CONCEPTUALIZING AND ESTIMATING UNION FORMATION RATES
In this appendix, we present the technical details of our strategies for estimating the effects of various parent and young-adult factors on marriage and cohabitation. In this work, our basic phenomenon to be analyzed is the monthly probability of making a union transition (H1). This transition probability (H1) can represent any of our conceptualizations of marriage and cohabitation, including marriage as a contrast to both being single and cohabiting; marriage and cohabitation as equivalent contrasts to being single; marriage and cohabitation as independent alternatives (competing destinations) to being single; and marriage as a transition for people already cohabiting. Another phenomenon of interest is the choice between marriage and cohabitation (K) for people entering either marriage or cohabitation.
In our analyses we transform the probability of making a union transition (H1) with the equation Y = ln[H1/(1 − H1)]. We also examine the choice between marriage and cohabitation (K) for those who form a union by the same logit transformation. In our base analysis we estimate the equation
where Ai represents four age factors and G represents the gender of the respondent. βg represents the effect parameter of gender and βi represents the effect parameters of the influence of each month of age during each of the four intervals on the log odds of making a union transition or choice. Because of the intuitive difficulty of interpreting effect parameters (β) in the log odds metric, we modify these parameters in our presentation of results by using the following transformation: B equals [exp(β) −1]. These transformed coefficients express the proportionate shift in the union-formation rate associated with each unit of the explanatory