A. Initial Objectives

The immediate need is for aggregate projections for the biomedical/behavioral sciences and for estimates by major field, leading to recommendations of levels of training support.

Aggregate projections. The Committee believes it feasible to project demand for biomedical and behavioral research personnel, based on suitable assumptions concerning levels of research support, by allowing the analysis to be directed by demographic factors and by expert judgments concerning the emergence of new areas of knowledge and technology. In addition, it seems feasible in the initial period to develop econometric models of supply and demand in these fields, using national time-series data for the biomedical and behavioral sciences in the aggregate, and to test such models for their applicability. In particular, one of these models would involve the use of relative-salary data. Models of this type have been used with some degree of success to provide estimates for physics, engineering, and law. Following the strategy of previous scientific manpower model-building developed by Freeman and others, the Committee would seek to use available data on degrees, salaries, R&D expenditures, enrollment, etc., to forecast supply and demand contingent upon exogeneous policy variables.

It will be essential in making these projections to take into account changing opportunities and demand for women and members of minority groups within the Nation's biomedical/behavioral research programs.

Field projections. Since adequate studies exist on the numbers of doctoral degrees awarded by field each year, short-term projections can be attempted during the initial period of continuing study. The feasibility of making forecasts by field, as suggested under III-A above, has limitations. The narrower the fields covered, the greater will be the problem of substitutability. Another way of stating the problem is that the number of those with first training in a given field is an inadequate measure of the supply of persons capable of working in that field. Moreover, the further into the future that forecasts are made, the greater the barriers to measuring supply in narrowly defined fields.

The margin of error may be minimized through projecting by broad fields in contrast to aggregating subfields. Broad-field projections can be supplemented by judgments of expert panels to delineate rapidly growing fields and those showing a more modest or relatively slow growth or decline. Although not currently available, models analogous to those referred to above can be developed relatively early for selected fields and subfields and to a limited degree could be made to reflect the extent to which workers could shift among fields. This analysis would make use of field-switching data developed by the National Research Council and other organizations.

Levels of training support. Based on the foregoing projections and examination of current trends in supply---including the output of other training efforts---recommendations would be made for the levels of NIH/ADAMHA training programs in FY 1976. As in the case of the projections of demand, these would be made by broad field.

B. Longer-term Objectives

Special importance is attached to improvement of the data base and continued refinement of methodology. Over the next several years, the full potential of the plan diagrammed in Figure 1 seems likely to be realized and should provide the basis needed for increasingly detailed recommendations by the Committee. The usefulness of current studies, such as the NRC-maintained Roster of Doctoral Scientists and Engineers, could be significantly enhanced for purposes of the continuing study. A suitable enlargement of the survey sample size would help to ensure representativeness of employment data for a greater variety of subfields in the biomedical/behavioral sciences. With respect to methodology, a detailed analysis can be made of the specific features of the biomedical/behavioral manpower market and second-round model-building can be designed to incorporate these features. Studies of manpower flow in academic medicine need to be made. It would also be desirable to conduct a study comparing the number of the Ph.D.'s awarded as a percentage of the number of baccalaureate degrees in prior years to determine the effect of support programs. Other steps would involve bringing to bear on supply/demand problems the expertise of persons in the market (academic deans, department chairmen, students, and research directors). The continuing study would involve also a more careful analysis of the characteristics of different sectors of the market and modes of adjustment to changing market and support conditions than has been possible in the past. The use of non-salary adjustment mechanisms, such as postdoctoral fellowships, could be taken into account in the economic modeling.