Introduction

Spanish society is experiencing a process of transformation, which has brought about profound social changes over a short period of time (Dema-Moreno 2009). One of the most important of them is the increased participation of Spanish women in paid work. As in the majority of developed countries, this increased participation has not been balanced by a corresponding decrease in the time spent on household work (Álvarez and Miles 2006). The incompatibility of paid work and family life, from which the need to find a balance arises, is caused by a conflict of interests or demands between the workplace and the home; such incompatibility is especially acute where the conflict is to be resolved by the individual and couples involved in work (Papí 2005). The purpose of this paper is to explore the explanatory potential of three theoretical approaches—the traditional gender division, role-strain, and the resource-bargaining approach—so as to account for the emergence of different patterns in the division of unpaid work among Spanish dual earner couples.

The study provides a different cultural perspective from other studies that have focused on countries such as United States (e.g. Bianchi et al. 2000; Biernat and Wortman 1991; Mannino and Deutsch 2007), Mexico and United States (e.g. Pinto and Coltrane 2009) or Denmark (e.g. Wiesman et al. 2008) and from Spanish studies based in older samples of workers (Álvarez and Miles 2003, 2006; Balcells i Ventura 2009). In addition to previous literature, the article also aims to disentangle the mechanisms through which dual-earner couples share housework by combining three largely unconnected perspectives on participation in household activities. In light of this objective, the discussion draws on a representative sample of 2,877 workers in Spain, who are in stable relationships and where the other partner is also engaged in work outside the home. Since the dependent variable is ordered throughout the analysis ordered logit models are estimated.

The Spanish Context

Spain is moving rapidly towards advanced economic and social-democratic development; at the same time, however, it remains rooted in a cultural tradition that emphasizes different roles for women and men (Sánchez and Hall 1999). The Franco regime in Spain (1939–1975) defined a very specific model of womanhood, centring on the role of mother and subordinated wife, whose natural space was the home (Nuño-Gómez 2008). Throughout this period, the law established differences between men and women that influenced the latter’s ability to participate in politics and their participation in economic, social and labour life (Durán 2003). The democratic Constitution of 1978 established equality of rights between men and women, but only in the last two decades in particular has Spain experienced the type of social and political changes that make it a society similar to those in other European Union countries (Dema-Moreno 2009).

This historical context has a clear bearing on the incorporation of women into the labour market. The employment rate for Spanish women aged between 16 and 64 rose from 22.9% in 1964 to 31.5% in 1994, and to 53% in 2009 (Ministerio de Trabajo e Inmigración 2009). In a study based on 13,567 interviews conducted in twelve countries (1,781 in Spain), Sevilla-Sanz (2010) indicates that Spain occupies the eighth position in the classification of equality between men and women in developed countries, behind Sweden, Norway, the United Kingdom, the United States, the Netherlands and Ireland, and ahead of New Zealand, Japan, Germany, Austria and Australia. Moreover, Spain is the country with the most imbalanced model in Europe in terms of use of time (Nuño-Gómez 2008). As a consequence, Spanish men have the highest load of paid work, whereas women have the lowest. On the other hand, women devote three and a half hours more per day than men to domestic work (INE 2004).

To sum up, it may be said that Spanish society has undergone significant changes in recent years. Spain is regarded as a country with a masculine culture in which differences in gender roles are heightened (Silván-Ferrero and Bustillos López 2007). Although the dual earner family model is growing in importance (Alberdi 2003), the division of household work in these couples resembles the traditional family model to a great extent. In Spanish dual earner couples, 34% of men do not do any kind of housework and 65% share such work with their partners, as shown in the data from the Spanish Quality of Work Life Survey used in this paper. Therefore, the causes of such inequality would appear to merit some analysis.

The Division of Household Labor in Dual Earner Couples

In dual-earner couples where both partners are in paid work, household labor division is effected in line with different patterns. Gershuny et al. (1994) identified two models for the division of unpaid work in such families. In the dependent model, the man remains the main breadwinner, and the proportion of household activity that falls to the woman is not significantly affected by her involvement in paid work. In the adaptive partnership both women and men contribute to the domestic economy through paid work, and they share the responsibilities of unpaid work, albeit often with varying degrees of commitment. This model is similar to the “total integration of roles” model of the family developed by Parsons and Bals in 1955. Couples that adopt the dependent model are likely to suffer higher levels of stress due to an unequal distribution of activities than those who choose the adaptive partnership strategy around the world, like in the United States (Claffey and Mickelson 2009), and in Canada (Higgins et al. 1992; MacDonald et al. 2005). At the same time, adaptive partnership may not lead to a reduction in the total stress borne by the couple, but such stress—like the activity itself—may be distributed between the two partners in a more equitable way.

Research on the division of household labor has given rise to a number of theoretical perspectives to explain the allocation of family work. The most prominent are relative resources, socialization-gender role attitudes, and time availability-constraints (Pinto and Coltrane 2009). In the case of dual earner couples, time availability constraints emerge mainly from the role-strain associated with the fact that both partners have to fulfill the duties of paid and unpaid work.

Traditional Gender Division

One explanation for the strategy adopted with regard to the division of unpaid work within the family may be derived from the impact and influence that tradition may have on all aspects of social life. For many years, the traditional model of the family was the prevailing strategy, where the man was engaged in paid work (male breadwinner) and the woman was responsible for housework and the home (female household) (Knudsen and Waerness 2009). This division of work exists because it was usually assumed that a man’s needs are fulfilled to a great extent in the work role, whereas women’s need-fulfillment comes from the home (Kavanagh and Halpern 1977). The gender ideology reflected in the differences between values attributed to men and women has also been used to underpin this model of the family (Lewin-Epstein et al. 2006; Ross 1987; Stier and Lewin-Epstein 2007).

Families today are making decisions about division of household responsibilities against the backdrop of lifelong socialization in traditional sex-role beliefs (Álvarez and Miles 2003; Biernat and Wortman 1991; Ferber and Birnbaum 1977). The traditional framework of education has trained women to devote a great deal of their time to household responsibilities, whereas men have been educated to participate in the workplace. In this sense, as Alberdi (2003) has showed for a Spanish sample, reconciling work and family life is especially difficult for women, since they have been traditionally assigned responsibility for the care and attention of children. Since working mothers can devote less time to being with their children, they usually feel some distress because they may not be able to fulfil this duty as they would like.

In this regard, workers from different generations are currently employed in the labor market, and the definitions of the role of husband and father have changed over time. Nomaguchi (2009) found that in the 1970s in the United States, husbands had little or no internalized awareness of domestic responsibilities, whereas nowadays men in dual-earner couples have a more refined understanding of their responsibility to participate and cooperate in household work. In light of these observations, demographic features may be said to act as determining factors in the distribution of roles within the couple. Thus, the division of housework is influenced by two factors: age and education.

In Europe, older couples are more inclined to support the traditional model (Hank and Jürges 2007), whereas younger people tend to share a different vision, oriented to shared responsibility for housework. With this in mind and in line with the argument articulated by Knudsen and Waerness (2009), who hold that the traditional model of the family (male breadwinner, female household) has been in decline in recent generations, the hope may be that young couples nowadays are more conscious of the changes in values relating to the division of work within the family that have arisen in recent years. Thus, it may be more likely that younger couples adopt a more equal distribution of activities than older couples and, as a consequence, the older the woman, the greater her involvement in household work; at the same time, however, the older the man, the lower his involvement in household work. In line with this prediction, Voicu et al. (2009) found that in European countries older couples are more likely to support the traditional model for the division of household work. This observation is congruent with the lagged adaptation model, wherein the equitable division of household work between the genders is to take place in the long term, over a period of years and generations (Boje 2007).

Finally, given that education—and, in particular, higher education—leads to more liberal attitudes towards household work, it may be regarded as a further, relevant demographic variable that may be a contributory factor in the more or less equal division of such work between partners (Newcomb 1943). The effect of modern social values through education is reflected in greater support for shared responsibilities (Brines 1994). Hence, it may be said that the higher the educational level obtained by the individuals involved, the less traditional and the more egalitarian they are likely to be (Presser 1994), and thus, the higher the educational level, the greater the participation of men in household work and the lower that of women will be. In line with this argument, previous studies have shown that in the United States the participation of men in housework rises in line with educational level, and a corresponding, positive effect for women is found (Bianchi et al. 2000; Brines 1994; Shelton and John 1996). In Spain, Alberdi (2003) found that not only the couples with a higher educational level tend to prefer a more egalitarian family model, but they are also more likely to put it into practice to a greater extent.

Role-strain

The relationship between paid work and family life is two-way (Frone et al. 1992). On the one hand, work may interfere with family life (WIF); and, on the other hand, family may interfere with workplace activity (FIW). The conflict between family life and paid work may be explained by reference to the psychological theory of a conflict of roles. The conflict exists when one role requires time and specific behaviors and generates such strain that make it difficult to fulfill the requirements of the other role (Greenhaus 1988). Time is viewed as a limited resource, which is to be distributed by the individual (often in competition with others) between the demands of work, home and other spheres of life.

Moreover, in dual earner couples, time is a scarce and highly valued resource that has an impact on the division of household work (Voicu et al. 2009). The number of hours worked is one of the most common indicators of the individual’s work commitments (Gupta 2007; Parkman 2004; Voicu et al. 2009). For this reason, the member of the couple who spends more hours at work over the course of the working week and, as a consequence, has less free time, would spend fewer hours participating in the activities of the household.

Nevertheless, the number of hours worked is not the only significant requirement in a job description. The way in which such time is organized also has an impact on the worker’s availability to engage in unpaid work. Working the same number of hours, an employee on a nine-to-five timetable, for instance, saves time that is lost to an employee working on split-shifts. In the latter situation, the worker must spend more time commuting and on meals, thus leaving even less time for involvement in household work. It is clear, then, that when the time devoted to one activity, such as paid work, compromises the individual’s availability to commit to another activity, such as unpaid work (Bacharach et al. 1991), nine-to-five employees are likely to participate in a more equal division of household responsibilities than split-shift employees, as the latter simply have fewer hours in the day to devote to household work.

The flexibility the worker may avail of in relation to start-time, end-time or breaks is also relevant in this regard. In their analysis of the effect of certain management practices on the balance between paid work and family life, White et al. (2003) found for a British sample that such timetable flexibility brought about a significant reduction in negative spillover for women. Likewise, Pleck et al. (1980) concluded that timetable inflexibility may cause conflict between paid work and family life. In light of these arguments, with regard to the division of labor within the home, an individual whose paid work situation includes timetable flexibility is more likely to find a better balance between paid work and family responsibilities, and thus may commit him/herself to greater participation in household work.

Finally, in addition to the factors of time-organization in the workplace, the employee’s position in the company hierarchy also has a significant impact on the time (s)he may devote to family life as a whole. A management position involves a greater degree of responsibility, which in turn connotes a higher level of availability. Moreover, management staff and professionals are to put professional demands before the demands of family needs so as to better motivate those working for them (Fried 1998; Hochschild 1997). Thus, the higher the individual’s position within the company’s hierarchy, the more limited is his/her time outside the workplace, and thus the more difficult it is for him/her to devote time to household activity.

Resource-bargaining Approach

One theory that tries to explain the division of paid and unpaid work is the “new home economics”. The assumption of this theory is that the objective of families is to maximize their income; thus, family members who are likely to be more efficient relatively speaking in workplace activities would tend to spend less time than other members of the family on work outside this sphere, including, for instance, unpaid work in the home (Becker 1965). In other words, as a result of the comparative advantages each member may have in his/her activity, respectively, the new home economics regards the specialization of one family member in household work as the most efficient division of labor (Becker 1981).

The new home economics implies that there should be a strong association between the number of hours a woman works outside the home and the number of hours she spends doing household work, but it fails to explain why housework should remain primarily women’s work. The proponents of the resource-bargaining approach pursue a similar line of argument (Blood and Wolfe 1960; Hiller 1984), in regarding the division of household work as the result of implicit negotiation between the spouses over inputs and outcomes in the home. The two partners in the couple disclose their resources at the negotiating table and come to a decision about their responsibility for unpaid work.

The relative resources approach might account for the observation that housework remains women’s work by arguing that because women tend to bring fewer resources into the negotiation process, they are less able to produce an outcome in which housework is equally divided between themselves and their partner (Greenstein 2000). The practical reality of this hypothesis may well be reflected in the lower salaries paid to women for carrying out similar types of work (Euronline 2009; Gupta 2007; Oaxaca 1973), as well as their work in positions that involve a lower level of responsibility and time-dedication (European Working Conditions Observatory, 2007), despite the fact that the condition that the most highly qualified person be recruited for a particular post would nowadays be more commonly met by hiring a woman (De Luis-Carnicer et al. 2002). From this perspective, then, the partner who earns a higher salary from paid work, relatively speaking, and whose post involves a higher degree of responsibility should devote less time and commitment to household work and family responsibilities. In other words, the greater the economic dependence of one partner on the other, the greater his/her commitment and involvement in the time to be spent on unpaid work.

Overview and Hypothesis

The increased number of couples in which both members work outside the home prompts an interesting question regarding the factors that might explain why the division of unpaid labor is effected in an equitable way in some cases, while in others the workload of one of the partners is doubled. Based on the above literature review, the purpose of this paper is to explore the explanatory potential of three theoretical approaches—the traditional gender division, role-strain, and the resource-bargaining approach—so as to account for the emergence of different patterns in the division of unpaid work among Spanish dual earner couples. In more specific terms, four general research questions were derived, with additional hypotheses:

  1. 1.

    If the division of household labor is explained by the traditional gender labor division model, then the following hypothesis should be verified:

    1. a.

      Women do more home duties than men.

    2. b.

      While older women participate in household labor more than younger women, younger men participate in household labor more than older men.

    3. c.

      Women with lower levels of studies are devote more time to household duties than women with higher education levels, whereas men with high educational levels participate more in household work than men with less formal education.

  2. 2.

    If the role-strain approach explains the division of household labor, the next hypotheses should be verified:

    1. a.

      The lower the number of paid work hours, the greater the participation in household work.

    2. b.

      Individuals working continuous working day work in household duties more than those with split-shift timetable.

    3. c.

      Greater work time flexibility is associated with greater participation in household work.

    4. d.

      The lower the hierarchical position of the individual’s jobs in the organization, the greater the involvement in household work.

  3. 3.

    If the division of domestic labor is explained by the resource-bargaining approach the following hypotheses should be verified:

    1. a.

      The lower the occupation compared to the partner, the greater the participation in household work.

    2. b.

      The greater the economic dependence on the other partner, the greater the participation in household work.

  4. 4.

    Are the theories that explain the strategies of household work division in dual earner couples the same for men and women? That is, how well do the models work in explaining men’s contribution to housework relative to women’s?

Method

Database

The database comes from the 2001 to 2004 Quality of Working Life Survey (QWLS). This survey is conducted annually by the Spanish Ministry of Work and Social Affairs and covers 6,020 individuals each year, and is designed to obtain information regarding characteristics, work and family situation, as well as various aspects of work and personal life. Responses to the questionnaire are compulsory because it is funded with public money. The questionnaire is structured in four sections. The first section deals with employment and family situation. The second section contains information about quality of life and work, providing information on attitudes towards work, work organization, working time and reward. Section three comprises worker opinions and attitudes on the relation between work and free time. The last section collects socio-demographic data about the worker, such as gender and marital status.

Given the objective of this paper, the sample used in the analyses discussed here is composed of couples in which both partners work outside the home and in which the household work is done by the people themselves. Those who responded that household work was carried out by a third party were excluded from the sample, as this situation produces a different dynamic in relation to the division of household labor (Ross 1987). Thus, the final sample comprises 2,877 observations, encompassing employees working in both the public and private sectors.

Sampling Design

The QWLS sample is representative of those in employment (employees and self-employed) during the fieldwork period covered. The QWLS sample is stratified according to region and size of municipality. Another variable, the number of habitants in each census section, is used in the sample selection. The QWLS sample follows a multi-stage stratified design with a random walk procedure for the selection of the respondents at the last stage.

The sampling design has three stages. In the first stage, stratification of primary sampling units according to census section is done. In the second stage, the units are the familiar houses that appear in the register. In the third stage the units are individuals aged at least sixteen. In each region, stratification according to the size of the municipality is done. The selection of home and respondent is random.

The data are collected by means of face-to-face interviews with individuals aged at least sixteen, working and living in Spain. Interviewers visit the homes of those in the sample between six and ten in the evening, in order to avoid localization problems among working people. The number of interviews by house is one, essentially because of the time needed in the work field. The sample size is of 6,020 individuals.

Measures

The objective of the paper is to discern what factors help to explain the division of household tasks between partners, for the general sample as a whole, as well as for the subsamples of women and men. Therefore, the dependent variable denotes the partner who does the housework. In specific terms, participants were asked the question: Are you the person in charge of household labor, such as cleaning, cooking, washing, etc? (¿Es Vd. la persona responsable de realizar las tareas domésticas en su hogar, es decir, limpiar, cocinar, lavar, etc?). The answer takes value one when the individual surveyed carries out some housework activities, value two when (s)he shares such activities with his/her partner, and value three when (s)he carries out all the housework activities.

In line with the theories outlined above, the independent variables of work are classified in three groups, as follows. The first category comprises the personal characteristics of the individual surveyed. These are the variables that explain the division of housework activities according to the traditional gender division model. Gender is the first characteristic considered. A binary variable, which has value one when the individual interviewed is a man and value zero in the case of women, is used (Crompton et al. 2005). Age is the second independent variable considered. Age range in this sample is from 19 to 65 and is represented as a continuous variable. The educational level or achievement of workers is gathered by means of three binary variables that have value one when the individual has reached a given level of formal education, and value zero in other cases (Álvarez and Miles 2003; Crompton et al. 2005). Participants answered the question: what level of education have you completed? (¿Cuál es el nivel de estudios más alto que ha alcanzado Vd.? Si está estudiando ahora, codifique el último completo). The possible answers are: primary education, secondary education (category omitted) and university education.

The second group of independent variables draws on the role-strain approach, especially in relation to worker profession. The first variable in this category is the exact number of weekly working hours (Greenstein 2000; Gupta 2007; Parkman 2004). The type of working day, which is included as the second independent variable in this category, is measured using a binary variable which has value one when the person interviewed has a continuous working day and zero if (s)he works on a split-shift timetable (Question: ¿qué tipo de jornada tiene en su trabajo? (1) Continua, (2) partida). The worker flexibility in deciding when to start and finish the working day is also included in this category of independent variables; it is measured on a five-point Likert scale: value one is assigned to individuals who have no flexibility; value three to those who enjoy some flexibility; and value five to individuals who have total flexibility. Finally, the worker’s position in the company is taken as independent variable (Biernat and Wortman 1991). Three dummy variables, with value one when the worker occupies the position under consideration and zero in other cases, are used to measure the mode of professional occupation. Such variables indicate whether the individual surveyed holds a management position, a supervisory position (category omitted), or an employee position (Question: ¿su puesto es de dirección, de supervisión o de otros trabajadores o de empleado? (1) Dirección, (2) supervisión, (3) empleado).

The final category of independent variables in the explanation of the division of household work between the partners in a couple relates to the resource-bargaining theory. The comparison between the employment position held by the individual interviewed and that held by her partner is carried out by means of three dummy variables (Evertsson and Nermo 2007). The variables indicate whether the employment position of one partner is lower than that of the other, both partners are on an equal professional standing (category omitted), or one individual occupies a higher employment position than his/her partner. These variables have value one when the circumstance indicated occurs, and value zero in other cases. In order to define these variables, employment positions (both for the individual surveyed and his/her partner) are structured in three groups: managers, technicians and professionals (value three); clerical workers, service workers and skilled workers (value two); and blue-collar and unskilled workers (value one). The dummy variables are established by subtracting the employment position of the individual interviewed from that held by his/her partner. So as to limit the number of fields involved, the variables arising from such subtraction are grouped in three dummy variables which indicate if the individual surveyed holds a higher, equivalent or lower employment position than his/her partner. Finally, a variable that indicates the economic dependence of the people interviewed on their partners is also included; as proposed by Sørensen and McLanahan (1987), this variable has been used in previous studies (Brines 1994; Evertsson and Nermo 2007; Greenstein 2000); economic dependence = (earning self–earning partner)/(earning self + erning partner). The potential values of this measure range from −1, which indicates that the respondent is completely dependent on her partner, to +1, meaning that the respondent provides complete earned-income support to her partner. A value of 0 on this measurement scale indicates that neither partner is economically dependent on the other; that is, they earn equal incomes.

As in existing empirical research literature dealing with cross-sectional data for several periods of time, the year the survey was conducted is included as a control variable. The survey iterations included in this discussion are 2001, 2002, 2003 and 2004 (reference year).

Results

This section is structured as follows: first, the basic characteristics of the sample are described; thereafter, the factors that explain the division of housework activities are analyzed using multivariate analysis techniques, thus enabling verification of the three theories offered to account for such division. In order to test the hypotheses and the complementarities of the theories, four models are estimated. In all cases, ordered logit models are estimated, since the dependent variables are ordered scale.

Descriptive Statistics

Table 1 presents the mean descriptive of the dependent and independent variables. Chi-square tests were carried to analyze gender differences among categorical variables, and one overall MANOVA was carried out to analyze overall gender differences in any continuous variables.

Table 1 Descriptive statistics and gender differences for dependent and independent variables.
Full size table

It is clear that there is a wide range of differences in the division of housework activities. In the general sample, more than half of the population says that they share the housework with their partner, which is almost equal to the combined number of respondents who say that they do nothing or everything at home. Nevertheless, these figures change significantly when the sample is split into male and female subsamples. In this case, the number of women who say they do nothing at home is infinitesimal, and the number of men who say that they do everything is similarly low. The percentage of women who say they share the housework with their partners is similar to the percentage of those who say they do everything. As for men, while 64.87% say that they share the housework with their partners, 33.29% admit to doing nothing at home. The significant difference in the amount of household work carried out by women and men suggests that an analysis of the factors that influence the division of household work not only in the general sample, but also in the subsamples of women and men, may well be pertinent.

With regard to independent categorical variables, the number of men and women are reasonably well-balanced in the sample; thus, the results found in the subsamples may be correctly compared and contrasted. The most common education level in the general sample is secondary education, whereas in the subsamples, women tend to have a higher average educational level than men. However, no gender difference arises. Few gender differences exist regarding the type of working day. The number of women saying they work a continuous working day, as opposed to split-shifts, is approximately 6% higher than the figure for men. In spite of the higher average educational level among women, the number of women in managerial and supervisory positions is considerably lower than the corresponding figure for men, and this difference is significant. In relation to the partner’s occupation or employment position, more than 50% of the women and men surveyed say that they have an equivalent employment position to that held by her/his partner. However, there are differences between the numbers of women and men who respond that they have a higher or lower position as compared with her/his partner. The number of women who say they have a lower employment position than that held by their partners is considerably higher than the equivalent figure for men saying their employment position is lower than that of their partners. This situation is reversed in the case of women—and men—who respond that they occupy a higher employment position than their partners. Thus, there are differences in partner’s occupation between the genders.

A MANOVA to simultaneously test for gender differences among all of the continuous variables revealed an overall gender effect. In terms of age, the average age of the general sample is 38.78 year old, while the average age of women is a bit lower than that of men, existing gender differences in terms of age. The average number of weekly working hours is higher for men than for women. On average, men tend to work 5 hr a week more than women; this difference is statistically significant. While men tend to have greater flexibility in deciding when to start and finish the working day, no gender differences occur in terms of flexible schedule. The sample considered in this discussion confirms the conclusion that the average wage for women is lower than that for men, a consensus position established in most existing research literature on the matter. In specific terms, it is clear from the data that more women are economically dependent on their partners, while more men provide earned-income support to their partners. Gender differences exist in economic dependence.

In addition, it should be mentioned that before doing the multivariate analysis, condition indices and variance inflation factors were calculated. They show values below the usual thresholds of 30 and 5 respectively (Judge et al. 1988), indicating that multicollinearity is not a problem in our models.

Division of Housework

Tables 2, 3 and 4 present the models framed to analyze the factors that may explain the division of household between the partners in a couple, both of whom work outside. Table 2 shows the models estimated for the general sample, Table 3 comprises the analysis for the subsample of women, and Table 4 displays the results for the subsample of men.

Table 2 Factors that affect the division of housework for total sample.
Full size table
Table 3 Factors that affect participation in housework for women.
Full size table
Table 4 Factors that affect participation in housework for men.
Full size table

As is clear from the tables, four models have been framed. The models estimated in each table differ in relation to the independent variables included. The variables included in first model are congruent with the traditional (gender) division of household work. The second model incorporates the variables referring to employment position or occupation; such variables enable explanation of the division of housework according to the theory of role-strain. The third model replaces those variables with comparative figures based on the activities of both partners in the couple, which relates to the resource-bargaining approach. The last model takes all the variables listed above into account.

General Sample

In the first model displayed in Table 2, the variables included are those that according the traditional gender division explain the sharing of household activities among the members of a dual earner couple. Where personal characteristics are presented as factors accounting for the division of household work, the mean characteristic that explains such division is gender. We find support for hypothesis 1a, since men are more likely to share responsibility for the housework with their partners than to do nothing or everything at home. The other personal characteristics included in the model are not significant or are of relatively little significance. In this regard, only those who have university education are less likely to participate in housework, in comparison with those who have secondary education. The gender of the individuals involved appears to be of primary significance in the category of personal characteristics, which is why separate analyses of the subsamples of women and men are likely to yield clearer conclusions with regard to the division of housework.

The second set of independent variables deal with the role-strain approach. Here, we find that the higher the number of working hours, the lower the individual’s involvement in housework. This result is in the direction expected according to hypothesis 2a. The type of working day also influences the division of housework. Individuals with a continuous working day are more likely to participate in housework than those working split-shifts; thus, we find support for hypothesis 2b. As flexibility in schedule doesn’t affect the participation of respondents in household labor, we do not find support for hypothesis 2c. We find mixed support for hypothesis 2d, whereby the individual’s occupation affects his/her participation in household labor. Individuals in the employee group are more likely to participate in household work than those in the supervisor group. There are no significant differences between individuals in managerial and supervisory positions.

Concerning the variables related to the resource-bargaining approach (the third model), the relative occupation in the firm and the wages earned by the respondent and his/her partner affects the division of work in the home. In specific terms, lower the employment position of the individual related to the employment position of the partner, the lower his implication in household activities. Then, we find some support for hypothesis 3a. Moreover, we find support for hypothesis 3b, since the lower the economic dependence of the respondent (that is, the higher his/her wage is in comparison to that of his/her partner), the lower the participation in housework. This is in line with the results in previous research literature, which show that both the educational level and the wage earned by the partner influence the division of housework (for a general overview, see Coltrane 2000; Shelton and John 1996). Likewise, in line with Becker’s theory (1965) regarding time distribution, according to which the principal goal of the family is to maximize income, the more efficient family member in workplace terms should spend less time on activity outside work than the family-member who is comparatively less efficient in professional terms.

An analysis of the fourth model discloses that most of the variables discussed in relation to the first three models are also relevant in the final model. The effects noted here are congruent with those traced in the other models, but are of relatively little significance. In this case, men are less likely to participate in housework than women. This result supports the traditional gender division of housework, according to which women are principally responsible for housework and men are the main breadwinners (Knudsen and Waerness 2009), and hypothesis 1a. Likewise, according to the role-strain model and to hypotheses 2a and 2b, those who work fewer hours in paid work and continuous working hours are more likely to be involved in household activities than those who work more hours in paid work and split working day. Finally, economic dependence also affects the division of housework, as hypothesis 3b predicted: as economic dependence on the partner decreases, participation in housework also decreases. This conclusion reflects the results from previous studies, which suggest a bargaining process between the partners, in which the individual earning a higher wage or occupying a higher employment position has greater negotiating power (Greenstein 2000).

Subsamples

Analyses of the subsamples of women and men disclose that the variables affecting the division of household responsibilities are not the same for women as for men; therefore, the factors that explain the participation of women in housework are different to those that explain the participation of men. With the analysis of these subsamples, we can test all the hypotheses, except hypothesis 1a.

With respect to the age of women, as hypothesis 1b predicts according to the traditional model, the older the women, the more likely it is that they will participate in housework to a greater extent. The situation in relation to men is the reverse: the older they are, the less likely it is that they will participate in housework. Thus, we find support for hypothesis 1b. The results relating to educational levels mirror the results for age and the prediction of hypothesis 1c. In comparison with women who have completed secondary education, those who completed primary education only and those who have carried out studies at university are more and less likely, respectively, to participate in housework. In other words, the higher the educational level of the women involved, the less likely they are to participate in housework. Again, this situation is reversed in relation to men. In comparison with men who have completed secondary education, those who completed primary education only and those who have carried out studies at university are less and more likely respectively to participate in housework. As previous research literature has suggested, university education has a liberalizing affect on attitudes towards housework (Newcomb 1943). Thus, the conclusions that older women with lower educational levels do more at home and older men with lower educational levels do less at home is consistent with traditional (gendered) division of labor, whereby men are defined as breadwinners and women as housekeepers (Hank and Jürges 2007). Changes have taken place in more recent generations, during which women have been able to study to higher levels of educational achievement than in the past: equality has begun to emerge in relation to the division of housework between women and men, since younger women with higher educational levels tend to participate less at home than older women with lower educational levels (Hersch and Stratton 1994), and because younger men do more at home than older men. Such observations reflect an overall change in the cycle of life.

With respect to the number of working hours, the greater the number of hours worked by men, the lower their participation in housework. The time-frame of the working day, as hypothesis 2b predicts, affects men and women in the same way which is the way expecting according to the role-strain model. As may be expected, men and women with a continuous working-day schedule participate to a greater extent in housework than men and women working split-shifts. These two results shown that as the role-strain model argues, time is a valuable and a limited resource that influences the division of household activities. Then, we find support for hypotheses 2a and 2b, but not for hypothesis 2c, since a flexible schedule does not have any effect on the division of household labor. In relation to employment status or position, hypothesis 2d, the explanation for the division of housework is different for men as compared with women. According to the role-strain model, women in the employee group are more likely to participate in housework than women in the supervisor group; but no discernible differences exist between women in supervisory and managerial positions. For men, the position occupied in the company hierarchy has no bearing on their participation in housework; and there are no differences between men, irrespective of the employment group to which they belong—manager, supervisor or employee.

Taking into account the resource bargaining approach, the relation predicted between partners employment position in the firm and the participation in household activities, shown in the third model outlined in Tables 3 and 4, is not found. We don’t find support for hypothesis 3a since the participation of women and men in household activities are not affected by the employment position in the firm of the partner relative to her/his own employment position in the firm. Finally the lower is the economic dependence on the partner, the lower the participation in housework as much for women as for men. This result regarding individuals who earn lower wages is consistent with the resource-bargaining account of the division of household labor within the couple (Blood and Wolfe 1960; Hiller 1984) and with the hypothesis 3b. When economic dependence on the partner is lower—that is, the greater the difference in wages with respect to the partner—the more bargaining power an individual may have (Brines 1994; Oppenheimer 1997).

Almost all the variables described in the first three models are significant in the fourth model, which underscores the overall significance of the model itself and enables explanation of the factors that affect participation in housework in relation to both women and men. This reveals that the three theories are complementary as long as they contribute to explain differently the division of household work.

The older a woman is, the greater her participation in housework. With regard to educational level, in comparison with women who have completed secondary education, those who completed primary education only and those who have carried out studies at university are more and less likely, respectively, to participate in housework: that is, educational level has a negative effect on women’s participation in housework. These results are consistent with the traditional gendered division of labor and with hypotheses 1b and 1c. We find some support for hypotheses 2b and 2d for women. According to these and to the role-strain approach, women who work a continuous working day participate more in housework than those who work split-shifts. Likewise, the position women occupy in the company hierarchy has little or no bearing on such involvement: only women in employee positions are more likely to participate in housework than women in managerial positions. Finally, as the hypothesis 3b argue according to the resource-bargaining approach, the economic dependence on the partner is also relevant to the division of housework for women. Thus, women who are more economically dependent on their partners are more likely to participate in housework than women who are less economically dependent on them.

According to the traditional gender division of labor and to hypotheses 1b and 1c, while age has a negative effect on male participation in housework, educational level has a positive impact. With regard to paid work and according to hypotheses 2a and 2b relating to the role-strain approach, it has been shown that lower the number of hours that men expend in work, greater their participation in housework activities and, also, that men who have a continuous working day, as opposed to working split-shifts, participate in housework to a greater extent. As may be expected according to hypothesis 3b relating to the resource bargaining approach, men who are economically dependent on their partners are more likely to participate in housework.

Discussion

The proportion of women in the workforce in Spain rose from 20.7% in 1978 to 41.1% in 2002. However, this change in the composition of the paid workforce was not mirrored by an equivalent rise in the proportion of men participating in unpaid work; only 12% of the women surveyed said that they share household responsibilities equally with their partners, whereas the European average is 25% (Álvarez and Miles 2006). This is a significant point because sharing housework was identified by 35% of women and 28% of men in Spain as the most important aspect in the achievement of equality of opportunity (European Commission 1998). In light of the significance of the division of unpaid work for society and the fact that the number of couples in which both partners work outside the home continues to rise, this paper has analyzed the factors that have a bearing on the division of household work between the individuals in the couple. Three theoretical frameworks have been drawn on in an effort to account for the existing inequalities: traditional gendered division, work–family conflict, and the resource-bargaining approach.

According to the traditional gender division theory, the weight and influence of tradition in all areas of social life prompt women to take responsibility for unpaid work, and thus to carry out most household activities. While the amount of time devoted by women to such work has been in decline in recent decades, such activity is still more likely to be carried out by women than by their partners (Bianchi et al. 2000; Shelton and John 1996); as a result, therefore, the traditional model of the division of household work persists. This situation is reflected in the data discussed in this paper: most women responded that they took care of all the housework or shared the responsibility for such work with their partners, whereas a tiny minority of men responded that they took on full responsibility for household activity (hypothesis 1a). The fact that women nowadays continue to carry out most of the housework, a situation shown in other research studies on the matter, may imply an interest in accepting the current division so as to avoid conflict, whereby the cost of the unequal division of household work is perceived as lower than the cost of confronting the partner over how such work might be distributed in a more equitable way (Wiesman et al. 2008; Kluwer et al. 1997). Moreover, the traditional distinction between gender-roles is also reaffirmed by the analysis above, especially with regard to the impact of age and educational level on the situation (hypothesis 1b and 1c). In this regard, older women and those with a lower educational level, who tend to figure most commonly in the traditional model, are most likely to be involved in unpaid work. Likewise in the traditional model, and by contrast, older men and those with a lower educational level are least likely to do so. That an increased educational level has a liberalizing effect on the traditional model, and thus on relative participation in unpaid work, is shown in this paper; along with the conclusion that the weight and influence of the traditional model may well be in decline as further advances in education are brought about.

The role-strain approach, the second theoretical account of the division of paid and unpaid work between partners addressed here, frames the relation between paid and unpaid work as a conflict shaped by the time and pressure demands that one sphere (work or family life) makes on the other (family life or work). Those whose paid work involves greater dedication in terms of time or energy (hypothesis 2a, 2b and 2c) or higher levels of responsibility (hypothesis 2d) will have less time available to participate in unpaid work. Some support for this general account of the distribution of household work has been discerned in this paper. Women with a continuous working-day schedule participate more in housework than women working split-shifts. The same holds true for woman’s employment position. The lower a woman’s position on the hierarchical scale, the greater her participation in household activities. Likewise, the higher the number of hours worked by men, the more limited their involvement in household activities, and men who work a continuous working day are more likely to participate in household work than those on split-shifts. Given that the conditions of paid work have a significant bearing on participation and distribution of unpaid work, a more equitable balance might be struck between the two by means of policies regarding the time demands and timetabling of paid work.

The resource-bargaining approach, the third theory adduced to the explanation of the division of unpaid work, posits a negotiation process between the partners regarding the inputs and outputs of the household, in which the partner who generates more resources from paid work, relatively speaking, has greater bargaining power and, as a consequence, a correspondingly lower commitment to unpaid, household work. One assumption underlying this sociological perspective on relative resources is that individuals strive to avoid household responsibilities (Blood and Wolfe 1960); thus, the more resources an individual has in relation to his/her partner, the more power and the more effective his/her negotiation of the distribution of unpaid work. Therefore, the lower the difference between the wages earned by the partners, the more equitable the division of household work (Presser 1994; Bianchi et al. 2000). The hypothesis that financial resources condition participation in unpaid work (hypothesis 3b) is confirmed in all cases, in so far as individual tend to carry out less housework as their salaries rise in relation to the wages earned by their partners.

Sociologists, along with an increasing number of economists, hold that the main reason economic models fail to account for the division of unpaid work is due to the significance of structural identity, social norms and traditional attitudes to gender roles as factors in the process (Álvarez and Miles 2006). Given that the sample analyzed here comprises men and women in couples where both partners work outside, paid work, and that the sample has been controlled by a range of personal and work characteristics, the difference in behavior between men and women in relation to household activities is difficult to explain (Parkman 2004).

In line with previous research studies in this field, this paper reflects the fact that while both men and women acknowledge in theory that men’s participation in household work should increase as women become more involved in paid work (Ferber 1982), such beliefs are not wholly effected in practice. The increased presence of women in the workforce has led to a corresponding decrease in the amount of time women devote to housework, but has not prompted an equivalent increase in the commitment among their male partners to household activities (Condran and Bode 1982; Geerken and Grove 1983; Knudsen and Waerness 2009).

At the same time, while the role of women in household work may be explained by reference to the three theories deferred to in the discussion above, men’s lower contribution would appear to be almost entirely attributable to the traditional gendered division of paid and unpaid work (hypothesis 4). The tradition and customs that encourage women to develop a special commitment to the home, and men to the workplace (Gwartney and Stroup 1973), remain in force to the present day. This situation is confirmed by the results from the general sample analyzed here, wherein the gender of the individual worker is the main variable in accounting for the division of unpaid work in the home. Thus, as previous studies have likewise shown, the roles of men and women are still determined by tradition to a very significant extent (Ferber and Birnbaum 1977), and women continue to shoulder greater responsibility for household work.

In summary, although it has been shown that tradition still plays an important role in household work division in dual earner couples, the other two theories used in the paper also help to explain the strategies followed in the division of domestic work. Therefore, although in general the three theories has been considered independently in the literature as substitute explanations, it has been found that in fact they are complementary, since relevant factors considered in one theory are not considered by the others.