MSU RETURNING MATERIALS: PIace in book drop to “33,0455 remove this checkout from 4--5--_ your record. FINES wiII be charged if book is returned after the date stamped beIow. 4:315 88% i I DETERMINANTS OF EMPLOYED WIVES' PERCEPTIONS OF THEIR CONTRIBUTION TO THEIR HUSBANDS' FINANCIAL SUPPORT By Cynthia A. Cameron A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Family and Child Ecology 1985 x723 / D .- N L K a "‘ “i ,"irt ‘, VT -.\ t I To Helen L. Cameron and Frank W. Cameron ii ACKNOWLEDGEMENTS It is vital that I acknowledge those who provided assistance and support in the writing of this thesis. Dr. Vicki R. Schram encouraged me to pursue graduate work and directed my growth as a scholar by always expecting the best. Dr. Margaret Bubolz graciously accepted the responsibility of directing my research. Dr. Robert Griffore found time in a hectic schedule to advise me on my statistical analysis. Dr. Marjorie Kostelnik served on my committee and offered constructive comments on methodology and writing style. These people have provided me with support throughout my graduate program. I also would like to express my gratitude to friends and family. C. Julie Hales unhestitatingly offered me emotional support and acted as a sounding board for my ideas. My husband, Tom Kopenhafer, who is both friend and family, listened patiently while I enthusiastically explained my latest accomplishment or talked of my feelings of frus- tration. My children, Laura and Jacob Youngdahl, learned to adjust their noise level to the degree of stress I was feeling. My parents. Frank and Helen Cameron provided me with a fine example of egalitarian marriage which has influenced my areas of interest in the study of family economics. TABLE OF CONTENTS Page LIST OF TABLES ........................ vi LIST OF FIGURES ....................... viii CHAPTER I. INTRODUCTION .................... 1 Theoretical Basis .................. 3 Relevant Research .................. 4 Ecological Perspective ............... 7 Significance and Generalizability .......... 7 The Research Data .................. 7 Research Questions and Objectives .......... 8 Research Variables ................. 11 II. METHODOLOGY ..................... 16 Research Design ................... 16 Development of the Instrument ............ 16 Sampling Procedures ................. 17 Techniques of Data Collection ............ 18 Response Rate . . . . ................ 19 Coding Procedures .................. 20 Data Analysis Procedures .............. 20 Relative Frequencies .............. 20 Crosstabulations . . . -. ............ 20 Multiple Regression ............... 21 III. 'ANALYSIS OF DATA .................. 24 Selection of the Study Sample . . . ......... 24 Description of the Sample .............. 26 Research Objectives ................. 36 Research Objective 1 . . ............ 36 Research Objective 2 . . ............ 41 iv CHAPTER Page Multicollinearity .............. 41 Regression Analyses ............. 41 Hypothesis 1 .............. 44 Hypothesis 2 .............. 44 Hypothesis 3 .............. 49 Hypothesis 4 .............. 49 Hypothesis 5 .............. 52 Hypothesis 6 .............. 52 Hypothesis 7 .............. 52 Hypothesis 8 .............. 52 Hypothesis 9 .............. 53 IV. LIMITATIONS, SUMMARY, DISCUSSION, AND IMPLICATIONS FOR FUTURE RESEARCH ......... . . . . . . . 54 Limitations ..................... 54 Summary . . ....... . . . ........... 54 Discussion of Findings ............... 56 Implications for Future Research .......... 59 APPENDICES .......................... 61 APPENDIX A. Portions of the Questionnaire Used in this Study . . 61 B. Calculation of Response Rate ............ 67 C. Additional Frequencies of Independent Variables . . . 69 BIBLIOGRAPHY . . .‘ ......... . ........... . 71 Table 100on 10. 11. 12. 13. 14. I 15. 16. LIST OF TABLES Crosstabulations of Perceived Support Provided by Employment Status ............ Frequency Distribution of Employment Status ..... Frequency Distribution Frequency Distribution Respondents Frequency Distribution Level of Respondents Frequency Distribution Frequency Distribution Frequency Distribution Frequency Distribution Income Adequacy Frequency Distribution of Support Provided by Frequency Distribution Support Provided .................. Frequency Distribution Egalitarian Family Style Frequency Distribution Egalitarian Family Style Frequency Distribution Responsibility Style Frequency Distribution Responsibility Style Frequency Distribution of Rural Status . . I . . . of Age of of Number of Children of Wife's Income Level . . . . of Family Income Level . . . . of Perceived Wives ........... of Perceived of Ideal of Perceived of Ideal of DIFFSCORE ....... vi Page 28 29 29 3O 3O 31 32 33 34 35 37 38 4O Table Page 17. Pearson Correlation Coefficients for Independent Variables .............. 42 18. Summary of Multiple Regression Analysis #1 for the Prediction of Perceived Support Provided by Selected Socio-economic and Socio-psychological Variables ............ 45 19. Summary of Multiple Regression Analysis #2 for the ‘Prediction of Perceived Support Provided by Selected Socio-economic and Socio-pyschological Variables ...................... 46 20. Summary of Multiple Regression Analysis #3 for the Prediction of Perceived Support Provided by Selected Socio-economic and Socio-psychological Variables Substituting Ideal Egalitarian Family Style for Perceived Egalitarian Family Style . . . . 47 21. Summary of Multiple Regression Analysis #4 for the Prediction of Perceived Support Provided by Selected Socio-economic and Socio-psychological Variables Including Discrepancy between Ideal and Perceived Egalitarian Family Style ....... 48 22. Summary of Multiple Regression Analysis #5 for the Prediction of Perceived Support Provided by Selected Socio-economic and Socio-psychological Variables including Ideal Responsibility Style and Discrepancy between Ideal and Perceived Responsibility Style ................ 50 23. Multiple Regression Analysis for the Prediction of Perceived Support Provided by Selected Socio-economic and Socio-psychological variables including Ideal Responsibility Style and Discrepancy between Ideal and Perceived Responsibility Style ................ 51 24. Frequency Distribution of Husband's Income Level . . 70 25. Frequency Distribution of Wife-Husband Income Ratio .................... 71 vii LIST OF FIGURES Figure Page 1. Classification of Respondents from Role Minimization to Role Maximization Using DIFFSCORE .................... 39 viii ABSTRACT DETERMINANTS OF EMPLOYED NIVES' PERCEPTIONS OF THEIR CONTRIBUTION TO THEIR HUSBANDS' FINANCIAL SUPPORT By Cynthia A. Cameron This study examined the way in which employed wives perceive their role as provider. The data were gathered by survey research in the Agricultural Experiment Station regional research project NC 164, Stress, Coping and Adaptation in the Middle Years of the Family. This study used a sub-sample of 53 cases from the Michigan data. There were two objectives of this study. One objective was to determine if employed wives could be placed on a continuum of role minimization to role magnification according to how they view their provider role. More than 50% of respondents minimized their role as provider and a little less than 10% magnified the provider role. The second objective was to identify the major variables that predict and explain employed wives' perceptions of their contribution to their husbands' financial support. Four major variables were identified as major predictors: discrepancy between ideal and perceived responsi- bility style (E$.Ol), wife-husband income ratio (23505), number of children, and wife's level of income. CHAPTER I INTRODUCTION Prior to the industrial revolution, the home was the center of economic production, and men, women and children worked together in order for the family members to survive (Reimer and Fout, 1980). In peasant families, fathers and sons performed the field work, while mothers and daughters gardened, tended the barnyard and prepared the food (Scott and Tillyu. 1975). Although men worked outside the home and women worked in or near the home, the labor of each was equally important for survival. After the industrial revolution, work shifted from the home to the factory (Reimer and Fout, 1980). At first, entire families worked to- gether in the factories and earned a family wage. After the passage of protective labor laws, women and children were forced to leave the fac- tories and return home. It was then necessary for men's wages to in- crease, as they had to provide for the entire family. During this time, societal expectations changed concerning men's and women's work roles. In the past, men and women were expected to contribute equally to the family economy. After the industrial revolu- tion, men's role became that of sole provider and women's work role changed to caretaker of the home and the family members. These roles commonly are referred to as traditional gender roles. 2 Married women in the United States remained in caretaker roles from the late nineteenth century until World War II. During the war, women were accepted in the paid work force in order to fill the void left by men away at war. Although it was expected that women would discontinue working outside the home after the men returned, this did not happen. Instead, female labor force participation continued to rise. The increase in the work force participation of married women has been dramatic. Married women tripled their work force participation rate between 1947 and 1976 (U.S. Bureau of Labor Statistics, 1977). In 1982, approximately 51% of married women were employed outside the home (U.S. Department of Labor Statistics, 1983). Wives in two earner families earn slightly more than 38% of the family income (U.S. Department of Commerce, 1984), and women's wages are instrumental in increasing the standard of living in dual-worker families (Oppenheimer, 1977). With the increase in the number of wives working outside the home, gender role attitudes started to change. Although societal expecta- tions still reinforced traditional gender roles, egalitarian gender role attitudes were adopted by some who perceived that both men and women should share the provider and caretaker roles. Little research has been conducted to determine if employed wives perceive their wage-earner role as providing financial support for family members. It seems likely that wives who hold traditional gender role attitudes will perceive their wage—earner role differently than wives who hold egalitarian gender role attitudes. The purpose of 3 this study is to identify the major variables that predict and explain the variance in employed wives' perceptions of their contribu- tion to their husbands' financial support. THEORTICAL BASIS Role theory predicts that the positions, or roles, occupied by individuals are determined by societal norms, demands and rules. Role performance is also affected by actions of persons in like roles, by reactions of observers to how the role is performed, and by a person's unique capabilities and personality (Thomas and Biddle, 1966, p. 4). General roles are "patterns of behavior that characterize members of a general position" (Biddle, 1979, p. 297). These patterns of behavior are determined by societal norms, demands and rules. Gender roles are general roles that are defined by the gender of the actor. That is, male and female roles are determined by societal expectations. Although societal norms reinforce traditional gender roles, observa- tions of women's actions (over 50% of married women work outside the home) have aided the development of egalitarian gender role attitudes. Employed wives may hold egalitarian gender role attitudes (i.e. males and females as co-providers and co-caretakers), traditional gender role attitudes (i.e. males as providers, females as caretakers), or a combination of egalitarian and traditional gender role attitudes (i.e. males and females as. co-providers, females as caretakers). Employed wives with egalitarian gender role attitudes will be likely to perceive themselves as sharing the provider role with their husbands. They may perceive their contribution to their husbands' support as close to the actual support they provide (role approxima- tion). They also may perceive their contribution to be equal to that of their husbands' if they perceive the provider role to be shared 4 equally, even though they earn less than 50% of the family income (role magnification). Employed wives with traditional gender role attitudes may perceive their husbands to be sole providers and under- estimate their own contribution as a provider (role minimization). Role theory is used as the basis for predicting that variables related to gender role attitudes will influence employed wives' per- ceptions of their contribution to their husbands' financial support. RELEVANT RESEARCH This section presents a review of existing literature concerning the variables that may influence employed wives' perceptions of their contribution to the financial support of their husbands. Several socio-economic and socio-psychological variables are discussed. Bird (1979) found that the way dual-earner couples view the pro- vider role is related to gender role attitudes. She identified four styles used by couples in allocating the income of the wife. One spending style allocates the wife's earnings as pin money. In these couples, the wife's income is kept separate from the husband's and the wife may spend it as she pleases. This allows both spouses to minimize the importance of the wife's wages. Earmarker Couples use the wife's income for special expenses. This allows them to live on the husband's income alone, while the wife helps out, thus allowing the husband to retain his provider status. Pooler Couples put both incomes together to use in whatever way is necessary, while in Bargainer Couples, the perception of the wife's income approximates the actual level of her contribution. Bird found that the husband in Pin Money Couples is likely to disapprove of his wife working outside the home, that the wife in Earmarker Couples is likely to be traditional in her 5 gender role attitudes, and that Pooler and Bargainer Couples are likely to hold contemporary, or egalitarian, gender role attitudes. Age has been found to influence gender role attitudes, and there- fore seems likely to influence wives' perceptions of their contribution to their husbands' financial support. Conflicting findings have been reported by researchers who have used different age groups to test this relationship. It has been found that younger women have more egali- tarian gender role attitudes than older women (Scott and Morgan, 1983). It has been hypothesized that this is due to a greater flexibility in behavior and beliefs in the younger women (Duncan, 1982). It also is true that younger cohorts have had more exposure to egalitarian ideas during years when they were forming attitudes than have women in older cohorts. Spitze and Huber (1980) hypothesized that middle-aged women may be more traditional than younger or older women because they were addlescents during the 1950's when the tradiational gender roles were idealized by the mass media. Duncan (1982) reported that in 1953 recent cohorts were less egalitarian than older cohorts, while in 1976 recent cohorts were more egalitarian than older cohorts. Mason, Czajka and Arber (1976) found that older women held more egalitarian gender role attitudes than younger women. This may be the result of older women becoming assertive as older men become more family oriented, allowing for an equalization of roles in the family (Lowenthal and Chirboga, 1972). Researchers have reported a positive relationship between women's educational levels and egalitarian gender role attitudes. The more education women have, the more likely they are to lean toward egalitar— ianism (Mason, et al, 1976; Scott and Morgan, 1983). Bird (1979) found 6 that highly educated women are more likely to be a member of Pooler and Bargainer Couples using their earnings to help pay for living expenses of both partners. Perceived income adequacy influences wives' perceptions of their husbands as providers. Wives who feel the family income is not ade- quate to meet family needs perceive their husbands to be less than adequate in fulfilling their role as providers (Keith and Schafer, 1980). This could influence wives' perceptions of their own roles, i.e. employed wives may be more likely to feel they are co-providers if family income is inadequate. The level of wives' and husbands' incomes has been found to influence wives' gender role attitudes. Scanzoni (1978) found that 55% of employed wives felt it was their duty to act as co-provider, and that the most important predictor was level of wives' earnings. The higher the income of wives, the more likely they were to consider them- selves as co-providers. Conversely, the higher the husbands' incomes, the less likely were the wives to consider themselves as co-providers (Scanzoni, 1978). There is a positive relationship between wife—husband income ratio and wives' perceptions of their duty to co-provide (Scanzoni, 1978). Peterson and Maynard (1981) reported that wives became more egalitarian the closer their income level approached their husbands' income level. In Becoming a Two Job Family, Hood (1983) describes the changes in husband and wife roles in 16 families after the wives obtained employment. She observed that the proportion of the family income earned by the wives seemed to influence their position as co- providers. 7 In summary, several variables are likely to influence employed wives' perceptions of their contribution to their husbands' support. However, there is little previous research directly related to this research problem. ECOLOGICAL PERSPECTIVE Each individual is part of an ecosystem consisting of the natural environment, the human constructed environment and the human behavioral environment (Bubolz, Eicher, and Sontag, 1979). This study proposes to ascertain the importance of one segment of the human constructed envir- onment, socio-cultural values, i.e. gender roles, in determing em- ployed wives' perceptions of their contribution to their husbands' support, a part of the human behavioral environment. SIGNIFICANCE AND GENERALIZABILITY In view of what has been presented above, this study examined employed wives' perceptions of how they contribute to their husbands' financial support, and will add knowledge to a little studied area. The findings of this study will add to the growing body of literature on dual-earner families and to the understanding of this family type. The findings of this study are generalizable to similar populations of employed wives. THE RESEARCH DATA The data for this research were gathered as part of a regional research project supported by the Agricultural Experiment Stations of the North Central Region of the United States and the Cooperative ' Research Service of the United States Department of Agriculture. The project is NC 164,5tress, Coping and Adaptation in the Middle Years of the Family. Nine states are involved in the project: Indiana, Iowa, 8 Kansas, Kentucky, Louisiana, Michigan, Minnesota, Missouri, and Nebraska. Michigan project directors are Dolores Borland and Margaret Bubolz. Only Michigan data were used in this study. RESEARCH QUESTIONS AND OBJECTIVES Research questions for this study were developed by the researcher as a result of familiarty with the data gained through serving as coding supervisor for the Michigan NC 164 project. During the coding process, patterns were observed in the way wives reported the percent of support they provide to household members. It was noted that both employed wives and full-time homemakers stated they provide 100% of support to children living at home. This was in contrast to the re- sponses of women reporting the percent of support they provide to their husbands. Full-time homemakers generally reported they provide 0% of support to their husbands, while some employed wives reported they provide no support to their husbands and others stated a specific per- cent of support. These observations led to the development of two ' research questions. 1. Can employed wives be placed on a continuum ranging from role minimization to role maximation by examining the responses to the percent of support they provide to their husbands? 2. What socio-economic and socio-psychological variables are related to the responses given by women regarding the percent of support they provide their husbands? One objective of this research is to classify the respondents according to their perception of their contribution to their husbands' financial support. This will be accomplished by determining if the respondent reports that her contribution to her husbands' support is 1. less than the actual percent of the family income she earns (role minimization); 9 2. within i10% of the actual percent of the family income she earns (role approximation); 3. more than the actual percent of the family income she earns (role magnification). The second objective of this study is to identify the major vari- ables that predict and explain the variance in employed wives' percep- tions of their contribution to their husbands' financial support. More specifically, the second purpose of this research is to test the hypotheses listed below. H01 There is no relatidnship between egalitarian family style and employed wives' perceptions of their contributiontxitheir husbands' financial support. H011 There is a positive relationship between egalitarian family style and employed wives' perceptions of their contribution to their husbands' financial support. This hypothesis is suggested by the work done by Bird (1979) who found that wives in couples who use the women's earnings to pay for living expenses are more likely to have egalitarian gender role atti- tudes. H02 There is no relationship between age and employed wives' perceptions of their contribution to their husbands' financial support. H012 There is a curvilinear relationship between the age of em- ployed wives and their perceptions of their contribution to their husbands' financial support. Research conducted by Duncan (1982), Spitze and Huber (1981), and by Mason, et al, (1976) indicates women in themiddle years are more likely to hold traditional gender role attitudes, while younger and older women are more likely to hold egalitarian gender role attitudes. H03 There is no relationship between the educational level of employed wives and their perceptions of their contribution to their husbands' financial support. 10 H013 There is a positive relationship between the educational level of employed wives and their perceptions of their contribution to their husbands' financial support. Research findings reported by Mason, et al (1976), and by Scott and Morgan (1983), stated that the higher the woman's level of educa— tion, the more egalitarian are her gender role attitudes and suggest the previous hypothesis. H04 There is no relationship between number of children of employed wives and their perceptions of their contribution to their husbands' financial support. H014 There is a negative relationship between the number of children of employed wives and their perceptions of their contribution to their husbands' financial support. There was nothing found in the review of relevant research on which to base this hypothesis. Scanzoni (1975) and Scott and Morgan (1983) did report that there is a negative relationship between women's egalitarian gender role attitudes and the number of children they plan to have. It seems logical that the more children women have the more likely they will be to identify with the caretaker role rather than the provider role. H05 There is no relationship between perceived income adequacy and employed wives' perceptions of their contribution to their husbands' financial support. H015 There is a negative relationship between perceived income adequacy and employed wives' perceptions of their contri- bution to their husbands' financial support. ' This hypothesis is suggested by Keith and Schafer (1980) who reported that women who perceive family income to be inadequate also perceive their husbands to be poor providers. H06 There is no relationship between level of wives' incomes and their perceptions of their contribution to their husbands' financial support. 11 H07 There is no relationship between level of husbands' incomes. and employed wives' perceptions of their contribution to their husbands' financial support. H017 There is a negative relationship between level of husbands' incomes and employed wives' perceptions of their contribu- tion to their husbands' financial support. These hypotheses are suggested by Scanzoni (1978) who found that level of wives' incomes and level of husbands' incomes influence employed wives' perceptions of their duties as co-providers. H0 There is no relationship between wife-husband income-ratio and employed wives' perceptions of their contribution to their husbands' financial support. H018 There is a positive relationship between wife-husband income ratio and employed wives' perceptions of their contribution to their husbands' financial support. Scanzoni (1978), Peterson and Maynard (1981) and Hood (1983) reported that women are more likely to see themselves as co-providers when their income level approaches that of their husbands'. An additional hypothesis was developed during the statistical analysis phase of this study. This hypothesis was formulated as the researcher examined the results of the first three multiple regression analyses, the raw data and the questionnaires of the study sample. H0 There is no relationship between the degree of discrepancy in ideal egalitarian family style and perceived egalitarian family style and employed wives' perceptions of their con- tribution to their husbands' financial support. H019 There is a positive relationship between the degree of dis- crepancy in ideal egalitarian family style and perceived egalitarian family style and employed wives' perceptions of their contribution to their husbands' support. RESEARCH VARIABLES This section conceptually and operationally defines the dependent and independent variables. The dependent variable in this study is employed wives' perceived contribution to their husbands' support. 12 Independent variables include perceived and ideal egalitarian family style, age, educational level, number of children, perceived income adequacy, level of wife's income, level of husband's income, wife- husband income ratio, and discrepancy between ideal and perceived egalitarian family style. Perceived contribution to husbands' support: Perceived contri- bution to husbands' support is conceptually defined as the extent to which the respondent feels she provides financial support to her husband. This variable was measured by asking the respondent to state the percent of support she provides to each member of her household (Appendix A, question 3). Egalitarian family style: Egalitarian family style is concep- tually defined as a family style in which all family members have input into family decisions and roles and responsibilities are shared and flexible. Ten statements that are included in FACES II, an instrument developed by Olson, Partner and Bell (1983) to measure family adapta- tion and cohesion, were used to measure egalitarian family style. (Appendix A, pp. 63-65, 013 is the FACES 11 scale. The ten items used to measure egalitarian family style are asterisked.) Statements were chosen by the researcher that appeared to be related to egalitarian family style. Nine of the ten statements measure family adaptability which is defined as “the ability of a marital or family system to change its power sturcture, role relationships and relationship rules in response to situational and developmental stress (Olsen, et al, 1983, p. 3). One statement measures cohesion (Appendix A, question 13, statement m) and was used to determine input of family members in decision making. Since no measure of reliability or validity was 13 available in the use of these ten statements to measure egalitarian family style, the ten statements were presented to a panel of five experts in the area of marriage and family for determination of face validity. All of the experts concluded that the ten statements appear to measure egalitarian family style. Respondents were asked to indicate the frequency of events in their families that are representative of egalitarian family style and the ideal frequency of these events. A response of almost always to statements b, d, f, j, n, p, v and 2 indicates an egalitarian family style. For these statements a response of almost always received a score of 5 and a response of almost never received a score of 1. A response of almost never to statement bb indicates an egalitarian family style. For this statement a response of almost never received a score of 5 and a response of almost always received a score of 1. The scores were averaged to determine an aggregate score for perceived family style and ideal family style. The highest possible score is 5 and indicates the highest egalitarian family style and the lowest pos- sible score is 1 and indicates the lowest egalitarian family style. An additional measure of perceived and ideal egalitarian family style was computed from the scores of two of the ten statements. These statements refer to the shifting and sharing of family responbilities (Appendix A, 013, j and v). It was felt that since co-prdviding can be viewed as a shifting from traditional to egalitarian roles and a sharing of the responsibility of breadwinner, these statements might be closely related to the dependent variable. The statements were scored as above, with the scores from the statements of "how you describe your family now" indicating perceived responsibility style, and the scores 14 of "how you would like your family to be" representing ideal responsi- bility style. Educational level: Level of education was measured by asking the respondent to indicate the number of years of schooling completed. Partial years were excluded (Appendix A, question 1). Age; The respondent's age was determined by asking her to indi— cate the month and year of her birth. Her actual age in years was cal- culated by subtracting her year of birth from the year in which she responded to the survey, 1983 (Appendix A, question 1). Number of children: Number of children was measured by asking the respondent to give information about each of her children from oldest to youngest (Appendix A, question 2). Perceived income adequacy: Perceived income adequacy is concep- tually defined as respondent's perception of the degree of adequacy of her household income to meet the needs of that household. This vari- able was measured by asking the respondent to indicate to what degree family income is adequate. A response indicating the lowest degree of adequacy was coded as 1, while the highest degree of income adequacy was coded as.5 (Appendix A, question 26). Level of wife's income: Level of wife's income was measured by asking the respondent to state her annual income for the preceding year (Appendix A, question 23). Level of husband's income: Husbands were asked to indicate their annual income for the preceding year in order to measure this variable (Appendix A, question 23). Wife-husband income ratio: Wife-husband income ratio was measured by using the respondent's annual income and her husband's 15 annual income in ratio form. Discrepancy between ideal and perceived egalitarian family style: This variable was calculated by subtracting the perceived egalitarian family score from the ideal egalitarian family style score. CHAPTER II METHODOLOGY This chapter presents information on the research design, devel- opment of the instrument, sampling procedures, response rate and coding procedures. Brief explanations of the statistical techniques used in data analysis also are included. RESEARCH DESIGN This study is exploratory in nature and used survey research design to identify major variables that predict and explain the variance in employed wives' perceptions of their contribution to their husbands' financial support. It is a cross-sectional study and is non- experimental in nature. The unit of analysis is employed wives in dual- earner couples. Data were collected using a self-administered questionnaire. DEVELOPMENT OF THE INSTRUMENT The instrument for the NC 164 research project was developed and compiled by the representatives of the nine states participating in the project. Representatives formulated sections of the instrument related to their specific interest areas. In addition, FACES 11 (Family Adapt- ability and Cohesion Evaluation Scales) as developed by Olson, Portner, and Bell (1983) was used in the instrument. 16 17 SAMPLING PROCEDURES The population chosen for the NC 164 study consists of middle-aged families in the state of Michigan. Because it would be an expensive and overwhelming task to draw a random sample from families in the entire state, representative areas were selected. One SMSA and four rural counties were chosen as the representative populations from A which samples were drawn randomly. I Flint was the SMSA chosen as representative of other Michigan SMSA's on the dimensions of education and income levels. Also on the dimensions of education and income levels, the following counties were chosen as representative of rural counties in Michigan: Alpena, Charlevoix, Dickinson, Hillsdale. The Donnelly Corporation (a company which collects and distributes lists of names for businesses and other organizations across the nation) was contacted to provide a thorough list of families which met the criteria that the mother in a two-parent family was between the ages of 35 and 65 years. From the provided lists, random samples were drawn for each county or city using a list of random numbers. In an attempt to receive a sample of 100 respondents from the urban area, 300 families were selected from within the city limits of Flint. In an attempt to receive a sample of 100 rural farm families and 100 rural nonfarm families, 600 families were randomly selected from the rural counties. For the rural counties, the number of families selected from each county was determined proportionally with equal weight given to population of the counties and number of farms 18 in the counties. The number of families selected from each rural county was as follows: Alpena, 138; Charlevoix, 84; Dickinson, 48; and Hillsdale. 330. (Fora more detailed explaination of sampling procedures, see Smith, 1985). The sample used in this study is a sub- sample of the NC 164 sample and is clarified in Chapter III. TECHNIQUES OF DATA COLLECTION The method of data collection was a modification of the Dillman procedure (Dillman, 1978). The data were collected using a self- administered questionnaire. A set of questionnaires, one for each spouse, was mailed to each household in the sample, along with a letter of explanation and a self-addressed stamped envelope. Each question- naire carried the designation husband or wife and an identification number to enable the researchers to determine who had returned the questionnaire. After one week,a postcard reminder was mailed to those who had not responded, and after four weeks an additional set of questionnaires was mailed to non-responding households with a letter encouraging participation in the study. RESPONSE RATE The overall response rate for the Michigan sample of NC 164 was 35.9%. As stated above, the sampling frame was provided by the Donnelly Corporation, which guarantees the accuracy of their lists at a certain percentage rate. Although only the "A", or most accurate, list had been requested by the NC 164 project, it was discovered after the questionnaires had been mailed that the sampling frame also included the less accurate "B" and "C" lists. When questionnaires were returned from couples who did not fit the age requirements for the project, it became clear that the inaccurate sampling frame could have 19 affected the response rate. It was possible that couples eliminated themselves from the study becasue they were not in the middle years of family life. In order to determine if this was so, the more accurate "A" list was requested from the Donnelly Corporation. Separate response rates were then calculated to determine if the "A" list re- sponse rate was higher than that of the "B" and "C" lists (Appendix B). Indeed, the response rate for the "A“ list names was 42.4%, while the response rate for the "B" and "C" list names was 34.0%. It is felt 'that the 42.4% response rate more accurately represents the true response rate for the sample and, although not optimal, it is an acceptable response rate when length and complexity of the question- naire are considered. CODING PROCEDURES The data of the Michigan NC 164 project were coded by five student coders who were supervised by this researcher. Accuracy and congruency of coding were controlled in the following ways. 1. Coding was done only in the presence of the coding super- visor. Coders were instructed to ask for clarification whenever they had questions. 2. Specific instructions for each question were included with the codebook. These instructions included descriptions of problems that had been identified and how they were to be coded. 3. Each question had a corresponding problem sheet included with the codebook. If there were no specific instructions and the coding supervisor could not answer a question, the question was brought before project leaders and a decision was made on how to code the response. 4. After the coders were familiar with the coding process, the coding supervisor and all coders coded the same questionnaire. The code sheets were then checked for inter-coder reliability. 20 5. One work week was used to check coding sheets for errors. Since error rate was less than 2%, it was decided by project leaders that no further checking would be done, due to financial and time constraints. After the coding was completed, the data were keypunched and entered onto the computer system. The data were cleaned by the coding supervisor. DATA ANALYSIS PROCEDURES Relative Frequencies Relative frequencies were computed for each of the independent variables, the dependent variable, rural status, employment status, and family income. These frequencies are used to describe the partici- pants of the study and to determine if the assumption of normality of the distribution of the dependent variable is met as if necessary for multiple regression analysis. Crosstabulations Crosstabulations are joint frequency distributions of two or more variables. In this study, a 6 X 10 crosstabulation table was computed to aid in understanding the way in which respondents interpreted the question pertaining to the dependent variable. This was necessary because the question did not specify financial or emotional support (Appendix A, question 3). Multiple Regression Multiple regression analysis is a statistical technique used to determine the relationship between a dependent variable and a set of independent variables. Using multiple regression, the researcher can determine the best prediction equation for the dependent variable. 21 The multiple regression equation is '= A + lel + 82X2 + . . . Bka where Y' is the estimate of the dependent variable, A is the Y' inter- cept and Bi are the values by which Xi are multiplied to obtain the best possible prediction line for Y. The equation describes a line which most closely represents the points on a scatterplot for each case. The distance between the actual values for Y for each case and the estimated point on the regression line (Y') is the residual. The regression line is calculated to minimize the sum of the squared resi- duals. The standard error of estimate is the square root of the squared residuals and gives an indication of the average error in predicting Y from the regression line (Kim and Kohout, 1975). Rg'hsthe proportion of the variance in the dependent variable explained by the independent variables. Each independent variable entered into the regression equation will increase 3?. The law of diminishing returns takes effect, however, and each additional variable explains less of the variance. When variables that account for little variance in Y are entered into the equation, 3? will rise slightly, but so will the standard error of estimate. The best prediction equation is at the point of the lowest standard error of estimate. In this study, the variables included in the best prediction equation will be identified as major variables in predicting the dependent variable. Step-wise multiple regression analysis is used to determine the . optimal prediction equation and identify the variable related most strongly to the dependent variable. Step-wise multiple regression with forward inclusion allows for independent variables to be entered into the equation from the single best predictor to the least effective 22 predictor (Kim and Kohout, 1975). The first independent variable to enter the equation is the one that explains the greatest amount of variance in the dependent variable. Each additional variable entered will be in the order that allows for the greatest additional explana- tion of the remaining unexplained variance in the dependent variable. With the addition of each variable, 3? will continue to rise as more of the variance in the dependent variable will be explained. The standard error of estimate will decrease as long as variables entered into the equation improve the predictability of the dependent variable. Variables will continue to enter the equation until the previously determined level of f_to enter is not met. In this study the default value of f.to enter was used so that the order of the variables in several regression analyses could be compared. In addition to determining the best prediction equation, it is possible to test for significance of the relationship between the dependent variable and each independent variable by using the [rtest. The Ertest was used to determine if it is possible to reject the proposed hypotheses in this study. Multiple regression analysis is most accurate when there is a high degree of correlation between the dependent variable and indepen- dent variables and a low degree of correlation between the independent variables. The use of Pearson Correlation analysis allowed for determinination of relatedness of the independent variables. The Pearson product-moment correlation coeffieients indicate the degree to which change in one variable is related to change in another. The coefficient describes the strength and the direction of the 23 relationship and ranges from -1 to +1 with values close to zero indicating little or no relationship. CHAPTER III ANALYSIS OF DATA This chapter will describe the results of the data analysis procedures. Sections included are selection of the study sample, description of the study sample and multiple regression analyses. SELECTION OF THE STUDY SAMPLE Cases for this study were selected by the following criteria: 1. The respondent was employed outside the home either full or part-time; 2. The respondent reported personal income for the previous year. The cases that met these criteria were screened to try and determine how the respondent interpreted the question that provided the data for the measurement of the dependent variable. Question 3 requested the respondent to indicate percent of support provided to household members (Appendix A). Since financial or emotional support was not specified, there was concern about the accuracy of the data. Crosstabulations were run on the dependent variable by employment status (Table 1). By examining the results, it can be noted that 27 of the 35 fUll-time homemakers reported they provide 0% of support to their husbands and eight responded with 100%. While a response of 0% of support indicates the question was interpreted as financial support, a response of 100% indicates the question was interpreted as emotional support. Since it was impossible to determine how the three full-time and one part-time 24 25 sHH NH N NH H H a m a N ON HaHoN asaHoN N o o o o o o o o o N aogHo m o o o o o o H o H m aaNoHasoca N o o o o o o o o o N aaLHHoN mN H N N H o N H N H NH oEHH-Haaa eoNoHaeN ca N 0 HH 0 H a H N o NH aEHH-HH=c aoNoHaeN mm m o o o o o o o o NN Laxasoso; asHH-HH=N NooH NmN Nam Noe Nmm Non NmN NON NoH No HaHoN zom umuN>oNa ugoaazm um>wmucma maumum pcmsonQEm magnum acmeonaEm An umuN>oN¢ acoaaam um>wmucma mo meowumpanmummocuii.H anm» 26 employed respondents who reported they provide 100% of support to their spouse interpreted the question, these four cases were dropped from the study sample. An additional nine cases were eliminated because of missing data or the reporting of a negative personal income. The final sample for this study consisted of 53 employed wives report- ing they provide from 0% to 50% of their husbands financial support. DESCRIPTION OF THE SAMPLE Of the 53 cases included in this study, almost 65% were wives who are employed full-time, while slightly more than 35% are employed part-time, which is defined as less than 35 hours per week (Table 2). A little more than 65% are from rural Michigan (Table 3). The age of the women in the sample ranged from 36-64 years, with a mean age of 46.1 years (Table 4). Only one respondent had less than a high school education, while almost 50% had at least some college education. The mean years of education for the study sample is 13 years (Table 5). Almost 80% of respondents had from one to three children with the mean number of children being 2.8 (Table 6). The mean annual income of these women is slightly more than $11,000, although almost 55% earned less than $10,000 annually (Table 7). The mean family income is slightly more than $35,000 annually (Table 8). Almost 80% of the women reported that their family is able to afford at least some wants (Table 9). Family income was calculated by summing the wife's and husband's annual income. The percent of the family income earned by the women ranged from 3% to 74% with a mean of 31.3% (Table 10), slightly below the national average of 38% for dual-earner couples. The calculated percent of income earned differed considerably from the perceived 27 Table 2.--Frequency Distribution of Employment Status (n=53) Employment Status Frequency Percentage Employed full-time 34 64.2 Employed part-time 19 35.8 Total 53 . 100.0 Table 3.--Frequency Distribution of Rural Status (n=53) Rural Status Frequency Percentage Rural 36 67.9 Urban 17 32.1 Total 53 100.0 28 Table 4.--Frequency Distributidn of Age of Respondents (n=53) Age (in years) Frequency Percentage 35-39 13 24.5 40-44 13 . 24.5 45-49 9 17.0 50-54 9 17.0 55-59 7 13.2 60-64 ' 2 3.8 Total 100.0 29 Table 5.--Frequency Distribution of Education Level of Respondents(n=53) Education Level Frequency Percentage Less than 12 years 1 1.9 12 years 27 50.9 (high school) Greater than 12 years 12 22.7 less than 16 years (some college) 16 years 8 ° 15.1 (college degree) Greater than 16 years 5 9.4 Total 53 . 100.0 Table 6.--Frequency Distribution of Number of Children (n=53) Number of Children Frequency Percentage 1 4 7.5 2 20 37.7 3 18 34.0 4 6 11.3 5 4 7 5 6 1 19 Total 53 100.0 30 Table 7.--Frequency Distribution of Hife's income level (n=53) Nife's income level Frequency Percentage Less than $10,000 29 54.7 $10,000-19,999 15 28.3 $20,000-29,999 9 17.0 Total 53 100.0 Table 8.--Frequency Distribution of Family Income Level (n=53) Family income level Frequency Percentage Less than $10,000 1 1.9 $10,000 - 19,999 3 5.6 $20,000 - 29,999 16 30.2 $30,000 - 39,999 15 28.3 $40,000 - 49,999 8 15.1 $50,000 - 59,999 5 9.5 $60,000 - 69,999 5 9.5 Total 53 100.0 31 Table 9.—-Frequency Distribution of Perceived Income Adequacy (n=53) Perceived Income Frequency Percentage Adequacy Income does not meet needs 1 1.9 Income meets needs only 10 18.9 Income meets some wants 31 58.5 Income meets most wants 7 13.2 Income meets wants with 4 7.5 some leftover Total 53 100.0 32 Table 10.--Frequency Distribution of Calculated Percent of Support Provided by Wives (n=53) Calculated Percent of Frequency Percentage Support Provided Less than 10% 3 3.8 10-19% 5 11.3 20-29% 18 34.0 30-39% 13 24.5 40-49% 9 15-9 50-59% 2 3.8 60:69% 2 3.8 70-79% 1 1.9 80-89% 0 0.0 90-100% 0 0.0 Total 53 100.0 33 TableIJ.--Frequency Distribution of Perceived Support Provided (n=53) Perceived Support Frequency Percentage Provided 0% 28 52.8 10% 1 1.9 20% 4 7.5 25% 1 1.9 30% _ 5 -9.4 33% 1 1.9 40% 1 1.9 50% 12 22.7 Total 53 100.0 34 Table 12.--Frequency Distribution of Perceived Egalitarian Family Style (n=53) Family Style Score Frequency Percentage 1.0-1.4 0 0.0 1.5-1.9 1 1.9 2.0-2.4 O 0.0 2.5-2.9 7 13.2 3.0-3.4 17 32.1 3.5-3.9 21 39.6 4.0-4.4 7 13-2 4.5-5.0 O 0.0 Total 53 100.0 35 Table u1--Frequency Distribution of Ideal Egalitarian Family Style Score (n=53) Ideal Family Style Frequency Percentage Score 4 1 0-1.4 O 0 0 l 5-1.9 O O 0 2 0-2.4 0 0 O 2 5-2.9 0 O O 3 0-3.4 4 7 5 3 5-3 9 23 43 4 4 0-4 4 19 35 9 4 5-5.0 7 13 2 Total 53 100.0 36 contribution to support they stated they provide to their husbands, as over 50% reported they provide 0% of support (Table 11). Although it was difficult to determine the degree of egalitarian- ism within the families of women studied, it was possible to determine that many women view the ideal family style as more egalitarian than the way their family is now. It was possible to draw this conclusion by comparing the scores for ideal egalitarian family style (IDFAMSTY) and perceived egalitarian family style (FAMSTYLE) in Tables 12 and 13 and the scores for ideal responsibility style (IDRESP) and perceived responsibility style (RESP) in Tables 14 and 15. RESEARCH OBJECTIVES Research Objective 1 By computing the difference between the perceived percent of support provided and the actual percent of family income earned, it was possible to place the respondents on a continuum from role minimization to role magnification (Figure 1). The frequency distri- bution for DIFFSCORE is reported in Table 16. Slightly more than 50% of respondents minimized their role as income providers with a dif- ference in perceived support and actual support from -11% to -70%. Less than 10% of the respondents were role magnifiers with a differ- ence in perceived support provided and actual support provided ranging from +11% to +40%. Almost 40% of the sample reported per- ceived percent of support provided within -10% and +10% of actual support provided and can be placed in the category of role approximation. 37 Table 14.--Frequency Distribution for Perceived Responsibility Style (n=53) Responsibility Frequency Percentage Score 1 0 0.0 2 3 5 7 3 6 11:3 4 3 5 7 5 7 13.2 6 11 20.7 7 11 20.7 8 4 7 5 9 5 9.4 10 3 5.7 Total 53 100.0 38 Table 15.--Frequency Distribution for Ideal Responsibility Style (n=53) Ideal responsibility Frequency Percentage score 1 O 0.0 2 0 0.0 3 0 0 O 4 1 1 9 5 3 5 7 6 6 11.3 7 10. 18 9 8 13 24 5 9 11 20.7 10 9 17.0 Total 100.0 mmoummmmo ou mcwucouum :owumuNmNcmmz o_om ca :oNHmNNEN:Nz mpom Eocm mucwucoammm ac :oNHmUNNNmmm—u .H mcamwm 39 cmwwou_a_cmmz coNHmaNxocaa< cowum~_s_cwz m mpom mpom ooH om om oN om om cc om om oH o oH- om- om- ow- cm. on- oN- om- om- ocHn . I. . N m e m o N m m 9H HH 40 Table 16.--Frequency Distribution for DIFFSCORE (n=53) DIFFSCORE Frequency Percentage -61% - -70% ’ 1 1.9 -51% - -60% 0 0.0 -41% - -50% 1 1-9 -31% - -40% 11 20.7 -21% - -30% 8 15.1 -11% - -20% 6 11.3 - 1% - -10% 8 15.1 0% I 4 7.5 +1% - +10% 9 17.0 +11% - +20% 3 5.7 +21% - +30% 1 1.9 +31% - +40% 1 1.9 Total 53 100.0 41 Research Objective 2 Several multiple regression analyses were used to determine the major predictor variables for perceived support provided. Multiple regression analysis assumes a normal distribution for the dependent variable. In this study, the dependent variable was not normally distributed as can be seen in the frequency distribution in Table 11. Because this assumption of multiple regression analysis was not met, the dependent variable was dichotomized. All responses of more than 0% were recoded as "1" and responses of 0% were coded as "0". This allowed for the determination of the predictive value of the indepen- dent variables in determining group membership. Although recoding of a continuous variable to a dichotomous variable causes information to be lost, it is sometimes necessary when the assumptions of multiple regression cannot be met (Kerlinger, 1973, p. 644). Multicollinearity. If the independent variables in a multiple regression analysis are intercorrelated at the .80 level and above, the problem of multicollinearity exists (Kim and Kohout, 1975). Pearson correlation Coefficients were computed in order to examine the degree of correlation between the independent variables (Table 17). Since the highest correlation between any two variables entered into the same equation was .491, it was decided that multicollinearity was not a major problem in this study. Regression Analyses. The first multiple regression analysis included AGE2, education, number of children, wife's income level, husband's income level, wife-husband income ratio, perceived income adequacy and perceived egalitarian family style (Table 18). AGE2 is a created variable that allows for the prediction of a curvilinear 42 ooo.H com. owe. 0mm. mmm.- c-.- oNN.1 Hco.i omH.u smo.i emo.u mHN.1 ammmmme ooo.~ mqfi. mac. mew. m¢N.1 mmo.u ocm. moH. m¢~.- fiao. wom.i mmmmoa ooo.fl mmm. moo.n mwN.i mm~.1 nw~.i mm~.u oeo.i mnmci emu. hmzhmzhmzkmz acmucmamncH com muzmwuNmmmou copaopmcgou comcmmaii.NH o—an 43 relationship between age and the dependent variable. A second multi- ple regression-analysis was performed using AGE in plae 0f AGE2. Because of the limited range of ages of the sample, it was not known if the predicted curvilinear relationship would act as a better predic- tor than a linear relationship between the independent variable age and the dependent variable (Table 19). As can be seen from comparing the two analyses, AGE2 did not enter the equation because the level of f_ to enter was not met. In the second analysis, AGE did enter the equa- tion, although it is not a major predictor variable. Because AGE did appear to be a slightly better predictor of the dependent variable than AGE2, it was used in place of AGE2 in further regression analyses. The third regression analysis was done in order to determine if ideal family style (IDFAMSTY) would be a better predictor of perceived contribution to support than perceived egalitarian family style (FAMSTYLE). This variable was a better predictor as it entered the equation in Step 3 (Table 20). By comparison, perceived egalitarian family style entered the second regression analysis in Step 5, and explained very little of the variance in group membership (Table 19). Table 21 is the summary table for the fourth multiple regression analysis. In this instance, the discrepancy score computed by sub- _tracting FAMSTYLE from IDFAMSTY was used in the analysis. This discrepancy score (DISFAMST) entered the equation in the second step ' and proved to be a better predictor than either FAMSTYLE or IDFAMSTY. The final regression analysis replaced IDFAMSTY and DISFAMST with IDRESP, ideal responsibility style, and DISRESP, the discrepancy score between ideal and perceived responsibility style. Because perceived and ideal egalitarian style included the statements representing 44 perceived and ideal responsibility style, it was decided that the analysis would not contain a combination of both sets of variables. DISRESP entered the analysis in Step 1 and proved to the major predic- tor of the variance in the dependent variable (Table 22). The hypotheses proposed were tested using the preceding regres- sion analyses. The significance of each variable was determined using the §_test. HO There is no relationship between egalitarian family style and employed wives' perceptions of their contribution to their husbands' financial support. This hypothesis was not rejected in any of the regression analyses. When the variable was entered as FAMSTYLE (perceived egali- tarian family style) it was the fifth variable to enter the equation (Table 19), and added little to the predictability of the dependent variable as the change in R? was only .004. When ideal egalitarian family style (IDFAMSTY) was used, it entered the equation in Step 3 of the analysis and explained a larger portion of variance in group mem- bership with a change in B? of .019 (Table 20). In multiple regression analysis #5, egalitarian family style was measured by ideal responsi- bility style (IDRESP). In this analysis, it was the last variable to enter the equation (Table 22). Although ideal egalitarian family style was a better predictor than perceived family style or ideal responsi- bility style, it was not significant at the .05 level, and this variable was not considered to be a major predictor of the dependent variable. HO2 There is no relationship between age and employed wives' perceptions of their contribution to their husbands' financial support. Two hypotheses were tested using age as an idependent variable. The proposed hypothesis using AGE2 predicting a curvilinear 45 .oowuoocm mop Noyce no: woo Amsoucw we Hm>mp m.c:oom:;vqu: coo mmo< mmpnwwgm> usha AoHooLoa .o HNHNNmzoao oHNHm moo. ooo. oNH. ooN. NHHan oaHLaHHHaoN .o Homo Hoo. ooo. oNH. ooo. ooHHaoooo co Ho>oH .o Hoszo HosoH ooo. oNo. oHH. NoN.H oeoooH m.oNH3 .m HoHHzoo ooo. omo. Noo. moo.H ooaoHHoo No Loosoz .N Hoszzv oHHaL ooo. ooo. ooo. NoH.m oooooH ooaomoo-ocHz .H cogcm ocoucmum mmcmgu mm mm cmucmvou a. Humcmucm opoowco> omum on och>ocm usoaoom .mmpoowco> pouwmopogoxmoioouom oco UNEoooumiopuom umuumpmm om>Nwocmo mo :oNuuNomgo mop cow Ho mwmx—oo< oopmmmcmmm m—opupoz mo xgosszmii.mH «Home 46 .cowpoocm mop gmucm poo oNo Amsoucw ac Hm>mp m.ocoom=gv qu: mpoowgo> mop H Hooov moo. Hoo. oNH. NNo. oNHz No moo .N AoHooaoa .o HNHNNmzoaV oHNHm moo. ooo. oNH. ooN. NHHsaN oaHLaHHHaoN .o Homo Hoo. ooo. oNH. ooo. ooHHaoooo co Ho>oH .a . . Hoszv HosoH ooo. NNo. oHH. NoN.H osoooH m.oNH3 .o HoHHzov ooo. ooo. Noo. moo.H oaLoHHoo No Loaooz .N Hoszzv oHHaN ooo. ooo. ooo. NoH.o oeoooH ooaoaoo-ocHz .H Noccm ncoocoum mmcogu mm mm cmuzm o» o Homgmucm mpoopco> amum .Nmm< sow omuouwumoom uw< coo: mwpoowso> NouwmoHoguxmoioNuom coo oweocoumiopuom umuumpmm Ho omoo>ogo «Nooozm vm>wmucmo ac copuuNnmNo ms» com No mpma—oc< copmmmgmmm oHoNu—az No zgosszmiu.mH «poo» 47 .cowuozcm mg“ cmpcm Ho: own o mst Auznzv pm>mp mmo. Hoo. NmH. omo. msoucw m.ocoom:: .o Homo wNo. «Ho. omH. wa. :owuoooum wo Hm>mH .m Aquzv Hm>m_ NNo. wmo. oNH. moo.H wsoucw momwwz .e HNNmzoaoHo «Noam HHwEow Hmo. omo. NeH. cem.m cowgmuwwomm HomuH .m AoHquv owe. «mo. moo. mmm.H omgopwgu wo cmossz .m Hoszzv owoac «mo. mmo. mmo. NoH.m msoucw noonmzcumwwz .H Noccm ocoogoum mucosa mm Logo“ on a Hcmgmucm mpoowco> omam .mpxum »p_5om cowcouwpomm om>wmucmo cow m—xum xpwsow cowcoawpomm HomoH mcwuouwumozm mmpoowco> HouwmoHoguxmoiowuom woo uwsoooumiowuom umpumwmm an omow>oco ucoooom um>wmucmm wo cowpuwomco on» sow mw mwmzpo=< oowmmmcmmm mpowupoz wo agmssomui.om m—oow 48 oNo. Hoo. moN. HNo. Hoosz Noaooooo osoooH .o oNo. Noo. NmN. ooo. Hoovoowoaoooo wo Ho>aH .o Hoszv Ho>oH ooo. Noo. ooN. ooH. osooow m_oooomo= .N moo. oHo. ooN. ooo. Hmoov owwz wo oo< .o HNHmzoaoHo mpxgm waEow moo. NNo. oNN. ooo.H ooHaoHHHooo HoooH .o Hoszv Ho>oH ooo. moo. HHN. NNN.N oeoooH a.ow_2 .o HoHHooo oNo. moo. ooH. moo.N ooLoHHoo wo consoz .m szEmHov 2.3m xpwsow cowcouwpomm um>wmusmo woo Pomuw Hoo. moo. MNH. oHN.o ooozooo Noooooaomwo .N Hoszzv oHHoL ooo. ooo. ooo. NoH.o oeooow ooaomoo-osz .H coccu oconaoum mmcosu mm gmucu on u umswuom opoowgo> awum ' I [II '1 .mprm xpwsom cowcouwpmmm cm>wmucmo.oco HomoH :mmzumo Noncomgumwa mcwoapucH mmwoowgo> HouwmopouXmoiowuom coo uwsocoumiowuom umpumpmm on ooow>ogo pcoaoom om>wmocmo wo cowpuwomgo as» cow oo mwmapo=< :owmmmcmmm mpowapoz wo acoesomii.HN epoch 49 relationship was tested in the first regression analysis, and the variable did not enter the equation using the default level of f_t0 enter (Table 18). In the second regression analysis, the hypothesis predicting a postive. linear relationship between AGE and the dependent variable was tested. In this analysis, AGE was the last variable to enter the equati.n with an f_to enter of .028 and a change of R? of .001 (Table 19). In testing both hypotheses it was found that the age of employed wives was not an important variable in predicting the variance in the dependent variable. In the final multiple regression, analysis, AGE was the fifth variable entered in to the equation and accounted for a change in R? of .007 (Table 22). HO There is no relationship between level of education and employed wives' perceptions of their contribution to their husbands' financial support. It was not possible to reject this hypothesis in any of the analyses. In Table 20, education entered the analysis in Step 5 with a change in R? of .014. This was the largest change in R2 accounted for by level of education. H0 There is no relationship between number of children and employed wives' perceptions of_their contribution of their husbands' financial support. The proposed hypothesis predicted a negative linear relationship between number of children and perceived support provided. In fact, the analyses showed a postive linear relationship between these two variables (Table 23). Although number of children was not a signifi- cant predictor at the .05 level and the hypothesis could not be rejected, it does appear to be a major predictor of the dependent variable as it is included in the best prediction equation in each of 50 Ho.vmaa mo .Vmco HamoooHo oHNHa NNo. Hoo. HNN. oNo. NHHHHonooomoo HoooH .o ooo. Noo. oNN. omH. Homvoowaooooo wo Ho>oH .o ooo. Noo. ooN. HoH. HoozHo Hoaooooa osoooH .N HozHov Ho>oH ooo. ooo. ooN. Noo. ooooow m.oooomo= .o ooo. Noo. ooN. ooo. Hoooo asz wa ooo .o Hoszo Ho>oH Noo. oNo. NoN. ooo.H osoooH m.awH3 .o HoHHzov Noo. Noo. oNN. moo.N ooNoHHoo wo Loosoz .N «Hoszzv oHHaN ooo. ooo. oNH. ooo.o oeoooH oooomoo-owH= .N waHamNomHoo mpxpm xawpwowmcoommc um>wmugmo oco Homow ooo. oNo. oNo. ooo.o ooozooo HooaaooomHo .H cogs“ ococcoam mmcosu mm gmpcu ow w cwcmucm mpoowco> awpm . .mpzum zpwpwowmcoommm um>wmucmo uco pomcH :mmZHmn aucoomcumwo coo mpxum xpwpwowmcoomma HomoH mcwuzpuoH mmpoowgo> Houwmoposuzmoiowuom uco uwsocoumiowuom omuumpmm on omow>oco ucoaoam ow>woucwo wo cowuowomco ms“ cow mo mwmxpoc< cowmmwgmmm mpowuooz wo x:a§5=m--.~m m_now 51 Ho.vhmxa mo.V.m x NmH. mNH. mom.H oHo. oac. quz omm. coo. oHN.m coo. mHH. oHqu oom. ooo. NN~.o ooH. mmm. «ozHIZ mam. moo. NHw.m Hmo. moo. aaommmmmo powwowwwmou cowmmmcowm xpwpwooooco a coccm Aoumm umuoewummv powwowwwmou mpoowgo> omeocooooam mcoocoum :owmmmcmmm umuwucoccmumcz .mpxp zuwpwnwmcoommm om>wmucmo ocm HomoH :mmzumo zucoomcomwo woo mezum HHwHwowmcoommm .HomoH wwwoopucw mmpoowco> Pmuwmopocozmoiowuom coo uweocoumuowuom umuumpmm on vouoowgucou ugooaom om>wmocmo wo cowuowomco mg» cow mwmxpoo< cowmmmgmmm mpowupaz i-.m~ «Home 52 the five analyses. In the fifth analysis, it accounted for a change in 3? of .047 (Table 22). HO5 There is no relationship between perceived income adequacy and employed wives' perceptions of their contribution to their husbands' financial support. This hypothesis was not rejected nor was the variable an impor- tant predictor of perceived contribution to support. It never appeared in the best predictor equation and the largest change in 2 was .003. This may be due to the fact that only one respondent viewed family income as not meeting family needs. HO6 There is no relationship between income level of employed wives and their perceptions of the1r c0ntr1but10n to the1r husbands' financial support. A positive linear relationship was predicted between level of wives' income and perceived contribution to support. Although the level of wives' income was not a significant predictor of the dependent variable and the hypothesis could not be rejected, it does appear in the best prediction equation in each analysis and accounted for up to 4.5% of the variance in group membership (Table 21). Level of wives' income appears to be a major variable in predicting the dependent variable. H0 There is no relationship between level of husbands' income and employed wives' perceptions of their contribution to their husbands' financial support. The proposed hypothesis predicted a negative linear relationship between level of husbands' income and perceived contribution to support. This hypothesis was not rejected and the level of husbands' income was not included in any of the best prediction equations. HO8 There is no relationship between wife-husband income ratio and employed wives' perceptions of their contribution to their husbands' financial support. 53 The ratio (If wife's income to husband's income was found to be a major predictor of perceived contribution to support. In all but the final regression analysis, this variable entered on Step 1 of the analysis and explained 5.8% of the variance in group membership (Tables 18 through 21). In the fifth multiple regression analysis, wife-husband income ratio entered on Step 2 with a change in R? of .098 and a significance level of .018. This hypothesis was rejected. H09 There is no relationship between the degree of discrepancy in ideal and perceived egalitarian family style and employed wives' perceptions of their contribution to their husbands' financial support. The summary table of regression analysis in Table 21 reveals that the discrepancy between ideal and perceived egalitarian family style is an important predictor variable and enters the equation at Step 2. This variable explains 6.5% of the variance in group member- ship. Although the variable is not significant, its entry into the" equation before ideal egalitarian family style is important to note. In Table 22, ideal responsibility style and the discrepancy between ideal and perceived responsibility style are substituted for the corresponding family style variables. This equation was the best prediction equation of all the analyses, with an R? of .25 at the point of lowest standard error of estimate. DISRESP accounted for 7.8% of the variance in group membership and was significant at the .01 level which supports the rejection of this hypothesis. CHAPTER IV This final chapter includes the limitations of the research, a summary, a discussion of the findings and implications for further research. LIMITATIONS The major limitation of this study is related directly to survey methodology. One of the problems was in determining the meaning of the data for the dependent variable. Although an attempt was made to drop cases that reported percent of emotional support, there was no way to determine if any of those responding they provide 0% of support did so because they feel both spouses are self-supporting. Data gathered by survey research are always subject to the interpretation of the researcher. An additional limitation is related to the use of ten statements from FACES II as a measure of egalitarian family style. 'Although this use of FACES II has face validity, it has not been tested for other types of validity or reliability. 'Another limitation is small sample size which limits generalizability. ms}. The first objective of this study, that of determining if respon- dents can be placed on a continuum of role minimization to role magni- fication, was met. The difference between the perceived contribution to husbands' support and calculated percent of family income earned by the wives ranged from -70% to +40%. It appears that most wives minimize 54 55 their role as provider (50.9%) and few magnify that role (9.5%), while 39.6% view their role as provider as approximately equal to the percent of family income they earn. The second objective of this research was to identify the major variables that predict and explain employed wives' perceptions of their contribution to their husbands' support. Since the dependent variable was not normally distributed, it was necessary to dichotomize the variable. Women Who reported they provide a specific percent of sup- port were placed in one group, while those reporting 0% of support were placed in a second group. The objective became that of identi- fying the major variables that predict and explain group membership. Through a series of five multiple regression analyses, f0ur major variables were identified as major predictors of group membership: wife-husband income ratio, level of wife's income, number of children and discrepancy between ideal and perceived responsibility style. The importance of the variables wife-husband income ratio and wife's level of income supports the previous research of Scanzoni (1978), Peterson and Maynard (1981) and Hood (1983). A positive relationship between number of children and wives' perceptions of contribution to their husbands' financial support was not expected. A negative relationship was predicted as it was felt that the more children, the m0re likely wives would be to identify with the homemakerrole. In light of the fact that discrepancy between ideal and perceived responsibility style was the major predictor variable, the positive relationship seems logical. The more children there are in a family, the more reponsibil- ities there are to be shifted and shared and the greater the chance for discrepancy between the ideal and perceived responsibility style. 56 DISCUSSION OF FINDINGS) Although four major predictor variables were identified, they accounted for only 25% of the variance in group membership. This is not too surprising since the independent variables were selected to reflect egalitarian gender role attitUdes, which was not a major variable in predicting perceived contribution to support. The propo- sition on which this research was based, that of a relationship between egalitarian role attitudes and perceived contribution to husbands' support, proved to be inaccurate. It appears that wives' perceptions are based on their view of their husbands' role performance. This too is supported by role theory. Burr, Leigh, Day and Constantine (1979) have set forth several propositions that relate to the dis- crepancy or dissatisfaction in responsibility style. Proposition 1: The perceived quality of role enactment in a relationship influences the satisfaction individuals in the relationship have, and this is a positive linear relationship. Proposition 1a: The quality of alter's role enactment influences ego's satisfaction, and this is a positivelinear relationship. Proposition 3: The greater the relative deprivation of one's situation as a whole, the less one's satisfaction with the situation. These three propositions can be applied to the employment of wives and the discrepancy between ideal and perceived responsibility style. In this instance, wiv‘es experience role strain from their roles as providersand homemakers. Since it is not possible to alter their roles outside the home, they may feel their husbands should take on more of the family responsibilities. If the husbands are not willing to do so, and research findings generally support that men with employed wives do very little more of household work than do men with wives who are 57 full-time homemakers (Moore and Hofferth, 1979; Pleck, 1981), then wives will feel husbands are not performing their roles adequately. This situation may arise when wives become aware that they are working outside the home and still retaining their duties for household work, while their husbands' roles remain that of provider. It seems likely that wives may feel relatively deprived when comparing their leisure time to that of their husbands. These propositions explain the dis- crepancy, or dissatisfaction, between the ideal and perceived responsi- bility style. The fact that this discrepancy can be used to predict whether or not employed wives' view themselves as providing financial support for their husbands' makes an additional proposition necessary. It seems that the cognizance of the amount of time needed to perform their homemaker roles in comparison '03 the amount of time spent by husbands in performing household tasks, makes wives more aware of their role as provider. Wives may question why their husbands have more leisure time, and determine it is because husbands are performing one role, that of provider, while wives are performing both homemaker and provi- der roles. The general proposition can be stated as The degree of relative deprivation, or dissatisfaction, influences ego's perception of her own role. The specific hypothesis fer this study is The greater the discrepancy between wives ideal and perceived responsibility style, the more likely they are to identify with the provider role. If the independent variables in this study were selected to reflect the above hypothesis, it is pdssible that a greater amount of the variance in group membership could be explained.' 58 The significance of this study is in the additional information it provides in understanding dual-earner couples. This has been accomplished by identifying the relationship between the discrepancy between ideal and perceived responsibility style and wives' perceptions of their contribution to their husbands' support. The next question that must be addressed is "What are the implications of these findings?" The answer depends on how these wives view the dissatisfaction in their relationship. Burr, et al, (1979) state that the more important the role expectation, the greater the effect on satisfaction. It may be that the dissatisfaction is not viewed as important to the mari- tal relationship and will have little effect on the relationship. However, if employed wives view the expectation for the husbands to shift and share household responsibilities as important, it is possible that this discrepancy will effect the marital relationship as a whole. If the marital relationship is less satisfactory due to the discrepancy between ideal and perceived shifting and sharing of respon- sibilities, the way in which dual-earner couples resolve this dissatis- faction is important. First of all, it is possible that wives will come to expect less of their husbands in sharing responsibilities and the dissatisfaction will decrease because the expectations are brought more in line with reality. A change in husbands' behaviornas a result of negative sanctions imposed by wives also will result in a decrease in dissatisfaction. The third possibility is that dissatis- faction will increase over time. If dissatisfaction with sharing of responsibilities increases and wives are well aware of their ability to provide TIDPDClally for themselves and other members of their 59 families, there are negative implications for these marital relation- ships. Unless there are other highly vauled positive factors in the relationships, the increase in dissatisfaction may lead to a decrease in marital stability. IMPLICATIONS FOR FUTURE RESEARCH Additional research is needed to clarify the above implications and the data from NC 164 allow for further analysis of this research problem. The NC 164 project is a longitudinal project with two surveys spaced two years apart. The following research questions could be answered using the data from the regoinal project. 1. Can the finding of this study be replicated using data from the additional eight states involved in NC 164? Does the discrepancy between ideal and perceived responsibil- ity style change over time? If so, does it decrease due to a change in husbands' behaviors or wives' expectations? Does it increase? Are any changes in discrepancy over time related to changes in marital satisfaction as a whole? Are there any wives who reported they provide 0% of support to their husbands in the first survey who report they provide a specific percent of support in the second survey? If so, is this related to a change in discrepancy between ideal and perceived responsibility style? Additional independent variables could be selected that reflect satisfaction with household responsibilities and marital satisfaction to see if a greater amount of variance in the dependent variable can be explained. An additional discrepancy score could be computed to compare husbands' ideal responsibility score with wives' to determine if that allows for better prediction of the variance in perceived con- tribution to husbands' support. The discrepancy between ideal and perceived responsibility style also could be treated as a dependent 6O variable to see what factors explain variation in this phenomenon. The findings of this research have raised many questions that need to be answered. 00 women first get jobs, then become dissatisfied with the way in which household responsibilities are shared,then become aware of their role as providers, then view divorce as a way to end their dissatisfaction? If so, a causal relationship between wives entering the work force and the rising divorce rate could be estab- lished. If such a causal relationship exists, it is important that it be identified through further research, so that family professionals and families can be aware of the risks and rewards of dual-earner marriage. APPENDIX A PORTIONS OF THE QUESTIONNAIRE USED IN THIS STUDY 61 MICHIGAN STATE UNIVERSITY COLLEGI OF HUMAN ECOLOGY EAST LAASIING ' MICHIGAN ' £8024 WAITMINT Of FAMILY AND CHILD ECOLCXEY 4 February I983 The strength of America's families will mold its future. We at Michigan State University have studied our state's families closely, but, as you know, society is changing quickly. Teachers. counselors, and lawmakers ' need to learn much more--especially about the way families cope with modern-day stresses, in order to be of help to families. Your family is one of only a few families in the "middle years" being included in our study of family stress in Michigan and eight other states. Therefore, it is very important that your opinions be known to us. We want you to answer the questions in the enclosed booklets. One booklet is labeled for the wife, and one for the husband. It is very important that husband and wife both answer and that they complete the booklets without talking to each other. He have included separate postage (prepaid envelopes) for each of you to return the booklets. - Your answers are private. The booklet has an identification number for mailing purposes only. This is so we can check your name off the mailing list when you return it. Your name will never be placed on it. You can receive a summary of the results by marking the box "if you would like a copy of the results of this study" at the end of the booklet. He would be most happy to answer any questions you might have. Please write or call. The telephone number is 517-353-5248. Remember, there are no right or wrong answers. The only right answers are the ones which truly describe your unique situation. Thank you for your assistance. Sincerely, Dr. Margaret M. Bubolz Dr. Dolores Borland Dr. Vicki Schram Professor Assistant Professor Assistant Professor MMszus enc 1 as u res "5L. ‘3 I. .Wmflv -“"’I- Eat-all Oppnfl-Iulv ("lulu-m 62 1 We would like to have some background information about your family to help us in our study. Please fill in the following information about each member of your household, identifying each person by their relation- ship to you. , First, think about yourself. 01 Sex: M F Mo. a Yr. of Birth Yrs. of School Completed (circle one) Marital Status If Married, Mo. 5 Yr.of Marriage -Mext, think about each of your children, starting with the oldest child. He will be asking questions about each of your children later in this questionnaire. Please make sure your answers are from oldest to youngest in each instance. Birth ‘Yrs. of Living at (If N0) Date 1 0f Q2 Sex DMgerr. 231$, rated '17:?” Mo fizg‘soggfor fifth. lgspggvide a. Child 1 H F 6. Child 2 M F c. Child 3 M F a. Child 4 M F o. Child 5 M F t. Child 6 M F (add on if necessary) Finally, think about each other member of your household. Yrs. of x of Relationship Sex Birth Date School Marital Support Q3 to You M or F Mo. Yr. Completed Status You Provide Q13 63 11 Please read each statement and circle the response number which best describes your family for each item. Family members are support- ive of each other during difficult times In our family, it is easy for everyone to express his/her opinion . It is easier to discuss problems with people outside the family than with other family members . Each family member has input in major family decisions . Our family gathers together in the same room . Children have a say in their discipline . Our family does things together Family members discuss problems and feel good about the solutions In our family, everyone goes his/her own way How would you How would you describe like your family your family now? to be? 0 0 N N x a x ‘0 a;* a aS a» * -% 1* + '5 A? 4? " ‘3 27 Z? 4? ‘3 ”'5? .. «Ln-Sago?!» ~~S~§5~ ‘0 “#9, ‘0 “#9! ‘éi o Io-é? A? «o i? g? ~§§8~ ~§§s~ it cs 0: k. V? ‘t :5 o, 9- it H N w a U1 H N w .p U'l l 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 S 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 l 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 This Question Continues On The Next Page *j' *n. q. We shift household responsibilities from person to person Family members know each other's close friends . It is hard to know what the rules are in our family Family meebers consult . other family members on their decisions Family meters say what they want . He have difficulty think- ing of things to do as a family . In solving problems, the children's sugges- tions are followed Family memers feel very close to each other . Discipline is fair in our family Family members feel closer to people outside the family than to other family members '64- p... to How would you describe your family 52!? H N U How would you like your family to be? 65 13 How would you describe your family Egg? 3" .. if ,9 ge”c: a$§§a 0 O- §§§$§ c. Our family tries new ways of dealing with problems 1 2 3 4 5 u. Family members go along with what the family decides to do 1 2 3 4 5 *4. In our family, everyone shares responsibilities 1 2 3 4 5 V. Family members like to spend their free time with each other 1 2 3 4 5 x. It is difficult to get a rule changed in our family 1 2 3 4 5 ‘ y. Family members avoid each other at home 1 2 3 '4 S *2. Hhen problems arise, we compromise 1 2 3 4 5 aa. He approve of each other's friends 1 2 3 4 5 ’56. Family members are afraid to say what is on their minds 1 2 3 4 5 cc. Family members pair up rather than do things as a total family 1 2 3 4 5 dd. Family members share interests and hobbies with each other I 2 3 4 5 How would you like your family to be? 1 2 3 4 5 Q23 66 Next, think about ur rsonal income before taxes for 1982. Be sure to include all sources of income that you receive personally; such as earned income, investments, social security, your own business, job-related benefits, welfare benefits, rent and so on. If you farm or have your own business, we would like you to indicate your net farm or net business income before taxes. 3 I (nearest $1,000) 024 Now, think about your total famil income for 1982. This is total income before taxes for all memBErs of your family, including yourself and ygur children. Be sure to include all sources ofTincome; suc as earne ncome, investments, social security, your own business, job-related benefits, wel- fare benefits, and so on. If your family farms or has its own business, indicate net farm or net business income before taxes. S (nearest $1,000) 026 To what extent do you think your income today is enough for you to live on? a. can't buy some necessities b. can meet necessities only c. can afford some of the things we want but not all we want d. can afford about everything we want e. can afford about everything we want and have some left over IAPPENDIX B CALCULATION OF RESPONSE RATE 67 CALCULATING RESPONSE RATE NC 164 MICHIGAN Sampling Frame Number of households on "A" list 1903 Number of households on not on "B" list 8728 Total number of households on original 10631 list received from Donnelly Sample Number of households in random sample 149 drawn from "A" list Number of households in random sample 751 not drawn from "A" list Total number of households in random 900 sample Response Rate - "A" List Total number of households that received 139 questionnaires and met criteria Households responding with refusals 2 Households not responding 78 Households completing questionnaire 59 139 Households completing questionnaire Reigogse = Total number of households that = 42.4% received questionnaire & met criteria 68 Response Rate - Households not on "A" List Total number of households that received 642 questionnaires and met criteria Households responding with refusals 28 Households not responding 396 Households completing questionnaire ‘218 642 Households completing questionnaire = Total number of households that received questionnaire & met criteria Response rate Response Rate - Total Households Total number of households that received 781 questionnaires and met criteria Households responding with refusals 3O Households not responding 474 Households completing questionnaire 277 Reggggse = ifiiiihiliiefimé‘i‘EEESEhSiSEtIfiSEa'“ = 35.97. received questionnaire and met criteria APPENDIX C ADDITIONAL FREQUENCIES 69 Table 24--Frequency Distribution of Husband's Income Level (n=53) Husband's income level Frequency Percentage Less than $10,000 3 5.7 $10,000 - 19,999 15 28.3 $20,000 - 29,999 22 41.5 $30,000 - 39,999 9 17.0 $40,000 - 49.999 1 1.9 $50,000 - 59,999 3 5.7 Total 53 100.0 70 Table 25 --Frequency Distribution of Wife-Husband Income Ratio (n=53) Wife-husband Frequency Percentage income ratio Less than 10% 2 3.8 10 - 19% 4 7.5 20 - 29% 7 13.2 30 - 39% ' 11 20.7 40 - 49% 7 13-2 50 - 59% 6 11.3 60 - 69% 6 11.3 70 - 79% 1 1.9 80 - 89% 2 3.8 90 - 99% 2 3.8 100% and above 5 9.4 Total 53 100.0 BIBLIOGRAPHY Biddle, Bruce J. Role Theory: Expectations, Identities, and Behaviors New York, Academic Press, Inc. 1979. Bird, Caroline. Two Paycheck Marriage. New York, Rawson, Wade Publishers, Inc. 1979. Bubolz, Margaret M., Eicher, Joanne B., and Sontag, M. Suzanne. "The human ecosystem: A model," Journal of Home Economics. Spring, 1979, 28-31. Burr, Wesley R., Leigh, Geoffrey K., Day Randall 0., and Constantine, John. "Symbolic Interaction and the Family," in Wesley R. Burr, Reuben Hill, F. Ivan Nye and Ira L. Reiss, eds., Contemporary Theories About the Family. Volume 2, New York, The Free Press, 1979. Cherlin, Andrew and Walter, Pamela Barnhouse, "Trends in United States men's and women's sex role attitudes: 1972-1978," American Sociologi- cal Review. 46, August 1981, 453-460. Dilman, Don A. Mail and Telephone Surveys: The Total Design Method. New York, John Wiley and Sons, 1978. _ Duncan, Otis Dudley, "Recent cohorts lead rejection of sex typing," Sex Roles. 8, 2, 1982, 127-133. Hood, Jane. Becoming a Two Job Family. New York, Praeger Publishers, 1983. Huber, Joan and Spitze, Glenna. "Wives' employment, household behaviors and sex-role attitudes," Social Forces. 60, 1, 1981, 150-169. Johnson, Beverly L. "Marital and family characteristics of the labor force, March 1979,“ Monthly Labor Review. April 1980, p.48. Keith P. and Schaffer, R. "Role strain and depression in two-job families," Family Relations. 29, October 1980, 483-488. Kerlinger, Fred N. Foundations of Behavioral Research. New York, Holt, Rinehart and Winston, 1973. 71 72 Kim, Jae-On, and Kohout, Frank J. "Multiple Regression Analysis," Statistical Package for the Social Sciences, Second Ediction, New York, McGraw-Hill, 1975. Lein, Laura.' "Male participation in home life: Impact of social supports and breadwinner responsibility on the allocation of tasks," The Family Coordinator. October 1979, 489-495. Lowenthal, Marjorie Fiske, and Chiriboga, David. "Transition to the empty nest," Archives (Hi General Psychiatry, 26, 1, 1972, 8-15. “Mason, Karen Oppenheim, Czajka, John L. and Arber, Sara. "Change in U.S. women's sex-role attitudes, 1964-1974," American Sociological Review. 41, August 1976, 573-596. Moore, Kristin A. and Hofferth, Sandra L. "Women and their Children," in Ralph E. Smith, ed., The Subtle Revolution: Women at Work. pp. 125-157. Washington, D.C., Urban Institute, 1979. Olsen, David H., Portner, Joyce and Bell, Richard. "FACES II: Family Adaptability and Cohesion Evaluation Scales," unpublished working paper, Department of Family Social Science, University of Minnesota, 1983. Oppenheimer, Valerie Kincade. "The sociology of women's economic role in the family," American Sociological Review. 46, August 1-81, 453-460. Peterson, Larry R. and Maynard, Jody L. "Income, equity and wives' housekeeping role expectations," Pacific Sociological Review. 24, 4, 1981, 87-105. Pleck, Joseph H. "The work-family problem: Overloading the system," in Barbara L. Forisha and Barbara H. Goldman, eds., Outsiders on the Inside: Women and Organizations, pp.239-254, Englewood Cliffs, N.J. Prentice-Hall, 1981. Reimer, Eleanor S. and Fout, John C. European Women: A Documentary History, 1789-1945. New York, Schocken Books, 1980. Scanzoni, John. Sex Roles, Women's Work and Marital Conflict. Lexington, Mass., Lexington Books, 1978. Scanzoni, John. Sex Roles, Life Styles, and Childbearing. New York, The Free Press, 1975. Scott, Joan W. and Tilly, Louise. "Women's work and the family in nineteenth century Europe," Comparative Studies in SOCiety and History. 17, 1, 1975, 36-64. Scott, Wilbur J. and Morgan, Carolyn Stout. "An analysis of factors affecting traditional family expectations and perceptions of ideal fertility," Sex Roles. 9, 8, 1983, 901-914. 73 Smith, Dwayne. "The relationship of family type to perceived family satisfaction and family health: A study of family stress," unpublished doctoral dissertation, Michigan State University, 1985. Spitze, Glenna and Huber, Joan. “Changing attitudes toward women's nonfamily roles," Sociology of Work and Occupations. 7, 3, 1980, 317-335. Thomas, Edwin J. and Biddle, Bruce J. "The nature and history of role theory," in Role Theory: Concepts and Research. New York, John Wiley and Sons, Inc. 1966. U.S. Bureau of Labor Statistics. U.S. Working Women: A Data Book. Washington, D.C., Government Printing Office, 1977. U.S. Bureau of Labor Statistics. Women at Work: A Chartbook. Washington, D.C., Government Printing Office, 1983. U.S. Department of Commerce, "Earnings in 1981 of married couple families by selected characteristics of husbands and wives," Current Population Reports. Series P-23, 133, Government Printing Office, March 1984.