Demographic Models

Introduction

In 1994, the Committee on National Needs for Biomedical and Behavioral Research Personnel published Meeting the Nation's Needs for Biomedical and Behavioral Scientists for the National Institutes of Health (NIH). This was the tenth volume in a series responding to the request by the Congress that the Secretary of Health and Human Services arrange a continuing study of the national need for biomedical and behavioral scientists, to be conducted by the National Research Council (NRC). This volume addressed, as had previous volumes, the nation's future need for biomedical and behavioral research scientists and the role the National Research Service Awards (NRSA) program can play in meeting those needs.

In addition to guidance about the future demand for researchers relative to the current supply, NIH also directed the NRC to review the mathematical projection models of supply and demand used by previous NRC study committees and to establish their adequacy in addressing "national needs" issues in the 1990s. To review these supply and demand models, the committee convened a Panel on Estimation Procedures. The panel carried out an analysis that was used by the committee in preparing its report to NIH. In addition, the panel, the work of which continued beyond that of the committee, undertook further studies that provide the basis for this report.

The panel found that most of the models employed by previous study committees were inadequate and recommended that these models not be used in the future as the basis for developing recommendations. In particular, the panel was skeptical about the possibility of generating useful forecasts of "demand"—defined as employment.

Forecasts of demand for scientific personnel used in the previous reports for NIH were based on functions relating employed scientists to student enrollments and size of funds available for research. They assumed that student-faculty ratios were affected only by biomedical R&D expenditures and that the number of dollars required to support one scientist was fixed. However, student-faculty ratios also depend on other factors, such as salaries of faculty, which in turn reflect market conditions. The relationship between research dollars and Ph.D.s employed also depends on many other factors, such as the cost of non-faculty researchers, facilities and equipment, other inputs to the research process, and the technology of this process, none of which have been included in the models. In addition, the models were employed to project demand forward only five years. Yet these demand forecasts are being used to assess the adequacy of a supply of doctoral



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 1
Demographic Models Introduction In 1994, the Committee on National Needs for Biomedical and Behavioral Research Personnel published Meeting the Nation's Needs for Biomedical and Behavioral Scientists for the National Institutes of Health (NIH). This was the tenth volume in a series responding to the request by the Congress that the Secretary of Health and Human Services arrange a continuing study of the national need for biomedical and behavioral scientists, to be conducted by the National Research Council (NRC). This volume addressed, as had previous volumes, the nation's future need for biomedical and behavioral research scientists and the role the National Research Service Awards (NRSA) program can play in meeting those needs. In addition to guidance about the future demand for researchers relative to the current supply, NIH also directed the NRC to review the mathematical projection models of supply and demand used by previous NRC study committees and to establish their adequacy in addressing "national needs" issues in the 1990s. To review these supply and demand models, the committee convened a Panel on Estimation Procedures. The panel carried out an analysis that was used by the committee in preparing its report to NIH. In addition, the panel, the work of which continued beyond that of the committee, undertook further studies that provide the basis for this report. The panel found that most of the models employed by previous study committees were inadequate and recommended that these models not be used in the future as the basis for developing recommendations. In particular, the panel was skeptical about the possibility of generating useful forecasts of "demand"—defined as employment. Forecasts of demand for scientific personnel used in the previous reports for NIH were based on functions relating employed scientists to student enrollments and size of funds available for research. They assumed that student-faculty ratios were affected only by biomedical R&D expenditures and that the number of dollars required to support one scientist was fixed. However, student-faculty ratios also depend on other factors, such as salaries of faculty, which in turn reflect market conditions. The relationship between research dollars and Ph.D.s employed also depends on many other factors, such as the cost of non-faculty researchers, facilities and equipment, other inputs to the research process, and the technology of this process, none of which have been included in the models. In addition, the models were employed to project demand forward only five years. Yet these demand forecasts are being used to assess the adequacy of a supply of doctoral

OCR for page 1
and post-doctoral trainees whose education may span a far longer period. Similarly, the panel was uncomfortable with the approach taken by previous studies to assess job openings created by turnover. Turnover rates based on net (rather than gross) flows were estimated using cross-sectional (rather than time series) data. The validity of such an estimation technique as a basis for forecasting requires that the components of these single-year rates be stable over time. No effort was made to validate this assumption.1 Having rejected the previously used models, the panel explored the feasibility of developing projections of the size and characteristics of the future workforce of biomedical and behavioral scientists employing demographic techniques through the use of a life-table model. The data upon which a life-table model is based are the known characteristics of an existing population. In this case, characteristics of interest include age, employment sector, and employment status. Changes are projected in these characteristics based on the life history of the members of the population. Rates of transition (from employed to unemployed by age group or death by age group, for example) are calculated and applied to the behavior of the population in future years. Rates of entry to the system are also an important element of the model. With these techniques, the model can be used to answer the following questions: What will the characteristics of the labor force be in five years? What will the retirement rate be in the near future? How many new openings will there be if the behavior of the population in the future is the same as it was in the past? Although life-table models can be used to obtain estimates of new openings, it is important to note that they should not be used, mechanically, to perform the task NIH set for the study committee: that is, to estimate future needs for biomedical and behavioral research scientists and the role the National Research Service Awards (NRSA) program can play in meeting those needs. The committee arrived at its recommendations through a process that involved gathering a large amount of both quantitative and qualitative information about future needs for research personnel in different fields in the biomedical and behavioral sciences. In future reports, the panel felt that life-table models could be useful in estimating an "other things equal" baseline.2 In this exploratory effort, the panel developed a life-table model using a database that could be used by NIH were it to adopt the model as an aid in formulating its policy decisions with respect to training. A primary objective of this paper is to determine whether the database is adequate to support future modeling efforts and to 1   A demographic model that employed similar estimation procedures was developed by Xie (1995) to track student commitment to science and engineering careers as they flow through the educational system. 2   The panel felt, however, that the estimation of such a baseline would be more helpful than earlier work that yielded estimates based on models that were both overly simple and reliant on strong, hidden assumptions.

OCR for page 1
recommend steps to be undertaken in the future to develop successfully such a model. This report begins by providing an overview of demographic modeling, including the strengths and weaknesses of this technique. Then the elements of the model are developed (e.g., workforce inflows and outflows are defined) and the data set used for estimation of the model—the Survey of Doctorate Recipients (SDR)—is described. Next, the calculation of transition rates for various life changes (e.g., retirement, changing profession) is described. At this point, the model is used to make retroactive projections that are tested against real data from the SDR. Finally, the panel presents recommendations for further work that is needed to develop the model so that it can be used to project the needs for biomedical and behavioral scientists.

OCR for page 1
This page in the original is blank.