the birth certificates, parent interviews, and resident and nonresident father questionnaires.
Information was also collected through direct child assessments, examiners’ observations,
and parent–child videotapes. The ECLS-B sample size was 10,700 (rounded to the nearest 50),
representing almost 4 million children born in 2001. Researchers designed sampling pro-
cedures to approximate a nationally representative population of infants. In addition, they
oversampled specific subgroups, such as twins, premature infants, and some racial/ethnic
groups, including American Indians and Asians.