Although in the Netherlands most mothers stayed at home full-time to look after their children in the past, these days working mothers are no longer exceptional in Dutch society. Labour market participation is growing, in particular among mothers with young children. By the end of the 1970’s, of the women who gave birth for the first time only 9% continued their employment. By the end of the 1980’s this proportion had risen to 29% (CBS 1992). In 1997, 55% of the women with a child between zero and four years of age were gainfully employed (SCP 1998).

One peculiarity of the Dutch labour market is that, compared to other Western European countries, we find here the highest rate of women working in part-time positions (Blossfeld & Hakim 1997, p. 5). In 1995, 67% of the women working in the Netherlands have a part-time job compared to, for example, 25% in the USA and 41% in Sweden (Eurostat 1997, Blossfeld & Hakim 1997). Many women in the Dutch labour market have a job of 12 to 23 hours (31%) or of 24 to 34 hours (22%). Most of the men on the other hand (90%) work in occupations of 35 to 44 hours (Hooghiemstra 1997). More women than men, however, want to increase their number of working hours, whereas more men than women want to decrease their working hours. Of all the Dutch employed women, only 18% prefer to work less whereas 28% prefer longer working hours. For men, 30% prefer to work fewer hours and only 9% would prefer to work longer (ibid., p.75).

Accordingly, the percentage of women who work fewer hours than their husband is large (61% in 1994), although in 1986 it was even larger (68%). The percentage of husbands working fewer hours than the wife has been increasing during that time from 4% to 8% (ibid., p. 58). Still, part-time employment remains a preferred strategy of Dutch wives and mothers (and very slowly, increasingly of husbands and fathers) to combine their family responsibilities with their labour market career.

In section 1.3 of Chapter 1 it is described how dynamically the labour market participation of Dutch women has developed. Women between 25 and 55 years are the most prone to have children at home or family responsibilities. Over the last two decades we observe the highest increase regarding the employment rate of these groups of women. The most recent development is a shift from the 20- to 24-year-old women to 25- to 30-year-old women, as the group with the highest female employment rate.

The most important reason for this shift is that many women invest in a prolonged formal education, postponing their labour market entry, and marriage and childbirth as well. With more women prolonging their formal education, we can expect the employment rate of mothers to further increase in the future. The reason is that women’s educational levels are positively related to their likelihood of remaining in paid employment (CBS 1994).

This chapter deals with the influence of the mother’s employment on children’s education. Children’s educational success is a vital topic in studies on social stratification research. Usually, the indicators of family background to predict the educational success of a child are the education and occupation of the father (e.g. Blau & Duncan 1967, Hauser & Featherman 1976, Blossfeld & Shavit 1993, Rijken 1999). Researchers have paid less attention to the influence of the mother’s compared with the father’s socioeconomic background. It is argued that the socioeconomic status of the father determines the market position of the family and that the mother is usually dependent on the socioeconomic resources of her husband (Goldthorpe 1983, 1984). However, when we observe the above-mentioned growth in mothers’ labour market participation, it also becomes interesting to study the effect of women’s status on the child’s process of status attainment. Studies in North America show that the mother’s socioeconomic background affects the educational location of her children (Treiman & Terrell 1975, Stevens & Boyd 1980, Heyns 1982).

One question that continues to occupy public and political debate is the effect of the growing labour market participation of mothers on the school performance of children. In the USA, during the 1980’s a lively discussion arose about whether out-of-home employment improves a child’s chances in life or affects these adversely (Kamerman & Hayes 1982, Milne et al. 1986, Heyns & Catsambis 1986, Nock & Kingston 1988, Desai et al. 1989, Hoffman 1989, Scarr et al. 1989). However, concern about the effects of maternal employment is not specific to the 1980's as the following passage shows:

"A question arises in the minds of many women [...] as to the effects of out-of-home employment upon the conditions in the home and the child’s attitudes toward these conditions. Sound mental health is likely to develop more successfully in the fertile soil of a happy well-ordered home life, free from any excess of irregularities, disturbances and flurry" (Mathews 1934).

Here we examine whether children gain more from full-time mothers than from working mothers in terms of educational performance. On the one hand, one can argue that due to additional resources children of working mothers achieve a higher educational performance (Kalmijn 1994, Dronkers 1995). On the other hand the lower educational attainment of some children with working mothers may be due to the restricted time mothers have at home (Milne et al. 1986, Desai et al. 1989).

Dronkers (1992) shows for the Netherlands that children from mothers with a working-class occupation have educational attainment levels below those of children whose mothers are housewives. All other job categories of mothers have positive effects on children’s educational attainment. Yet, Dronkers (1992) does not introduce a separate control for the mother’s time budget restrictions. We study not only the influence of the mother’s job status but also whether or not she works full-time. It may be the case that the employed mother’s transfer of status resources may be inhibited by her restricted time at home.

Dronkers and Doornik (1996) investigated what influences the child’s school-related behaviour. They found that maternal working-class employment increases the chances of behavioural problems for the child inside and outside the school; children have fewer behavioural problems if their mothers are full-time housewives. Only upper-class maternal occupations significantly reduce the likelihood of the child having school-related behavioural problems.

A recent small scale study by Van der Slik and Felling (1999) underlines that the mother’s working hours, when their child is ten years old, have a negative influence on sons’ secondary school performance, but only if the mother uses several child minders. Daughters’ educational attainments, according to this study, are not affected. However, they do not consider that there are more and less sensitive phases for the intellectual development of a child. In pedagogical research the preschool age has been identified as a particularly sensitive phase for the intellectual development of a child (Groenendaal et al. 1996). In fact, the influence of the mother’s time restrictions on children’s educational attainment later in life is distorted by other factors. For instance, when the child enters the educational system, other factors, like the quality of the school, teachers, friends, neighbours, etc. enter the setting. Therefore, unlike Van der Slik and Felling (1999), we also focus on the effects of the mother’s out-of-home employment in the preschool years of the child.

As we are interested in the effects of mothers’ labour market participation, we formulate exclusively expectations regarding her influence. Because previous studies show that the father’s socioeconomic background has an influence on his child’s education, we include his traits in our study as control variables. In this way we prevent an overemphasis of the mother’s influence. The question answered is the following: How heavily do the time restrictions caused by the mother’s employment and her occupational resources influence children’s educational attainment?

One issue regarding maternal employment is whether a mother ought to quit her job after childbirth and not re-enter the labour market until the child is less dependent on her. The theoretical argument underlining the possible negative consequences is that if the mother is in paid employment, the competing demands for her scarce time resources imply that she devotes little time to the child (England & Farkas 1986, Desai et al. 1989). Because of the reduced time available, where working mothers are concerned, the family environment is probably of a lower quality since the child is given assumably less stimulation and support (England & Farkas 1986). We call this reasoning the time budget argument.

Additionally, certain aspects of the developmental theory also suggest that the current increase in the numbers of working mothers with small children may have negative implications. Developmental theory holds that it is within the first few years after birth that an infant learns to trust its parents. What is more important, the infant also learns to rely on himself or herself — an important personal trait for the process of learning further in life (Groenendaal et al. 1996, p.182). When the child is older and more independent and he or she has been looked after, for instance, by persons other than the mother, he or she can relate to such persons without any negative effects on his or her intellectual development.

Some researchers have emphasised, based on empirical evidence, that maternal employment has negative effects on the child’s intellectual abilities and school career. Gold and Andres (1978a, 1978b) found support for their hypothesis that maternal employment is negatively related to boys’ cognitive performances at nursery school age and to language ability scores of 10-year-old boys, compared to girls in the same group. As an explanation for this finding they state that "[...] the greater role similarity between the mothers and fathers when the mothers are employed should broaden the daughters’ conception of their own identity but should cause problems for sons in establishing a separate masculine identity" (1978b, p.75). Only with respect to adolescent children were no differences discovered in academic performance, neither within the population of children of employed nor non-employed mothers, nor between boys and girls (Gold & Andres 1978c). Desai et al. (1989) also discovered negative effects of maternal employment on the performance of 4-year-old boys (but not girls) in a language ability test. However, this effect was significant only if mothers in higher income families were gainfully employed during the early years of the child.

The time budget argument also entails that not only having a paid job in the first few years after childbirth, but also the number of hours worked should influence the child’s school success. Milne et al. (1986) empirically support this idea with their results. They show that net of mothers’ homework monitoring, family income and the like, the working hours of the mother negatively affect the educational performance of primary and secondary school students. Thus, the number of maternal working hours and the child’s educational level may be linked in a negative sense. If the assumption holds that in the sensitive phase during childhood the mother should stay at home, the effects derived from the time budget argument are most likely to appear in the first few years after the child is born and before it enters elementary school. Based on this, we formulate the following two time budget hypotheses:

1. If a mother does not quit her job during the preschool years of her child, she will have an educationally less successful child than a mother who has resigned her job during that time.
2. The more time the mother spends working in the first few years after childbirth the lower the child’s educational attainment.

In contrast to the above argument the following section argues that children of working mothers may have additional resources at their disposal compared with children of non-working mothers. For instance, a mother’s high status occupation normally implies that she has to keep up to date with the latest developments in her profession. The image of a frequently studying parent is, of course, a very good role model for a child’s academic progress.

A mother’s employment also enhances a degree of independence in the child, and familiarizes it with home organisation schedules and ideas originating from persons other than their parents. These traits impress teachers and help children get on at school (e.g. Gold & Andres 1978a, 1978b). In addition to the job status itself the job’s income is implicitly included in the term ‘occupational resources’. However, income is of course highly related to the occupational status of the incumbent. Regarding maternal employment we believe that her occupational status, labour market experience and career ambitions stimulate and support the child in its school career (Heyns & Catsambis 1986, Dronkers 1992, 1995, Kalmijn 1994).

When children are born into a family, very often the mother continues to have the prime responsibility for their upbringing. The additional resources from her employment enable her to invest more into her children’s educational attainment than if she was a homemaker. Perhaps her occupational status transfer to children’s education outweighs the implied negative effects of the employed mother’s time constraints, as posed in the time budget argument. We call this second rationale the resource argument. It leads to the expectation that a mother’s labour market participation has a positive effect on the educational attainment of her child.

We also find empirical evidence underlining the positive influence of the mother’s occupational status on the educational results of the child. The higher the occupational status of the mother, the better the children score on a test measuring their academic achievement (Dronkers 1992). It may be the case that mothers with a higher socioeconomic status begin to stimulate their child’s intellectual development from an early age and may be more eager to educate their child at home. Here two issues are raised that are related to the mother’s occupational resources: (a) her occupational status when the child is about to make its transition into secondary school and (b) her career patterns.

When the child is approximately thirteen years old, within the Dutch educational system parents have to take an important decision about the child’s further schooling. Children’s schooling is channelled into either a vocational or an academic education. Educational decisions are seldom reversed. For the USA, Kalmijn (1994) showed that maternal occupational status later in life has an impact on the child’s chances to make a transition into higher secondary education.

Regarding the second issue - the mother’s career pattern - usually the likelihood of a woman gaining promotion is reduced if she works intermittently, which implies that also her career pattern may be related to the child’s educational success. Because an employer assumes that a woman with an interrupted employment history is less reliable, it is unlikely that important tasks will be assigned to her. Thus, her career pattern is also important. Heyns & Catsambis (1986) find that a mother’s interrupted or intermittent employment history has harmful effects on the child’s schooling in contrast to the situation for non-working or continuously working mothers.

It may be the case that women who interrupt their career reduce the specific human capital needed to obtain high-status occupations and thus diminish the resources they can transfer to their child. We derive the following two resource hypotheses:

1. The higher the occupational status of a mother (before the child enters its secondary education), the higher the educational attainment of her child.
2. The more intermittent the labour market participation of a mother, the lower her child’s educational attainment.

The present analysis uses data from two surveys: the ‘Households in the Netherlands 1995’ (HIN95) (Weesie et al. 1995) and the ‘Netherlands Family Survey 1992-1993’ (FE92-93) (Ultee & Ganzeboom 1993). Both surveys contain a representative stratified random national sample, where couples are oversampled.² In both surveys the two partners answer questions on their educational, vocational, and employment history.

We exclude single parents since these groups are special cases where the influence of working mothers was found to follow a different mechanism (Milne et al. 1986). The HIN95 includes information only on the first child’s education. Therefore, the firstborn child’s educational level is used in both surveys to enhance the comparability of the results from the two sets of data. To ensure that the child completed or could have completed primary school, only families having a child of at least 13 years of age are selected. After the selection, 804 valid cases remained.

We examine how the eldest child’s educational success depends on the mother’s time budget and socioeconomic background. As the age of the dependent child ranges between 13 and 47 years, several problems had to be solved. Obviously some teenagers would be continuing their education, after their parents had been interviewed. Furthermore, the educational level of a teenage child had to be comparable to that of a 20-, 30- or even a 40-year-old one. The solution was to rank the various educational levels by generation, as a percentage on a scale from 1 to 100 (Norusis 1990).³ The result of this ranking procedure is presented in Table 3.1. The ranked education of the first child of the respondents is the dependent variable in the analysis (see for another example: Niehof 1997, p. 24ff.). Higher scores show a higher educational level for a child in comparison to his or her peer year group. For instance, a 13-year-old receives 20 points, whereas a 33-year-old receives two points for completing primary school. The more children in one of the educational categories, the higher their scores are. The further away the educational level of children from the mean of the peer year group, the greater their score differences are.

Calculating the ranking scores on the level of all respondents of both surveys (N=3146) resolves the problem of the decreasing number of cases within older age groups. We then ‘borrowed’ these scores and applied them to the children’s file, for persons 25 years and older. The focal set of independent variables include time budget, resource and control variables.

• A mother’s working hours when her child is a preschooler are averaged during the period when her child is between zero and four years old, that is before it enters primary school. Her average working hours are divided into two categories: part-time employment with up to three full days and part-time or full-time work with four and more days of employment. Mothers who do not work during the preschool years of their child serve as reference category.

• Maternal occupational status is measured when the child is 13 and about to enter secondary education. Included are occupations coded by the occupational classification of the Central Statistical Office (CBS 1984). We recoded the CBS occupational classification using the International Socioeconomic Index of Occupational Status (ISEI) (Ganzeboom & Treiman 1996). The ISEI scale has the advantage that it considers income when measuring occupational status. To measure the ‘added value’ of the mother’s continued employment we form three categories for the occupational status of the mother when her child was 13 years old. The category ‘low occupational status’ includes occupations between 10-30 points. A ‘medium occupational status’ ranges between 31-50 points and occupational scores above 50 points are grouped together as ‘high occupational status’.
• Number of intermissions in working career measures the extent to which the mother’s career after childbirth has been intermittent or permanent. For every period of leave longer than one year the score of this variable increases by one point. This variable measures interruptions until the child turns 18. Of course, if children in the data set are younger than 18 years it counts the number of interruptions of the maternal career only up to the child’s age at the interview.

• The effect of the mother’s occupational status before childbirth is controlled by a continuous variable. Her ISEI score before childbirth is divided by ten. Effects of non-working mothers are singled out by mean substitution (Cohen & Cohen 1976, p.274ff.). Additionally, if they never were employed before childbirth mothers are coded ‘one’ on a dummy variable. Therefore, the main effect of homemakers displays itself on a dummy variable, while being included in the analysis. Principally any ISEI value or even a zero value could have been assigned to mothers who never worked before childbirth. Applying a mean value is only one possible choice. The father’s occupational status before childbirth is constructed in the same way as the mother’s occupational status before childbirth.⁴
• Mother’s age is measured at the time of the interview and controls changes that are related to the birth cohort of the mother.
• We control the effect of both parents’ educational level. Mother’s and father’s educational level range between one and eight (see footnote 2). The same categories as the variable for the educational level of the child are included, but of course not ranked by their age.
• Father’s working hours when his child was 0-4 years old serve as a control variable because we filter the effect of the mother’s working hours net of his effect. A father’s working hours enter as a continuous variable.

Except for the mother’s occupational status before childbirth, for all other control variables we substitute any missing value by their estimated regression value, based on the education and the occupational title of the mother or the father. In a second instance, if the result of the former substitution still yields a missing value, we use a ‘mean substitution of subgroups’, again based on the educational and occupational title of the mother or the father. Because of the substitution of missing values, the database contains 804 respondents.

The descriptive statistics for all the variables in the model are given in Table 3.2. The ranked formal education is the dependent variable in the analysis. As it is based on the average achievement compared to the peerage group of the child, the average ranked education of the child bears little information because, by definition, it has to be around 50%. The reason for the ranking being slightly above 50%, that is 52%, is because we used for children 25 years and older the ranking scores of the entire database. More information can be obtained from the variables from which the ranked formal education of the child is constructed.

One expected outcome is that the child’s average educational level of 4.04 surpasses that of both of the parents. The mother’s educational level of 3.12 is lower than the father’s educational level of 3.67. The children in the data are on average 24 years old with a standard deviation of seven and a half years. Mothers are on average 49 years old, with a standard deviation of about eight years.

By the time their first child is four years old, 25 percent of the mothers return to the labour market, of which 13% work part-time and 12% work full time. This leaves us with 75% of the mothers who return to the labour market after their child enters elementary school or who never return. Of the mothers who work when their child is 13 years old and about to make its transition into secondary school, 13 percent hold a low, 10 percent a medium, and 16 percent a high occupational score. Thus, 61 percent of the mothers in the database do not work while their child makes its transition into secondary school. On average only every second mother has one intermission during her career. This suggests that mothers do not often interrupt their careers. They either quit or continue their employment steadily.

The average occupational status score of mothers before childbirth is 42 points on the ISEI scale (4.21*10). Only 6% of the mothers never worked before childbirth. The father’s average occupational status score before childbirth is higher than that of the mother. Most of the fathers in the data work full time and longer, their average working hours ranging at 39 hours, with a standard variation of 16 hours. If mothers return to work in the preschool years of their child, than they hold a job with on average 26 hours a week (table not shown). This is in line with recently published numbers (Hooghiemstra 1997).

Table 3.3 shows how the percentage of homemaking mothers has changed over the years, before and 13 years after their first childbirth. Throughout all the age groups only 6% of the mothers were never in paid employment before childbirth. This number varies very little between the age groups. Thirteen years after their first child is born the labour force participation of the mothers, shows a distinct pattern, related to their age group. The older the mother is at the time of the interview, the more likely she is a homemaker. The younger the mother, the more likely she is to be in paid employment. Only 47% of the women between 32-40 years old, but more than 79% of the mothers between 61-70 years, are homemakers when their first child is a teenager. Thus, the data confirms the overall found trend that the employment rate of mothers is highly related to their birth cohort.

* = p<0.05; ** = p<0.01

In neither of the analyses carried out was the time restriction of the mother’s employment significant. This means that the educational attainment of children from employed mothers does not differ significantly from the educational attainment of children whose mothers stayed at home and took care of the child.⁵ We do see, however, that the coefficient for mothers who work up to three days is negative, whereas the coefficient of mothers who work four and more days has a positive sign. Therefore, in a second instance (table not shown), we have checked whether the effect of up to three days of employment differs significantly from the effect of four and more days of employment. Now the reference category consists of mothers working up to three days. On a 10% significance level, children of mothers who work four and more days had a higher educational attainment than children of mothers who work up to three days.

The second argument made in this chapter is that, according to the resource argument, the mother’s occupational resources support the educational attainment of her child. As the most important time to transfer the mother’s job resources we have identified the period when the child is 13 years old and about to make its transition into secondary education. For this argument we find no support. Only if the mother helds a low occupational status when her child was 13 years old did it negatively influence her child’s educational attainment. Otherwise, the effects of her occupational status remain insignificant. Also, the mother’s number of career intermissions does not significantly relate to her child’s educational attainment.

If we enter both, the time budget and the mother’s occupational resources, into the analysis simultaneously, we observe that the effects of the two blocks of variables essentially do not change. It cannot be maintained that the effects of the mother’s employment, neither her time budget nor her occupational status, balance each other. Essentially, only mothers who reenter or remain at a low occupational level throughout their childrearing years have a significantly negative influence on their child’s educational attainment.

The ‘additional-worker’ argument perhaps best explains the latter result. If the husband’s employment is insufficient to sustain the family, it may result in a forced labour market entry by the mother to obtain a second paycheck. Often the need to find a job quickly, however, does not combine well with the quality and the status of the occupation found.

In the set of control variables we observe that if the mother was not employed before childbirth, it results in a large and significantly negative effect for her child’s educational attainment. On the other hand, the occupational status of the mother before childbirth also has a large, but this time significantly positive effect on her child’s educational attainment. For every 10 points on the occupational status (remember that the ISEI scale was divided by 10), the child gains about two points for its educational ranking score. The mother’s and father’s educational level influence the child’s educational level in the expected, significantly positive way. The father’s employment resources and time budget are not significant for the child’s educational attainment.

A word of caution may be at order. We here apply a very strict concept for the measurement of the influence of the mother’s occupational resources, if we control for her occupational status before childbirth. Of course, a mother does not suddenly ‘lose’ her occupational resources by giving birth to a child. The question here would be whether it is better for a mother to stop working and take care of the child at home. As the working hours of the mother in the preschool years of her child are not harmful for the educational attainment of either boys or girls, another interpretation of the above results of model (A) is possible. It can be posed that if the mother continues her employment after childbirth this does not influence her child’s educational attainment negatively.

We started this investigation with the question whether the restricted time available to working mothers’ influences the educational achievement of children in a negative sense or whether through her additional resources gained at work her child’s educational attainment is influenced positively. The answer we can give to this question, based on the above evidence, is that only if the mother re-enters or remains at a low occupational status position her continued labour market participation is harmful for the child’s educational attainment.

Looking at the results in detail, we found no support for the time budget argument. Neither her absence from home when she was employed during the child’s early childhood years, nor the extent of her employment is negatively related to the child’s educational performance. Mothers who are homemakers after childbirth do not enhance the educational performance of a child more effectively than working mothers.

With respect to the resource argument, the mother’s occupational status is not influential at the time when the child makes the transition into secondary education as long as she does not work in a low status occupation. Earlier, Dronkers has argued a similar point (1992): "[...] if she is employed at the working-class level her paid employment affects the educational chances of her children negatively".⁶ Therefore, it is the occupational status level at which the mother is employed and not her employment per se that matters for the educational attainment of her children.

In a sense indirect support for the resource argument exists. The additional resources, which a mother acquires from her paid employment before childbirth, help her child to get on in school. They serve the child’s educational performance better than if she was never employed before. Again, we can only emphasise the point that mothers who remain or reenter at a low occupational status level negatively influence the educational attainment of their child. The latter results may underline the strength of the resource argument regarding the entire socioeconomic position of the child’s family. The employment of the mother at a low status level, when the child is about to enter its secondary education, may indicate that the family is lacking enough economic resources to sustain the family by the job of the father. Therefore, our resource argument perhaps should be called a ‘lacking resource argument’: If the mother is forced to add to the family’s socioeconomic resources by being employed at a low occupational level, the educational attainment of her child is influenced negatively.

Obviously, our conclusions differ from those of Van der Slik and Felling (1999). They showed that the mother’s working hours had a negative impact on boys’ educational attainment if the childcare facilities used by the family varied a lot. However, it might be the case that the parents could not afford proper childcare facilities continuously. The quality of childcare facilities may partly explain why some parents would switch more often from one facility to the next than others. Yet, Van der Slik and Felling were unfortunately unable to retrieve a measure for the quality of the childcare employed by the parents.

Another important conclusion is that many standard measurements of social stratification have a distinct effect. The educational levels of both parents explain a great deal of the school performance of children. Yet, as the mother’s occupation displays a strong additional effect with respect to the children’s schooling, taking her occupational traits into consideration in future studies on educational inequality will be necessary.

For future investigations it will be interesting to see to what extent the increasing participation of fathers in the upbringing of children must be considered for future studies on this subject. If the child is taken care of by the father while the mother is working, the time budget hypotheses are not valid anymore. Another question is whether the mother’s occupational status is effective not only for her children’s education but also for their occupational choices. In Chapter 4 we will see whether primarily the daughter’s or the son’s first occupational attainment is influenced by maternal occupational status, or maybe both sexes equally.

Footnotes:

1. This chapter has been presented at the Sociologen Marktdagen (Utrecht 1997) and recently been submitted to the Netherlands Journal of Sociology, together with Tanja van der Lippe.
2. Both married and unmarried couples living together were included in the sample.
3. Originally, educational level was split into eight degrees (values in parentheses): completed or currently still in elementary school (1), lower vocational school (2), lower secondary school (3), intermediate secondary and vocational school (4), higher secondary school (5), higher vocational school (6), a university degree (7), or a doctorate (8).
4. We also controlled father’s occupational status when his child was 13 years old, but found no additional effect.
5. We also tested for significant differences in educational attainment between boys and girls and found no indication for such an effect.
6. "[…] indien het een arbeidsterberoep is, heeft het werk buitenhuis een negatieve effect op de onderwijskansen van haar kinderen" (p.154).