National Academies Press: OpenBook

Using Human Resource Data to Track Innovation: Summary of a Workshop (2002)

Chapter: Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen

« Previous: Appendix B: Workshop Participants
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×

Appendix C Using Human Resource Data to Illuminate Innovation and Research Utilization

PAULA E. STEPHAN

Andrew Young School of Policy Studies

Georgia State University

Issues paper prepared for a workshop of the Board on Science, Technology, and Economic Policy

National Research Council

Grant Black provided valuable research assistance on this project.

SECTION I. INTRODUCTION

Substantial evidence exists that widespread changes are occurring in patterns of innovation. One consequence of these changes is that traditional measures, such as patent counts and research and development expenditure data, are increasingly unable to illuminate R&D activity in the United States. “Without substantial change in the content and coverage of data collection, our portrait of innovative activity in the U.S. economy is likely to become less and less accurate.” (Mowery, 1999, p. 46). A key element of this change, although not the only element, is the increased incidence of collaboration both of a formal and informal nature that is occurring across institutions. By their very nature collaborative arrangements blur the boundaries between organizations and make it difficult to relate inputs, such as R&D expenditures of a firm, to outputs, such as patent counts.

This paper explores the use of human resource (HR) data concerning scientists and engineers to illuminate innovation and research utilization. For policy purposes it is important to gain an understanding of the extent to which this can be done, not only because of the aforementioned failure of current measures to capture changes occurring in the system but also as a way of gaining better insight into the social return to investments in science and engineering. For example, the federal government invests billions of dollars annually in funded research. Because grants are often

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×

awarded to individual principal investigators, human resource data and links of human resource data to outcomes, such as patents, make it possible to ascertain the eventual outcome of the investment and consequently make some inference concerning its economic impact. Similarly, federal labs train a significant number of young investigators who eventually leave to work elsewhere. The subsequent performance of these trainees could provide one measure of the contribution that federal labs make to overall performance.

The plan of this paper is as follows. Section II summarizes changing patterns of research and development and comments on gaps in our ability to measure innovative activity. Section III defines what is meant by human resource data and describes the data that are readily available. Examples of what can be learned from the use of human resource data to illuminate changes in innovation as well as provide insight in areas concerning the innovative process where substantial gaps exist are drawn by using the Survey of Doctorate Recipients (SDR). The section concludes by examining what could be learned if the HR data that are available were to be linked with other databases. Section IV looks at lessons learned from studies of biotech firms. Section V examines citation analysis.

In this discussion we are particularly interested in (1) what we can learn from data concerning the deployment of human resources as well as what we can learn from indicators such as publications that have individuals as a fundamental unit of analysis; (2) what can be learned from studying collaborative and sequential1 relationships that often are identified by using human resource data; and (3) what we could learn if we had better human resource data or the ability to link together two or more existing databases, such as the SDR with firm data.

SECTION II. CHANGING PATTERNS OF RESEARCH AND DEVELOPMENT

At least four broad changes have occurred in the structure and organization of innovative activity in the United States: ( 1) a decreased role for federal funding of R&D; ( 2) a change in the industrial distribution of innovative activities; ( 3) a shift of resources toward development activities and away from basic research; and ( 4) a change in the organization of research.

1  

Here we use the term sequential to connote either the source of an innovation or the impact the innovation has on subsequent innovation.

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Change in the Public/Private Mix of R&D

Industry has been the largest source of R&D funding in the United States for nearly two decades and its share continues to grow, approaching two-thirds of all R&D expenditures by 1998. This trend reflects both the increased growth of industrial R&D and a decrease in federal spending on R&D in terms of constant dollars in the late 1980s to the early 1990s. The decrease at the federal level since the mid-1980s has been largely a result of declining expenditures on R&D for defense. Federal civilian expenditures for R&D, on the other hand, which were relatively stable in the early 1980s in constant dollars, have increased in the late 1990s. This is largely a result of increased spending on health-related R&D, primarily through the National Institutes of Health.

Changes in the Industrial Mix of R&D

The most striking change in the industrial mix of innovation in the United States is the increased role played by the service sector. In the early 1980s R&D performance by nonmanufacturing industries made up less than 5 percent of total industry R&D performance. But beginning in the early 1980s this began to change dramatically. By 1991 the service sector accounted for nearly 20 percent of all the R&D performed in the manufacturing and service sector of the U.S. economy. Since then the service share has declined a bit but is still more than triple what it was only 15 years ago. The situation is, however, more dramatic than these statistics indicate since in “many nonmanufacturing industries that are essential to the development and diffusion of information technology, R&D investment is difficult to distinguish from operating, marketing, or materials expense” (Mowery, 1999, p. 46). This is reflected by the fact that although nonmanufacturing firms account for about 25 percent of the industrial R&D expenditures, they employ something like 45 percent of scientists, engineers, and S&E managers in industry.

A Shift Toward Development

Another trend observed in R&D data is a shift away from research toward development. “The upturn in real R&D spending that has resulted from more rapid growth in industry-funded R&D investment is almost entirely attributable to increased spending by U.S. industry on development, rather than research.” (Mowery, 1999, p. 45). By 1997, eight out of every $10 spent by industry on R&D were directed toward development (National Science Board, 1998, Tables 4-6 and 1-17).

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Changes in the Organization of Research

The organization of industrial research in the U.S. has also undergone substantial change. Four trends characterize the change: (1) increased reliance on external R&D, such as that performed by universities, consortia, and government laboratories (Mowery, 1999, p. 44); ( 2) increased collaboration in the development of new products and processes with domestic and foreign competitors and customers (Mowery, 1999, p. 44); ( 3) a decentralization of in-house R&D activities (Merrill and Cooper, 1999); and ( 4) the movement of innovative activities to functions in the firm typically not thought of as being drivers of innovation. The latter is fueled in part by the development of technologies that, as noted above, impact the operation and marketing of the firm’s production. Because these changes contribute to the growing inadequacy of traditional measures to describe innovative activity, we examine each in some detail.

The trend toward increased reliance on external R&D is undeniable.2 Firms outsource R&D to other firms or, in a more aggressive mode, acquire R&D through acquisitions.3 Cisco Systems is an oft-cited example of the latter, acquiring start-up companies as an R&D strategy. One consequence of this strategy is that the role of scientists in the firm is not to perform R&D but to assess the R&D capabilities of possible acquisitions. Again, standard R&D data fail to classify such individuals as engaged in R&D and thus undercount R&D activity in the United States.

Since the passage of the National Cooperative Research Act (NCRA) in 1984, nearly 600 formal research joint ventures (RJVs) have been filed with the U.S. Attorney General and the Federal Trade Commission (Leyden and Link, 1999, p. 575). A not insignificant number of these RJVs include federal laboratories as partners. Link and Leyden estimate that overall the rate is 8.7 percent (p. 577). As a result of the growth of RJVs traditional R&D data portray innovative activity less accurately today than in the past. The fact that associated firms typically allocate a portion of a scientist’s time to the collaborative project suggests that our knowledge of innovative activities resulting from RJVs could be enhanced by focusing on human resource deployment.

U.S. companies also enter into numerous collaborations with foreign companies. Most of these international alliances are with Western European companies although alliances between U.S. and Japanese companies are also

2  

See Badaracco, 1991; Hagedoorn, 1993; Hamel, 1991; Saxenian, 1994.

3  

Friar and Horwitch (1986, p. 77) studied 10 leading U.S. R&D companies. Although none was planning to extend its internal R&D activities, five intended to increase the acquisition of technology acquired through licensing, by forming joint ventures, or by fully acquiring firms with the needed technology resources.

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×

widespread (Mowery, 1999, p. 44). While domestic consortia of firms focus their efforts most consistently on research, many of the international alliances focus on joint development, manufacture, or marketing of products.

Firms also outsource to universities. In recent years the proportion of university research that is funded directly by industry has grown from approximately 2 percent in 1960 to over 7 percent by 1997 (Mowery, 1999, p. 45; National Science Board, 1998, Table 4-6).4 Trends in outsourcing to universities are clouded, however, by the fact that although industry funds an increasing share of university research, contributions to universities are not a significantly growing share of industry-funded R&D (Merrill and Cooper, 1999).

Universities have been not only a source of innovative ideas for established firms but also midwives to new firms formed by faculty members joining with venture capitalists. Nowhere has this been more evident than in the area of biotechnology,5 although other examples, for instance software and lasers, exist. University founders and researchers often have their cake and eat it too, maintaining their university jobs while they work in industry. In other instances they move back and forth between universities and firms, taking sabbaticals at companies (Powell and Owen-Smith, 1998, p. 263). Numerous federal programs also exist promoting cooperation between universities and industry. For example, the National Science Foundation (NSF) has programs that promote university-industry collaboration, and in some instances funding requires that NSF-supported centers have an industrial component (Powell et al., 1998, p. 256).

Changes in the law have also encouraged federal labs to develop alliances with industry and develop arrangements whereby R&D activity can be outsourced to private labs. Federal labs also join research consortia that involve firms in the private sector. The Clinton administration’s 1993 “defense conversion initiative” opened up formerly off-limits defense-related research to commercialization (Powell et al., 1998, p. 256).

Organizational changes have also occurred within the firm as some firms have shifted away from the central R&D lab model choosing not only to outsource research but in many instances to locate research activities at the plant level. This adds to the fuzziness of current R&D data since

4  

Universities are not only receiving a larger and larger proportion of their research funds from industry. University faculty increasingly are seeking patent protection for research performed within the university. One contributor to this is the eightfold growth in less than two decades in the number of university technology transfer offices.

5  

Audretsch and Stephan (1999) find that 50 of the 101 scientific founders of the 52 biotech firms they study were in academe at the time they founded the firm. At the time the firm went public, 35 of these founders remained working full time in academe.

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×

the location of where the actual innovation is developed less and less corresponds to corporate headquarters. Moreover, the growth of mergers and acquisitions makes it increasingly difficult to associate R&D activity with firm output. This is because the survey instrument that collects R&D data is fielded to the firm rather than the business unit. This results in attributing all of the firm’s R&D spending to the firm’s industry classification. Thus, if 51 percent of the firm’s business is in computer sales and 49 percent is in computer services, all of the R&D expenditures are attributed to the former category. Human resource data may help solve the “location” problem because they usually contain the address of the individual. HR data could overcome the line of business problem if the industry code of the plant or other establishment were made part of their record.

A final organizational change occurring within firms is the movement of innovative activities to functions in the firm not typically regarded as drivers of innovation. One example, given above, is the assignment of scientific personnel to evaluate and seek R&D through mergers and acquisitions. Another example is the involvement of technically teamed personnel in marketing and distribution. The important innovations that firms make in these areas are generally missed in standard measures of R&D. HR data could provide insight into these innovative activities by examining the deployment of S&E-trained individuals in non-R&D jobs.6

More generally, the improved competitive performance of many U.S. industries has come not only from the development of new technologies but also from “the more effective adoption and deployment of innovations” (Mowery, 1999, p. 42). These capacities are not measured by traditional R&D indicators such as patent counts and expenditures. They include “ investments in human resources and training, the hiring of consultants or specialized providers of technology-intensive services, and the reorganization of business processes.” (Mowery, 1999, p. 46). Creative uses of human resource data could illuminate industries that have high technology absorption capacities and could aid in our understanding of the strong performance enjoyed by a number of industrial sectors in recent years.

Powell and Owen-Smith (1998, p. 266) argue that some of the structural changes enumerated above have occurred because knowledge is increasingly located in networks of relationships and access to such networks is a key to competitive survival. In other work Powell and coauthors have shown that firm value is positively and significantly related

6  

Of course, it does not necessarily follow that individuals trained as scientists and engineers working outside of R&D in industry are using their training on the job. They may have accepted these jobs in the absence of employment opportunities in research.

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×

to network access. Traditional R&D indicators fail to measure these networks as well as the access that individual companies have to networks.

These structural changes mean that traditional indicators of R&D as well as the traditional unit of analysis, the firm, are becoming less relevant to the study of innovation. But that is not the whole story. It is not just that the formal organization of R&D is changing. Ample evidence exists that knowledge spillovers play an important role in innovation and that traditional measures fail to capture the effects of these spillovers.7 These data inadequacies are becoming apparent at a time when we tout the economic growth enjoyed by the United States as being “knowledge-based.” We are attributing growth to inputs that, because of organizational change occurring within firms and the development of outsourcing and collaborative ventures, are increasingly difficult to measure accurately. In short, the changes outlined above result in a blurring of boundaries and a blurring of roles. Measures of innovation designed when firms were discrete firms and universities were strictly universities fail to portray these changes adequately.8

Even without the changes noted above the traditional measures of innovative activity, namely patent counts and R&D expenditures, reveal little to investors and analysts concerning the knowledge base of firms. As Lev (1999) notes, firms report nothing on a regular basis other than their R&D expenditures. This makes it difficult to evaluate companies, particularly companies that are knowledge-based. There is no way, for example, of determining the closeness of the science link or to evaluate the quality of the link.

SECTION III. USE OF HUMAN RESOURCE DATA

Human Resource Data: Definition and Availability

Broadly speaking, human resource data refer to data collected on individuals either working in or trained in the field of science and technology. Although such data can and are collected on a case-by-case basis as well as by professional societies and universities, six primary sources for HR data in S&E exist in the United States.9 These are briefly summarized

7  

Spillovers are often examined by studying the relationship between a measure of innovative activity of the firm and the research expenditures of universities and other organizations in close geographic proximity. The rationale for expecting them to be bounded is that tacit knowledge is difficult to communicate in writing but is facilitated through face-to-face communication. See, for example, Jaffe, 1989; Acs et al., 1992.

8  

Powell and Owen-Smith (1998, p. 266) do a good job of summarizing these changing boundaries.

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×

here along with the target population that each addresses. We then use one of them, the SDR, to illuminate several of the trends discussed above and as a way of exploring how, with certain linkages and additions, the data could be used to illuminate other trends existing in patterns of innovation.

The NSF directs considerable resources towards gathering information on the scientific and engineering workforce who have completed their PhD training in the United States. These data are collected in two complementary ways. First, the Survey of Earned Doctorates is administered to all individuals receiving a doctoral degree in the United States, regardless of field. This survey, begun in 1958, is administered by the awarding university and forms the basis of a census of all individuals who received their doctoral training in the United States. The census, referred to as the Doctoral Records File (DRF), was begun early in the twentieth century and was originally constructed from administrative records. Since 1957 survey data have been available on field of training, financial support during graduate education, employ-ment plans, and an array of demographic characteristics including date of birth, marital status, education of parents and geographic location of high school. The SDR is a biennial survey of a sample of individuals whose records are contained in the DRF and who indicated at the time that they received their doctoral degree that they intended on staying in the United States.10 The intent is that the data be longitudinal and that individuals remain in the frame until the age of 75. The data capture individuals trained as scientists and engineers who are working outside their field of training as well as, until quite recently, individuals with doctoral degrees outside S&E. Thus the linguist who received a PhD in English but is now working in an information tech-nology field was included in the survey until financial considerations recently led National Endowment for the Humanities (NEH) to dis-continue their support for sampling the humanities.

Individuals who have not received training at the doctoral level clearly contribute to innovation. This is particularly the case in the areas of engi-

9  

A University of California study has tracked the career paths of PhDs in biochemistry, computer science, electrical engineering, mathematics, political science, and English over a ten-year period. The study group is composed of all of those receiving a PhD between July 1982 and June 1985. The results of the study for mathematicians and biochemists were reported by Nerad and Cerny (1999).

10  

Many more scientists and engineers indicate that they have plans to work abroad than actually do. In an exceptionally creative use of data linkages, Michael Finn and coauthors (1995) matched the SED records to Social Security records to estimate the number of individuals who say they have plans to leave but do not actually leave or who subsequently return. They used a similar procedure to examine whether individuals who said they definitely planned to stay in the United States actually stayed.

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×

neering and computer information and technology. Human resource data are collected on nondoctoral-trained individuals (as well as doctoral-trained individuals) through two additional surveys: the National Survey of College Graduates (NSCG) and the National Survey of Recent College Graduates (NSRCG). Since 1993, both surveys have become biennial. The sampling frame for the NSCG is drawn from all college-educated individuals in the most recent decennial census regardless of occupation reported in the census. Follow-up biennial surveys include college-educated individuals trained and/or working in science and engineering. Thus, for example, the Russian physicist who immigrates to the United States but cannot find a job in S&E is included in the NSCG as is the physicist who works on Wall Street.11 The sample also includes the linguist who works in information technology. By using the census for the basis of the sampling frame, the methodology includes individuals working in the United States who received their training outside the country.12

The NSRCG provides informa-tion about individuals who recently obtained bachelor’s or master’s degrees in S&E. The population surveyed includes all individuals under the age of 76 who received bachelor’s or master’s degrees in an S&E field within a 2-year period prior to the survey reference date from a U.S. institution. In addition to information concerning education and employ-ment status, the survey collects data on such variables as primary work activity, occupation, and salary. Information from these three surveys (SDR, NSCG, and NSRCG) has been integrated into the SESTAT database, available on the web or on CD-ROM. The SESTAT database allows for analyses of different components of the S&E population.

A question that readily arises when using these or any other HR data sources is, “Who constitutes the S&E workforce?” If analysis is restricted to individuals trained in S&E, the linguist who makes the transition to an S&T occupation is missed but the individual trained in physics working on Wall Street is included. If the analysis is restricted to those working in S&E, the linguist is included but the physicist missed. There appears to be no ready answer to this question of definition, but users are cautioned

11  

The predecessor to the NSCG was the 1982 Post Censal Survey. The sampling frame for this survey was different from the 1993 NSCG, however, being drawn from individuals identified as being in scientific and engineering occupations in the 1980 Census. The 1993 NSCG sample, by contrast, was drawn from all college-educated individuals regardless of occupation reported in the 1990 census.

12  

Given that the sampling frame is based on college education, the NSCG also includes individuals who received their doctoral training in the United States but left, only to return. It also includes medical doctors who, unless they receive a joint MD-PhD, are excluded from the SDR.

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×

to be aware of underlying definitions when using the data. For example, the NSCG could be and has been used to study individuals trained out of the field working in information and technology as well as the proportion of the highly trained financial community who received their training in science. Furthermore, to date none of the NSF surveys tracks individuals who are technically trained but do not have a baccalaureate degree.

Information on those working in S&E occupations can also be obtained from two databases collected by the Bureau of Labor Statistics (BLS). The best known of these databases is the Current Population Survey (CPS), a monthly survey of approximately 50,000 households. Sampled units are asked for basic demographic information concerning all persons residing at the address and detailed labor force information for all persons 15 or over.13 Included are questions related to level of education as well as detailed occupational codes. From the CPS one can obtain information on individuals working in S&E occupations by level of training. One cannot, however, identify individuals trained in S&E.

The Bureau of Labor Statistics also collects data concerning employment by occupation from establishments. Known as the Occupational Employment Statistics (OES) program, these data are collected yearly on wage and salary workers in nonfarm establishments to produce employment and wage estimates for over 750 occupations. The OES program surveys approximately 400,000 establishments per year, taking 3 years to collect the entire sample of 1.2 million establishments. Data are released at the aggregated level.

Broadly speaking, from the six sources described above we are able to obtain information on the training and deployment of individuals working in S&E occupations as well as individuals trained in S&E occupations. Where the surveys are longitudinal, we are able to observe changes over time and thus can examine mobility and earnings patterns over the life cycle. As a general rule, these sources contain little if any information on output measures other than salary or on patterns of collaboration with others working in the field of science and engineering.

13  

To improve the reliability of estimates of month-to-month and year-to-year change, 8 panels are used to rotate the sample each month. A sample unit is interviewed for 4 consecutive months, and, after an 8-month rest period, for the same 4 months a year later. Each month a new panel of addresses, or one-eighth of the total sample, is introduced.

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
The Survey of Doctorate Recipients

Several of the trends discussed in Section II are apparent from an examination of human resource data. For illustrative purposes we use data from the SDR, that span the period 1973 to 1993.14 We use these data to illuminate three trends: (1) the increased importance of industry as an employer of PhDs in the United States; (2) within industry, the decreased importance of the manufacturing sector as an employer; and (3) and the increased deployment of industrial PhDs in non-R&D (or R&D management) positions.

Table C-1 presents the data by field and by year in summary form. The specific categories of interest to us for this study are (a) the percent in industry; (b) the percent in manufacturing of those in industry; (c) the percent in services and “other” of those who report jobs in industry; and (d) the percent in R&D or R&D management of those with positions in industry.15 Because deployment varies considerably by field of training, we present the data for six fields.16 The trends noted above hold in almost all instances across fields. In many instances the change is most noticeable in the early 1980s. Irrespective of field, we find an increase in the percent of PhDs working in industry. In math by 1993 21.7 percent of those trained in the field reported holding a position in industry, almost a three-fold increase over the 20-year period. In engineering and chemistry approximately one out of two PhDs worked in industry by 1993 and the proportion approached this in computer science. This increase in the deployment in industry has come largely at the expense of employment at PhD-granting institutions and reflects in part the poor job market conditions in the academic sector over much of the period.

14  

These data were available at Georgia State University under a licensing agreement at the time this paper was written. The 1995, 1997, and 1999 SDR data are available at the national level and could be incorporated into this analysis.

15  

Changes in survey design and execution affect the types of comparisons that can be performed and the interpretation of observed differences. For example, beginning in 1991 the sample size of the SDR was reduced due to funding constraints and more effort was invested in follow-up. This resulted in a reduction in non-response compared to the 1980s. These modifications and the low-response rate, particularly during the 1980s, compromise the robustness of comparisons that can be drawn over time. At the same time, the SDR is the primary source of national-level employment information for PhDs educated and working in the United States.

16  

The data are restricted to include those with good responses located in the United States. We exclude individuals reporting military employment and those who report that they are retired or out of the labor force.

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×

TABLE C-1 Summary of Doctorate Recipients Data by Field and Year Characteristics of Employed Scientists and Engineers

Year

% in Industry

% in Manufacturing of Those in Industry

% in Service and “Other” of Those in Industry

% in R&D and R&D Management of Those in Industry

Life Science

1973

11.3

74.5

23.5

75.8

1979

13.7

61.9

34.7

59.6

1983

17.0

54.3

41.3

57.8

1989

21.0

53.9

39.2

61.3

1993

24.4

53.3b

Chemistry

1973

39.4

90.9

6.6

85.8

1979

43.9

81.8

14.2

80.0

1983

46.5

72.6

23.0

75.8

1989

48.8

77.2

19.9

75.8

1993

49.7

62.9b

Physics and Astronomy

1973

23.4

69.9

17.2

90.5

1979

28.6

65.2

27.5

78.3

1983

34.8

49.7

31.3

77.2

1989

34.4

48.8

38.3

76.5

1993

30.7

53.3b

Math

1973

07.9

59.9

21.6

80.1

1979

14.2

45.0

36.0

66.5

1983

19.4

39.8

39.9

67.9

1989

20.9

35.3

46.2

59.4

1993

21.7

35.2b

Computer Science

1973

a

a

a

a

1979

a

a

a

a

1983

45.1

49.0

45.3

83.1

1989

43.6

42.3

46.0

75.0

1993

44.8

38.9

Engineering

1973

44.0

73.1

17.6

83.8

1979

48.9

65.9

25.1

74.7

1983

52.2

53.7

37.8

70.3

1989

51.3

53.7

35.2

70.4

1993

51.8

56.3b

a Unweighted cell size below 30.

b The 1993 data are not comparable to the earlier years because counts include only those engaged in R&D; R&D management as a work activity was not distinguished from a generic management category in 1993.

SOURCE: Survey of Doctorate Recipients, NSF, 1973–1993.

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×

Consistent with the patterns noted above, we see that among those working in industry the percent in manufacturing has declined over the period of observation.17 For example, in physics and astronomy two out of three PhDs in industry were working in manufacturing in 1973; by 1989 that share was about one-half. The trend is even stronger in math, engineering, and computer science. Even in chemistry, where the vast majority of PhDs working in industry were in manufacturing in the 1970s, we see a decline, although the decline had begun to reverse itself by 1989.

The industrial sectors where the increase has been dramatic include service and “other,” combined here because of the ambiguity of definitions used across the years.18 In all but the life sciences, the propor-tion in industry working in these combined sectors more than doubled during the 16 year interval portrayed in Table C-1. In the life sciences the percent grew by more than 60 percent.

The final trend shown in Table C-1 is the decline in deployment, among those working in industry, in R&D and R&D management activity. This pattern is consistent with the observation that innovative activity is increasingly moving out of the lab and into other positions within the firm. Unfortunately, the 1993 survey failed to collect data on R&D management and thus the data for 1993 are not comparable with that for 1973-1989.

R&D activity, as measured by expenditures, is heavily concentrated in a small number of states (National Science Board, 1998, pp. 4-30). For example, one-half of the $177 billion spent on R&D in the United States in 1995 was expended in just six states: California, Michigan, New York, Massachusetts, New Jersey, and Texas (National Science Board, 1998, pp. 4-30). The top 11 states (adding Illinois, Pennsylvania, Maryland, Ohio, and Washington) perform two-thirds of all R&D. By contrast, the bottom 20 states produce less than 5 percent of the R&D conducted nationwide in 1995.

17  

Unfortunately, the 1993 survey did not collect information on industrial classification.

18  

The industrial classification used by the SDR changed substantially between 1983 and 1989. In particular, in 1973 and 1979 three-digit SIC industry codes were used. In 1983 four-digit codes were used. These were simplified considerably in the 1989 questionnaire where only two-digit codes were used. The tremendous increase in the proportion of PhDs reporting that they worked in the service sector between 1983 and 1989 and the decrease of those reporting working in the “other” sector suggests that when the classification was simplified, a considerably larger proportion of PhDs classified themselves as in the service sector. The sectors supposedly excluded from “other” across all periods are construction, manufacturing, mining, transportation, communication and utilities, wholesale and retail trade, and finance, insurance, and real estate. In 1991 the industrial classification was no longer done by coders but instead by the respondent. In 1993 the survey collected no information on industrial classification directly although the name of the employing institution is part of the record.

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×

TABLE C-2 Geographic Distribution of Doctorate Scientists and Engineers in the Labor Force

 

All Fields

 

Percentage of Labor Force in Selected Years

Top Eleven States

1973

1979

1983

1989

1993

California

11.08

12.05

12.61

13.67

13.82

Illinois

04.34

04.39

04.44

04.22

04.07

Massachusetts

04.29

04.24

04.19

04.76

04.63

Maryland

05.46

05.73

05.48

05.92

06.23

Michigan

03.43

03.22

03.17

03.26

02.91

New Jersey

04.69

04.66

04.91

04.72

04.84

New York

09.67

09.10

08.50

07.85

07.27

North Carolina

02.81

Ohio

04.47

04.07

04.20

04.05

03.63

Pennsylvania

05.42

04.74

04.81

04.81

04.64

Texas

05.15

05.59

05.74

05.69

06.11

Virginia

02.60

02.92

02.98

02.94

 

SOURCE: Survey of Doctorate Recipients, NSF, 1973–1993.

The deployment of PhDs is also heavily concentrated, although not to the extent that R&D expenditures are. In 1993, for example, the top six states employed 43 percent of all PhDs; the top 11 employed 60.7 percent and the bottom 20 states 8.6 percent. Little change in the concentration of PhDs’ employment, at least as measured by the SDR, has occurred over the 20-year period 1973–1993. In 1973 the top 6 states employed 41.3 percent, the top 11 employed 60.7 percent, and the bottom 20 employed 8.5 percent. These figures, however, mask certain changes that have occurred in deployment among the top states. Most notably, as we see from Table C-2, the gap between the top employing state, California, and the second state, New York, has widened considerably during the 20-year interval.

These data provide one means of examining geographic deployment—focusing on the U.S. doctoral-trained labor force—but fail to reveal the deployment patterns of foreign-trained PhDs. More importantly, they tell us nothing about employment patterns of the nondoctoral S&E workforce. Many of the changes noted above, however, undoubtedly reflect changes in the deployment of this portion of the workforce and we could undoubtedly learn something about these changes if these data were readily available by geographic location.

The blurring of boundaries between industry and academe and the extent of knowledge spillovers from academe to industry and vice versa make

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×

TABLE C-3 Percentage of Doctorate Scientists and Engineers Employed in Industry Who are Located in the Same State as their PhD Institution

All Fields

1973

1979

1983

1989

1993

All Doctorates

25.6

25.0

25.9

26.1

26.2

5–6 Year Cohort

25.2

28.3

27.2

27.7

29.3

 

SOURCE: Survey of Doctorate Recipients, NSF, 1973–1993.

the geographical deployment of scientists and engineers of particular interest. This is especially the case in situations where tacit knowledge plays a key role. One way in which knowledge transfers are fostered between industry and academe is through the placement of graduate students. Not only do new PhDs bring new ideas, but they help to build and maintain effective networks between industry and academe.

This raises the question of whether data on PhDs can shed light on changing patterns of deployment of new PhDs. We are particularly interested in changes in the percent of PhDs that accept industrial em-ployment within the same state of training.19Table C-3 summarizes these data. We see that while there has been no change for all doctorates when analyzed jointly, there is a slight increase for the 5-6 year out cohort, suggesting that the proclivity of newer PhDs for working for industry in their state of education has increased.

Linking HR Data with Other Data

As revealing as these data are concerning changes in innovation practices, they leave numerous unanswered questions that might be answered if available HR data were linked to other databases. Consider what is known versus what could be learned regarding scientists and engineers working in industry. Usually only one question is asked concerning industrial classification and this question is not consistently ascertained over time nor is it coded consistently over time. Furthermore, no infor-mation is available on characteristics of the employer. In contrast, we know much more about characteristics of employers for PhDs working in academic in-

19  

This is obviously an imprecise measure given that many cities spill across state lines.

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×

stitutions. Much of this information (PhD-granting, Carnegie classification, etc.) comes not from the respondent but instead from matching the code of the employing academic institution with data collected on educational institutions. A similar matching could occur for industry employment. For example, respondents in both the SED and SDR surveys are asked to supply the name and location of the employing oganization. By matching this information with Census establishment data we could get more detailed information on firm characteristics including size.20 Information provided by such a link could illuminate the relationship between firm size and innovation as well as the source (external vs. internal) of innovation in small firms. Such data could also go a long way to addressing the “line of business” issue mentioned earlier.21

Only one measure of output has been consistently collected in human resources databases of scientists and engineers and that is salary. We have little information on non-salary components of income, including stock options. More importantly, we have no indication of the respondents’ productivity, as measured either by article counts and the citations associated with these articles or by patent counts.22 Neither do we have information on the productivity of the firm for which the in-dividual works, as measured by traditional indicators of firm per-formance. Information on such dimensions could potentially be obtained by linking HR data with other databases. For example, Levin and Stephan (1991) had the SDR linked with publication data from the Institute for Scientific Information (ISI). Similar linkages could be made concerning patents.

Not only do we know very little about output measures; we also know little about inputs. The individual-based data that are collected provide minimal information about resources that are available to scientists and engineers, for example in the form of research grants from funding agencies or research budgets of firms. In theory, such information could be obtained by linking HR data to databases such as the Computer Retrieval of Information on Scientific Projects (CRISP) file of the National Institutes of Health which provides information on grants.

Another deficiency is that the existing HR data do not allow us to examine the degree to which scientists wear “multiple” hats in the sense that they work for more than one institution—an increasingly common

20  

Firm size has been collected in the SDR since 1995.

21  

Respondents are asked to supply the actual location where they work, which is often different from that of the corporate headquarters. Employer name is not coded for respondents working in industry but is retained in the record.

22  

The exception is that the 1995 SDR asked a question with regard to the number of papers and articles authored since 1990. It also asked a question with regard to whether the respondent had been named as an inventor on any patent application since 1990.

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×

phenomenon given the amount of outsourcing from industry to academe and the number of academics who work with start-up firms.23 Perhaps more importantly, we are unable to determine from survey data with whom scientists collaborate and thus cannot use the data to study characteristics of collaborative patterns such as geographic proximity and fields of complementarity. Nor can we determine how individuals are recruited into collaborative relationships, especially those in industry. Data linkage might enable us to trace how investments made by government labs in training researchers spill over to other sectors as the trainees leave to take positions outside the government. Data on patterns such as these would not only give us insight into changing patterns of innovation. They would also help us evaluate the impact of government programs designed to train scientists and foster the productivity of existing scientists and engineers.

SECTION IV. LESSONS LEARNED FROM THE STUDY OF BIOTECHNOLOGY

By far the best example of what can be learned by examining linkages based on human resource data comes from the study of biotechnology firms. Zucker and Darby have contributed a great deal to our under-standing of what can be learned through the careful analysis and linkage of data. Stephan has also examined several issues that can only be studied by linking individual with firm level data. Here we summarize the approaches and results of each.

Zucker and Darby have constructed a rich database and used it to advance our understanding of how and why new firms in biotechnology are established and locate in certain areas. Their work also informs our understanding of how firms and scientists outside these firms benefit from collaboration. The construction of this database has several elements. A key component was the construction of a measure of intellectual capital in biotechnology. This was done by identifying leading researchers, termed “stars,” on the basis of the number of genetic sequence discoveries reported up to 1990 for which they were an author. Characteristics of these scientists, such as employing institution, were determined as well as characteristics of their coauthors regardless of whether the latter qualified to be stars (Zucker et al., 1994). Zucker and Darby have also collected considerable data on biotechnology firms and on research resources located in the same geographic area of the firm. Included in the latter data are the proximity of highly rated university departments and the level of spending on R&D.

23  

Since 1993, however, the SDR has asked for information on a secondary employer.

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×

The construction of this type of database has contributed to a number of insights into the development of the biotechnology industry and the human resource dimensions of this development. Key findings include, but are not limited to, the following:

  • Over time the proportion of stars and active collaborators working primarily in universities rather than firms has declined significantly, from nearly 100 percent initially to less than 50 percent by 1989.

  • The growth and geographic location of intellectual capital was the principal determinant of the growth and location of the biotechnology industry (Zucker et al., 1994, p. 29).

  • The extent of collaboration by a firm’s scientists with stars is a powerful predictor of firm success as measured by products in development and on the market as well as the number of employees (Zucker and Darby, 1995b).

  • Commercial involvement by stars is associated with increased research productivity as measured by article citations (Zucker and Darby, 1995b, p.18).

  • The higher the quality of the star, the shorter is the time that the star remains at a university before moving into the biotechnology industry, other things being equal (Zucker et al., 1997).

Stephan’s research focused on biotechnology firms that made an initial public offering during the hot market of the early 1990s. She used the prospectuses of the 50-odd firms to ascertain the names of the in-dividuals with scientific training affiliated with the firm. Although especially interested in scientists who give an academic address as their primary employer, she also collected data on full-time employees of the firm who are listed in key positions as well as the names of founders of the firm. Stephan then determined the citation counts of the scientists and key demographic information such as date of birth, country of birth, educational training, work history, geographic location of primary employer, and whether or not a Nobel Prize recipient. Various databases were linked and, in some instances, the scientists were asked for missing pieces of information. Stephan used the CRISP data to measure stock prices over time and the Investnet CDA database to determine the extent to which “insiders” engage in trades and profit taking.

Key findings from this work, which could only be ascertained through a human resources lens, include the following:

  • Although a substantial number of university-based scientists participate in networks that are geographically bounded, approximately 70

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×

percent of the links between biotechnology companies and university-based scientists are nonlocal (Audretsch and Stephan, 1996).

  • Reputation of the scientists is positively related to day-one performance of the initial public offering. Proceeds raised also relate positively to reputation (Stephan, 1999).

  • Scientific founders who come from academe are older and more highly cited than those who come from drug companies (Audretsch and Stephan, 1999).

  • Approximately 10 percent of the university-based scientists held sufficient amounts of stock to qualify as “insiders” by the Securities and Exchange Commission. Among this group, it was not uncommon to engage in stock market activity that yielded handsome capital gains (Stephan and Everhart, 1998).

The work of Zucker and Darby and of Stephan shows the richness of results that can be obtained by linking data on scientists to indicators of their productivity such as citations and then linking this information either directly or by geographic indicator to firm data. No other industry appears to have garnered such attention and for no other industry have such intricate linkages based on human resource data been constructed.24

SECTION V. CITATION ANALYSIS

Publications

On a much larger scale, the bibliometric work of Hicks and her coauthors provides insights into changing patterns in innovation and, although the data are not based on human resource survey data, scientists and engineers generate the data for this work on authorship patterns. Hicks and coauthor Katz, through the use of bibliometrics, identify several changes occurring in the production of scientific papers. Key to the methodology

24  

Sleeper (1998) examines characteristics of founders of de novo firms in the laser industry. She finds that one-quarter of the founders came from university and government labs; close to half from the laser industry; a sixth from industry outside of lasers and an eighth are not identified. De novo firms have a higher exit rate than established firms that went into laser production. Within the de novo class of firms, those with founders from the laser industry have the highest survival rate followed by firms established by founders from universities and government labs. Sleeper is able to determine the department of origin of 27 of the 36 university founders. She finds that the firms founded by the 12 scientists coming from physics departments have a higher survival rate than other firms, including those that originated as laser spin-offs.

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×

they employ is the fact that articles list names of authors and addresses for authors. Hicks uses the address information to classify by sector each of the 376,226 papers indexed in the Science Citation Index (SCI) with a United Kingdom address published during the 11-year interval of 1981–1991. A related methodology is used by CHI to analyze publishing patterns for Science and Engineering Indicators.

Prominent trends include the following:

  • Increased multiple authorship. Although the growing importance of coauthorship has been observed for a long time,25 Hicks and Katz (1996b) document that the increase is attributable to growth in papers with four or more authors; the proportion of papers with one or two authors was in decline during the period of study while the share of papers with three authors remained steady. Such detail is not readily available for U.S. articles but the incidence of multiauthored papers has risen from 45 to 56 percent during the period 1981–1995 (National Science Board, 1998, Table 5-53).26

  • Increased intersectoral collaboration. Hicks and Katz find that during the 1980s U.K. papers published by authors located at a single institution did not grow while the number published by authors working in more than one institution rose steadily. Specifically, by the end of the period of study the proportion of collaborative papers rose from 28 percent of all U.K. papers to 41 percent. Increased intersectoral collaboration is occurring in the United States as well (National Science Board, 1998, pp. 5-38).

  • Increased collaboration between industry and universities. During the 1981-1991 period, the percent of U.K. industry papers that included a university address rose from slightly less than 20 percent to slightly less than 40 percent (Hicks and Katz, 1997a, p. 138). Similar trends are occurring in the United States (National Science Board, 1998, pp. 5-38).

  • Change in publishing patterns within industry. Hicks and Katz document that manufacturing is not the only sector of the U.K. industry that publishes. Indeed, taking the Times 1000 companies as their base they find that the only sector in which all companies (in this case 10) published during the period was “water.” Viewed from the perspective of the percent of all industrial publications, several nonmanufacturing sectors—e.g. oil, gas and nuclear fuels, engineering, electricity—produce a sizeable pro-

25  

de Solla Price (1986) analyzed the number of authors on papers listed in Chemical Abstracts from 1900 to 1960 and found that the proportion of single-authored papers began decreasing in the 1920s.

26  

Strickly speaking the comparison is for the period 1981–1985 and 1991–1995.

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×

portion of industrial publications (Hicks and Katz, 1997a, p. 32). In the United States industry publications almost doubled between 1981 and 1995 in clinical medicine and tripled in biomedical research. U.S. industry publications in physics, chemistry, technology, and mathematics all declined during the 1990s (National Science Board, 1998, pp. 5-38).

  • Increased international collaboration. Hicks and Katz also find that for articles having at least one U.K. address the average number of countries per article increased during the period 1981 to 1991, from 1.17 to 1.25 (1996a, p. 390). The increase in U.S. scientists’ participation in international collaborative research is seen by the fact that the proportion of articles with one or more U.S. addresses along with a non-U.S. address rose from 9 to 16 percent during the period 1981–1995 (National Science Board, 1998, Table 5-53).

Bibliometric research holds remarkable promise for using human resource data to study innovation if links can be made between biblio-metric information and data collected in such surveys as the SDR with files from funding agencies concerning the amount and source of research support. For example, if we were able to make such linkages we would gain insight into whether collaborations stem from attendance at the same graduate school, work with a dis-sertation advisor, work as a post doc, or work with a former employer. Understanding how these collaborative relationships are formed is crucial given the increasing importance of networks and the great difficulty faced by researchers in tracking informal relationships. Of course, not all informal relationships produce papers, but papers are one indication that a relationship exists.27

Patents

Patent applications in the United States include references to U.S. and foreign patents as well as “other” references, many of which are published articles. In recent years there has been guarded but growing interest in using these citations to study the process of innovation.28 This interest is

27  

Zucker and Darby quote a manager as saying: “ Copublishing is about as good an indicator as you can get of commonality of interests between [the company] and an academic collaborator. Although formal relationships are on a publicly available list, many relationships are not publicly acknowledged.” The investigators continue, “ In this and other fieldwork we have repeatedly validated the usefulness of linking academic scientists to firms by bibliometric research on patterns of co-publication… This concept of linkage is powerfully predictive of firm success when academic star scientists are involved.” (Zucker and Darby, 1995, p. 22).

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×

driven in part by the fact that the use of patent counts as an indicator of innovation has not lived up to expectations because of what Trajtenberg (1990, p. 172) refers to as “the enormous variance in their ‘importance’ or ‘value.’” Weighting patent counts by the citations received in subsequent patent applications, however, provides a means of attributing value to the patent. Trajtenberg (1990) demonstrates that this weighted measure performs much better as an indicator of the value of innovations than does total patent count.29 Hall, Jaffe, and Trajtenberg (1998) provide evidence that a citation-weighted patent measure is contemporaneously associated with market value. To date no one has attempted to link authors of cited and citing patents to analyze characteristics of the spillover process that could be learned through this human resource link.

Considerable information can also be gleaned from articles cited in patent applications.30 Narin, Hamilton, and Olivastro (1997) use these cites to determine the origin of the basic science that underlies the patent. Their analysis shows that 73 percent of the references to published articles were to “public” science—that is science authored at academic, governmental, and other public institutions. They also find that the number of references to public science nearly tripled during the 6-year period studied. The research also indicates that NSF was the most widely acknowledged support agency in cited chemistry, physics, and engineering papers. NIH was the most widely cited in biomedical papers.

More recently, Narin collaborated with Deng and Lev (Deng et al., 1999) to demonstrate how the use of patent citation information adds to our understanding of the performance of firms in capital markets using such measures as stock returns and market-to-book ratios. Three mea-

28  

Hall, Jaffe, and Trajtenberg (1999, p. 5) discuss the substantial “noise” in patent citation data. In addition to cites that are included in the patent document because of the knowledge linkage, citations are also made for legal reasons. There are also what they call “after the fact” citations—those added to the document after the actual invention and what they call “teaching” cites, those that everyone considers basic. In addition, the patent examiner may also require the addition of relevant citations to “further bound the scope of intellectual property rights conferred by the patent, even though the inventor may not have been aware of the patent to which the citation is added.”

29  

The dependent variable in the Trajtenberg study is the gains accruing to the representative consumer as well as the total gains to all consumers. Trajzenberg focuses his study on the CT scanner industry.

30  

Deng, Lev, and Narin (1999 p. 21) report that a typical U.S. patent cites about eight earlier U.S. patents and one to two foreign patents. In addition, the typical patent cites one or two nonpatent references, the majority of which are science references. In recent years, there has been a steady increase in the number of patents referenced and the citations to science, the latter having grown from an average of .3 per patent application in 1985 to 2.0 in 1997 with a tripling of citations to U.S. scientific papers in the 1988–1994 span.

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×

sures derived from the citations are constructed: one measures the importance of the patent through the use of forward citations to the patent; another measures the link between the patent and science by measuring the number of references to scientific papers; and a third indicator measures the median age of patents referenced in the application and is developed as a measure of how quickly a technology is evolving.

The authors find the three patent measures as well as a measure of patent counts to be significantly related to market-to-book value with the expected sign. Of particular interest is the fact that none of the three patent variables contain information that is obtainable in the companies’ financial reports. The stock return results behave somewhat similarly, although the association between patent attributes and subsequent stock returns is generally weaker than the association between patent attributes and market-to-book value.

The patent studies are indicative of the richness of insights that can be gained by using existing information in creative ways. Additional insights could undoubtedly be gained by linking these data with human resource data. The funding link established by Narin is suggestive, but substantially more could be learned if we were able to make human resource links. For example, one could establish the educational origins of individuals authoring highly cited patents and more readily determine sources of funding for the research of cited articles by using the name as the link between publication and funding agency. Similar knowledge could be gained if we were able to establish human resource links on a regular basis with publication and citation information, as collected by ISI.

SECTION VI. CONCLUSION

Four broad changes have occurred in the structure and organization of innovative activity in the United States: (1) decreased role for federal funding of R&D; (2) a change in the industrial distribution of innovative activities; (3) a shift of resources toward development activities and away from basic research; and (4) a change in the organization of research. The latter change reflects an increased reliance on external R&D, increased collaboration in the development of new products and processes, a decentralization of in-house R&D activities, and the movement of innovative activities to functions in the firm typically not thought of as being drivers of innovation. These changes mean that traditional indicators of R&D as well as the traditional unit of analysis, the firm, are less relevant to the study of innovation than they once were.

Here we have explored how human resource data can be used to illuminate patterns of innovation and resource utilization and, perhaps more

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×

importantly, what could be learned if the data that we already have were linked, providing insights into networks and the avenues by which collaborative ventures are formed and knowledge moves across boun-daries. Such links would provide a clearer picture of the process of innovation and the causes of economic growth. A nontrivial benefit from such research is that it offers the possibility of providing a clearer under-standing of how the investments of government and the nonprofit sector contribute to economic growth.

Our preliminary investigation shows that certain changes in the structure and organization of innovative activity can readily be seen by using HR data. For example, the deployment data show a change in industrial mix that R&D data support but fail to fully capture. These trends could be more clearly discerned if the firm address were carefully coded.31 We also find that the HR data provide insight into the movement of innovative activities to non-R&D functions in the firm. The HR data also allow us to see how the geographic distribution of scientists and engineers changes over time.

Much more could be learned if we were able to link this HR data to other already established databases. For example, we know a great deal about firms but have made no effort to link information on firms to HR data collected by such agencies as NSF. Publication and citation data are readily available going back for a number of years. In addition to containing author name and journal name, the SCI data also include address information for authors. These data have rarely been linked to HR data. The lack of linkage means that we have forfeited information on the nature of collaboration between industry and academe as well as the way in which firms based on new technologies are founded. A similar case can be made for patent data. We have long used patents as an indicator of innovation but only recently have become interested in using patent citations to learn about the science linkages of inventions as well as to measure the importance of the patent. In the past linkages such as these appeared to be luxuries; but as boundaries and roles continue to blur they cease to be luxuries. In a world where the process of innovation is radically changing, valuable information is being lost by our failure to create and analyze HR linkages in a systematic way.

31  

NSF reports that in many instances the employer address data (city, state, zip) are missing and that the employer name is reported in unclear acronyms. This suggests that an effort would need to be made to get cleaner information from respondents if matching were to occur in the future.

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×

REFERENCES

Acs, Zoltan J., David Audretsch, and Maryann Feldman. 1992. “Real Effects of Academic Research: Comment.” American Economic Review, 82(l): 363-67.

Audretsch, David, and Paula Stephan. 1996. “Company-Scientist Locational Links: The Case of Biotechnology.” American Economic Review, 86(3): 641-52.

_____. 1999. “Knowledge Spillovers in Biotechnology; Sources and Incentives.” Journal of Evolutionary Economics, 9(1): 97-108.


Badaracco, Joseph. 1991. The Knowledge Link. Boston: Harvard Business School Press.


de Solla Price, Derek. 1986. Little Science, Big Science and Beyond. New York: Columbia University Press.

Deng, Zhen, Baruch Lev, and Francis Narin. 1999. “Science and Technology as Predictors of Stock Performance.” Financial Analysts Journal, 55(3): 20-32.


Finn, Michael, Leigh Ann Pennington, and Kathryn Hart Anderson. 1995. “Foreign Nationals Who Receive Science or Engineering PhDs from U.S. Universities: Stay Rates and Characteristics of Stayers.” Oak Ridge, TN: Oak Ridge Institute for Science and Education.

Friar, J., and Horwitch, M. 1986. “The Emergence of Technology Strategy: A New Dimension of Strategic Management,” in Mel Horwich, ed. Technology in the Modern Corporation: A Strategic Perspective. New York: Pergamon Press.


Hagedoorn, John. 1993. “Understanding the Rationale of Strategic Technology Partnering.” Strategic Management Journal, 14: 371-385.

Hall, Bronwyn, Adam Jaffe, and Manuel Trajtenberg. 1998. “Market Value and Patent Citations: A First Look.” Working paper, University of California at Berkeley.

_____. 1999. “Market Value and Patent Citations: A First Look.” Unpublished paper.

Hamel, Gary. 1991. “Competition for Competence and Inter-Partner Learning within International Strategic Alliances.” Strategic Management Journal, 12: 83-103.

Hicks, Diana, and J. Sylvan Katz. 1996a. “Science Policy for a Highly Collaborative Science.” Science and Public Policy, 1: 39-44.

_____1996b. “Where is Science Going?” Science, Technology & Human Values, 21(4): 379-406.

_____1997a. “The Changing Shape of British Industrial Research.” STEP Special Report no. 6.

_____1997b. “A National Research Network Viewed from an Industrial Perspective,” Revue D’Economie Industrielle, 79(l): 129-142.


Jaffe, Adam B. 1989. “Real Effects of Academic Research.” American Economic Review, 79(5): 957-70.


Lev, Baruch. 1999. “R&D and Capital Markets.” Journal of Applied Corporate Finance, 11(4): 21-35.

Levin, Sharon G., and Paula E. Stephan. 1991. “Research Productivity Over the Life Cycle: Evidence for Academic Scientists.” American Economic Review, 81: 114-32.

Leyden, D.P., and A.N. Link. 1999. “Federal Laboratories as Research Partners.” International Journal of Industrial Organization, 17(4): 575-592.


Merrill, Stephen A., and Ronald S. Cooper. 1999. “Trends in Industrial Research and Development: Evidence from National Data Sources” in National Research Council. Securing America’s Industrial Strength. Washington, D.C.: National Academy Press.

Mowery, David C. 1999. “America’s Industrial Resurgence: How Strong, How Durable?” Issues in Science and Technology, 15(3): 41-48.


Narin, Francis, Kimberly S. Hamilton, and Dominic Olivastro. 1997. “The Increasing Linkage between U.S. Technology and Public Science.” Research Policy, 26(3): 317-330.

National Science Board. 1998. Science & Engineering Indicators1998. Arlington, VA: National Science Foundation.

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×

Nerad, Maresi, and Joseph Cerny. 1999. “Postdoctoral Patterns, Career Advancement, and Problems.” Science, 285: 1533-1535.


Powell, Walter W., and Jason Owen-Smith. 1998. “Universities and the Market for Intellectual Property in the Life Sciences.” Journal of Policy Analysis and Management, 17(20): 253-277.

Powell, Walter W., Kenneth W. Koput, Laurel Smith-Doerr, and Jason Owen-Smith. 1998. “Network Position and Firm Performance: Organizational Returns to Collaboration in the Technology Industry.” Unpublished paper.


Saxenian, Annalee. 1994. Regional Advantage. Cambridge, MA: Harvard University Press.

Sleeper, Sally. 1998. The Role of Firm Capabilities in the Evolution of the Laser Industry. The Making of a High-Tech Market. PhD Dissertation, Department of Social and Decision Sciences, Carnegie Mellon University.

Stephan, Paula. 1999. “Capitalizing on Human Capital: The Case of Biotechnology.” Unpublished paper.

Stephan, Paula, and Stephen Everhart. 1998. “The Changing Rewards to Science: The Case of Biotechnology.” Small Business Economics, 10: 141-151.

Stephan, Paula, and Sharon Levin. 1991. Striking the Mother Lode in Science: The Importance of Age, Place, and Time . New York: Oxford University Press.


Trajtenberg, Manuel. 1990. “A Penny for Your Quotes: Patent Citations and the Value of Innovations.” Rand Journal of Economics, 21(l): 172-187.


Zucker, Lynne, and Michael Darby. 1995a. “Present at the Revolution: Transformation of Technical Identity for a Large Incumbent Pharmaceutical Firm after the Biotechnological Breakthrough.” NBER, Working Paper No. 5243.

_____. 1995b. “Virtuous Circles of Productivity: Star Biochemists and the Institutional Transformation of Industry.” NBER, Working Paper No. 5342.

_____.1997. “Academic Scientists-Entrepreneurs and Commercial Success in Biotechnology.” Unpublished paper.

Zucker, Lynne, Michael Darby, and Marilyn Brewer. 1994. “Intellectual Capital and the Birth of U.S. Biotechnology Enterprises.” National Bureau of Economic Research, Working Paper No. 4653.

Zucker, Lynne, Michael Darby, and Maximo Torero. 1997. “Labor Mobility From Academe to Commerce.” NBER, Working Paper No. 6050.

Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 43
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 44
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 45
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 46
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 47
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 48
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 49
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 50
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 51
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 52
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 53
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 54
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 55
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 56
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 57
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 58
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 59
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 60
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 61
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 62
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 63
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 64
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 65
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 66
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 67
Suggested Citation:"Appendix C: 'Using Human Resource Data to Illuminate Innovation and Research Utilization' by Paula Stephen." National Research Council. 2002. Using Human Resource Data to Track Innovation: Summary of a Workshop. Washington, DC: The National Academies Press. doi: 10.17226/10475.
×
Page 68
Using Human Resource Data to Track Innovation: Summary of a Workshop Get This Book
×
Buy Paperback | $29.00 Buy Ebook | $23.99
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

Despite the fact that technology is embodied in human as well as physical capital and that interactions among technically trained people are critical to innovation and technology diffusion, data on scientists, engineers and other professionals have not been adequately exploited to illuminate the productivity of and changing patterns in innovation. STEP convened a workshop to examine how data on qualifications and career paths, mobility, cross sector relationships, and the structure of work in firms could shed light on issues of research productivity, interactions among private and public sector institutions, and other aspects of innovation.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    Switch between the Original Pages, where you can read the report as it appeared in print, and Text Pages for the web version, where you can highlight and search the text.

    « Back Next »
  6. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  7. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  8. ×

    View our suggested citation for this chapter.

    « Back Next »
  9. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!