Table of Contents Table of Contents
Previous Page  190 / 126 Next Page
Information
Show Menu
Previous Page 190 / 126 Next Page
Page Background

190

E

UR

A

MERICA

of 14,660. There are 8,148 whites, 3,168 African Americans, 2,302

Hispanics and 1,042 Asian Americans. There were originally 1,582

Asian American youths interviewed in Wave 1, but only 1,037 were

re-interviewed in Wave 4. Using the self-reports on Asian

background, three Asian subgroups were created: East Asians

(n=369, including Chinese, Japanese, and Koreans), Filipinos

(n=411), and South Asians (n=257, including Indians, Vietnamese,

and other Asians). Five cases lacked a valid report to this question

and were dropped from analyses, along with a few cases with

missing values on maternal education and generational status. The

final analytical sample consists of 14,618 respondents of all races,

of which 1,037 are Asian Americans.

C. Variables and Measurements

The outcome variables for the three sets of analyses are: (1)

dichotomous variables of ever cohabited or ever married by Wave 4; (2)

dichotomous variables of whether a first cohabitation ended in

marriage, or whether a first marriage was preceded by cohabitation; (3)

the race of first cohabiting or marital partners. In addition, covariates

of age, sex, family structure, maternal education, and generational

status were also included in the models. These variables all come from

the Wave 1 survey, except for age. Given that Add Health is a

multi-cohort survey, respondents’ ages at Wave 4 (between ages 25 to

32) were controlled. A dummy variable of male was created for sex.

Family structure is considered because prior research has pointed out

that children with divorced or step-parents are more likely to form

marital or cohabiting unions earlier than those from intact

two-biological-parent families (Michael & Tuma, 1985). This study

identified two-biological-parent families, single-parent families,

step-families, and other families (i.e., children living with relatives or

grandparents). In addition, maternal education is also controlled

because children of better educated mothers have fewer transitions to

cohabitation and marriage in young adulthood among white and