Overview 
Survey Design
Survey Quality Measures
Trend Data
Availability of Data
1. Overview (2007 survey cycle) 
a. Purpose
The NSF-NIH Survey of Graduate Students and Postdoctorates in Science and Engineering (also known as the Graduate Student Survey or GSS), an annual survey of academic institutions in the United States, provides data on the number and characteristics of graduate students, postdoctoral appointees, and doctorate-holding nonfaculty researchers in science and engineering (S&E) and health fields. The National Science Foundation (NSF) uses the results of this survey to assess shifts in graduate enrollment and postdoctoral appointments and trends in financial support.
b. Respondents
NSF distributes the survey to school coordinators at GSS-eligible institutions. Data are collected separately for each reporting unit (academic department or program, research center, or health facility). School coordinators may choose to delegate data collection activities to respondents in these units or to other institution personnel, as needed.
c. Key variables
Count data are available for the following groups by the following characteristics:
Graduate Students[1]
- Enrollment status: first-time; full-time; and part-time; [2]
- Sex
- Race/ethnicity
- Citizenship
- Primary source of support (e.g., by specific federal agency)
- Primary mechanism of support (e.g., fellowship, research assistantship)
Postdoctoral appointees
- Primary source of support (e.g., fellowship, research grant)
- Sex
- Citizenship
- Whether appointee holds a first professional doctorate in health-related fields
Doctorate-holding nonfaculty researchers
-
Sex
- Whether researcher holds a first professional degree in a medical or related field (MD, DDS, DO, DVM) or foreign equivalent to a U.S. doctoral degree.
The following information is also available on the academic institution:
- Academic institution name
- School name[3]
- Unit name and field
- Geographic location
- Highest degree granted by all GSS-eligible units (doctorate/master's)
- Institutional control (public vs. private)
- Type of academic institution (e.g., historically black college or university, land-grant institution, medical school)
2. Survey Design
a. Target Population and Sample Frame
The target population for the 2007 GSS is all academic institutions in the United States and its territories that grant master's degrees or research doctorates, appoint postdocs, or employ doctorate-holding nonfaculty researchers in S&E and health fields as of fall 2007. The survey includes data for branch campuses, affiliated research centers, and separately organized components, such as medical or dental schools, schools of nursing, and schools of public health.
b. Sample Design
The 2007 GSS is a census of all eligible institutions as described above. The 2007 survey universe consisted of 582 institutions, including 375 doctorate-granting institutions and 207 master's-granting institutions.[4] There were 700 schools affiliated with these institutions: 493 at doctorate-granting institutions and 207 at master's-granting institutions. Of the 12,629 units associated with these institutions, 8,261 granted doctoral degrees.
c. Data Collection Techniques
The GSS data collection enumerates graduate students, postdocs, and doctorate-holding nonfaculty researchers by field as of the fall term of the academic year. Institutions select a GSS school coordinator for each reporting school. GSS school coordinators identify all eligible units (departments, programs, research centers, and health facilities).
The school coordinator submits the data for all units to the survey contractor. A Web survey system, introduced in 1988, has become the primary mode of data submission. New procedures designed to improve coverage of GSS-eligible units were introduced in the 2007 survey cycle. There had been little modification of the reporting unit listing in recent years, leading to a suspicion of undercoverage of GSS-eligible units. Increased emphasis was given to the update of the reporting unit list, the primary activity in the completion of Part 1 of the survey, and the Web survey was redesigned with this purpose in mind.
In 2007, revisions were made to the presentation of the list of GSS-eligible fields of study provided to school coordinators in the survey materials and through the Web survey. An exhaustive list of GSS-eligible programs within existing GSS fields was provided to better assist the school coordinators. In previous years, only a representative list of GSS-eligible programs was associated with each GSS field, which may have resulted in school coordinators not reporting on all eligible reporting units.
The set of GSS-eligible fields was also modified in 2007. Three science fields became newly eligible: communication, family and consumer science/human science, and multidisciplinary/interdisciplinary studies. Two programs were reclassified as fields of study, and each of these was provided with its own GSS code. In prior years, neuroscience was classified as a program within the health field neurology; in 2007 it became a science field. Architecture, which had previously been classified as a program within the field of civil engineering, was reclassified as its own separate engineering field. In addition, some programs that were listed in previous years were mapped to different GSS fields in 2007. For some programs, particularly in health fields, master's degrees were determined to be practitioner degrees, as opposed to research-oriented degrees; school coordinators were instructed to exclude master's students in those fields in 2007. Finally, 34 programs became ineligible and were no longer listed.
Throughout the data collection cycle, school coordinators and unit respondents could call a toll-free telephone number or send an e-mail to receive assistance from help desk staff. Post–data collection activities included providing extra assistance to school coordinators who had not submitted their school's data by the second and final deadline. Extensions were granted on a case-by-case basis. Also, follow-up telephone calls were made when a unit's reported counts of graduate students, postdocs, and/or nonfaculty researchers fluctuated sharply from the previous years' counts and an explanation had not been provided in the comment box provided in the Web survey.
d. Estimation Techniques
Because the survey is a census of eligible units, no weighting for sampling is necessary. Imputation rather than weighting is used to adjust for unit nonresponse; imputation is also used for item nonresponse.
3. Survey Quality Measures
a. Sampling Variability
Because the GSS survey was distributed to all institutions in the sample frame, there was no sampling error.
b. Coverage
A review of the 2007 frame indicated that there may be considerable undercoverage of GSS-eligible institutions. The review process identified 605 institutions that might be within the scope of GSS. However, many of these institutions offered a graduate degree in only one GSS-eligible field. In conjunction with the 2008 GSS, a pilot test of approximately 80 of these institutions is being conducted to assess the feasibility of determining the eligibility of these institutions, gaining cooperation, identifying school coordinators, and collecting data from the eligible subset of these institutions.
With the new data collection procedures introduced in the 2007 survey cycle (see
Data Collection Techniques), the number of responding units increased substantially (1,273 units were added, as compared with only 328 new units in 2006). The increase in the number of units added in the 2007 suggests that there was undercoverage of GSS-eligible units in previous survey years.
c. Nonresponse
- Unit nonresponse. Of the 582 institutions eligible in 2007 survey cycle, 552 (95%) were complete respondents, 10 (2%) were partial respondents, and 20 (3%) were nonrespondents. Of the 700 eligible schools, 668 (95%) provided complete data; 10 provided partial data (1%); and 22 (3%) were nonrespondents. At the unit level, 11,020 of the 12,629 units (87%) were complete respondents; 1,290 (10%) were partial respondents; and the remaining 319 units (3%) were nonrespondents.
The method for calculating the response rate for units (i.e., departments, programs, research centers, and health facilities) was changed in 2007. As in prior years, classification was based on responses to the survey's three data-collection tables (graduate student counts by race/ethnicity; graduate student counts by sources/mechanisms of support; and counts of postdocs and nonfaculty researchers), but the criteria for classifying a unit's response as complete were more stringent than in previous years:
- Units that required no imputation for any of the three tables were counted as complete respondents
- Units that required partial imputation for any of the three tables were considered partial respondents
- Units that required imputation for all cells were counted as nonrespondents
- Item nonresponse. Of the 201 items collected in the 2007 GSS, the mean item nonresponse rate was 6.6%, ranging from 2.7% to 9.8%. All missing data were imputed. Imputation was based on data reported by the institution within the previous 5 years, when available. Otherwise, imputation was based on data provided by similar units at a peer institution.
d. Measurement
Cognitive interviews, site visits, and other communications with school coordinators and unit respondents have pointed to a number of possible sources of measurement error. These are discussed below, along with steps taken to minimize their impact on the data.
First, although instructions emphasize that each individual should be enumerated only once, there is anecdotal evidence that some individuals have been counted twice by different school coordinators at the same institution or at institutions offering a joint program. In an attempt to prevent double counting, the Coordinator Contact Information screen in the 2007 Web survey provided names and contact information for all school coordinators at the institution.
Data on the race and ethnicity of graduate students also appears to be subject to some measurement error. The Office of Management and Budget standards treat Hispanics as an ethnic group rather than a racial group. Following these standards, "Hispanic" is not to be counted as a race in GSS. Cognitive interviews with respondents have revealed that this is a source of considerable confusion. For example, black Hispanics and white Hispanics may be counted as "Hispanic—More than one race" rather than "Only one race—Hispanic." (In 2008 these two Hispanic categories will be collapsed into one, "Hispanic/Latino ethnicity (one or more races)." This proved easier for cognitive interview respondents to comprehend.) Also, increasing numbers of students are choosing not to report their race to their institution, leading to growth over time in the "Unknown/race not stated" GSS category.
Data on financial support are sometimes difficult for school coordinators to collect and report accurately, in part because the information is usually not stored in one centralized database for the institution. Also, types of support that are not channeled through the institution, such as self-support, may be underreported. Foreign sources of support may not always be known. School coordinators and unit respondents may also have difficulty breaking down financial information by field, such as when a student is enrolled in one unit but receives support from another. Finally, institutions define mechanisms of support differently (e.g., fellowships vs. traineeships) and may report students according to the institution's definition rather than the definition provided by GSS.
In the 2007 survey cycle, some unit respondents provided notes indicating that although their units did have postdocs, they were unable to provide data for them. This reinforced reports from site visits, cognitive interviews, and other correspondence about the difficulty of providing this information.
4. Trend Data
NSF has collected data on graduate S&E and health enrollment and postdoctoral appointees since 1966. From 1966 through 1971, NSF collected data from a limited number of doctorate-granting institutions through the NSF Graduate Traineeship Program, which requested data only on those S&E fields supported by NSF. Beginning with the 1972 survey, NSF assigned this data collection effort to the Universities and Nonprofit Institutions Studies Group and gradually expanded the effort during the period 1972–75 to include additional S&E fields and all institutions known to have programs leading to a doctoral or master's degree. Due to this expansion, data for 1974 and earlier years are not strictly comparable with 1975 and later data. NSF has made no attempt to inflate the data for 1966–74 to reflect universe totals.
In 1984 the survey design was changed to a stratified random sample with a certainty stratum that included all doctorate-granting institutions; all master's-granting, historically black colleges and universities; and all land-grant institutions. The remaining master's-granting institutions were divided into two sample strata, based on enrollment size. Enrollment data for 1984–87 have been adjusted to reflect universe totals.
In 1988, surveying the entire eligible survey population was resumed for the first time since 1983–84. Since 1988 GSS has attempted to cover all institutions with doctoral or master's-granting programs in S&E or health fields and has excluded institutions that do not have any such graduate programs.
Also during the 1988 survey cycle, NSF tightened the criteria for including departments in the survey universe and reviewed all departments surveyed. NSF considered those departments that were not primarily oriented toward granting research degrees as no longer meeting the definition of "S&E." As a result of this review, NSF determined that a number of departments, primarily in the field of social sciences, not elsewhere classified, were engaged in training primarily teachers, practitioners, administrators, or managers rather than researchers; consequently NSF deleted these departments from the database. NSF continued this process during the 1989–2006 survey cycles and expanded it to ensure trend consistency for the entire 1975–2006 period. As a result, these changes reduced total enrollments and social science enrollments for all years.
During the 1992 survey cycle, NSF revised the definition of "medical schools" to include only those institutional components that are members of the Association of American Medical Colleges (AAMC). Prior to 1992, tables that reported on medical schools included schools of dentistry, nursing, public health, veterinary medicine, and certain other health-related disciplines. Beginning with the 1992 data, the term "medical schools" means only AAMC members. All tables with historical data on medical schools for 1991 and earlier were revised to include data based on AAMC membership criteria. Consequently data on medical schools in reports starting in 1992 do not match data from reports in earlier years.
In 2007 a comprehensive review of the GSS-eligible fields led to several changes to the classification scheme. The GSS-eligible programs (subfields) were updated from the NCES 1990 Classification of Instructional Programs (CIP) taxonomy to the NCES 2000 CIP taxonomy. Three new science fields became eligible. Two programs (one engineering and one health program) were reclassified as science fields. Many programs were eligible or were explicitly listed in the taxonomy for the first time. For some fields, master's degrees were determined to be practitioner degrees and, therefore, not eligible for GSS. Some programs were determined to be ineligible. Other programs were reclassified from one field to another. The 2007 GSS emphasized updating the list of eligible departments, programs, research centers and health facilities. New survey procedures were introduced that helped the school coordinators identify all GSS-eligible organizational units. School Coordinators were asked to review and update each organizational unit's assigned field of study. Updating the list of eligible units was assigned a separate deadline from the GSS survey and school coordinators had to confirm that the unit list was complete and that appropriate GSS field codes had been assigned to each unit.
Due to these adjustments to the taxonomy and other methodological changes introduced in 2007, data for 2007 are not directly comparable with data from previous years. However, for the purpose of trend analyses, the 2007 Detailed Statistical Tables provide estimates of the counts that would have been collected in 2007 if the 2006 methodology had been employed.
In addition to these changes to the survey universe and GSS-eligible fields, not all data items were collected from all institutions in all survey years. For these reasons, use only the latest trend data in historical analyses. NSF encourages analysts intending to do trend analyses not covered in this report to contact the project officer for additional information.
5. Availability of Data
a. Publications
NSF releases the data from this survey annually in Graduate Students and Postdoctorates in Science and Engineering and includes information from this survey in the Division of Science Resources Statistics (SRS) publications Science and Engineering Indicators and Women, Minorities, and Persons With Disabilities in Science and Engineering. NSF includes selected data items from this survey for individual doctorate-granting institutions in SRS's Academic Institutional Profiles series (http://www.nsf.gov/statistics/profiles/).
b. Electronic Access
Data for the years 1972–2007 are available in public use format and from the WebCASPAR data system.
c. Contact for More Information
To obtain additional information about this survey, contact
Footnotes
[1] To be included in the survey, graduate students must be enrolled for credit in any S&E or health master's or PhD program in fall 2007. Candidates for MD, DO, DVM, or DDS degrees; interns; and residents are counted if they are concurrently working on a PhD as part of a joint medical/PhD program or working on another S&E master's or PhD degree.
[2] Each institution reports full-time and part-time students according to its own policies and definitions.
[3] In this report, the term "school" refers to a graduate school, medical school, dental school, nursing school, or school of public health; an affiliated research center; a branch campus; or any other organizational component within an academic institution that grants an S&E or health degree, appoints postdocs, or employs doctorate-holding nonfaculty researchers.
[4] Institutions are classified as doctorate-granting institutions if at least one GSS-eligible unit confers doctoral degrees. Otherwise, institutions are classified as master's-granting institutions.
