Survey Methodology, FY 1995

Survey Population and Sample Design

Originating in fiscal year (FY) 1954, the National Science Foundation's (NSF) Survey of Scientific and Engineering Expenditures at Universities and Colleges (R&D expenditures survey) has been conducted annually since FY 1972. The population of institutions surveyed in most years has consisted of the 500 to 700 universities and colleges that grant a graduate science or engineering degree and/or annually perform at least $50,000 in separately budgeted research and development (R&D).1 These institutions have traditionally expended more than 95 percent of the Nation's academic R&D funds. In addition, the survey population includes all federally funded research and development centers (FFRDCs) that are academically administered and engage in basic research, applied research, development, or the management of R&D activities, as well as all historically black colleges and universities (HBCUs) that perform any separately budgeted R&D in the sciences or engineering.

The R&D expenditures survey is conducted as a population survey every 5 years and as a sample survey in each of the 4 intervening years. A new sample is drawn for the first sample survey and the same sample is maintained for all 4 years. Institutions in the survey population can be placed in one of two categories: the certainty category or the probability category. During a population survey, all institutions in both the certainty and probability categories are surveyed. During a sample survey, all institutions in the certainty category and a random sample of 25 percent of institutions in the probability category are surveyed.

Institutions in the certainty category include the following: Institutions in the probability category comprise the following:

A full review of institutions in the population occurs every 5 years. Prior to the start of each population survey, all academic institutions not already in the survey population are reviewed to identify those institutions that are likely to meet the survey criteria. Those institutions are added to the survey population temporarily. Several sources of information are used to select those institutions most likely to meet the survey criteria. Degree data from the National Center for Education Statistics (NCES) Integrated Postsecondary Education Data System Survey (IPEDS) are reviewed to add institutions that currently grant doctoral degrees in S&E. The NSF Survey of Federal Support to Universities, Colleges, and Selected Nonprofit Institutions (Federal support survey) is reviewed to add institutions that received at least $50,000 in Federal R&D obligations in S&E during the most recent 3 fiscal years. In addition, since service academies are not included in the Federal support survey, they are temporarily added to the population and surveyed. Once the survey has been conducted, the institutions are reviewed and only those meeting the certainty or probability criteria remain in the final population. Institutions not meeting the criteria are dropped entirely.

Characteristics of the institutions, such as highest degree granted or school closings and mergers, are reviewed throughout the course of the surveys to determine if changes have occurred, and the institutions are updated accordingly.

In FY 1995, there were 698 institutions in the R&D expenditures survey population. Of those 698 institutions, only 517 were sampled (499 academic and 18 FFRDCs). Of the 517 institutions in the sample, 18 were FFRDCs, 315 were doctorate-granting institutions (excluding those that were also HBCUs and system schools), 65 were HBCUs, 56 were system schools, and 63 were master's-granting institutions or below. (See list of institutions sampled during FY 1995.)

[ Survey Population and Sample Design | Data Collection and Processing Activities ]
[ Response Rates and Imputation for Nonresponse | Data Weighting and Standard Errors of Measurement ]