Abstract: | The effect of being Catholic on first-birth timing (hereafter, birth timing) varies according to how the dependent variable is measured. When birth timing is measured as age at first birth, Catholics' is slower than non-Catholics'; measured as duration from marriage, Catholics' is more rapid. Structural equation models for discrete data showing these findings explain that Catholics marry later than non-Catholics. The direct effect of being Catholic is to speed birth timing, while the larger, indirect effect (through marital status) is to slow it. |