Patterns of childlessness among Catholics and nonCatholics in the United States

Data from the 1976 US National Survey of Family Growth was used to test the hypothesis that ever-married Catholic women are less likely to be childless than ever-married non-Catholic women. Several previous investigations have detected a convergence in the fertility of Catholic and non-Catholic women in the US. This convergence claim can be substantiated only if the level of childlessness among Catholic and non-Catholic women is similar. Childlessness decisions were considered to be fundamentally different than decisions to limit family size and it was therefore expected that the strong pro-family orientation of the Catholic Church would make Catholics less predisposed than non-Catholics to remain childless. In the 1976 survey information was obtained from 8611 ever-married and childed single women of reproduction age on marital and pregnancy histories contraceptive usage birth intentions and social and demographic characteristics. For the present study the sample was restricted to ever-married nonpregnant white women aged 35-44 years. The dependent variable was childlessness and the independent variables were religious affliation high or low religious activity and educational attainment. Log-linear techniques were used to analyze the data. These techniques were considered more appropriate than multiple regression techniques because the dependent variable childlessness was dichotomous and rare. 14 models were developed and the model which provided the most parsimonious description of childlessness was identified. The chi-square value of this model indicated that the direct effect of religious affliation on childlessness was minor and an analysis of odds-ratios also indicated that the probability of being childless was higher for Catholics than for non-Catholics. These findings appeared to contradict the hypothesis; however when voluntary and involuntary childlessness were taken into account the results of odds-ratios analysis strongly supported the hypothesis. Criteria based on childbearing intentions and contraceptive usage were used to categorize women as either voluntarily childless involuntarily childless and temporarily childless. Findings revealed that the probability of a Catholic being voluntarily childless was extremely small compared to the probability of a non-Catholic being voluntarily childless even when religious activity and educational levels were controlled. Among women with 2 or more years of college the probabilities of childlessness were closer for the 2 groups of women but remained lower for Catholic women. When the data was further analyzed in reference to the 4 categories of voluntary childlessness involuntary childlessness temporary childlessness and childed the hypothesis was even more strongly supported. Catholic women compared to non-Catholic women were 3.6 times more likely to be in the childed category than in the 3 other categories. These differences persisted when education and religous activity levels were controlled although again the probabilities were closer among Catholic and non-Catholic women with 2 or more years of college. The results of the analysis were provided in tabular form.