Prern C. Saxena*
& Habbouba Y. Aoun**

1. INTRODUCTION

Researchers have argued that women's economic activity leads to a decrease in fertility of those working in white-collar (higher prestigious) occupations due to the conflict between the mother and work roles which is perceived less by women working in blue collar (lower prestigious) occupations. A sociological orientation hypothesis is drawn on the role incompatibility. The emphasis is placed on the woman as a mother and as a worker. The more incompatible the two roles are, the more negative the employment-fertility relationship will be. If a woman finds difficulty in combining child rearing and work, an inverse relationship may result. Bhargava and Saxena (1986), on the basis of data collected in Bombay, India, noted that the inverse relationship between work and fertility was more pronounced in the case of women engaged in white-collar jobs. The authors attributed the decrease in fertility to maternal role incompatibility. Their model is based on the hypothesis stating that if her two roles-as mother and as worker-are not compatible, an inverse relationship exists between work participation and fertility. However, if the two roles are compatible, work may have little effect on fertility. According to El-Safety (1993), a working mother finds it difficult to match the traditional expectations of rearing children and the newly acquired role. In the author's words "... The difficult situation of a working wife mother is aggravated by the absence of facilities that can help her dual performance...".

Two factors seem to interact when assessing the role incompatibility. First, the organization of production and second, the availability of inexpensive and reliable parental surrogates to working mothers. Due to fixed time schedules and no feasible way to raise children at the same time in the case of white-collar working women, a negative relationship between work and fertility exists. However, where the mother can leave her child with relatives or in a day care center, the inverse relationship is weakened to some extent. On the other hand, women working in agriculture or self-employed may find little problem in raising their children. They can carry their young children to their work place or may adopt flexible working hours. Such arrangements make the mother and worker roles congruent and the relationship between work and fertility may be further weakened.

Contrary to the experience of many developing countries, Lebanese women engaged in blue-collar (lower prestigious) occupations seem to experience more conflict in their two roles as a mother and as a worker, since they cannot bring their young children with them to the work site and at the same time they cannot make alternative arrangements for child care due to their meager income. Parental surrogates are also not available in the majority of cases in this society. In a recent study it has been observed that only 9% of the households in Lebanon are extended family households and around 4% are joint family households (Ministry of Social Welfare, 1996).

The primary objective of the present paper is to investigate how women's education and work status affect the fertility of Lebanese women. Also, the study attempts to assess the extent of maternal role incompatibility experienced by Lebanese women engaged in higher and lower prestigious occupations and its effect on their fertility.

2. METHODS AND MATERIALS

The study has been based on a sample of women working in a Lebanese private school. A list of the school staff, academic and non-academic, by their marital status was obtained. The list included 596 names of married men and women working at the school. Subsequently, these names were classified according to the level of education. Those who had completed secondary or more were placed in one list and those who did not complete secondary were placed on the second list. The former category of workers are hereafter called 'higher educated' and the latter 'lower educated' workers. The higher educated list of the school workers contained 360 workers: 284 female and 76 male workers. The list of lower educated contained 140 female workers and 96 male workers. A 33% sample from each of the list of higher and lower educated workers (females and males together) was drawn through a systematic random sampling procedure. Non-working wives of males working at the school were included in the sample, which comprised 222 currently married women. Eighty-nine of the responding women were higher educated working women and 20 were higher educated nonworking. Forty-three were lower educated working and 70 were lower educated non-working women.

The data was collected through the personal interview method. The sampled women were asked about their socio-economic and demographic characteristics such as current age, education, work status and occupation if working, husband's income and their complete birth histories. The assessment of the extent of maternal role incompatibility was made through questions on the type of arrangements the working mothers made for child-care. The satisfaction derived from such arrangements has been used as a proxy measure for maternal role incompatibility.

To examine the influence of women's education, work status and maternal role incompatibility on fertility, hierarchical analysis of variance (ANOVA) and Multiple Classification Analysis (MCA) have been used. The MCA is a special case of multiple regression with dummy variables. The predictor variables can be nominal and a non-linear relationship between any predictor and the dependent variable is permissible. The main advantage of the MCA technique is that it provides the grand mean of the dependent variable as its constant term and a set of category means for each factor expressed as deviations from the grand mean as main effects. Expressed in deviation form, category means reflect the magnitude of each category of a factor. Two models have been considered for the present analysis. Below is given the description of the variables taken in each of the two models:

2.1 Model 1

Dependent variable

Explanatory variables

1. Women's education (below secondary; completed secondary)
2. Work status (lower prestigious occupation; higher prestigious occupation; non-working)
3. Age at first marriage of the woman (< 17; 18-21; 22-25; 26-29; 30+)
4. Husband's income (Less than US $375; US $375-less than US $750; US $750 and above).

Covariates

5. Current age of the women
6. Square of current age of the woman.

The rationale for taking the current age of the woman as a covariate concurrently with the explanatory variables is that it helps in understanding the effect of women's education, work status and maternal role incompatibility on fertility after adjusting for differentials in fertility caused on account of these variables.

2.2 Model 2

In addition to the predictors and covariates considered in Model 1, the husband's education, and satisfaction derived from child care arrangements, are assumed to influence the dependent variable (number of children ever born). Husband's education appears in two categories of education: below secondary (lower educated) and secondary and above (higher educated). Again, the satisfaction derived from child care arrangements when the mother was at work, taken as the proxy for maternal role incompatibility, is considered as a dichotomous variable.

In both the analyses, the order chosen for explanatory variables reflects an implicit causal ordering where each of the variables is adjusted for the previous variable or variables.

3. RESULTS

Table 1 presents the results of the hierarchical analysis of variance of the children ever born to women, on their current age and duration of marriage squared as covariates, and the other variables as explanatory variables. It indicates that the education of women explains a significant variation in their fertility. It seems that the work status of the women is also an important predictor explaining fertility differentials. The analysis reveals that age at marriage affects fertility; husband's income also seems to have a marginal effect on the variation in fertility. Two way and higher order interactions are found insignificant. This ensures the applicability of multiple classification analysis (MCA).

The corresponding MCA output is presented in Table 2, which is given in absolute form. The hierarchical analysis gives the effect of a particular variable controlling for the effects of all the preceding variable(s) and the covariates taken as the adjusted values. For example, the effect of women's work participation on fertility is reflected by the fact that higher prestigious workers had borne, on average, 2.4 children, the lower prestigious workers 1.9 children and the nonworking women 2.9 children after controlling for the effects for the effects of age at marriage, education and the covariates (current age and its square) (Table 2).

Grand Mean = 2.48

As expected, both women's education and age at marriage are found to be universally related to fertility. Age at marriage seems to have greater effect on fertility as compared to education. For example, women who married between age 14 and 17 had on average 3.4 children, whereas those married between age 26 and 29 had borne on average 1.8 children.

It appears also that income had a positive effect on fertility. For instance, women having husband's income less than US $375 had on average 2.1 children and those more than US $750, 2.6 children. Further, it may be mentioned that the effects of all the predictors, except the income, on the dependent variable (children ever born), are found to be highly significant (Table 1).

3.1 Effect of Maternal Role Incompatibility

As stated earlier, in the second analysis (Model 2), husband's education and satisfaction derived from child care arrangements when the mother was at work are included as predictors to the set of explanatory variables of the first model. The latter predictor added to the analysis is taken as the proxy for maternal role incompatibility and is assumed to influence the dependent variable (number of children ever born). Results of the second analysis are presented in Tables 3 and 4 as ANOVA and corresponding MCA outputs, respectively.

The results of ANOVA show that all the explanatory variables taken in the model had a significant effect on the number of children ever born (fertility). Table 4 reveals that as the level of education of the husband increases, fertility shows a slight decrease. But increase in women's education seems to have greater influence in the reduction of fertility. As before, for the age at marriage the average number of children ever born decreases significantly with the increase in the age at marriage.

Grand Mean = 2.46

Women's work status is found to have greater impact on fertility. Nonworking women had, on average, more children as compared to working women after adjusting for the effects of covariates and other predictors, namely, husband's education, women's education and women's age at marriage. Interestingly enough, among the working women, those engaged in higher prestigious occupations had higher fertility than those working in lower prestigious occupations. The former have had on average 2.1 children as against 1.7 borne by women working in lower prestigious occupations. It may be noted that the women who were satisfied with the type of care they provided for their children when they were at work, had more children that those who were not satisfied. The "satisfaction" has been taken as a proxy measure of maternal role incompatibility. Thus, it may be inferred that the women who experienced maternal role incompatibility had lower fertility as compared to those women having their two roles-as a mother and as a workercompatible. Non-working women, who did not experience role conflict, had higher fertility-such women had borne, on average 3.2 children.

The predictors and covariates of the model explain 53% of variation in the dependent variable as shown by the value of multiple R 2 . It may be noted that the inclusion of the variable "satisfaction" in the second model has dramatically increased the value of multiple R 2 to 53.0 from 21.7 in case of the first model. This clearly shows the importance of maternal role incompatibility in explaining the fertility of working women.

4. DISCUSSION AND CONCLUSION

In the sample under study, higher educated working women engaged in higher prestigious occupations were working out of career aspirations while the lower educated women were driven to participate in the labor force out of economic need and were in lower prestigious occupations. For the child-rearing practices, leaving the baby 'with the parents' of the mother remains the most popular arrangement to take care of the infant when the mother is at work in the case of lower educated women, while the 'nursery' is the preferred solution in the case of higher educated women. This difference in child-rearing practices leads to a variation in the degree of satisfaction derived from the child-care arrangements which, in turn, may strengthen or weaken the relationship of role incompatibility with fertility in two different groups of women. Seventy percent of the women .working in lower prestigious occupations felt discomfort in leaving their children to be cared for by their parents, while the higher educated working women revealed that they were satisfied with leaving their child at a nursery. If perceived dissatisfaction with child care arrangements is considered as the proxy measure for role incompatibility between roles of motherhood and work, it can be said that role incompatibility, though present in both the higher and lower educated group of working women, is experienced more in the lower educated category. Actually, in the present sample, women working in lower prestigious occupations had borne on average 1.7 children as compared to 2.1 children born to women working in higher prestigious occupations. Non-working women had borne on average 3.2 children.

Moreover, on returning home after a day's work, women in lower prestigious occupations have to complete the household work themselves. In fact, they put their young child aside and do the household work. As for women working in higher prestigious occupations, they may employ a maid to assist them in household work while they nurture their child. Thus, the effect of women's maternal role incompatibility on fertility in the case of women engaged in lower prestigious occupations seems to be stronger than those working in higher prestigious occupations, resulting in lower fertility of the former. The results of ANOVA have also shown that maternal role incompatibility had a significant effect on the number of children born (fertility). The results show that nonworking women who did not have the problem of role conflict had borne, on average, more children. Women engaged in higher prestigious occupations who were satisfied with the type of care they were providing for their children when they were at work, had borne, on average, more children as compared to women in lower prestigious occupations who seemingly experienced greater role conflict. The inclusion of the variable "satisfaction with the child care arrangements" taken as a proxy for maternal role incompatibility in the second analysis, had considerably increased the predictive power of the model. Although the study is based on data collected from one private school only, the results have clearly demonstrated a greater impact of maternal role incompatibility as compared to education in the reduction of fertility of working women.

The effect of maternal role incompatibility in the family building of working women can analogously be viewed as the function of a resistance in an electrical circuit. If resistance is high, the current in the circuit would be low. The impact of maternal role incompatibility on fertility is similar. The three factors, namely, the possibility of raising children while at work, the availability of parental surrogates and income enabling a woman to put her child in a nursery/creche or to employ a maid, act as a shunt which short-circuits the resistance and weakens its effect resulting in increase of current in the circuit (see Fig. 1). Similarly, these factors enfeeble the influence of maternal role incompatibility in the work-fertility relationship and thus promote higher fertility.

BIBLIOGRAPHY