| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
THE EFFECTS OF RELIGIOUS FACTORS ON CHILDLESSNESS: THE CANADIAN CASE VIJAYA KRISHNAN Department of Sociology, University of Alberta, Edmonton, Alberta, Canada Summary Using the Canadian Fertility Survey of currently married or cohabiting women in the age group 35-44, this study evaluates the role of selected demographic and socio-economic factors on childlessness (voluntary, involuntary, and temporary). The effects of such factors on the incidence and types of childlessness were determined, and how they differ between women classified by generation of Canadian residence was determined. The three types of childlessness and the two generations differ in their responses to these factors. The findings show that, contrary to earlier assertions, Catholics are more likely than non-Catholics to be childless. Women who attend church services quite frequently are less predisposed to choose to be childless. Also, the results indicate that the higher the woman’s wage, the more likely she is to remain childless. Religious factors make a particularly important contribution to the incidence of childlessness among first- and second-generation women. Introduction In many Western countries much of the decline in fertility in the post-baby boom era stems from increases in childlessness due to voluntary factors; involuntary childlessness does not appear to have changed markedly in the past decade or so.1 An increasing proportion of married women in Canada is committed to remaining childless, estimated at about 10% of all women in Canada in the near future.2 The Quebec Survey of 1980 also indicated that almost 10% of those married in 1974-75 expressed their intention to remain childless3, but very little is known about the characteristics of women which may have urged them to choose voluntarily to have no children. Previous studies on childlessness in Canada often did not make any distinction between voluntary and involuntary childlessness. The purpose of this paper is to examine the effects of selected socio-economic and demographic variables (e.g. age at marriage and cohabitation, religion, income) on childlessness (voluntary, involuntary, and temporary) among married or cohabiting women aged 35-44, using data from the 1984 Canadian Fertility Survey (CFS). In addition, it explores differential childlessness among two groups of women, classified by generation of Canadian residence in order to see whether generational status accounts for any observed effect. Material and Method Data used in this paper were drawn from the Canadian Fertility Survey (CFS) conducted in April-June, 1984. The CFS was Canada’s first national fertility survey, using telephone conversations for both household selection and detailed questioning of the selected respondents, namely, women in the reproductive ages of 18-49 years. The total of 5,315 women of all marital statuses who were included constitutes a nationally-representative sample. The overall response rate was 70 per cent. For the present analysis, the sample was restricted to women in the age-group 35 to 44 who were still married to or living with their first husband or partner at the time of the survey, resulting in a sample of 2,807. This was necessary to control for differential rates of marriage, marital disruption, and remarriage, each of which can affect women’s decisions regarding (non) motherhood. Furthermore, by focusing on the age-group 35-44, only those women were included whose decision not to have children can be considered permanent. The childless women in the sample are classified into voluntary, involuntary, and temporary categories (Table 1). This classification is based on the women’s responses to questions dealing with expected family size, fecundity status, contraceptive use, pregnancy history, and so forth. Of the total sample, 91 are voluntarily childless, 27 are involuntarily childless, 84 are temporarily childless, and 2605 have children. In addition to types of childlessness, separate analyses of women of first and second generations were deemed necessary due to the interaction between immigrant status and fertility4, and possible minority group status effects on childlessness5. The data were analysed by logistic regression, chosen for two major reasons. First, the dependent variable has an extremely skewed distribution and secondly, the maximum likelihood logit model is capable of handling both categorical and continuous independent variables.6,7,9 Results Table 1 provides a brief description and Table 2 provides means or percentages and standard deviations of the variables used in the analyses. Table 2 shows that childless women were married or started living with a partner about two years later than women with children while there were no significant differences in age at marriage or cohabitation between the voluntarily and involuntarily, and temporarily childless women. Focusing on socio-economic status, childless women in general have more years of schooling than those with children but the differences are small. There are virtually no differences in educational attainment between the involuntarily childless and those with children. The mean wage of childless women is typically higher. The single exception of note is husband’s or partner’s income, where income of husbands/partners of the childless is much lower. This is consistent with the view that wife’s wage is directly related to childlessness while husband’s income is inversely related.6
With regard to religious affiliation, about half of the women with children are Catholics whereas more than half of the childless women are Catholics. Focusing on religious assiduity, generally women who have children are somewhat likely to be more religious than their counterparts who are childless. Generational differences in childlessness are most pronounced in the case of involuntary childlessness.
* For categorical variables, the figures correspond to percenta Percent for Catholic b Percent for the category high c Percent for first-generation women; the figures in parentheses are for second-generation women Effects of the determinants of childlessness Tables 3 and 4 present the results of the logistic multivariate analyses for the three types of childlessness – voluntary, involuntary, and temporary – and for the two groups of immigrant generations – the first-generation and the second-generation. Education was dropped from the analyses as it showed high zero-order correlation with wage. The tables contain the estimated coefficients and to assess their significance the t-values (the ratios of the coefficients to their asymptotic standard errors). The data give partial support to previous findings6,8 that there is an association between age at marriage and childlessness. On the whole, the coefficients for all the three types of childlessness show a remarkably similar effect, an increase in the probability of childlessness as the age at marriage increases. The coefficients for religious preference are unexpected. The coefficients are positive in all but temporary childlessness. The finding that Catholics are more likely than non-Catholics to be childless is at variance with previous research for Canada6,8 and for the United States.1 However, given the extremely low levels of fertility in the province of Quebec, where a large proportion of the population is Catholic, it is likely that the probability of childlessness is slightly higher among Catholics than non-Catholics. Grindstaff, et al.9 reported that generally, those women who profess religious affiliation have the lowest levels of childlessness, but at the older ages (40 and over) the differences are not meaningful. In other words, there may be a curvilinear relationship between age and childlessness, with older cohorts of Catholic women showing a greater probability of childlessness. The possibility of postponement of marriage and fewer legal marital unions for reasons of maximizing upward social mobility should not be ruled out. It seems likely that if a trend towards convergence in childlessness between Catholics and non-Catholics emerges it will become evident in the older (35 and over) women first.
Religiousness has a negative effect on all types of childlessness. The effect is significant for voluntarily and temporarily childless women. These results are consistent with the hypothesis that a commitment to parenthood occurs when a couple is more religious and attends church services very often. The effects of wage on childlessness are positive and statistically significant; the higher the women’s wage, the more likely they are to remain childless. These results support the findings of other researchers.6 It seems that there is little compatibility between choice of motherhood and woman’s participation in the labour force. In addition to testing all main effects, interaction of education and religious factors were examined, for it is possible that the effects of religion on childlessness are different at different levels of education and religious assiduity. None of the interactions were found to be statistically significant. Therefore religious preference does not have different effects on different levels of education and religious assiduity. Differential effects of determinants of childlessness across immigrant sub-populations The proportion of childless women is lower among second-generation than first-generation Canadians. However, this issue must be examined in the context of other factors that might be expected to influence the childlessness behaviour. The estimates within generations for the three types of childlessness combined are presented in Table 4. The difference in the constant term suggests that the two generations differ in some important aspects. For the first-generation women, there are slight and insignificant effects of religion (negative) and religiousness (positive). For the second-generation the effects of both are negative and stronger but still insignificant. Perhaps the most important conclusion is that when other variables are taken into account, first and second generations are not similar in their response to religious factors in childlessness. Although more conclusive evidence is needed, the data suggest a high probability of childlessness among first-generation women with high levels of religious participation while highly religious second-generation women are less likely to be childless. This finding confirms the bivariate relationships observed earlier. The coefficients of age at marriage are very similar in the two generations, reaching significance in the first. The standardised coefficients (not reported) indicate that religiousness ranks as the most important predictor for both first- and second-generation women. It seems (Tables 3 and 4) that the link between socio-economic and demographic factors and childlessness depends in part upon whether childlessness is voluntary, involuntary or temporary. Discussion The results show that there is some variation in the effects of demographic and socio-economic variables among the subpopulations. Voluntarily, involuntarily, and temporarily childless women are dissimilar in their response to religious factors. The findings are twofold. Most important, the finding that Catholics tend to be more likely than non-Catholics to have no children at all helps to explain why fertility rates of Catholics and non-Catholics have tended to converge in recent decades. Secondly, the differences between first- and second-generation women in terms of determinants of childlessness, are shown to lie mainly in religious factors, which are more important predictors of childlessness among the second-generation than among the first-generation women. Putting this finding in the context of the first, differences in probability of childlessness found between immigrant generations may be due to socio-cultural differences between women of the three types of childlessness. By aggregating immigrant generations, previous studies of childlessness have not discerned that they respond differently to the relevant factors. It is clear that the voluntary-involuntary-temporary distinction is crucial in childlessness research. The causes for the generation differences in childlessness are unknown yet it is crucial to find out what role these differences play in growing or lessening fertility differentials. Since fertility rates are on the decline, it is important to assess how far variations in childlessness – whether caused by cultural preferences, immigrant or minority position, or other structural factors – actually contribute to fertility differentials among population subgroups. Acknowledgements The author is grateful to Dr Karol J. Krotki, co-director of the Canadian Fertility Study, for permission to analyse the Canadian Fertility Survey data. The author would also like to express her appreciation to Dr Bob Arnold and Ms Margaret King for their helpful comments on earlier drafts. References 1. Poston, D. L., Jr. and K. B. Kramer. 1986. Patterns of childlessness among Catholic and non-Catholic women in the U.S.: a log-linear analysis, Sociological Inquiry 56(4), 507-22. 2. Veevers, J. E. 1980. Childlessness by Choice. Toronto: Butterworths. 3. Romaniuc, A. 1984. Fertility in Canada: from baby-boom to baby bust. Current Demographic Analysis. Catalogue 91-524 Occational. Ottawa: Statistics Canada. 4. Krishnan, Vijaya. 1988. Occupational status, earnings and fertility expectations: development and estimation of a causal model. De Economist 136(3), 358-82. 5. Rao, K. V. 1987. Childlessness in Ontario and Quebec: Results from 1971 and 1981 census data. Canadian Studies in Population 14(1), 27-46. 6. Amemiya, T. 1981. Qualitative response models: a survey. Journal of Economic Literature 19. 1483-536. 7. Tomes, N. 1985. Childlessness in Canada 1971: a further analysis. Canadian Journal of Sociology 10, 37-68. 8. Aldrich, J. H. and Forrest, D. N. 1987. Linear probability, logit and probit models. Sage University Paper Series on Quantitative Applications in the Social Sciences, 07-045. Beverly Hills and London: Sage Publications Ltd. 9. Grindstaff, C. F., Balakrishnan, T. R. and G. E. Ebanks. 1981. Sociodemographic correlates of childlessness: an analysis of 1971 Canadian census. Canadian Journal of Sociology 6, 337-51. | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||