalance .Independent Variables: Spouse Presentor Absent • .Independent Variables: Child CareArrangements . . . ..Independent Variables: Frequency of andReliance on Child CareControl Variables: Satisfaction withDivison of Household LabourControl Variables: Finances243334• . . . 36• • . . 39• • • . 40ivDiscussion: Satisfaction ScaleDiscussion: Amount of Housework Doneby the RespondentDiscussion: Shift and/or WeekendWorkIV.Table of Contents(continued)ResultsDemographic Summary of the SampleUnivariate Distributions: SatisfactionScaleUnivariate Distributions: Amount ofHousework Done by the RespondentUnivariate Distributions: Whether theRespondent Works Shift and/or Weekendor]cUnivariate Distributions: Number ofHours the Respondent Works (Outsidethe Household) Per Week424242434445• . . . • . 4545 4646• 47474766697072Univariate Distributions:Present or Absent • .Univariate Distributions:ArrangementsUnivariate Distributions:Child CareUnivariate Distributions:Child CareUnivariate Distributions:Household Labour . .Univariate Distributions:Hypotheses Testing .Additional Findings •V. DiscussionSpouseChild CareReliance onFrequency onDivision ofIncome .73VTable of Contents(continued)Discussion: Number of Hours theRespondent Works (Outside theHousehold) Per Week 76Limitations and Strengths 83Implications 87References 89AppendixA 94AppendixB 100AppendixC 101viList of TablesPageTable 2.1 Four Typologies Representing the Conceptof Work/Family Balance 8Table 2.2 Child Care Arrangements: Married MothersVersus Single Mothers 17Table 3.1 Weighted Results Using the Variableof Household Type 24Table 3.2 Reliability and Factor Analysis onSatisfaction Measures 28Table 33 Factor Analysis for Scales and VariablesUsed to Measure Work/family Balance 31Table 3.4 Frequencies of Child Care Arrangementsby Household Type 36Table 3.5 Frequency Listings for the Variables ofReliance on Child Care and Frequency ofChild Care Use 39Table 4.1 ‘Child Care Arrangements and Singles VersusCouples with Children 46Table 4.2 Correlation Matrix of The Independentand Control Variables 50Table 4.3 Regression Effects of Family Care andControl Variables on the Four DependentVariables 53Table 4.4 ANCOVA Analysis of Household Type withControl Variables on the Four DependentVariables 56Table 4.5 Regression Analysis of Relative Care andControl Variables on the Four DependentVariables 58Table 4.6 ANCOVA Analysis: Reliance and FrequencyVariables with the Control Variables onthe Four Dependent Variables 61viiList of Tables(continued)PageTable 4.7 ANCOVA Analysis: Household Type and Typeof Child Care Arrangement with CovariatesAgainst the Four Dependent Variables 65Table 4.8 Summary Table: Regression Analysis ofIndependent and Control Variables on theFour Dependent Variables 68Table 5.1 Summary of Hypotheses 82Chapter IIntroductionNorth American society, in the last decade, has seen a shiftin the labour force. Today women are no longer expected to solelyperform the roles of mother and homemaker. Statistics Canada(1990) reports that women with young children are becoming thefastest growing working group to enter the labour force in Canada.The percentage of women from 1950 to 1987 with children under theage of six who were employed outside the home increased from 12% to65% (Kamerman & Kahn, 1987). As such these women with youngchildren are expected to balance the multiple roles of worker,mother and perhaps, wife.In North America, female—headed households have become thefastest growing form of family (Geoff, 1990). Some researcherspredict that 40% of all current and potential marriages among youngwomen now in their late twenties and thirties will end in divorce(Hughes & Galinsky, 1990). Between 1970 and 1983 the female—headedhouseholds increased by 100%, while dual-headed householdsincreased by only 12% in the same period (Geoff, 1990). Maritalbreak—ups account for many of these female—headed householdsbecause women usually obtain joint or sole custody of theirchildren. Statistics Canada (1990) reports that 72.4% of separatedor divorced women had obtained sole custody of their children;while only 11% of men obtained sole custody of their children in2the same period. This rapid growth of the single mother familysuggests more encompassing role demands for these women.Single mothers are more likely to experience work/familyconflict because of the problems of low income and lack of childcare arrangements. For example, low income may result frominsufficient or inconsistent child support payments. As Fletcher(1989) points , “... current levels of child support payments arefar from adequate and fall short of the principle of equity betweenfathers and their children” (p. 416). Therefore, women who are thesole heads of their household are often forced into balancing a lowincome and, perhaps, facing the effects of poverty.In addition to this problem of low income, the child careissue for single mothers is salient. In Canada, 1.9 millionchildren under 13 years of age are in need of licensed day care(National Council of Welfare, 1988). There are presently only1,318 licensed spaces available. The largest need for child careis for children in the newborn to 3 years of age group because ofthe influx of young women with young children into the labourmarket. Single mothers may be more affected by this shortage oflicensed spaces since the problem of affordability and adequacy ofchild care are factors in their options of child care. Singlemothers are under great pressure to balance the dual roles ofworker and mother because the family is dependent on her income andalso her ability to find adequate care for her children.For both single and married mothers, child care and workdemands are problematic. Kamerman (1985) has found that, “When3queried as to the problems they (workers) face in managing work andfamily life, or the barriers they face in entering or advancing inthe labour force, women are almost unanimous in defining childcare...as the single most important problem they face” (p. 259).Other researchers also view the importance of adequate, affordableand dependable child care as a critical variable in establishing abalance between work and family life (Elman & Gilbert, 1984; Fine,Donnelly & Voydanoff, 1986; Hum & Balgopal, 1990; Kurdek, 1988).Research that investigates both child care and work/family rolebalance or conflict considers how child care arrangements affectwork/family role balance. This includes ease of scheduling,arrangements the child care facility makes when the child fallsill, and proximity of the facility to home and work. Yet, what isabsent in the literature is research that investigates therelationship of different types of child care arrangements and theresulting work/family role balance that may occur. Single motherswith their greater maternal role demands, lower income, andinadequate availability of alternative care, would seem to be atgreater risk of work/family imbalance than married mothers.The purposes of this study are to investigate therelationships and/or differences that exist between mothers’ childcare arrangements, presence or absence of a spouse, and greater orlesser work/family role balance. Empirical work to date suggeststhat there may be a difference between certain child carearrangements and greater work/family role balance (Fine, Donnelly& Voydanoff, 1986; Hofferth, Brayfield, Deich & Holcomb, 1990; Hum4& Balgopal, 1990; Kamerman, 1985; Kurdek, 1986; Johnson, 1986;Turner & Smith, 1983).5Chapter IIReview of the LiteratureWithin the literature addressing work/family role balancethere is a lack of continuity in the terminology that refer towork/family imbalance or the clashing of work/familyresponsibilities such as; work/family role conflict (Baruch &Barnett, 1986; Goff, Mount & Jamison, 1990); work/family rolestrain (Holahan & Gilbert, 1979; Hochschild, 1989; Keith & Schafer,1980; Voydanoff, 1987) and role spillover (Bolger, Delongis,Kessler & Wethington, 1989). The majority of research focuses onwork/family conflict which is the “simultaneous occurrence of twoor more sets of pressures such that compliance with one would makedifficult compliance with the other” (Goff et al., 1990, p. 795).Thus, the imbalance between work and family is the emphasis foundin most of the literature reviewed with the exception of roleenhancement theorists such as Marks (1977). This view seems to beboth pessimistic and deterministic in that work/family issues areportrayed as a negative type of interaction.Tiedje, Wortman, Downey, Eminons, Biernat, and Lang (1990) havewritten an article that concisely discusses the empirical focus onwork/family role balance. This literature which has investigatedthe maintenance of multiple roles (such as worker, mother and wife)has usually been guided by one of the three overarching conceptualframeworks: conflict theory, role enhancement theory and balancetheory.6Role conflict theory, according to Tiedje et al. (1990),“views energies of individuals as finite and role demands asinfinite. Role conflict, then, becomes an inevitable, normal, andexpected consequence of multiple roles” (p.64). Stress theoristssuch as McCubbin (1979, 1980) and Kessler & McCrae (1982) exemplifythe use of this type of theory. Researchers who discusswork/family role conflict in terms of role allocation would alsofit into this type of theoretical framework. Role allocationassumes that families allocate roles in mainly two types of ways:traditionally, where the woman does the stereotypically “feminine”chores within the household and the man does the “masculine”chores. A traditional arrangement is exemplified in the woman whowould be expected to cook, clean and care for the children, whilethe husband would be expected to perform the gardening, maintainthe automobiles, and the exterior of the home. The other way thatroles may be allocated within the household is symmetrically, wheretask assignment is perceived in a more equitable division (Berk,1985; Hill, 1987; Hochschild, 1989; Hughes & Galinsky, 1988; Nock& Kingston, 1988; Perry-Jenkins & Crouter, 1990; Repetti, 1989).Therefore, role conflict assumes that the roles of mother, wife andworker are constantly competing.Role enhancement theory “emphasizes the potential benefits ofmultiple roles (Marks, 1977)” (Tiedje et al., 1990, p. 64).Stephen Marks (1977) uses role strain theory to explain why someindividuals performing multiple roles experience role strain whileothers do not. Marks has found that under certain conditions7multiple roles result in energy creation rather than energydepletion. He sees the relative degree of commitment in variousroles as a critical determinant of role strain. When there is avariation in the level of positive corrunitment to different roles,the energy and time allocated for overcommitted interests expandand encroach on the time and energy available for undercommittedtasks. Yet, he and other researchers (Holahan & Gilbert, 1979;Pleck, 1987; Voydanoff, 1987; Voydanoff, 1990) argue that multipleroles may positively enhance self esteem, self worth and a feelingof purpose. This enhancement comes from an individual realizingthat both work and family life are equally beneficial andimportant.The last and most important theoretical framework to bediscussed is a theory that was postulated by Tiedje et al. (1990)called balance theory. These researchers propose that bothconflict and enhancement theories alone do not adequately explainparents’ experience with balancing work and family simultaneously.Instead of conceptualizing the work/family issue as a continuumwith conflict on one side and enhancement on the other, Tiedje andher colleagues suggest that one can view conflict and enhancementas “simultaneously” occurring. These researchers found that thework/family role issue can be scored both on enhancement andconflict concurrently. In addition, the variable that mostinfluences their “balance” measure is the score of conflict.Regardless of perceptions of enhancement, “women who perceivedtheir roles as conflicting were more depressed and less satisfied8as parents” (Tiedje et al., 1990, p. 70). This quote emphasizesthat although parents can experience and report that they areenhanced by having these two roles, they can at the same time alsoexperience work/family conflict. Depending on the amount ofwork/family conflict and enhancement the parents report, they arecategorized into four different typologies as seen in Table 2.1.Table 2.1Four TypolocTies Representing the Concept of Work/Family Balance**ENHANCEMENTLOW HIGHHIGH Low Enhancement High EnhancementHigh Conflict High Conflict(Conflicting) (Imbalance)CONFLICTLOW Low Enhancement High EnhancementLow Conflict Low Conflict(Imbalance) (Optimum Balance)**From “Women with Multiple Roles:Role-Compatibility Perceptions,Satisfaction, and Mental Health”by Tiedje et al., 1990, Journal ofMarriage and the Family, 52, p. 65The balance of conflict and enhancement for managing work andfamily roles is key to understanding how the multiple roles ofworker, mother, and perhaps wife coexist. Ultimately, Tiedje(1990) view optimum health and satisfaction when conflict islow and enhancement is high. One can envision the possibility that9work may act as a support network for a single mother while it canalso cause friction within her family role as mother. It is thebalance of work enhancing her life more than adding friction to itthat exemplifies the conceptual framework of work/family rolebalance.Empirical LiteratureResearch on Sincle MothersStatistics Canada (1990) reports that by the year 2010, womenwill be the sole heads of their families in 25 to 30% of Canadianhouseholds. With lack of social or spousal supports, lack ofsupport payments from the children’s father(s) (Fletcher, 1989),paired with inadequate and unaffordable child care arrangements(National Council of Welfare, 1988), single mothers are finding itincreasingly difficult to manage a balance between the dual rolesof mother and worker. Researchers such as Hum and Balgopal (1990)agree that single parent mothers experience multiple role imbalanceas they “must manage the home, children and work simultaneously”(p. 247). Child care in such households affects the daily life insuch ways as timing of the day’s events, the activities undertaken,the amount of money spent, and even the development of interactiverelationships between family members. “Within the single parentfamily, this interactional effect between child care and the familybecomes more relevant as the source of care becomes an extendedpart of the family experience” (Hum & Balgopal, 1990, p.248).10Hum and Balgopal (1990) studied 10 fathers and 10 mothers whowere single parents. The salaries of the participants ranged from$8,000 to $35,000 (U.S.$). Fathers tended to be in a higher incomebracket with a higher status profession. Ages of children in careranged from 8 months to 17 years. Mothers tended to have youngerchildren, both male and female, while fathers had older childrenwho were almost all male. Mothers sought advice from family orfriends for their child care decisions and usually utilized onetype of child care consistently. Occasionally this care wassupplemented by family and friends on a limited basis. Theresearchers found that 40% of the sample used a day care center,30% used a family day care and 30% used in—home care.These researchers gathered qualitative data through interviews anddiscovered that single mothers saw child care as a stabilizingforce in their family’s life. In other words, when child care wasadequately provided it mediated the tension between work and familydemands. It is also important to ask what kinds of mediatingeffects different types of child care arrangements may have onwork/family demands.ICurdek (1988) studied 35 white custodial mothers and theirchildren who had a mean age of 11 years and ranged from 6 to 17years of age. The mothers’ mean age was 35 years with a meanmonthly income of $1,169.00 (U.S. $) with 80* of the sample workingfull-time. Of interest in this study was the way in which socialsupport mediated stresses of being a single parent. One of theareas that was focused on for social support was child care. For11single mothers, the most used supports for child care wererelatives, own parents, and ex—spouses. Also, 31% of the samplesaid they had used their own children as sources for child care.A limitation of this study is that children’s ages were notcollected. This information would be useful in viewing anotherangle of child care sources. What this study emphasizes is thelack of outside support these single mothers have and, therefore,must rely on their own children for child care support.Another study by Johnson (1986) conducted a systematic randomsample from divorce abstracts of 1,994 divorced women from countieswithin Ohio state. A sub—sample of 381 women who were divorcedbetween 6 months to 9 years, with a range in education from 9 yearsto doctoral and/or professional degrees was selected. Theresearcher looked at sources of support available to these divorcedwomen in managing conflicts between work and child careresponsibilities. The results of the study found that regardlessof the many support systems these women had, they usually chose tomanage the employment—child care conflict themselves. Johnson(1986) has speculated the reason women attempt to manage this typeof conflict:The mother’s handling these employment child careconflict situations may be a reflection of a normativeexpectation rather than a preferred response. As womenentered the labour force, they added employmentresponsibilities to their existing familyresponsibilities. Not only were they supposed to handleboth of these responsibilities, but they were expected tolet family responsibilities have precedence overemployment ones when there was a conflict (Johnson &Firebaugh, 1985; Pleck, 1977). The traditional patternof a divorce settlement reinforced the division of12responsibility: men as providers through maintenance andchild support payments; women as the caretakers, withcustody of children. In actuality, with reduced ornonexistent financial support available, the custodialparent, usually the mother, assumed both responsibilities(Johnson, 1986, p. 101).This example suggests the conflicting pressures women areexperiencing as sole heads of their families. Johnson (1986)suggests that normative societal rules dictate that a mother’sallegiance should be to her family first and paid work second.This can give rise to much inner conflict for the mother andconflict between the two spheres of work and family.Literature that investigates single parents emphasizes anongoing struggle for these parents to maintain balance betweentheir roles as mother and worker. These studies operationalizework/family balance as the “managing of home, children and worksimultaneously” (Hum & Balgopal, 1990, p. 247). Althoughresearchers investigate how child care may be a variable thatmediates work/family role conflict or balance, they do not considerdifferent types of child care specifically, and the differingeffects that could influence work/family role balance directly.A large majority of the research on work/family role balanceconsiders dual earners or married couples where the wife may not beemployed outside the home. The literature operationalizes socialsupport frequently using spousal support (Baruch & Barnett, 1986;Bohen & Viveros-Long, 1981; Elman & Gilbert, 1984; Gray, Lovejoy,Piotrkowski, & Bond, 1990; Houseknecht & Macke, 1981; Rudd &McKenry, 1986; Thompson & Walker, 1989). Berk (1985), Bohen, and13Viveros-Long (1981), and Hughes and Galinsky (1988) have found thatwomen are most satisfied with their roles when the dual roles ofworker and wife/mother are supported by an outside agent, forexample, a boss or spouse. For example, when a spouse takes onsome of the household chores, women are more satisfied with boththeir work and family roles (Menaghan & Parcel, 1990). The abovestudies emphasize the lack of support that single mothers sometimesface. Since it is not possible to measure spousal support forsingle mothers, other supportive networks must be explored.Friends, clergy, counsellors, relatives, ex—spouse, ex—spouse’sparents, own parents, and people within their social networks(i.e., at a job or day care) are some examples of supportivenetworks (Kurdek, 1988). In fact, Fine et al. (1986) suggest thatsingle parents may use child care workers and other support systemssuch as friends and family in lieu of spousal support. Therefore,child care may become a form of support for the single parent whichthen may help alleviate work/family role imbalance and managework/family role balance. Since there is little research that hasstudied single parents, and in particular, single mothers,generalizations can only be inferred from the large majority ofwork/family role conflict research collected on marriedindividuals.14Research on Child Care ArrangementsIn 1988 the National Council of Welfare investigated theexisting arrangements for child care with a special emphasis on lowincome families. British Columbia is one of the top threeprovinces that has major disparities between numbers of childrenthat need child care and numbers of licensed spaces available.Licensed care refers to the provincially set standards that a childcare facility must meet in order to be provincially approved.Areas such as health, child’s safety, group size of the carefacility, area of the care facility and staff—child ratios areinvestigated for approval. Government approval is necessary forobtaining grants and subsidies from governmental agencies.Families with low incomes who qualify for subsidized child careunder the existing system may not get services to which they areentitled because of a shortage of spaces in licensed child care.According to a study conducted by Hofferth, Brayfield, Deich,and Holcomb (1990), child care for pre—schoolers falls into fivedifferent categories, each of which will be discussed.Center—based programs or institutional care refers toestablished settings where children are cared for within a groupthat is away from the child’s home. These centers can be run byreligious groups, nursery schools, cooperatives, or may be workrelated.Family child care refers to child care where the child istaken outside their home to another wherein an adult cares for thechild. One finds that the care giver is often a mother who is also15caring for her own children. This type of care can be licensed orunlicensed and is left up to the discretion of the care giver. Thecare giver can be a friend, neighbour, or someone that the familydid not previously know.In—home care refers to the care giver coming to the home ofthe child. Sometimes the care giver may bring his/her own childrenwith them to this home. This type of care can also include nannies(or au—pairs) which families sometime share between households.Spousal/parental care refers to consistent care given byeither the mother or the father of the child. This could be doneespecially if working schedules do not overlap.Relative care refers to the care giver being related to thechild in some way. For instance, the child could be given care byhis/her grandmother/father, aunt/uncle, or cousin on a regularbasis in the child’s home or within the relative’s home.A National Child Care Survey (1990) was conducted by Hofferthet al. (1990) in Washington, D.C.. The survey representsapproximately 27 million U.S. households with children underage 13. For full—time employed mothers, the most common use ofchild care was a child care center (30.9%), followed by family care(23.0%), parental care (21.5%), and finally relative care outsidethe child’s home (11.0%). In-home provider care was used by only3.4% of the sample of working mothers. Mothers who were workingonly part—time used parental care the most (43.6%), followed bychild care centers (19%), and family child care (12.5%).16Hofferth et al. (1990) found that employed mothers and mothersof pre—schoolers are more likely to use non—parental care than arenon—employed mothers and mothers of school age children. In theUnited States, center based programs have increased in use for preschool age children of employed mothers over the past 25 years(from 6% in 1965 to 28% in 1990). There has also been a decline inthe use of in—home providers and relatives, while the use of familychild care has stayed about the same over the past 25 years.Probably, the most crucial finding is the comparison betweensingle and married mothers which is presented in Table 2.2.Hofferth et al. (1990) found that single mothers are more likelythan two—parent families to use a center—based child care for theirpre—school child (38% versus 26%). Table 2.2 also shows thatmother-only families use relative care either within the child’shome or another’s home more than two—parent families (26% versus18%, respectively). Single mothers use their spouse (or ex—spouse)less than married mothers as would be assumed. Also, Table 2.2shows that married mothers use family child care more than singlemothers (26% versus 19%, respectively).Table 2.217Child Care Arranamnts Married Mothers Versus Single Mothers________—r\\ \\\—---II II\‘I \\\\ICenter Based Care Family Ci!d C Reive Child CrSinie Mothers Mcrried MothereChild Care Arrangements: Single Versus MarriedSingle Mothers Married MothersCenter Based CareFamily Child CareRelative Child Care38.00%19. 00%28.00%28.00%28.00%18. 00%2 —18Another study that demonstrates the importance of child careas a mediating variable to work/family balance is that of Turnerand Smith (1983). Their sample consisted of 252 single parents(218 females, 26 males, and 8 whose sex was unknown) with custodyof one or more minor children living in or near Albuquerque, NewMexico. The demographic data approximated the national and statepopulation of single parents in terms of sex, age, education,religion and income. One significant finding was that only a smallpercentage of single parents used family child care. This wasconsistent with Johnson’s (1986), Hofferth et al. ‘s (1990) and Humand Balgopal’s (1990) research which found single mothers use daycare centers more than family or in—home child care alternatives.In terms of cost, in—home care is the most expensive childcare arrangement, followed by child care centers, family child careand finally relative care (which could be free). Subsidies forlicensed child care within Canada are available to low incomefamilies. Therefore, this type of child care would be the mostattractive yet difficult to obtain. Since single mothers usuallyhave a mean lower income than married mothers, they may have to usealternative care arrangements such as relative care which may bemore affordable and accessible. Single mothers may not have achoice in child care arrangements but use the arrangement that ismost affordable and readily available.19Literature which focuses on work/family role balance and thevariables that enhance or lessen this balance has been reviewed.Types of child care arrangements and the differing uses due tomarital status (or household type) has also been investigated.Single mothers are faced with a lack of spousal support (in somecases, lack of any social support), lower incomes than theirmarried counterparts, and child care arrangements that are at timesinaccessible and unaf fordable. Since these mothers need to work tosupport their families they must also find affordable child care.Canadian mothers who have low incomes indicate they do and preferto use licensed care as they will then receive financial assistancefrom the government. But there is a shortage of spaces within alltypes of licensed care which means single and married mothers areforced to choose alternative types of care such as unlicensedcenter—based care and family care or relative care. Thecoitthination of unavailability of center-based care paired withinadequate or unaffordable child care alternatives creates aconcern for the mother who needs child care. This may in turn,create an even greater work/family imbalance for the married motherbut, in particular, the single mother.20PurposesThe primary purposes of this study are:1) to investigate the relationship between a mother’s childcare arrangements and her work/family role balance2) to investigate differences that may exist between marriedmothers and single mothers’ child care arrangementsand their work/family role balance.HypothesesEmpirical studies suggest that the accessibility andaffordability of child care may have a mediating effect inwork/family role balance (Hum & Balgopal, 1990; Johnson, 1986;Turner & Smith, 1983). Also, Klein (1985) and Turner and Smith(1983) suggest that certain forms of child care could mediate moresuccessfully than others. Although, studies argue that family andfriends are a highly supportive network, there is still an ideal tohave an independent agency who supplies educational and emotionalsupport for the child. Therefore, the following hypotheses areproposed:El: The greater the use of family child care, the greaterthe work/family role balance.Investigations which contrast married parents and singleparents show that they differ in how they use child care (Hogan,Hoa & Parish, 1990; Hum & Balogopal, 1990; Turner & Smith, 1983).Single mothers tend to use center—based and relative child caremore than married mothers; while married mothers use family childcare more than single mothers do (Hofferth et al., 1990). Theunderlying assumption that needs to be confirmed within this study21is that married mothers have greater work/family role balance thansingle mothers.H2: Married mothers will have a greater use of familycare than single mothers.H3: Married mothers will rate their ability to balancework and family higher than single mothers.H4: The greater the use of relative care, the lowerthe work/family role balance.Hofferth et al. (1990), Hum and Balgopal (1990), Kurdek(1988), and Johnson (1986) suggest single mothers may have agreater reliance on their child care arrangements. In particular,Hofferth et al. (1990) and Hum and Balgopal (1990) suggest singlemothers may use more than one type of child care simultaneouslywhen compared to married mothers.H5a: Using one primary source rather than more than one type ofchild care will be accompanied by greater work/family rolebalanceH5b: An increased reliance on child care arrangements will beaccompanied by lesser work/family balance.The last hypothesis is a summation of all the individualhypotheses. This hypothesis restates that child care arrangementsmay be different for single and married mothers and therefore, mayresult in a difference in their work/family role balance.H6: The difference in child care arrangements of married andsingle mothers leads to a difference in work/family rolebalance between married and single mothers.22Chapter IIIMethodSample and Data CollectionThe data being used in this study has been collected byStatistics Canada in the Cycle 5 (1990) version of the GeneralSocial Survey (Family and Friends). The sample consisted of 13,495respondents who were interviewed by telephone. These respondentswere both male and female (5,967 and 7,528, respectively). Theweighted sample represented the overall Canadian population.A sub—sample of 560 respondents was selected out of the above13,495 subjects. This sub—sample has been screened to ensure thatall respondents are women and have children under the age of 5 thatare living with them. In particular, the two variables of sex ofthe respondent (DVSEX) and age of the child (CHILDAGE) were used toproduce a sub—sample of 560 respondents. The mothers may bemarried, common-law, cohabiting, single (never—married), separated!divorced or widowed. Income, education, number of children under5 years that reside with the mother, and other pertinent sociodemo—graphic information will be employed in the analysis to follow.SamplingWeighting the sample responses is done in order to ensure thatthe sample is representative of the larger Canadian population.Initially, the population weights (which are the frequencies listedthroughout for each variable) represent what would be seen in theoverall population. Variance weights are then computed by dividing23the population weights already calculated by Statistics Canada anddividing it by the mean weights of the 560 respondents. As seen inTable 3.1, the variable household type (HHTYPE) was manipulated inthis manner. The first group of frequencies in the table is theunweighted sample, the second frequencies are the populationweights and third frequencies are the variance weights. In thefirst frequency listing, the single parent with children categoryand the couple with children category respectively contain 82(14.6%) and 422 (75.4%) respondents. The frequencies which containthe variance weights for the single parent versus couple contain 46(8.1%) and 431 (76.9%) respondents. Since the single parentcategory is initially so small, it is important that the number inthis category when weights are applied goes down by 36 respondents.With such small numbers the question of external validity israised. Since the number of single parents is indeed small, theability to generalize outside the sample to which this analysis isbeing applied cannot be done with great confidence. As a result,using weights becomes less valuable since the reason the weightswere being applied primarily was to ensure that external validityto the Canadian population was possible. Therefore, weights willnot be applied and generalizability to an overall Canadianpopulation (or any sample outside this study) is not possible.24Table 3.1Weighted Results Using the Variable of Household Type.Frequency Valid PercentUNWEIGHTEDSingle Parent with Children 82 14.6Couple with Children 422 75.4POPULATION WEIGHTSSingle Parent with Children 60,571 8.1Couple with Children 572,633 85.6VARIANCE WEIGHTSSingle Parent with Children 46 8.1Couple with Children 431 76.9MeasuresDependent VariableWork/Family Role BalanceWork/family role balance is conceptualized by Tiedje et al.(1990). As previously mentioned, Tiedje et al. (1990) view optimumhealth and satisfaction when conflict is low and enhancement ishigh.A review of the work/family balance literature shows that themeasure of conflict (or one type of imbalance) is split into twocategories; subjective and objective measurement. Subjectivemeasures are usually “satisfaction with role” type of statements or25questions. For example, Tiedje et al. (1990) ask the respondent tostate on a 5—point Likert—type scale how strongly she agrees ordisagrees with the statement, “Having a career often causes me tobe tired, irritable or short tempered with my children.” Anexample of the enhancement subjective measure would be “Having acareer helps me to better appreciate the time I spend with mychildren.” Tiedje et al. (1990) found that conflict measures werehighly predictive of satisfaction with both work and family roles(the reliability measure using a Cronbach’s Alpha was .74 for theconflict measures).What is lacking in the research conducted by Tiedje et al.(1990) is that they do not consider the subjective measurestogether with objective measures. It is advantageous to look atobjective measures such as work schedules and amount of houseworkresponsibilities because one gets greater insight into the amountof work/family role balance that women experience. Anotheradvantage to using objective measures is that it can either addcredibility or decrease the validity of a subjective measure. Inother words, construct validity (which is concerned with thetheoretical constructs and, in particular, with how the measuresaccurately represent the construct) is being addressed by usingobjective measures since work/family role balance or imbalance canbe quantified. By using subjective data alone one can notgeneralize with great confidence since each person has their ownfeelings and emotions that are not so easily tapped or measured.26By looking objectively through time assessments of work andfamily management one can assess the objective part of thework/family role balance construct. For example, Baruch andBarnett (1986) look at time spent doing certain tasks such astaking the child to the doctor. It has been found (Baruch &Barnett, 1986; Bielby & Bielby, 1988; Bolger, Delongis, Kessler, &Wethington, 1989; Goff, Mount, & Jamison, 1990; Repetti, 1989) byresearchers that subjects’ time commitments are correlated withtheir perceived work/family role balance or imbalance. Forexample, if one works a 60 hour work week with no spouse present tolend support to the household chores, then, the mother would likelyrate her roles as highly conflicting with little enhancement.Therefore, the subjective and objective measures should combine toadequately represent work/family role balance.Within this study, 4 scales are used to measure work/familyrole balance: satisfaction, amount of housework done by therespondent, amount of shift and/or weekend work, and number ofhours the respondent works (outside the household) per week.Operationally, work/family role balance may be depicted in thesubjective and objective measures as a high score on overallsatisfaction, a middle score on amount of housework done by therespondent (which represents either a little more or a little lessthan 1/2 of the housework performed by the respondent), a low scoreon shift and/or weekend work, and a part—time work rating (roughly20 hours per week). This working model was drawn from literaturewhich suggests that if women are more satisfied (Baruch & Barnett,271986; Bolger, Delongis, Kessler & Wethington, 1989), share inhousework (Baruch & Barnett, 1986; Elinan & Gilbert, 1984; Gray,Lovejoy, Piotrkows]ci, & Bond, 1990; Hughes & Galinsky, 1988), worklittle or no shift and/or weekend work (Staines & Pleck, 1983), andwork part-time (Moen & Dempster-McClain, 1987; Shelton, 1990) theywill tend to have less work/family conflict, and may evenexperience enhancement from both roles. Therefore, these scoreswould represent the ideal of work/family role balance.The subjective measure initially started with 6 differentitems of satisfaction: satisfaction with spouse/partner or singlestatus; satisfaction with immediate family; satisfaction with theway housework is shared; satisfaction with job or main activity;satisfaction with balance between job and family; and finally,satisfaction with time and other interests. Using a 4—pointLikert—type scale (‘1’ very dissatisfied to ‘4’ very satisfied)respondents were instructed to use these numbers to indicate thelevel of satisfaction with the previously mentioned items. Thesescores were taken from the General Social Survey (Friends andFamily) Cycle 5 version and then summed and averaged to form one ofthe dependent variables, the satisfaction scale (SATSCAL). Thereliability analysis is measured using a Cronbach’s alpha of .65for this sample. A factor analysis of this scale indicates thatsatisfaction with family, satisfaction with job, satisfaction withwork/family balance and satisfaction with time and other interestsload on one factor while satisfaction with spouse and houseworkload onto another factor.28Table 3.2 presents results of the factor analysis performed.The variables, satisfaction with family (DVSATFAM), satisfactionwith job (DVSATJOB), satisfaction with work/family balance(DVSATBAL) and satisfaction with available time (DVSATTIM) arerelated to the first factor (alpha = .66). While satisfaction withspouse (DVSATMAR) and satisfaction with division of housework(DVSATHWK) load weakly on the second factor (alpha = .33).Table 3.2Reliability and Factor Analysis on Satisfaction MeasuresRotated Factor MatrixCommunality Factor 1 Factor 2Satisfaction MeasuresDVSATMAR .4660 .10132 .67508DVSATFAM .31775 .54796 .13225DVSATHWK.65171 .10172 .80085DVSATJOB .57625 .73891 —.17396DVSATBAL .65181 .77856 .21367DVSATTIM .58335 .69326 .32053N=494 Alpha = .65The weak alpha of satisfaction with spouse (DVSATMAR) andsatisfaction with division of housework (DVSATHWK) poses a strongargument for keeping the two variables in the satisfaction scale(SATSCAL). Since the reliability (including all the sixsatisfaction items) initially is .65, the scale has reasonable29reliability. Conceptually, it is clearer to keep all six itemstogether in one satisfaction scale. In this way, a more concisescale is formed by incorporating all types of satisfaction withinthe household and work domains.The objective measures that represent the work/family rolebalance concept are amount of housework done by the respondentwithin four predetermined household tasks: meal preparation, mealclean—up, laundry and household cleaning, and outside maintenance.Also, included in the objective measures is whether the respondentworks shift and/or weekend work and the number of hours therespondent works per week (outside the home). These measurementsare included to give an objective time assessment for household andwork management.The household task measurement of work/family role balancecontains four items for which the respondent said she had someresponsibility for performing. The items indicate the amount ofwork done by the respondent within the household. For example, oneof the items involves meal preparation (F3WORKO1). This item statesthat the respondent could have done from less than one quarter toall the meal preparation within the household (‘1’ less than aquarter to ‘4’ all). Finally, if the respondent was classified ina single person household she was scored an ‘8’. The assumptionbeing made by the researchers is that there was no spouse presentto help with the chores. A reliability analysis run on theindependent household items of: amount of meal preparation(F3WORKO1), amount of meal clean-up (F4WORKO1), amount of cleaningand laundry (F5WORKO1), and amount of outside maintenance30(F6WORKO1) indicates a Cronbach’s alpha of .98. When a factoranalysis was conducted on these four items it was found that allfour load on one factor (Eigenvalue of 3.852). Since the workmeasure shows such unity the four work items are summed and dividedto form amount of work done by the respondent or ANTWK1.Shift and/or weekend work are also used to assess theobjective measure of work/family role balance. It has been foundthat 164 or 44% of the respondents are working weekends and/orshift work. Following from these questions 215 or 56% of theworking respondents did not engage in weekend or shift work. Also,in this category 179 respondents are coded as missing.The number of hours the respondent works (other than houseworkor work in the house) per week is another variable used to assessthe objective measure of work/family role balance. Respondentswork an average of 35 hours per week. It should also be noted thatof the 540 respondents considered, 301 or 56% of the sample areworking 5 hours or less per week.A factor analysis of the number of hours the respondent worksper week (HRSLOAD) and whether the respondent works either shiftand/or weekends (SHIFT) was done. Both the variables of HRSLOADand SHIFT load onto one factor (Eigenvalue of 1.0324). Thisprovides support that number of hours the respondent works (outsidethe home) per week and whether the respondent works shift and/orweekend shifts are measuring one concept. Yet, these two measuresrepresent different areas of the work domain and therefore, will bekept as two separate dependent variables within the analysis.31A correlational matrix (see Appendix B) has analyzed dependentvariable against dependent variable and has found that only numberof hours the respondent works (outside the household) per week issignificantly correlated with the satisfaction scale , (n = 15)= -.22, p < .01. This lends support for examining the fourvariables separately.Table 3.3Factor Analysis for Scales and Variables Used to measureWork/family Balance.Communality Factor 1 Factor 2SATSCAL .58378 —.19760 —.73806ANTWK1.91502 .95189 .09444SHIFT.55889 —.37328 .64733HRSLOAD .34209 .26752 .52012As seen in Table 3.3, factor analysis was performed on all thenewly formulated measures: satisfaction scale (SATSCAL), amount ofhousework done by the respondent (AMTWK1), whether the respondentworks shift and/or weekend work (SHIFT) and number of hours therespondent works (outside the household) per week (HRSLOAD). Twofactors become evident when a rotated factor matrix is interpreted.HRSLOAD (number of hours the respondent works (outside thehousehold) per week) and SHIFT (whether the respondent works shiftand/or weekends) load on one factor with respective results of .52and .65. Also, (satisfaction scale) SATSCAL loaded on the samefactor as HRSLOAD and SHIFT, but it loads on the other side of the32factor with a highly negative result of —.74. Finally, (amount ofhousework done by the respondent) ANTWK1 loads on another factor(.95) separately from the other measurements or scales (see Table3.2). Conceptually, this would mean that the variables of numberof hours the respondent works (outside the household) per week(HRSLOAD), and whether the respondent works shift and/or weekendwork (SHIFT) are measuring roughly the same concept in the samemanner. Although, the satisfaction scale is also considering thesame concept, it is measuring it from the opposite spectrum. Inother words, although all three measurements are dealing with thesame concept, the satisfaction scale (which is an attitudinal typeof measure) is almost in direct contrast with number of hours therespondent works (outside the household) per week and whether therespondent works shift and/or weekend work (which are objectivebehavior types of questions). On a more practical level this mightindicate that the satisfaction scale is tapping a differentdimension than the variables of number of hours the respondentworks (outside the household) per week and whether the respondentworks shift and/or weekend work.Of course, the advantage to this finding is that a moreaccurate picture of the duality or multiplicity of the work/familyrole concept is achieved. Also, the amount of housework done bythe respondent (ANTWK1) did not load on the same factor as the restof the measures. Perhaps, this also adds an argument for includingsatisfaction with division of housework as a dependent variablesince not only is the amount of housework performed by the mothera key measure but, the satisfaction with the amount she is33responsible for may also be important. ANTWK1 loads on a differentfactor which suggests that this measure may be tapping into adifferent dimension that also measures work/family role balance.Since the variables load on two factors and are not highlycorrelated, these statistics suggest that each type of dependentvariable may be measuring a different area of the work/family rolebalance concept. Therefore, the satisfaction scale, amount ofhousework done by the respondent, shift and/or weekend work andnumber of hours the respondent works per week are included asseparate dependent variables in order to get the best overallrepresentation of the concept.Independent VariablesSpouse Present or AbsentThe work/family role balance literature considers the presenceor absence of the father in the family as a crucial variableinfluencing the concept. Researchers have either noted theimportance of spousal support or lack of support (Hum & Balgopal,1990; Kurdek, 1988; Rudd & McKenry, 1986). The assumption thatunderlies this research is that not only may a spouse help out withsome household chores and perhaps give moral support, he may alsobring a substantial income to the family. Therefore, singlemothers may experience a qualitatively different life because ofthe absence of the spouse when compared to married mothers.The measure that is used to assess this variable, spousepresent or absent, in the General Social Survey is “Household Type”(HHTYPE). This measure provides 18 categories to assess whetherthe mother is sharing residence with a person or is single. As34seen in Appendix A, the categories that measure spouse present orabsent is “single parent with children” and “couple with childrenonly”. In total, 82 respondents are single parents and 422respondents live as couples. The 56 respondents that did not fitinto these two types of categories are recoded into a residualcategory. Therefore, the total of the sample is 504 respondents.Child Care ArrangementsHofferth et al. (1990) investigated child care usages andfound that primarily, mothers made the decisions about the type ofchild care their child would receive. They also found that maritalstatus, that is whether the mother was single or married, alteredthe results of the mother’s choice of child care.Hofferth et al. (1990) gathered telephone interview data fromapproximately 27 million households with children under age 13, or“3 out of 10 U.S. households” (Hofferth et al., 1990, p.1). Thiswas a nationally representative sample of United States familieswith children under 13 years of age. This study focused on 7 areasof measurement some of which included:1) The kinds of arrangements parents make for their childrenand the amount of time children spend there,2) Expenditures for child care,3) Parental choice and satisfaction with child care, and4) Parents’ views of availability, affordability, andknowledge of their child care.Hofferth et al. (1990) then defined 5 types of child care withinwhich respondents were categorized. The typologies consisted of:center-based child care or day care, family child care, in-homecare, spousal/parental child care and relative care.35Initially, this variable is assessed using the General SocialSurvey’s child care typologies. The researchers who designed theSurvey had split child care into two areas: the child receives careeither within the home or outside the home. Therefore, if thechild is receiving care within the home, the care givers could bethe child’s grandparent, another relative, a sitter or nanny andfinally, someone else. If the child receives care outside the homethe options for child care are: a workplace day care center,another day care center, a sitter or neighbour’s house, agrandparent’s home, another relative’s home or some otherarrangement.To form the variables so they would match the definitions ofthe Hofferth et al. (1990) study and to achieve mutually exclusivecategories, respondents using a workplace day care center andanother day care center are summed and divided to form one day carecenter variable. The ‘sitter or neighbour’s house’ variable isrenamed “family care” to match Hofferth et al.’s (1990)terminology. The relative care variable is formulated exactly likethe day care variable. Therefore, people who use relative care areselected out from the rest of the sample and the two types ofrelative care are then summed and divided to form one variable.For those who arrange child care within the household, one relativevariable is again formed by summing and dividing the two types.The sitter or nanny category falls into Hofferth et al.’s (1990)“in—home” category which was not considered when the hypotheseswere formed and therefore, is not used in this study. T h enumber of respondents that used each type of care can be found in36the following Table 3.4. Only 266 respondents used child care withany regularity.Table 3.4Frequencies of Child Care Arrangements by Household TypeSingle Parent MarriedDay care Center 18 47Family Care 14 88Relative Care (outside) 13 40Relative Care (inside) 3 30N=253Frequency of and Reliance on Child CareFrequency of and reliance on child care are conceptualizedwithin Hofferth et al.’s (1990) study and also from research doneby Hum and Balgopal (1990). They suggest that although singlemothers may use center—based child care more than married mothers,the frequency or amount of use between these mothers differs.These studies suggest that married mothers are more stable and useone primary child care source more than single mothers. In fact,single mothers may vary their child care sources. For example, asingle mother may bring their child to a center one day, to thegrandmother the next and yet, even to a neighbour the next day.The variables have two themes underlying them: first, themothers have a low or high reliance on their types of child careand second, the mothers have a frequency of use of each type ofchild care that could range from one type of care to all four types37of care. The underlying assumption that is made here is that if amother is using child care because she works and has many childrento arrange care for then she is highly reliant on her type(s) ofchild care. Also, this reliance may dictate whether she uses oneor more types of care concurrently.RelianceWomen who are using child care on a regular basis (forexample, they need care for their full—time job) are selected forthe reliance variable. The respondents who state that this was thereason for their child care are scored a ‘2’. Also, the number ofchildren that the respondent needed care for was incorporated. Thetwo measures (reason for child care use and how many childrenrequire care) are then multiplied together to form the reliancevariable. Therefore, the variable of reliance is measuring thestrength of the need for child care. If the respondent has a highreliance for child care, she would score a ‘6’, which would beinterpreted as a mother with at least 3 children and a regular useof child care. As seen in Table 3.5, only 4 respondents score a‘6’, while 61 respondents have two children who need regular careand 200 respondents have one child who needs regular care. Again,an assumption has been made that those mothers with a greaternumber of children who have a regular need of child care would havegreater reliance than those mothers with 1 child and irregular (oreven regular) need of child care.38Freciuency of CareTo obtain the frequencies of child care, each child carevariable and the possible combinations of these variables areconsidered. For example, if someone uses only day care, they areselected out and a new variable is created called DAY. If therespondent reports that they use family care, relative care withinthe home and day care, a new variable is created for thiscombination. Twelve new categories (see Appendix C) are formedwith most cells containing either 0 or under 10 respondents (withthe exception of those cells which include only one type of carebeing used). A final category is formed where respondents eitheruse one or two (or more than two) types of care. This variable iscalled ‘Frequency’ and can be found as a total to the above 12categories in Table 3.5.39Table 3.5Frequency Listings for the Variables of Reliance on Child Care andFrequency of Child Care UseReliance on Child CareCategories Frequencies2 - Respondent has one childin need of regular care 2004 - Respondent has two childrenin need of regular care 616 — Respondent has three childrenin need of regular care 5N=266Frequency of Child CareCategories Frequencies1 — One type of child care 1642 - Two types of child careconcurrently 49N=2l3This variable is trying to tap frequency of use of child care.Researchers have found people may use more than one type of careregularly (Hofferth et al., 1990; Hum & Balgopal, 1990). Thereare 164 respondents who use only one type of care while 49respondents use at least two types of care.Control VariablesSatisfaction with Division of Household LabourWithin the work/family role balance literature, whenresearchers want to see how household task and duties are assignedthey would list many different household duties. The respondentwould then be asked to list who was largely responsible for40performing these duties. Studies vary on how many tasks and dutiesare listed and the detail to which the tasks are probed. Forinstance, Baruch and Barnett (1986) have used 20 items to measurehousehold tasks and child care, while Berk (1985) has used as manyas 45 items to measure just household tasks (not including childcare). In fact, research has consistently found that women areresponsible for the majority of household responsibilities even ifthey are employed outside the home.This variable will be considered as a control in the analysison the three objective measures. Perhaps, satisfaction with thedivision of household labour mediates the mother’s work/familybalance as her roles may be more enhanced. This variable isassessed by using the satisfaction question on division ofhousework (DVSATHWK). Again, this variable was measured using a 4-item Likert—type response (‘1’ very dissatisfied to ‘4’ verysatisfied).Finances and Child CareKoren (1984) reports that 75% of female and 57% of maleemployees reported difficulty in locating current child carearrangements. Scarr and McCartney (1989) state that “for parents,the most pressing issues are affordability and availability ofconsistent and dependable child care” (p. 136). Child care of goodquality and dependability is expensive as it can range from $450.00to $1000.00 per child, per month for full time care (NationalCouncil of Welfare, 1988). As these single mothers are usually ina lower educational level as well, employment can be difficult tofind as it may not pay enough for women to meet child care costs41and other expenses of living (B.C. Task Force on Child Care, 1991).In fact, “...sporadic unemployment and limited work experiencedeter women to take advantage of job promotions or trainingnecessary for advancement and thus, may have consequences forwomen’s professional success” (Joesch, 1991, p.l61).The financial differences between single mothers and marriedmothers are reported by the B.C. Task Force on Child Care (1991)who found that the average income of single mothers was $15,578while the average married family income was $49,442 in B.C. in1989. Thus, it is possible that income may directly influence thework/family role balance variable. The variable that measuresincome is the total family income reported by the respondent. Thisvariable is categorized into 9 levels of income starting at lessthan $5,000 to $80,000+ (C $). The mean for this variable is thecategory of $40,000 to $59,999 (C $) which closely approximates theB.C. Task Force on Child Care mean of $49,442.SummaryThe dependent variables [satisfaction scale, amount ofhousework done by the respondent, whether the respondent workedshift and/or weekend work, and number of hours the respondent works(outside the house) per week], independent variables (spousepresent or absent, child care arrangements, frequency of andreliance on child care) and the control variables (satisfactionwith division of household labour and finances) have beendiscussed. These variables are analyzed in terms of the postulatedhypotheses in the results section to follow.42Chapter IVResultsDemographic Summary of the SampleThe sample of 504 respondents is a sub—sample from the GeneralSocial Survey (Friends and Family) Cycle 5 (1990) version whichcontained 13,495 respondents who were interviewed by telephone.The age of the respondents varies from 16 to 47 years of age witha mean age of 28 years. The majority of the group (65%) fall intothe ranges of 24 to 32 years of age. Respondents marital status is398 respondents married, 51 respondents common—law, 7 are divorced,29 separated and 73 respondents are single (2 were not stated).There are 82 single parents, 422 respondents who fall into thecouples with children category. Sixty-eight percent of the sampleonly have one child in the household that is 5 years or younger,while 28% of the sample have two children 5 years or younger and 4%have 3 children 5 years or younger. The majority of the sample(70.3%) have graduated high school with a mean education level of5.77 or the category of ‘some university’. The sample has 46.6% ofthe respondents working at a job or business with a mean of 35hours per week. Respondents’ total family income ranges from lessthan $5,000 to $80,000+ (1990 C $). The mean family income of thesample falls in the $40,000 to $59,999 (C $) category.Univariate DistributionsSatisfaction ScaleThis measure is a sumxnated scale of six different satisfactionitems that could be scored on a Likert-type scale of 1 to 4 (‘l’=Very Dissatisfied to ‘4’= Very Satisfied). This measure has a mean43of 3.40 ( = .45). It is significantly skewed toward the answersof ‘somewhat satisfied’ or ‘very satisfied,’ yet this is expectedof Likert—type response scale satisfaction measures. People willtend to answer more positively to a satisfaction scale especiallywhen only four choices are given, this is known as the “positivitybias”.amount of Work Done by the RespondentThis measure is a summation of the four individual work itemsthat pertained to: amount of meal preparation done by therespondent, amount of meal cleanup done by the respondent, amountof cleaning and laundry done by the respondent and amount ofoutside maintenance done by the respondent. In each item therespondent could answer using five responses: less than 1/4 of thework, less than 1/2 of the work, 1/2 or more of the work or all ofthe work and finally, if the person is from a single personhousehold they are scored separately because it was assumed by theresearchers that there is no spouse present to help with thechores. For all the variables except outside maintenance, themean is almost 4.0 ( = 3.94, 3.90 and 4.0; p = 1.81, 1.87, 1.76).The outside maintenance mean is 3.87 (SD = 2.77). Therefore, forthe meal preparation, meal cleanup and cleaning and laundry, mostrespondents have stated that they do either 1/2 or more, or all ofthe work. In the outside maintenance work, the respondents’answers are more varied. In fact, 24% of the sample said they didless than 1/4 of the work, 16% of the sample said they did lessthan 1/2 of the work, 24% of the sample said they did more than 1/2of the work and finally, only 7% of the sample said they did all ofthe outside maintenance. This can be explained by understanding44the traditional roles that are played within the domestic householdscene. The first three items of meal preparation, meal clean—upand cleaning and laundry are seen as stereotypically “feminine”chores by most researchers when they classify chores into suchcategories as masculine and feminine.Since the literature suggests that women do most of thehousework it is not surprising that the respondents reported thatthey are responsible for more than 1/2 to all of the work (Bielby& Bielby, 1988; Gurken & Gove, 1983; Hochschild, 1989; Menaghan &Parcel, 1990; Voydanoff, 1987). Outside maintenance is seen as amore “masculine” type of housework and therefore, it was skewedtowards the smaller scores. This indicates that perhaps men orsomeone other than the respondent is largely responsible forgetting these tasks done. Also, 82 respondents are classified inthe single person household in each type of measure. When thesefour types of measures were collapsed into one type of scale themean is 3.25 (Q = 2.37).Whether the Respondent Works Shift and/or Weekend WorkWithin the sample 56% or 215 of the respondents do not engagein any weekend or shift work. Yet, 83 respondents or 22% of thesample work both shift and weekend work and a further 22% workeither shift or weekend work. This distribution is skewed becausemost of the sample falls into the category of respondent workingneither shift and/or weekends.45Number of Hours the Respondent Works (Outside the Household)Per WeekThis variable has a mean of 35 hours per week ( = 15.83).It was slightly skewed (but not significantly kurtotic) in thedirection of fewer hours. It should be mentioned that 20% of thesample were working 20 hours or less and 10% worked over 50 hoursper week. Within the sample, 31% or 174 respondents were notworking or working 0 hours per week. Since the children are 5years or younger, some women may have children who are infants andmay be on maternity or other types of leave of absences.Spouse Present or AbsentThe variable of Household Type is used to assess whether thespouse was present or absent. A score of 1I indicates that therespondent is a single person with a child, and ‘2’ indicates thatthe respondent is part of a couple with a child or children.Eighty-two respondents fit into category ‘1’ and 422 respondentsfit into category ‘2’.Child Care ArrangementsThis variable is organized into four different categories: daycare, family care, relative care (outside the household) andrelative care (inside the household). The following categories arepresented by household type in Table 4.0.46Table 4.1Child Care Arrangements and Singles Versus Couples with ChildrenDay care Family Care Relative (out) Relative (in)Single 27.6% (n=18) 13.7% (n=14) 24.5% (n=13) 10.0% (n=3)Couple 72.4% (n=47) 86.3% (n=88) 75.5% (n=40) 90.0% (n=30)N=2 53Reliance on Child CareAs presented in Table 3.5 (see previous chapter) the reliancemeasure is coded into 3 categories: ‘2’ which indicates that therespondent has one child who needs regular care, ‘4’ whichindicates that the respondent has two children that need regularcare and ‘6’ which indicates that the respondent has three childrenwho need regular care. When examining the frequencies of thisvariable one sees that 200 of the respondents fall into category 2while only 6 fall into category 6. Therefore, there are fewrespondents who fall into the highest reliance category although,61 respondents fall into the second category of reliance. Thiswould suggest that 42 respondents may have a higher reliance thanthe previous 200 respondents because the former group have morechildren who need regular care.Frequency of Child CareAlso, in Table 3.5, the frequency measure was split into twocategories. 164 respondents use one type of child care only, and49 respondents use more than one type of child care regularly.47Division of Household LabourThe measure of satisfaction with division of household labouris used. This variable could be scored on a 4 point Likert typescale (‘1’ Very Dissatisfied to ‘4’ Very Satisfied). The meanscore is 3.42 ( =.75). This measure is significantly skewed andits explanation can be found under the Satisfaction Scale section.IncomeThe measure of ‘total income of the respondent’s family’ isused. This variable is coded on a 9—point index. The mean incomeof the respondents fall between $40,000 and $59,999 (1990 C $).This measure is significantly skewed (but not significantlykurtotic) to the right of the mean. In other words, the curvetends to skew to higher incomes.For all dependent variables, studentized and standardizedresiduals were run to test for any outliers. All the dependentvariables span within the three standard deviation range (there areonly two cases of outliers in each dependent variable). Yet, allvariables do not conform precisely to the normal distribution forthe residuals. In fact, there are usually higher scores than thenormal distribution and therefore, the residual curve is slightlykurtotic.Hypotheses TestingTo present an overview, there are primarily six differentfindings that came out of the multiple regression analysis, chisquare analysis and ANCOVA procedures. The first significantfinding suggests that as family care, day care and relative care48increase, the respondents’ satisfaction decreases and they tend towork more hours per week. Also, the use of relative care seemed toshow a decrease in the shift and/or weekend work variable. Marriedmothers in this sample are more satisfied overall and work lessshift and/or weekend shifts than single mothers. In addition, ithas been found that as reliance on child care increases, the amountof housework the respondent does decreases.The control variables also have significant relationships withthe dependent variables. The analysis on the education variablesuggests that as education increases, the amount of housework therespondent does decreases and shift and/or weekend work and numberof hours worked per week increases. In addition, an increase inthe satisfaction with division of housework is accompanied by adecrease in shift and/or weekend work. Finally, as the number ofchildren living in the household increases, the amount of houseworkthe respondent does and the number of hours the respondent worksper week decreases. To see how these findings were obtained, eachhypothesis has been listed below with the appropriate statisticalprocedure explained in detail.Initially, a correlation matrix between variables other thanthe control or dependent variables (such as religion, age ofchild(ren) or occupational status) was run to see if significantcorrelations emerged with any of the dependent variables. Theanalysis shows that number of children living in the householdcorrelates with amount of housework done by the respondent r (n =178) = -.25, p < .001 and correlates with number of hours therespondent works (outside the household) per week r ( = 379) = -49.15, p < .01. Education is another variable that correlates withamount of housework done by the respondent ( = 178) = -.21, p <.01, correlates with whether the respondent works shift and/orweekend work ( = 374) = .14, p < .01 and also correlates withnumber of hours the respondent works (outside the household) perweek r ( = 379) = .18, p < .001. Therefore, as the number ofchildren living in the household increases, the amount of houseworkdone by the respondent decreases and the number of hours therespondent works (outside the household) decreases. In addition,as education increases, the amount of housework done by therespondent decreases, the respondent works more shift and/orweekend work and also increases the number of hours they work(outside the household) per week. Since these variables correlatedsignificantly with the dependent variables they will be includedwith the other control variables in the statistical analysis tofollow.50Table 4.2Correlation Matrix of The Independent and Control VariablesDependent Variables1 2 3 4Family care —. 15**—.07 .04Day care —.21** .14 —.02 .30**Relative care(Outside the household) -.13* .11 —.11 .12*Relative care(Inside the household) .03 —.08 .03 .09Household Type .18** -59** .01 —.11Satisfaction withDivision of Housework n.a. .10 —.11 —.07Education—.10 —.20* .14*Income .04 .08 —.01 —.04Number of Children .08 —.21**—.04Living in Household* p .05, ** p .01, *** p .001Dependent Variable Key: 1 = Satisfaction Scale2 = 1iuount of Housework Done By theRespondent3 = Shift and/or Weekend Work4 = Number of Hours the Respondent Works(Outside the Household) per week.Table 4.1 presents a correlation matrix which analyzes eachdependent variable, independent variables and designated controlvariables to investigate for high correlations. For thesatisfaction scale, day care (r = -.21, p <.001), family care ( =-.15, p <.001) and relative care (outside the household) (r = —.13,p <.01) but not relative care (inside the household) showsignificant correlations. The only other dependent variable that51has significant correlations with types of child care is number ofhours the respondent works per week: day care (r = .30, p <.001),family care ( = .29, p <.001), and relative care (outside thehousehold) (r = .12, p <.01). The household type variablecorrelates with the satisfaction scale ( = .18, p <.001) andamount of housework done by the respondent (r =—.59, p<.00l).Also, three control variables, satisfaction with division ofhousework, number of children living in the household and educationlevel are significantly correlated with the dependent variables.Hypothesis 1The greater the use of family child care, the greater thework/family role balance.Four different multiple regression equations with each of thefour dependent variables show that family child care is significantwith the satisfaction scale variable with an (1, 492) = 11.38, Esquared = .02, p = .0008. Also, family child care is significantwhen analyzed with number of hours the respondent works (outsidethe household) per week, (1, 558) = 49.39, R squared = .08, p =.0000.When the control variables are entered into the differentmultiple regression equations, different control variables becomesignificant for different dependent variables. The regressionequation found that when all the control variables (education,total income of the respondent’s family, and number of childrenliving in the household) are entered into the multiple regressionequation followed by the family child care variable, the variablesincrease explanatory power from 2.0% to a total of 2.6% of the52satisfaction scale’s variance [F (4, 487) = 4.34, p = .002]. Thefamily care variable still significantly explains a proportion ofthe variance in the satisfaction scale variable (Beta = —.13 p =.01). Table 4.2 shows that for the amount of housework done by therespondent [ (5, 242) = 7.26, p = .00011, the control variable ofeducation level is significant (Beta = -.22 p = .01). The variableof shift and/or weekend work [ (5, 359) = 2.52, p = .05) has thecontrol variables of satisfaction with division of housework (Beta= .10, p = .03) and education level (Beta = — .11, p = .04)significantly related to it. For number of hours the respondentworks (outside the household) per week [ (5, 523) = 18.42, p =.0001], the control variables of education level (Beta = .21 p =.0000) and number of children living in the household (Beta = —.15p = .0002) explaining a significant proportion of variance in thisdependent variable. Most importantly, the independent variable,family care (Beta = .23 p = .0000), is still significantly relatedto the number of hours the respondent works (outside the household)per week. A discussion on the findings of other types of childcare and their explanation of the variance in the dependentvariables will be discussed at the end of this chapter, but canalso be found in Table 4.8.In summation, the hypothesis is not supported. Although,some variance in the satisfaction scale and the number of hours therespondent works (outside the household) per week are significantlyexplained by family care, the direction of the relationships arenot supporting the hypothesis. For support, the family carevariable must be positively correlated to the satisfaction scaleand negatively correlated with the number of hours the respondent53works per week. In addition, the other two dependent variables arenot supported (amount of housework the respondent performed andshift and/or weekend work). Therefore, family care does not seemto help in attaining work/family role balance.Table 4.3Regression Effects of Family Care and Control Variables on the FourDependent VariablesDependent Variables1 2 3 4Independent Variables Beta Weightsand Control VariablesFamily Care —. 126**—.028 —.003 .234***Income .039 .055—.009 —.038Satisfaction withDivision of Housework N.A. .056 —.104*—.026Education Level —.082 —.217** .110* .209***Number of ChildrenLiving in the Household .067 —.185** .035F Values 11.38 7.26 2.01 22.94p = .0008 .0000 .05 .0000* p .05, ** p .01, p .001Dependent Variable Key: 1 = Satisfaction Scale2 = mount of Housework Done By TheRespondent3 = Shift and/or Weekend Work4 = Number of Hours the Respondent Works(Outside the Household) per week.54Hypothesis 2Married mothers will have a greater use of family child carethan single mothers.This hypothesis is tested with a chi-square. The hypothesisis questioning whether there is a significant difference betweensingle and married mothers when family child care is considered.In fact, this hypothesis is not supported [X (4, j = 504) = .61, p= .44]Hypothesis 3Married mothers will rate their ability to balance workand family higher than single mothers.Table 4.4 shows the four dependent variables which areanalyzed in an ANCOVA procedure with the variable of householdtype. The dependent variables of the satisfaction scale [ (4,440) = 10.06, p = .01], shift and/or weekend work [ (5, 322) =7.17, p = 008] and number of hours the respondent works (outsidethe household) per week { (5, 232) = 3.89, p = .0111 aresignificant.This hypothesis is substantiated by a significant differencein means between married and single mothers in the dependentvariables: satisfaction scale (married mothers were more satisfiedthan single mothers), number of hours the respondent works (outsidethe household) per week (married mothers work less hours thansingle mothers) and shift and/or weekend work. The amount ofhousework done by the respondent variable could only be assessed55for married mothers, as they were the only respondents allowed toanswer how much housework they performed. Therefore, a differenceon this variable could not be substantiated. Yet, three of thefour dependent variables showed a significant difference whichsupports this hypothesis.56Table 4.4ANCOVA Analysis of Household Type with Control Variables on theFour Dependent VariablesDependent Variables1 2 3 4Independent Variablesand Control Variables F ValuesHousehold Type 10.06** 7533.06*** 7.17* 3.89*Income .41 44.13***.07 .23Education 7.09** 449.38*** 395* 7•75*SatisfactionWith Divisionof Housework N.A. 3359*** 2.41 2.52Nu!nber ofChildren Livingin Household 3.23 302.51*** .88 .38DF (4, 440) (5, 226) (5, 322) (5, 232)* p .05, ** p .01, *** p .001Cell MeansHousehold TypeSingle Mothers 3.25 8.00 2.94 38.38(n=57) (n=6l) (n=34) (n=34)Coupled Mothers 3.43 2.94 3.38 36.30(n=388) (n=171) (n=294) (n=245)Dependent Variable Key: 1 = Satisfaction Scale2 = Anount of Housework Done By theRespondent3 = Shift and/or Weekend Work4 = Number of Hours the Respondent Works(Outside the Household) per week57Hypothesis 4The greater the use of relative care, the lower thework/family role balance.When the dependent variables are analyzed in a multipleregression without the controls, the variable of relative care(outside the household) explains a proportion of the variance inthe satisfaction scale [R squared = .14, F (1, 492) = 8.07, p =.005), shift work performed by the respondent [R scniared = .01, F(1, 379) = 4.69, p = .03), and number of hours the respondent worksper week [R squared = .01, (1, 558) = 8.28, p = .004). When eachof the four dependent variables are analyzed in a multipleregression with relative care and the controls, the satisfactionscale and the number of hours the respondent works (outside thehousehold) per week have variance that are explained by theindependent variable, relative care (outside the household).Table 4.5 shows that when control variables are enteredagainst the dependent variable of the satisfaction scale, theexplained variance in this variable increases from 1.4% to 2.7% f(4, 489) = 5.62, p = .001]. When stepwise method is employed,relative child care (outside the household) (Beta = —.13 p = .01)and education level (Beta = -.11 p = .01) are significantlyexplaining 2.7% of the variance within this dependent variable.This would indicate that as women’s use of relative care (outsidethe household) increases and their education level increases,women’s satisfaction levels would decrease. This relationship58supports that an increase in relative care would decreasework/family role balance.Table 4.5Regression Analysis of Relative Care and Control Variables onthe Four Dependent VariablesDependent Variables1 2 3 4Independent Variables Beta Weightsand Control Variables.Relative Care(Outside the household) —.125** .025 —.109* .081*Income.057 .058 —.011 —.042Satisfaction withDivision of Housework N.A. .061 -.105 -.046Education Level —.lll** —.220** .104* .258***Number of ChildrenIn the Household .076 —. 18l** .036 —.171***F Values 5.62 5.59 3.94 11.21p = .0009 .0001 .004 .0000* p .05, ** p .01, *‘* p .001Dependent Variable Key: 1 = Satisfaction Scale2 = Anount of Housework Done By TheRespondent3 = Shift and/or Weekend Work4 = Number of Hours the Respondent Works(Outside the Household) per weekWithin the variable of amount of housework done by therespondent, the three controls and the relative care (outside thehousehold) variables explain 8.5% of the variance in this dependent59variable [F (5, 242) = 5.59, p = .0001]. In particular, onlyeducation (Beta = -.22 p = .004) and number of children living inthe household (Beta = -.18 p .004) are significantly related tothe amount of housework done by the respondent.Using the dependent variable, shift and/or weekend work, thefour variables explain 2.7% of the variance in this dependentvariable [ (5, 359) = 3.94, p = .01]. After a stepwise method isperformed on the equation only relative care (outside thehousehold) (Beta = -.11, p = .04) and education (Beta = .11, p =.04) are left in. This suggests that as the use of relative caredecreases and education increases, shift and/or weekend work wouldincrease. This relationship does not support the hypothesis.Finally, the dependent variable of number of hours therespondent works (outside the household) per week is analyzed withthe control and independent variables. The six variables explain10.4% of the variance in this dependent variable [ (6, 524) =11.21, p = .0001]. When using a step-wise method, relative care(outside the household) (Beta = .08 p = .05), education (Beta =.26, p = .0001) and number of children living in the household(Beta = -.17, p = .0001) are now significant and explaining 9.7% ofthe variance in number of the hours the respondent works (outsidethe household) per week. This suggests that as relative care(outside the household) and education increases and the number ofchildren living in the household decreases, the number of hours therespondent works (outside the household) per week increases.Therefore, the hypothesis is initially supported as two of thefour dependent variables, the satisfaction scale and number of60hours the respondent works (outside the household) per week, aresignificantly explained by the relative care (outside thehousehold) variable. The amount of housework done by therespondent is not explained by relative care and does not conformto the hypothesis. In addition, although relative care (outsidethe household) does explain variance in the shift and/or weekendwork variable, the relationship is related in a direction that doesnot support the hypothesis.Hypothesis 5aUsing one primary source rather than more than one type ofchild care will be accompanied by greater work/family rolebalance.Hypothesis 5bAn increased reliance on child care arrangements will beaccompanied by lesser work/family role balance.In Table 4.6 an ANCOVA is used to analyze each dependentvariable against the reliance variable and frequency of carevariable. The reliance measure is significant [E (7, 87) = 5.81,p = .02] when analyzed against the dependent variable of amount ofhousework done by the respondent and against the dependent variableof number of hours the respondent works (outside the household) perweek [F (7, 178) = 3.65, p = .05]. When the covariates of income,education level, and satisfaction with division of housework areanalyzed along with the other independent variables within theANCOVA, reliance still is significant (at the same levels) in thetwo dependent variables discussed above. In addition, the controlvariable of education level is significant in two dependentvariables: the shift and/or weekend work variable [ (7, 162) =613.72, p = .04), and also within the number of hours the respondentworks (outside the household) per week [ (7, 178) = 7.36, p =.01).Table 4.6M4COVA Analysis: Reliance and Frequency Variables with the ControlVariables on the Four Dependent VariablesDependent Variables1 2 3 4STEP 1:Covariates F ValuesIncome .602 .376 .075 .001Education .009 2.35 3.72* 7.36**Satisfaction withDivision of Housework N.A. .231 .717 .080Number of ChildrenLiving in the Household 1.29 5.10* .423 3.21STEP 2:Main EffectsReliance .643 5.81* .530 3.65*Frequency .064 .297 1.15 .521STEP 3:2 Way InteractionsReliance * Frequency .200 2.02 .016 .813DF: (6, 176) (7, 87) (7, 162) (7, 178)* p .05, ** p .01, ** p .001Dependent Variable Key: 1 = Satisfaction Scale2 = Amount of Housework Done By theRespondent3 = Shift and/or Weekend Work4 = Number of Hours the Respondent Works(Outside the Household) per week62A multiple regression analysis investigated the dependentvariables that had the reliance variable significantly related tothem. In the amount of housework done by the respondent, frequencyor amount of use of care and the reliance variable explain 5.6% ofthe variance in the dependent variable [ (2, 129) = 4.91, p =.009]. Reliance is the significant variable with a Beta =— .25,p = .003. This suggests that as the married mother’s reliance onchild care increases, the amount of housework done by therespondent decreases. When controls are entered into thisregression equation, the adjusted R squared increases to 8% EF (7,124) = 3.45, p = .01] and education level has a Beta = —.19, p =.04. This also suggests that as the respondent’s educational levelincreases, the amount of work done by the respondent within thehousehold decreases.A multiple regression was also used to analyze number of hoursthe respondent works (outside the household) per week. Frequencyof use of care and reliance account for 2.2% of variance of thedependent variable [ (2, 262) = 3.98, p = .02]. The reliancevariable is significantly related to the dependent variable with aBeta = -.13, p = .03. This suggests that as the reliance on childcare increases, the number of hours the respondent works (outsidethe household) per week decreases. When controls were entered in,the group of variables explain 3.4% in the above dependent variable[F (7, 257) = 3.19, p = .01]. Reliance stays barely significant(Beta = —.12, p = .06) and education level becomes the mostsignificant variable (Beta = .18, p = .01). Therefore, as the63education level increases and the reliance decreases, the numberof hours the respondent works (outside the household) per weekincreases.A difference in the frequency of use of care and householdtype was not substantiated by this analysis. Therefore, hypothesis5a is not supported by the findings. Although, reliance is relatedto both amount of housework done by the respondent and the numberof hours the respondent works (outside the household) per week, thedirection of the relationship does not support the working modelproposed earlier and therefore, does not support the 5b hypothesis.Hypothesis 6The difference in child care arrangements of married andsingle mothers leads to a difference in work/family rolebalance between married and single mothers.Table 4.7 represents an ANCOVA procedure which is used toinvestigate for interaction effects. Each dependent variable wasrun with the household type variable and each type of child care.In addition, the covariates of income, education level andsatisfaction with the division of housework are also included. Themain effects for the independent and control variables almostperfectly match the results previously run in a multiple regressionanalyses. Significant interaction effects are found in only twocases. First, an interaction effect is found between household typeand relative care (outside the household) when run with thedependent variable of number of hours the respondent works (outside64the household) per week [ (7, 146) = 3.45, p = .05). Anotherinteraction is found between household type and day care when thedependent variable of number of hours the respondent works (outsidethe household) per week [ (7, 146) = 3.62, p = .05) is considered.Therefore, single and married mothers have different uses of bothrelative care (outside the household) and day care which effectsthe outcome of the dependent variable, number of hours therespondent works (outside the household) per week. In fact, singlemothers use relative care and day care more than married mothers.65Table 4.7ANCOVA Analysis: Household Type and Type of Child Care Arrangementwith Covariates Against the Four Dependent Variables.Dependent Variables1 2 3 4STEP 1: CovariatesIncome.142 68.97*** .004 .080Education 2.53 507.73*** 39.76*** 4.91*Satisfaction withDivision of Housework N.A. 32.61***.847 2.62STEP 2: Main EffectsHousehold Type 20.51*** 7666.78***.163 6.63**Day care l1.31*** 1.44 43.93** 1.20Family Care 2.39 .001 30.69*** .096Relative Care (Outside the household) 10.31***.037 555* 3.21Relative Care (Inside the household) .700 .115 6.09* .195STEP 3: 2 Way Interactions F ValuesHousehold X Day care .170 .307 1.033 3.62*Household X Family Care .977 .036 .852 1.084Household X Relative .083 .163 .432 3.450*(Outside the household)Household X Relative(Inside the household) .099 .000 .303 .940*p.05, ** p .01, *** p .001Dependent Variable Key: 1 = Satisfaction Scale2 = Amount of Housework Done By theRespondent3 = Shift and/or Weekend Work4 = Number of Hours the Respondent Works(Outside the Household) per week66These findings substantiate only one out of the fourdependent variables involved in the hypothesis. Therefore,although the hypothesis is supported in the number of hours therespondent works (outside the household) per week, it was notsupported in the satisfaction scale, amount of housework therespondent does and shift and/or weekend work.Additional FindingsSome interesting findings that are additional to the aboveproposed and tested hypotheses were found to be so significant thatthey must be addressed and are presented in Table 4.8.Within the satisfaction scale, when a multiple regression isrun and household type, all types of care, and the controlvariables are included [ (8, 485) = 7.06, p = .0001), day care(Beta = — .17, p = .0002) is found to be significantly related tothe satisfaction scale.When control and independent variables are analyzed withamount of housework done by the respondent [ (9, 238) = 21.34, p= .0001, R squared = .37], education level (Beta = -.12, p = .03)and number of children living in the household (Beta = -.12, p =.02) are significantly related to this dependent variable.Within the shift and/or weekend work variable, when controland independent variables are analyzed with this dependent variableonly 2* of the variance within shift and/or weekend work isexplained [F (9, 355) = 2.73, p = .05). Yet, the two variables[other than the relative care (outside the household) variable]that are significant after stepwise deletion are satisfaction with67division of housework (Beta = -.11, p = .04) and education level(Beta = .12, p = .05).Finally, when considering the number of hours the respondentworks (outside the household) per week, the control and independentvariables accounted for 19.7* of the variance within this dependentvariable [F (9, 519) = 17.78, p = .0000]. Using the stepwisemethod within the regression equation, day care (Beta = .22, p =.0000) education level (Beta = .17, p = .0000) and number ofchildren living in the household (Beta = -.12, p = .003) aresignificantly related to number of hours the respondent works(outside the household) per week.Therefore, as the use of day care increases, satisfactiondecreases and number of hours the respondent works (outside thehousehold) per week increases. As educational levels increase,amount of housework the respondent does decreases, the respondenttends to work more shift and/or weekend work and also increases thenumber of hours she works (outside the household) per week. As thenumber of children living in the household increases, amount ofhousework done decreases and number of hours the respondent works(outside the household) per week decreases. Also, as thesatisfaction with the division of housework increases, shift and/orweekend work decreases. These relationships will be furtherexplored and explained in the discussion section to follow.68Table 4.8Summary Table: Regression Analysis of Independent and ControlVariables on the Pour Dependent Variables.Dependent Variables1 2 3 4Independent Variablesand Control Variables Beta WeightsFamily Care —.123*—.020 .001 .234***Relative Care(Outside the house) -.130* .058 —.113* .056Relative Care(Inside the house) .049 —.048 .043 .061Day care —.169**—.014 —.033 .229***Household Type .149** —547***.007 —.033Income .038 .026 —.012 —.033Satisfaction withDivision of Housework N.A. .034 —.109*—.023Education Level—.078 —.123* .115* .181***Number of Childrenin Household .027 —.123* .023 —.118**F Values: 7.06 21.34 2.73 17.78p = .0001 .0000 .01 .0000* p .05, ** p .01, *** p .001Dependent Variable Key: 1 = Satisfaction Scale2 = Amount of Housework Done By theRespondent3 = Shift and/or Weekend Work4 = Number of Hours the Respondent Works(Outside the Household) per week69Chapter VDiscussionThis study has investigated the relationship between mothers’child care arrangements and their work/family role balance.Differences between married mothers and single mothers’ child carearrangements and their work/family role balance was alsoinvestigated. In addition, the relationship between controlvariables (income, education, satisfaction with the division ofhousework and number of children living in the household) and thefour dependent variables were examined.This study was guided by the conceptual framework ofwork/family role balance by Tiedje et al. (1990) paired withcurrent empirical research on single mothers and married mothers,as well as current literature on child care arrangements. Thesethree areas have conjoined to produce underlying assumptions whichare 1) parents who are employed in some capacity experiencework/family role balance or imbalance; 2) single and marriedmothers differ with respect to their work/family role balancebecause married mothers’ have a spouse who may be present to assistand ease conflict between work and family roles; 3) single andmarried mothers use different types of care and even if the sametype of care are used, the frequency of use are not the same forsingle and married mothers. It is believed that the difference inchild care arrangements between single and married mothers70influences their work/family role balance. The results from thisstudy will be discussed in terms of the four dependent variables.Satisfaction ScaleThe data indicate that contrary to what was predicted, familycare, day care, and relative care (outside the household) are allnegatively related to the satisfaction scale. That is, as the useof these types of child care increases, overall satisfactiondecreases. This finding is contrary to Hum & Balgopal’s (1990)data that suggest that types of care might mediate work/family rolebalance. Although, this finding supports Goff et al.’s (1990)conclusion that child care centers at work do not increasework/family role balance (Possibly, this conclusion may be inferredas the day care variable in the study included child care centersat the respondents’ place of work). The negative relationship mayresult because using any type of care at all may make a motherexperience guilt and conflict as she would like to be with herchild(ren) but needs (or wants) to go to work. In addition,Kamerman (1985) suggests these types of care may not be flexible tothe needs of mothers and may make work/family role conflict evenmore evident.Married mothers have significantly higher ratings ofsatisfaction than single mothers. Two possible explanations forwhy this difference may exist is explored. First, researcherssupport the assumption that women are more satisfied when a spousetakes on some of the household chores and perhaps, supports thewife in her work role (Berk, 1985; Bohen & Viveros-Long, 1981;71Elman & Gilbert, 1984; Hochschild, 1989; Holahan & Gilbert, 1979;Hughes & Galinsky, 1988; Menaghan & Parcel, 1990; Piotrkowski1987; Pleck, 1987; Thompson & Walker, 1989). Since singlemothers lack the supportive network that a partner might provide,single mothers must cope with work/family role conflict alone.This may give rise to the finding that single mothers have lesswork/family role balance than married mothers. Second, a lack ofoverall supports for single mothers (Fine, Donnelly & Voydanoff,1986) paired with the fact that single mothers tend to managework/family conflicts by themselves (Johnson, 1986) possiblycontributes to higher conflict and lower levels of enhancement fromwork and family roles when compared to married mothers. Thissignificant difference between married and single mothers and theirsatisfaction scores is supported by Hum & Balgopal (1990) whosuggest that single mothers experience more multiple role imbalancethan married mothers.The control variables of education, income and number ofchildren living in the household are not significantly related tooverall satisfaction. This finding is supported somewhat by Hansonand Sloane’s (1992) research which found that number of childrendid not significantly correlate with job satisfaction (one of themeasures within the satisfaction scale). Perhaps having one childor three children does not change the respondents’ satisfactionexperience. This may be because having one child has initiallyaltered the satisfaction level which may not be lowered with anyadditional children. Also, the finding that income and education72do not alter satisfaction ratings may possibly be explained byconstruct validity. That is, since the satisfaction questions donot enquire about satisfaction with income or education but ratherasks about satisfaction with spouse, division of housework,children, balance between work and family, and availability of timeto do other interests, the effects of income and education do notappear to have been captured. In addition, except for the work andwork/family questions, the other measures of satisfaction may besomewhat independent of one’s income or level of education.amount of Housework Done by the RespondentThe General Social Survey coded this variable on a 4-pointLikert—type scale that only married mothers could answer. If therespondent was a single mother she was placed in another category‘8’. Since the General Social Survey coded in such a manner it wasimpossible to investigate different patterns of amount of houseworkdone by single mothers versus married mothers. Furthermore, thefindings that result on this variable can only be explained to themarried mothers sample.The data indicate that only the control variables of educationand number of children are significantly related to amount of workthe married respondent does within the household. These results donot support the hypotheses on this dependent variable.Education is negatively related to the amount of work themarried respondent does in the house. The relationship is that asthe education level increases, the amount of housework the marriedrespondent does decreases. Therefore, the married respondent who73is highly educated and performs less housework may be explained bya lack of time to do the housework because of increased hours atwork (Staines & Pleck, 1983) and by a greater income (Kamerman,1985) that allows the possibility of purchasing help to maintainthe household.The number of children living within the household isnegatively related to amount of housework performed. As morechildren live within the household, less housework is performed.Berk (1985) suggests the relationship may be due to an absence ofavailable time to do the housework because of an increase in thedemand of time needed to tend to more children.Shift and/or Weekend WorkOnly the independent variable of relative care (outside thehousehold) is significantly related to shift and/or weekend work.The finding suggests that as relative care increases, shift and/orweekend work decreases. This was not the hypothesized relationshipbetween relative care and shift and/or weekend work. Rather, itwas predicted that as relative care increased, shift and/or weekendwork would have also increased. This finding may be understood bysuggesting that bringing a child to a grandmother/father,aunt/uncle or cousin on a weekend or at night may considerablyinterfere with the life of the sitter and the child. Therefore,the option of having a family member sit a child during shiftand/or weekend work may simply not be available.The other types of child care were not significantly relatedto this dependent variable. This relationship may possibly be74explained in reference to the relative care variable that is foundto be significantly related to shift and/or weekend work. Sinceother types of care are usually business agreements (such as daycare and family care) the ability to work shift and/or weekendsmight depend on whether the child care facility carries hours thatextend to these times. Also, relative care (within the household)is not significantly related because relative care that is providedinside the house could mean that the relative tended to live withthe child and mother. Possibly, this arrangement would neitherincrease nor decrease shift and/or weekend work but rather wouldprovide child care when needed. Research that considers therelationship between shift and/or weekend work and any child carearrangement is absent in the literature and therefore, theexplanations posed above can not be formally refuted or supported.Perhaps, this is an area that deserves more attention and research.Married mothers and single mothers significantly differ intheir rates of shift and/or weekend work. Married mothers workless shift and/or weekend work than single mothers. This findingmay be related to research by Staines and Pleck (1983) who foundthat wive’s work schedules were influenced by their husband’s workschedules. The husband working non—weekend and non—shift work waspositively correlated with the wife also working the same type ofshift. The inverse of this relationship was also supported byStaines and Pleck (1983). This suggests that married mothers wouldwork less shift and/or weekend work because of their husbands,while single mothers would engage in such work since they may need75the money (Kamerman, 1985) more than married mothers and are notinfluenced by any husbands’ work schedules.The control variables of satisfaction with division ofhousework and education level are significantly related to shiftand/or weekend work. The more shift and/or weekend work that isperformed, the less satisfaction with division of housework. Thisresult is supported by Staines and Plec]c (1983) who state thathusbands and wives experience work/family conflict when wives worknon-day shift. In addition, women still do the majority ofhousework in the household (Bielby & Bielby, 1988; Gurken & Gove,1983; Hochschild, 1989; Menaghan & Parcel, 1990; Voydanoff, 1987).Husbands and wives experiencing work/family role conflict pairedwith the husbands’ unwillingness to share in the housework wouldsuggest support for the finding that wives who engage in shiftand/or weekend work would have lower levels of satisfaction withdivision of housework.The higher the education level, the more shift and/or weekendwork that is performed. A possible explanation of this finding maybe that respondents who are professionals are interpreting thisquestion as work done in the office or at the home computer. Inaddition, this finding may also be explained by the professionalperson’s career which mandates that person to be on—call or to workweekends.Number of children living in the household is notsignificantly related to whether the respondent works shift and/orweekend work. This finding could be accounted for by the76relationship that number of children in the household does notgreatly impact on whether the mother engages in shift work for tworeasons. First, mothers who work shift and/or weekend work becausethey are employed in a professional career can do little to lessentheir work load unless they have sympathetic superiors (Galinsky &Stein, 1990). Also, these mothers tend to make a greater incomeand can afford alternative types of care such as nanny care.Second, mothers who work shift and/or weekend work because they areforced to for financial reasons have little choice to work shiftssince these mothers need an income to sustain their families.Number of children would not influence these mothers if the choicethey have is working weekends or not working at all.Number of Hours the Respondent Works(Outside the household) Per WeekContrary to what was predicted, both family care and day careare positively related to number of hours the respondent works.The positive relationship between relative care (outside thehousehold) and number of hours the respondent works is an expectedone. The findings suggest that as relative care, family care andday care increase, the number of hours the respondent works(outside the household) per week increases.It is difficult to understand the causal relationship betweentypes of care and number of hours the respondent works (outside thehousehold) per week. The fact that the respondent initially workedmany hours in a job may have made the respondent quickly seek outthese three types of care. Family care, relative care and day care77arrangements may have become known to the respondent because it wasclose to their job, part of a work benefit, had flexible hours, wasknown as a quality care alternative, had a familiar surrounding toboth the respondent and her child(ren) and, in the case of relativecare, may have had no or low cost and was an arrangement that gavecare during a crisis. The other explanation for this finding isthat these types of care facilitate the respondent’s working longhours during the week. In this interpretation, the search forchild care would come before the long hours started. Researchsupports that women who work many hours during the week tend to usefamily care, day care and some relative care arrangements due toboth reasons previously mentioned (Hofferth et al, 1990).Education is positively related to number of hours therespondent works per week. This finding may possibly be explainedby the relationship which suggests that people who have highereducational levels tend to be employed in professional careers thatmandate a greater number of working hours whether it be at theoffice or at home.Number of children living in the household is negativelyrelated to number of hours the respondent works per week. Perhaps,this is a result of the cost of care versus the financial benefitsof a job. For example, if a person has three children to providefull-time care at an approximate cost of $700.00 (C $) per child,the cost of care may exceed the financial rewards of a career.Therefore, the respondents may find that it is beneficial to timeshare their jobs, work part—time, or not even work at all.78Income and satisfaction with the division of housework are notsignificantly related to number of hours the respondent works.This may be because income is being expressed in the variable ofeducation since these two variables are sometimes difficult todifferentiate. Satisfaction with division of housework may not berelated to the number of hours the respondent works per weekbecause as Staines and Pleck (1983) point out, it is not the hoursworked per week, but, rather whether the wife’s hours worked matchthe husband’s hours per week. Staines and Pleck (1983) argue thatmatching husbands hours, in addition to a wife working little or noshift and/or weekend work, may slightly increase the amount ofhousework the husband does. Also, if the husband is supportive ofthe wife working, he would be more supportive with helping withhousehold chores, helping with infant care and understanding jobdemands (Gray, Lovejoy, Piotrkowski, & Bond, 1990). Therefore,satisfaction with division of housework is more dependent onwhether the husband is supportive of the wife’s job, and whetherthe hours worked are “normative” work hours than the number ofhours the respondent works per Se.Also, day care and relative care (outside the household) havean interaction effect with number of hours the respondent works(outside the household) per week. This finding states that marriedmothers use day care and relative care (outside the household) lessthan single mothers. This finding is supported by researchconducted by Hofferth et al. (1990) and Hum and Balgopal (1990).79Finally, the variables of reliance and frequency are discussedseparately. These variables were measured in an ANCOVA andmultiple regression procedure. The findings indicate that onlyreliance is related to the amount of housework the respondent doesand number of hours the respondent works (outside the household)per week. That is, as reliance increases, the amount of houseworkthe respondent does decreases and the number of hours therespondent works decreases. This relationship was unexpected andmay be a result of the inclusion of the variable “number ofchildren requiring care”. Possibly, this variable would be relatedto the control “variable of number of children living in thehousehold which has a negative relationship with the amount ofhousework done by the respondent. Therefore, it might be thatreliance would also have a negative relationship with the amount ofhousework done by the respondent. Another reason for this findingcould be that when one has many children and is needing full—timecare, the amount of housework that the respondent has time forwould be significantly different than the respondent who only hasone child and needs care only part—time. This relationship is alsocontrary to the model of work/family role balance that assumesbalance would be depicted in a greater partnership of housework.Throughout this study, a discovery was made that although therelationships may seem to confirm the model of work/family rolebalance, when the possible explanations for the relationships arediscussed, one sees the reasons behind them may not support themodel.80It is also unexpected that reliance is negatively related tonumber of hours the respondent works (outside the household) perweek. The finding from a multiple regression indicates that asreliance increases, the amount of work the respondent does (outsidethe household) per week decreases. It is possible that a personwith more than one child (in the high reliance category with threechildren) needing full—time care may not work more hours since itwould be costly to arrange care for two or more children. Onceagain, the variable of number of children requiring care, mayinfluence the reliance variable because both variables (number ofchildren living in the household and the reliance variables) havea negative relationship number of hours the respondent works(outside the household) per week.In terms of the purposes, the significant findings can begenerally concluded. It seems that the relationship between childcare arrangements and work/family role balance is a negative one.Child care arrangements overall do not seem to mediate work/familyrole balance. In particular, day care and family care have thestrongest negative relationship.It was also found that single mothers have lower work/familyrole balance than married mothers. In addition, single mothers userelative care and day care more than married mothers. Possibly, itcould be concluded that using certain child care arrangements (suchas relative care, family care and day care) while being a singleparent could influence negatively on work/family role balance.81Within these relationships between the dependent andindependent variables lies the answer to whether or not the sevenhypotheses previously mentioned are supported. Table 5.1 providesa summary of the seven hypotheses.82Table 5.1Summary of HypothesesDependent Variables1 2 3 4Hi: The greater the use offamily child care, thegreater the work/familyrole balance no(—) no no no(+)H2: Married mothers will havea greater use of familycare than single mothers noH3: Married mothers will ratetheir ability to balancework and family higherthan single mothers yes yes no yesH4: The greater the use ofrelative care, the lowerthe work/family balance—relative care (outsidethe household) yes(-) no no(—) yes(+)—relative care (insidethe household) no no no noH5a: Using one primary sourcerather than more than onetype of child care willbe accompanied by greaterwork/family role balance no no no noH5b: An increased reliance onchild care arrangementswill be accompanied bylesser work/family rolebalance no no(—) no no(—)H6: The difference in childcare choices of marriedand single mothers leadsto a difference in work/family role balancebetween married andsingle mothers no no no HHTYPExday care andrelative care(outside thehousehold)83The relationship between the dependent, independent, andcontrol variables are complex and diverse. The assumption that thefour dependent variables are a unidimensional measure of thework/family role balance concept may be too simplistic to gaintotal understanding. Since the dependent variables are somewhatcorrelated and account for most of the variance on one factor,these results suggest that work/family role balance may be a multidimensional concept. To date this is the first study to examinework/family role balance using both subjective and objectivemeasures. This approach may provide a more ‘exhaustive’ type ofanalysis which may refine the knowledge of work/family rolebalance.Limitations and StrengthsThe limitations will be discussed first. A primary limitationto this study was that the results are based on cross—sectionaldata. Longitudinal data would better answer the conceptual aspectsof the variables. For example, longitudinal data would betterindicate when a respondent was highly reliant on care, what type ofcare she would use over a five year period, and more reliablemeasures of satisfaction and number of hours worked per week. Inaddition, applying my concept to this data limited the sample thatcould be used for analysis. Therefore, findings could only begeneralized to the sample within this study and not to the generalCanadian population.The second limitation this study incurred was the problem ofusing secondary data. As has been previously discussed some84questions could have been interpreted in different contexts (forexample, whether the respondent worked shift and/or weekend workcould have been understood as straight time for which you are paidfor working, or could have been understood to mean time spent athome or in the office for which you are expected to work but maynot necessarily be paid for). Using secondary data also posed aproblem when applying the Tiedje et al. (1990) typology as theunderlying framework. The data did not allow for measuringenhancement and conflict as Tiedje et al. (1990) had done.A third limitation is directed at the formulation of some ofthe measures within the survey. For example, the questions onsatisfaction with the job, spouse, or family had very generalresponses. The responses were coded on a 4—point Likert—type scalethat went from “Very Dissatisfied” to “Very Satisfied”.Respondents had little room to express their satisfaction on sucha limited scale. In addition, the satisfaction scale did not askwhether the respondents were satisfied with their child carearrangements. This would have helped in trying to ascertainwhether satisfaction with child care arrangements effects overallsatisfaction. Also, the reliance question was a derived variablefrom the data and may have had a problem with construct validity.Perhaps, in another study a reliance question could be posed withinthe questionnaire.A fourth limitation is directed at the operational model ofwork/family role balance. That is, it was discovered that the fourvariables may not have been accurately representing all aspects of85the work/family role balance concept. Initially, it seemed thatthis model was quite clear and any deviation from this model wasunderstood to be imbalance. Yet, upon discussing the relationshipsand relating it back to conceptual literature the model began tobreak down. The goodness of fit of the model to the data began tobe questioned throughout the study. This limitation was a directresult of the questions that were posed and their inability, attimes, to tap the constructs in the manner in which they wereconceptually devised. Also, this limitation arose from an absencein the literature which never addressed work/family role balance insubjective and objective measures simultaneously. Therefore, thistype of measurement of the concept is the first of its kind.Although, the methodological concepts may not have ideallymatched the conceptual concepts, this study highlighted a multidimensional aspect to work/family role balance. This study alsohighlighted the need for a cohesive theory that empiricallymeasures the concept in both subjective and objective terms. Thefour dependent variables have only measured a segment ofwork/family balance. Therefore, the statistical analysisemphasizes that there are more variables that are needed to fullyassess the concept.The above limitations are direct results of using secondarydata. Yet, there were definite strengths that came from thisstudy. A strength of the survey is that it represents a largeproportion of Canada. Statistically it is impossible (due to thesmall numbers of single mothers) to weight back to the Canadian86population, nevertheless, the initial sub—sample was chosen fromthe Statistics Canada survey and represented Canadian mothers withchildren under the age of 5 years. In addition, Statistics Canada(1990) made a great effort to operationalize their measures to onespreviously used by the United States. This allows otherresearchers the opportunity to compare results using this databetween the United States and Canada.Another strength is seen in the finding that types of childcare are influencing work/family role balance. This supportscontemporary research that suggests that types of child care candifferentially affect a person’s work/family role balance ratings(Hum & Balgopal, 1990; Kamerman, 1985; Klein, 1985; Turner &Smith, 1983). Because each child care type may not influence eachdependent variable in the same manner an argument for a multidimensional construct begins to take form.In addition, another strength actually comes from a limitationwithin this study. By experiencing problems with the working modelthe complexity of this area is reinforced. The difficulties foundin the model may stem from problems within the work/family rolebalance literature. Most of the literature considers satisfaction,amount of housework done in the household, shift and/or weekendwork and working hours separately. Perhaps, when the model joinedthese four areas to create work/family role balance, the conceptbecame more complex and diverse because of the inter—connectionsthat underlie all four variables. This limitation shows87researchers that work/family role balance is very complex and mayneed to be researched at a multi—dimensional phenomena.ImplicationsThis study has attempted to assess several aspects of thework/family role balance concept. By using both subjective andobjective measures for work/family role balance this study hasrecognized and answered Tiedje et al. ‘S (1990) limitation of “beingdevoid of . . . objective stressors..” (p. 70). seeing the objectivemeasures of amount of work done by the respondent within thehousehold, performing shift and/or weekend work and calculating thenumber of hours the respondent works (outside the household) perweek emphasizes the work and family demands that women experiencebut can not describe in subjective evaluations.The theory of work/family role balance conveys a hopeful andpositive message unlike most of the other work/family theorieswhich emphasize the continual conflict and lack of positivereinforcement from these two roles. This theory may offer newinsight into managing the roles of worker, mother and, perhapswife. Some critics suggest that only the subjective measure needsto be considered when assessing work/family role balance. By usingboth objective and subjective measures, this study provides holismto the theory. That is, it is the combination of a person’s selfreports with time assessments and responsibility assessments forhousehold tasks that enrich and empirically base this concept.Possibly, more refined subjective and objective measurescould be used to assess the multi—dimensional nature of work/family88role balance. It is when the measures become conceptually andoperationally whole that the complexities of this concept might bebetter understood. This combination may come when all of thedimensions of work/family role balance are recognized andempirically established. In addition, future research shouldconsider collecting information from the spouse or significantothers which would give greater insight into this experience as itis for women. This study highlights the need for a theoreticalframework that more appropriately provides the criteria forassessing work/family role balance.89ReferencesBaruch, G. & Barnett, R. (1986). Consequences of father’sparticipation in family work. Journal of Personality andSocial Psychology, (5), 983—992.B.C. Task Force on Child Care. (1991). Showing We Care: A ChildCare Strategy for the 90’s.Berk, S. (1985). The Gender Factory. New York: Plenum Press.Bielby, D., & Bielby, W. (1988). She works hard for the money.American Journal of Sociology, (5), 1031—1059.Bohen, H., & Viveros-Long, A. (1981). Balancing Jobs and FamilyLife. New York: Temple University Press.Bolger, N., Delongis, A., Kessler, R., & Wethington, E. (1989). Thecontagion of stress across multiple roles. Journal of Marriageand the Family, j,, 175-181.Elman, M.R., & Gilbert, L.A. (1984). Coping strategies for roleconflict in married and professional women with children.Family Relations, fl, 317—327.Fine, M.A., Donnelly, B.W., & Voydanoff, P. (1986). Adjustment andsatisfaction of parents: A comparison of intact, single—parentand stepparent families. Journal of Family Issues, 2 (4),391—404.Fletcher, C. (1989). A comparison of income and expenditures ofmale headed households paying child support and female headedhouseholds receiving child support. Family Relations, (4),412—417.Galinsky, E., & Stein, P.J. (1990). The impact of human resourcepolicies on employees. Journal of Family Issues, 11, 368—383.Geoff, 5. (1990). Employer supported child care: Work/Familyconflict and absenteeism: A field study. Personnel Psychology,j (3), 793—809.Goff, S., Mount, M., & Jamison, R. (1990). Employer supported childcare, work/family conflict and absenteeism: A field study.Personnel Psychology, 41, 794-809.Goode, W.J. (1960). A theory of role strain. American SociologicalReview, (4), 483—496.90Gray, E., Lovejoy, M., Piotrkowski, S., & Bond, J. (1990). Husbandsupportiveness and the well—being of employed mothers ofinfants. Families in Society, fl (6), 332—341.Gurken, M., & Gore, W. (1983). At Home and At Work. California:Sage Publications.Hanson, S.L., & Sloane, D.M. (1992). Young Children and JobSatisfaction. Journal of Marriage and the Family, , 799—811.Hill, N. (1987). Sharing Childcare in Early Parenthood. New York:Routledge and Regan Press.Hochschild, E. (1989). The Second Shift. New York: PenguinPublishing.Hofferth, S.L., Brayfield, A., Deich, S., & Holcomb, P., (“TheUrban Institute”). (1990). National Child Care Survey, 1990.Washington: The Urban Institute Press.Hogan, D.P., Hoa, L., & Parish, W.L. (1990). Race, kin networks andassistance to mother-headed families. Social Forces, (3),797—812.Holahan, C.K., & Gilbert, L.A. (1979). Interrole conflict forworking women: Careers versus jobs. Journal of AipliedPsycholoqy, 64 (1), 86—90.Houseknecht, S., & Macke, A. (1981). Combining marriage and career:The marital adjustment of professional women. Journal ofMarriage and the Family, 43, 651—674.Hughes, D. & Galinsky, E. (1990). Balancing work and family lives.New York: Routledge and Regan Press.Hum, J.J., & Balgopal, B. (1990). Child care in an age of FamilyDifferences. Child and Adolescent Social Work, 2, (3),247—263.Joesch, J. (1991). The effect of the price of child care of AFDCand mother’s paid work behavior. Family Relations, 4Q (2),161—168.Johnson, P.J. (1986). Divorced mothers: Sources of support forconflicts in responsibilities. Journal of Divorce, 9 (4),89—105.Johnson, P.J. (1983). Divorced mothers’ management ofresponsibilities: Conflicts between employment and child care.Journal of Family Issues, 4 (1), 83-103.91Kamerman, S.B. (1985). Child care Services: An Issue for GenderEquity and Women’s Solidarity. Child Welfare, , 259-271.Kamerman, S.B., & Kahn, A.J. (1987). The Responsive Workplace. NewYork: Columbia University Press.Keith, P.M., & Schafer, R.B. (1980). Role Strain and depression intwo—job families. Family Relations, , 483-488.Kessler, R. & McRae, J. (1982). The effect of wive’s employment onthe mental health of married men and women. Research inCommunity and Mental Health, 2, 84-96.Klein, R.P. (1985). Caregiving arrangements by employed women withchildren under 1 year of age. Developmental Psycholoqy,(3), 403—406.Koren, S. (1984). In A.J. Kahn & S.B. Kamerman (Eds)., Child Care:Facing the Hard Choices. Dover: Auburn House Publishing.Kurdek, L.A. (1988). Social support of divorced single mothers andtheir children. Journal of Divorce, U (3/4), 167-187.McCubbin, H. (1980). Family stress and coping: A decade review.Journal of Marriage and the Family, , 855-872.McCubbin, H. (1979). Integrating coping behavior in family stresstheory. Journal of Marriage and the Family, Ai, 237—244.Marks, S.R. (1977). Multiple roles and role strain: Some notes onhuman energy, time and commitment. American SociologicalReview, 42. (6), 921—936.Menaghan, E., & Parcel, T. (1990). Parental employment and familylife: Research in the 1980’s. Journal of Marriage and theFamily, , 1079—1098.Moen, P., & Dempster-McClain, D.I. (1987). Employed parents: Rolestrain, work time, and preference for working less. Journal ofMarriage and the Family, .j, 579-590.National Council of Welfare. (1988). Child Care: A BetterAlternative. Minister of Supply and Services, Canada.Nock, S., & Kingston, P. (1988). Time with children: The impact ofcouple’s work-time commitments. Social Forces, (1), 59—85.Norusis, M.J. (1983). Introductory Statistics Guide SPSSX. NewYork: McGraw—Hill Book Company.92Perry-Jenkins, M., & Crouter, A. (1990). Men’s provider roleattitudes. Journal of Family Issues, 11 (2), 136—156.Peterson, R.R., & Gerson, K. (1992). Determinants of responsibilityfor child care arrangements among dual—earner couples. Journalof Marriage and the Family, , 527-536.Piotrkowski, C.S., Rapoport, R.N., & Rapoport, R. (1987). Familiesand work. InM.B. Sussman & S.K. Steinmetz (Eds.), Handbook ofMarriage and the Family (pp. 251-284). New York: Plenum Press.Pleck, J. (1987). Dual career families: A comment. CounsellingPsychologist, 15 (1), 131—133.Pleck, J. (1985). The work-family role system. In J. Pleck (Eds.),Working Wives, Working Husbands (pp. 45-65). California: SagePublications.Pleck, J. (1985). Facilitating future change in men’s family roles.Marriage and Family Review, . (3/4), 11—16.Powell, D.R., & Widdows, R. (1987). Social and economic factorsassociated with parents’ decisions about after—school childcare: An exploratory study in a medium sized community. Child& Youth Care quarterly, (4), 272-282.Repetti, R. (1989). Effects of daily workload on subsequentbehavior during marital interaction. Journal of Personalityand Social Psychology, 57 (1), 651-659.Rudd, N., & McKenry, P. (1986). Family influences on the jobsatisfaction of employed mothers. Psychology of Womenquarterly, 10, 363-372.Scarr, S., & McCartney, K. (1989). Dilemmas of child care in theUnited States: Employed mothers and children at risk. CanadianPsychology, (2), 126-139.Shelton, B.A. (1990). The Distribution of Household Tasks: DoesWife’s Employment Status Make a Difference? Journal of FamilyIssues, a, 115—135.Staines, G.L., & Pleck, J.H. (1983). The impact of work schedule onthe family. Ann Arbor, Michigan: The University of Michigan.Statistics Canada (1989). The Family in Canada. Ottawa: Minister ofSupply and Services, Canada.Statistics Canada (1990). Focus on Canada: Women and the LabourForce. Ottawa: Minister of Supply and Services, Canada.93Statistics Canada (1990) (Preliminary edition). General SocialSurvey Cycle 5: Family and Friends 1990 Codebook.Statistics Canada (1992). The Labour Force. Ottawa: Minister ofSupply and Services, Canada.Thompson, L., & Walker, A. (1989). Gender in families: Women andmen in marriage, work and parenthood. Journal of Marriace andthe Family, i, 845-871.Tiede, L., Wortman, G., Downey, C., Emmons, M., Biernat, M., &Lang, E. (1990). Women with multiple roles. Journal ofMarriage and the Family, , 63-72.Turner, P.H., & Smith, R.M. (1983). Single parents and day care.Family Relations, , 215—226.Voydanoff, P. (1980). Work roles as stressors in corporatefamilies. Family Relations, (4), 489—494.Voydanoff, P. (1987). Work and Family Life. California: SagePublications.Voydanoff, P. (1990). Economic distress and family relations.Journal of Marriage and the Family, , 1099—1115.94Appendix AList of the Subiective and some of the Obiective Questions Used in the Dependent Variableof Work/Family Role BalanceSatisfaction QuestionsK4. Are you satisfied or dissatisfied withIs that somewhat or very?Somewhat Verya) Your relationship with yourspouse/partner, or yoursingle status9 Satisfied . . . . 01 C] —------‘ 02 C] 03 C]Dissatisfied . 04 C] ---------‘ 05 C] osDNo opinion . 07 C]b) Your relationship with yourimmediate family? Satisfied ... 08 C] --—---i 09D 10 C]Dissatisfied . 11 C] ----—--÷ 12 C] 13 C]No opinion . 14 C]c) The way housework is sharedin your home Satisfied . . . 15 C] —--- 16 C] 17 C]Dissatisfied . 18 C] --—----‘ 19 0 20 0No opinion . 21 0d) Your job or main activity? ... Satisfied ... 22 0—‘ 23 0 24 0Dissatisfied . 25 0 —------ 26 0 27 C]No opinion . 28 0e) The balance between yourjob or main activity andfamily and home life9 Satisfied ... 29 0—i 30 0 31 C]Dissatisfied . 32 0-—-----+ 0 34 0No opinion . 35 0f) The amount of time you haveto pursue other interests? ... Satisfied ... 36 0 --—-----‘ C] 38 0Dissatisfied. 39 0 —---.--‘ io 0 41 0No opinion . 42 095Objective Questions: Amount of Housework Done by the RespondentSECTION F: Household helpFl. INTERVIEWER CHECK ITEM. Review GSS 5-1.Single person household 1 EJ—---+ GO TO F7Otherwise 2 ElF3. b) During the past 12 months, how much of the meal preparation did ... do? Was ItLess than 1/4 Less than 1/2 1/2 or more____?01 El 02 El 03 El 04 El 05 El06 El 07 El 08 El 09 El 10 EliiEl 12E] 13D 14[] 15[]16 El 17 El 18 El 19 El 20 ElNot applicable, no-one in household 21 El ------i GO TO F422 El Someone from outside householdF4. b) During the past 12 months, how much of the meal cleanup did ... do? Was itLess than 1/4 Less than 1/2 1/2 or more___23 El 240 250 260 27 El28 0 290 30 0 31 0 32 00 El 0 360 37 El38 El 0 400 41 0 420Not applicable, no-one in household 43 0------f GO TO F5El Someone from outside householdF5. b) During the past 12 months, how much of the cleaning and laundry did ... do? Was itLess than 1/4 Less than 1/2 1/2 or more0 46 El 0 48 El El50 [1 51 El 520 0 ElEl 560 57 El 580 59 El600 61 El 620 El ElNot applicable, no-one in household 65 0-—--- GO TO F666 0 Someone from outside household96F6. b) During the past 12 months, how much of the house maintenance and outside work diddo? Was itLess than 1/4 Less than 1/2 1/2 or more670 68 El 690 700 71 0720 730 740 750 760El 78 El 790 800 81 E]820Not applicable, no-one in household 87 0 ------ GO TO F788 0 Someone from outside household97List of the Independent Variables of Household Type and Child Care ArrangementsDescription of the Independent Measure “Spouse Present or Absent”or the Variable “Household Type”Code PopulationSingle Parent with Children 1 82Couple with Children Only 2 422Three Generation Family 0 56Comments: This is a derived variable. Household type is from the point of view of the referenceperson who is not necessarily the respondent.98List of Questions Measuring the Independent Variable of Child Care ArrangementsC8. Are there any children less than 15 years old living in the household?Yes iDNo 20—--‘ GOTOC16C9. The next questions refer to your children less than 15 years old living in the household.Gb. During the past 12 months, did any of your children receive childcare on a REGULARbasis? Exclude childcare provided by a family member living in this household.Yes 3 0 ------‘ How many?______No 4—.---’GOTOC16childrenCli. Did your child(ren) receive this care so that you or your spouse/partner couldYes NoWork atajob’?Study?Do volunteer work’?Provide care to a family member or friend’?Do something else’?oi0 020030 040050 060070 080oD ioD4-SpecifyC12. During the past 12 months, did .... (your youngest child) receive childcare OUTSIDE YOURHOUSEHOLD on a regular basis?Yes 3DNo 4D—----’GOTOC14Cl 3. Did ... (your youngest child) go to ... YesN4-SpecifyA workplace daycare center’?Another daycare center’?A sitter or neighbour’s home’?Grandparent’s home’?Another relative’s home’?Some other arrangement (outside your household)’?oiD 020030 040osD 060070 080090 ioDiiD 12099C14. During the past 12 months, did .... (your youngest child) receive childcare IN YOUR HOMEon a regular basis? Exclude childcare provided by a family member living in yourhousehold.Yes 51:No 6D------’GOTOC16C15. Who provided this care to ... (your youngest child)? Was itYes NoThe child’s grandparent9 01 1: 02 1:Another relative9 03 1: 04 1:A sitter or nanny9 05 [] 06 1:Someone else9 07 1: 08 1:‘I,Specify100Appendix BCorrelational Matrix for the Dependent VariablesDependent Variables --1 2 3 4Satisfaction Scale 1.000 —.1775—.0625 —.2192*Amount of HouseworkDone by theRespondent—.1775 1.000 —.1842 .1397Shift and/orWeekend work—.0625 —.1842 1.000 —.0045Number of Hours theRespondent Works(Outside the Household)per week —.2192* .1397 —.0045 1.000N=1 53* p < .01Dependent Variable Key: 1 = Satisfaction Scale2 Amount of Housework Done By theRespondent3 = Shift and/or Weekend Work4 = Number of Hours the Respondent Works(Outside the Household) per week101Appendix CThe Frequency Measure: Reliance and Frequency of Use Variables.Variable Name FrequencyReliance MeasureRespondent has one childin need of regular care 2 200Respondent has two childrenin need of regular care 4 61Respondent has three childrenin need of regular care 6 5Child Care UsageFour Types of Care 0Day care, family care and relative (dfrell) 2care (outside the household)Day care, family care andrelative care(inside the household) (dfrelat) 0Day care, relative care(outside the household) (drlvar) 9Day care, relative care(inside the household) (drvar) 7Famil( care, relative care(outside the household) (flvar) 16Family care, relative care(inside the household) (frvar) 7Relative care (both insideand outside the household) (rrvar) 8Day care only (day) 46Family care only (fain) 75Relative care(outside the household) (rell) 26Relative care(inside the household) (rel) 17


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items