3
SESTAT 2000 Decade Design Options
Determining the sample design for the 2003 SESTAT survey marks a critical decision point. The 2000 decennial census offers an opportunity to mitigate the limitations of the current sample by either refreshing or completely replacing the current sample. Another opportunity to do so may not occur for 10 years or longer.
The SRS developed three design options for the 2000 decade (Westat, 2002a); see summary in Table 3-1. The table contains a separate row for each population component of the target SESTAT population. The first eight rows show the population components covered in the current SESTAT sample by one of the past surveys. The next two rows address components of the target population that are currently not covered in SESTAT. The last row addresses scientists and engineers without a college degree.
The first column lists subgroups of the general population. The next two columns cover the current SESTAT that falls within these subgroups. The last four columns cover the three design options being considered. The two columns involving SESTAT indicate the frame from which the SESTAT survey was selected. The remaining columns show how each population component would be sampled and surveyed under each of the options.4
TABLE 3-1 Current SESTAT Design and Decade 2000 SESTAT Design Options
Decade 2000 SESTAT Design Options |
|||||
2000 Census Frame Option |
1990 Census Frame Option |
Hybrid Option Combining Elements of the 2000 Census Frame Option and the 1990 Census Frame Option |
|||
|
2000 postcensal |
1990 postcensal subsample NSRCG panel subsample (1991-1992 graduates) NSRCG panel subsample (1993-1994 graduates) NSRCG panel subsample (1995-1996 graduates) NSRCG panel subsample (1997-1998 graduates) NSRCG panel subsample (1999-2000 graduates) Regular NSRCG |
1990 postcensal subsample NSRCG panel subsample (1991-1992 graduates) NSRCG panel subsample (1993-1994 graduates) NSRCG panel subsample (1995-1996 graduates) NSRCG panel subsample (1997-1998 graduates) NSRCG panel subsample (1999-2000 graduates) Regular NSRCG |
|
2000 postcensal half sample |
2000 postcensal and April-June 2000 panel component Regular NSRCG |
|
2000 post-censal half sample and April-June 2000 panel component |
|||
SED 2000 postcensal foreign bachelors oversample 2000 postcensal |
SED 2000 postcensal foreign bachelors targeted subsample 2000 post-censal targeted subsample |
SED 2000 postcensal foreign bachelors oversample |
|||
Optional 2000 postcensal targeted sample |
2000 postcensal half sample |
||||
Current data only (CPS) |
Current data only (CPS) |
Current data only (CPS) |
The three options have some characteristics in common. First, the Survey of Doctorate Recipients would continue its current design and system of sampling and data collection under all of the options. Second, under all three options the sample of scientists and engineers would be updated biennially with new graduates from the NSRCG. Third, the SESTAT survey would be conducted biennially under each of these options.
2000 CENSUS FRAME OPTION: NEW NSCG SURVEY BASED ON 2000 CENSUS
Under the 2000 census frame option, a replication of the 1990s design, the Census Bureau would conduct a postcensal survey in October 2003 based on the 2000 census. This survey would include persons who, in April 2000, had received at least a bachelor’s degree, were 72 years of age or younger, were not institutionalized, and were living in the United States or overseas serving in the armed forces. In 2003, this sample would be contacted and interviewed for the National Survey of College Graduates. On the basis of the interview, those persons in the 2003 sample who (1) are college graduates with S&E degrees or (2) have a college degree and are working in S&E occupations would be screened into the SESTAT sample. With this approach, persons with foreign S&E degrees who were in the 2000 census as well as those with non-S&E degrees who are working in S&E occupations during October 2003 would be included (Westat, 2002a).
1990 CENSUS FRAME OPTION: A CONTINUATION OF THE 1990s PANELS
Under the 1990 census frame option, the current sample based on the 1990 census would continue, updating gaps in coverage where feasible. Nonrespondents from the original samples since 1993 would be traced in an attempt to decrease bias due to nonresponse. Targeted samples screened from the 2000 census would be used to update the sample of foreign-trained college graduates and those who work in S&E but do not have an S&E degree.
HYBRID OPTION
The hybrid option combines features of the 2000 census frame option and the 1990 census frame option. Part of the sample would be selected using the 2000 census frame option and the remainder using the 1990 census frame option. Under this option, the 2000 census would be used to draw a sample of college graduates of (nominally) about half the size of that planned under the 2000 census frame option. The subpopulations consisting of foreign-trained college graduates and those with non-S&E degrees who have moved into S&E occupations since April 1993 would be represented by this part of the total sample. The remaining portion of the sample would be derived from the existing SESTAT panel, which would be subsampled, bringing its size to (nominally) about half of the total current sample size (Westat, 2002a).