Overview
Survey Design
Trend Data
Availability of data
1. Overview
a. Purpose
The Higher Education Surveys (HES) system was established to conduct ad hoc surveys of higher education institutions
on topics of interest to Federal policymakers and the education community. Unlike the other surveys described in this
volume, the HES surveys are not recurring surveys.
b. Respondents
Survey coordinators at the responding institutions identify individuals at their institutions who are appropriate
to respond to a particular HES survey. Depending on the topic of a given survey, these could be department
chairpersons, registrars, financial aid directors, etc.
c. Key variables
The variables differ from survey to survey. Variables are typically limited to information likely to be
readily available to the institution.
2. Survey design
a. Target population and sample frame
The HES population consists of U.S. colleges and universities. Individual surveys may be targeted at
a subset of the larger population, e.g., 2-year and community colleges.
b. Sample design
In 1991 there were 1,134 institutions in the total HES sample. The HES sample is stratified by control
(private versus public), size of enrollment, and region of the country. Sample election within strata is
based on probability proportionate to the square root of enrollment. Further subsampling may be used to
conduct specific surveys within this series and/or some additional institutions may be added to the
population surveyed.
c. Data collection techniques
The surveys are conducted under contract to SRS by Westat. Questionnaires are mailed to institutional
coordinators who are responsible for identifying the appropriate respondents for the survey and collecting
the questionnaires from them. Follow-up is by phone.
d. Estimation techniques
Base weights are assigned to responses within sampling strata. These weights are the inverse of
sampling probabilities associated with the institution. An adjustment for nonresponse is made or each
sampling stratum by using a ratio of the sum of the number of responses plus the number of refusals to
the number of responses. The final weight used is the product of the base weight and the nonresponse weight.
e. Possible sources of error
(1) Sampling - The amount of sampling error varies from survey to survey, depending upon the subsample
methodology employed in the particular survey. Information about sampling errors is contained in the
survey reports.
(2) Coverage - Institutional coverage is excellent when all colleges and universities are included
in the target population. However, surveys targeted at subgroups of the institutional population are
more prone to coverage error because of the possibility that information on the sample frame that affects
the institution's in-scope classification has changed over time.
(3) Unit nonresponse - Response rates for this survey are typically over 90 percent. Sample weights
are adjusted for nonresponse within strata.
(4) Item nonresponse - Item response rates are generally above 95 percent.
(5) Measurement - To minimize nonsampling errors due to differential interpretation of items,
questionnaires are pretested with respondents similar to those who have completed the survey and the
questionnaire and instructions are reviewed by NSF. Manual and machine editing of the questionnaires
is conducted to check data for accuracy and consistency. Telephone calls are made to clarify cases
with missing or inconsistent items; data are rekeyed for 100 percent verification. Even with all these
checks, however, measurement error remains a possible source of error that is difficult to estimate.
3. Trend data
The nonrecurring nature of the surveys in this series precludes their use for trend analysis.
4. Availability of data
a. Publications
The data from this survey are published as Higher Education Surveys. Examples of recent reports
include Undergraduate Education in Electrical, Mechanical, and Civil Engineering, March 1994 (HES 16),
Survey on Undergraduate Education in Sociology, December 1992 (HES 15), Survey on Retention at Higher
Education Institutions, November 1991 (HES 14), and Plant Biology Personnel and Training at Doctorate-Granting
Institutions, December 1990 Information from some of these surveys is also included in
Women, Minorities, and Persons With
Disabilities in Science and Engineering and in Indicators of Science and Mathematics Education,
produced by NSF's Directorate for Education
and Human Resources.
b. Electronic access
Aggregate data from HES surveys is available in electronic format from
CASPAR.
c. Contact for more information
Additional information about this survey can be obtained by contacting:
Lawrence Burton
Senior Analyst
Science and Engineering Education and Human Resources Program
Division of Science Resources Statistics
National Science Foundation
4201 Wilson Boulevard, Suite 965
Arlington, VA 22230
Phone: (703) 292-7794
Internet: lburton@nsf.gov