Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Appendix A Use of SIPP for Research and Policy Analysis In assessing the cost-effectiveness of a survey such as SIPP, it is important to ask to what extent the data are being used outside the sponsor agency for research and policy analysis purposes. Immediate use will not follow initia- tion of a complex survey because of a time lag until data files become publicly available and a further time lag until users complete and publish their analyses. David and Robbin (1991:62-66, 78-79, 84-88) examine pub- lication activity for SIPP during the first 6 years after the data files from the 1984 panel became available (1985-1990), comparing it with the publica- tion activity over a comparable period (1971-1976) in the history of the Panel Study of Income Dynamics (PSID) and the National Longitudinal Surveys of Labor Market Expenence (NLS). Table A-1 provides counts of research and policy analysis papers by users of SIPP and NLS (excluding methodological papers and those pre- pared by the survey staff) for years 1 to 6. The SIPP counts are from Committee on National Statistics (1989:Table 3-1) and panel staff; the NLS counts are from bibliographies obtained and analyzed by David and Robbin (l991:Table 3.10. SIPP exhibits a rate of increase in use of the data over the first 6 years that compares well with the NLS. (Conferences sponsored _ 1The publication counts developed by David and Robbin for the PSID are not shown because they are not entirely comparable with those for SIPP and NLS: the PSID counts are limited to papers published in refereed journals, while the SIPP and NLS counts also include working papers and professional association presentations. 265
266 APPENDIX A TABLE A-1 Papers Produced from SIPP and NLS Data Over a 6-Year Period SIPP NLS Year (1985-1990) (1971-1976) 1 17 12 2 10 14 3 16 18 4 69 36 50 51 6 56 57 Total 218 188 NOTES: SIPP, Survey of Income and Program Participation; NLS, National Longitudinal Surveys of Labor Market Expenence. The period for each survey represents the first 6 years after data files became widely available. SOURCE: Data from Committee on National Statistics (1989:Table 3-1) for years 1-4 for SIPP; compilation by panel staff for years 5-6 for SIPP; David and Robbin (l991:Table 3.10) for NLS. David and Robbin (l991:Table 3.10) also provide counts of papers from SIPP, including those using data from the SIPP ACCESS system at the University of Wisconsin and those using data obtained di- rectly from the Census Bureau or other sources. However, the bibliography they developed to obtain the latter counts is known to include duplicates and may also include methodological pa- pers, as the combined totals appear unrealistically high for most years. The David and Robbin SIPP totals are as follows: 1985, 39; 1986, 21; 1987, 26; 1988, 68; 1989, 67; 1990, 74 (based upon pare-year bibliography). by the Social Science Research Council account for the relatively large number of SIPP papers in 1985 and 1988.) Table A-2 shows the number of dissertations by year that used data from the SIPP ACCESS system at the University of Wisconsin (which David and Robbin managed), the NLS, and the PSID. SIPP compares favorably on this dimension of data use with the PSID. The higher dissertation output of the NLS results from a Department of Labor grant program that sup- ported dissertation research with the NLS data. Turning to the subject areas for which SIPP data have been used in research and policy analysis studies, Table A-3 classifies SIPP papers issued in 1989 and 1990 by topic. Percentages add up to more than 100 because some papers were assigned to more than one category. Papers of Census Bureau staff as well as outside analysts are included.
USE OF SIPP FOR RESEARCH AND POLICY ANALYSIS TABLE A-2 Dissertations Produced from SIPP, PSID, and NLS Over a 6-Year Period SIPP PSID NLS Year (1985-1990) (1971-1976) (1971-1976) 1 0 2 1 2 1 0 3 3 3 2 3 4 1 1 8 5 4 1 8 6 4 4 11 Total 13 10 34 NOTE: SIPP, Survey of Income and Program Participation; PSID, Panel Study of Income Dynamics; NLS, National Longitudinal Surveys of Labor Market Experi- ence. SOURCE: David and Robbin (l991:Table 3.11). SIPP counts are of dissertations using data from the SIPP ACCESS system at the University of Wisconsin. TABLE A-3 SIPP Papers Issued in 1989-1990 by Topic Anneals) Papers TopicNumber Percent Assets, wealth18 14.4 Child care, children15 12.0 Disability10 8.0 Education2 1.6 Elderly20 16.0 Family change, living19 15.2 arrangements Health care, insurance10 8.0 Income, poverty32 25.6 Jobs, welfare-labor supply30 24.0 . . declslons Long-term care4 3.2 Migration3 2.0 Program participation27 21.6 Race, ethnicity7 5.6 Rural population1 0.8 Veterans2 1.6 Total number of studies125 NOTES: Numbers of papers add to more than 125 and percentages add to more than 100 because papers were assigned to more than one category as appropriate. Papers include those issued by Census Bureau staff and outside analysts. SOURCE: Compilation by panel staff. 267