National Survey of America's Families

Report No. 1 in the Round Two methodology series focuses on the 1999 NSAF questionnaire. The introductory chapter describes the household screener and extended interview. In addition, the chapter covers respondent selection, types of NSAF interviews, and the NSAF family definition so that the reader may gain a better understanding of the NSAF questionnaire. The second chapter describes differences in the survey instruments between the 1997 and 1999 NSAF surveys. The remainder of the report provides the full text of the 1999 questionnaire.

Report No. 2 provides a description of the 1997 and 1999 NSAF sample design for both telephone and in-person interviews. For more detailed information on the design, the reader is referred to Report No. 2 in this series for 1997. Particular emphasis is given to research on design options and adjustments to the nontelephone sample. In addition, there is a parallel discussion for the telephone sample, a discussion of how subsampling rates were developed for telephone households, and a review of survey yields. The final chapter offers some conclusions.

Report No. 3 focuses on the methods employed to produce estimation weights and the procedures to use these weights to make state and national estimates from the survey data. These weights were used to produce estimates in Snapshots of America's Families, the first reports released from the NSAF.

Report No. 4 describes the methods and estimated sampling errors for statistics from the 1999 National Survey of America's Families (NSAF). First, an overview of the sample design and summaries of the precision of the survey estimates for both children and adults. It also presents sampling error estimates for differences estimated from NSAF data collected in 1997 (Round 1) and 1999 (Round 2). Chapter 2 is a general review of the two main methods of computing sampling errors or variances of estimates from surveys with complex survey designs such as the NSAF. Chapter 3 discusses why the replication method of variance estimation was chosen as the main method for the NSAF and describes procedures for computing replicate estimates of variance from the data. Chapter 4 and its attachments show how software available for computing sampling errors can be used with the data. The final chapter summarizes the findings and methods used.

Report No. 5 describes processes used to complete the in-person component of the NSAF. The in-person component was designed to augment the telephone survey with a sample of households without telephones. The report outlines the pilot study, procedures and materials used to conduct the interviews, the component's management structure, field organization, recruiting, and training, response rates, and special issues/problems that arose in the field during data collection.

Report No. 7 collects occasional papers given at professional meetings and workshops at the Urban Institute on technical issues in the design, implementation, and operation of the 1999 round of the NSAF.

Report No. 8 provides information on the response rates obtained for the 1997 NSAF (taking the estimation weights into account) and explains the methods used to compute these rates. Tables are included of response rates for important subgroups, such as telephone and nontelephone households. The report reviews approaches used to increase the response rates, and concludes that the response rates for the 1999 NSAF were higher than those typical of this type of survey.

Report No. 9 describes methods employed to complete the telephone component of the 1999 NSAF including a list-assisted method to select the random digit dialing (RDD) sample of telephone numbers and computer-assisted telephone interviewing (CATI) for screening and interviewing. Subsampling of telephone households is described, as is how respondents were selected, information on the topics covered during the interview, and data quality control methods used. The report concludes that there were few difficulties with the telephone interviewing component of the NSAF. Those that did exist, primarily related to within household coverage issues and family definition problems.

Report No. 10 focuses on data editing techniques, including data processing, coding guidelines, how data errors were dealt with, and how edits were made including the editing that was done to protect respondent confidentiality. In addition, this report discusses the imputation procedures that were often used when values were missing and explains the impact these imputations will likely have on the estimation of variance.

Report No. 11 provides documentation for the 1999 NSAF Public Use Files, which include data on over 42,000 households, yielding information on over 100,000 people. This report gives an overview of the survey, describes the sample design and methodology used, and discusses the limitations on use of the survey data. After providing an overview of the NSAF data, it also describes how to use the data files, and offers instructions and examples on how to use the NSAF survey weights.