1 Introduction

Academic entrepreneurship—the establishment of new ventures to commercialize technology derived from university research—has long been recognized as a critical activity at research universities and as a fundamental contributor to regional economic development and dynamism. Though scholars have focused on academic entrepreneurship for nearly four decades (e.g., Doutriaux, 1987), we have little understanding about when and how this category of scholarship emerged over time.

The origins of technology transfer research can arguably be traced back at least to Gilmore (1969, p.2) who defined the concept as “a purposive, conscious effort to move technical devices, materials, methods, and/or information from the point of discovery or development to new users.” Table 1 includes several early definitions of technology transfer by academic researchers and policy organizations.Footnote 1

Table 1 Selected definitions of technology transfer (listed chronically)
Full size table

Given our interest in the origins of academic entrepreneurship,Footnote 2 we note that except for WIPO, the listed definitions in Table 1 do not describe technology transfer as an activity of focus for universities or other research institutions. Yet, one might also surmise from the definitions in the table that technology transfer does include elements—such as knowledge development (e.g., inventions or innovations), intellectual property (e.g., patents), and scientific outcomes (e.g., scientific publications)—that are constituent considerations within the academic entrepreneurship literature.Footnote 3

This paper therefore investigates the emergence of academic entrepreneurship within the context of the technology transfer literature. We began our study with the maintained assumption that researchers themselves are an important source, if not the most appropriate source, of information to rely on to characterize the scope of research related to technology transfer. The Journal of Technology Transfer (JTT) is arguably the premier academic journal primarily devoted to the study of technology transfer.

The JTT was established in 1977.Footnote 4 Its purpose was to publish conference proceedings from periodic meetings of the Technology Transfer Society; members of the Society primarily represented technology transfer personnel from federal laboratories located within the United States. Editorial responsibility for the JTT rotated among non-academic members until 1996,Footnote 5 when A. Link was invited to be an editor of the journal.Footnote 6 In 1998/1999 Link invited D. Siegel and M. Wright to join as co-editors, and Link became Editor-in-Chief. Also during this transition, the publication of the journal changed from a self-published society publication to an academic journal managed and published by Kluwer Academic Publishers. Kluwer Academic Publishers later merged with Springer-Verlag (now Springer). After 1998/1999, others were invited to join the editorial structure as co-editors and co-associate editors. The broadly defined focus of the JTT did not change with the new editorial structure, but the preferred methodology and methods of research did change to reflect less of a practitioner perspective and more of an academic perspective.

Given the journal’s history and evolution, we view the JTT as an appropriate source from which to obtain relevant information to address the research questions considered in this paper.

The remainder of this paper is organized as follows. In Sect. 2, we describe the data used to characterize the scope of published research related to technology transfer, and we identify keywords from the titles of published papers used by authors to define the activities of emphasis in their study. In Sect. 3, we posit an econometric model from which to investigate changes over time in the use of researchers’ keywords in the titles of their published papers that describe their published technology transfer research, and we discuss our findings. In Sect. 4, we offer a summary statement, and we conclude that if past is prologue, our analysis might represent a roadmap for future research.

2 The data

Springer, the publisher of the JTT, graciously provided a database from which we could extract the title of all published papers in the JTT, by calendar year.Footnote 7 The database begins with the first issue of the journal in 1977 and goes through mid-2022. Over these decades there were 1649 titled entries in the JTT. Given the purpose of our research, we deleted a number of these entries. Those deleted were book reviews, editorials, errata statements, and introductions to special or symposium issues. The resulting sampling population contains 1578 published papers. Based on a word search of published paper titles, the most commonly used terms by authors focus on dimensions of university activity. Using the designated keywords of university, universities, and academic, we arrived at 398 uniquely titled papers over the calendar years 1977 through mid-2022.Footnote 8 These 398 titled papers represent 25.2% of the sampling population of published papers.

Figure 1 shows, by calendar year, the total number of published papers and the number of published papers using at least one of the three designated keywords. The precipitous drop in both metrics in 2022 is due to data being available for only a part of that calendar year.Footnote 9 The trend in both metrics is visually positive, and the two lines of trend appear to move together.

Fig. 1
figure 1

Number of published papers and number of published papers with a Keyword of University, Universities, or Academic in the Title, by Calendar Year

Full size image

Figure 2 shows the proportion of published papers using at least one of the keywords of university, universities, or academic, by calendar year. The trend in this metric is also positive. Noticeable in the figure is that in 2001 over one-half of the titled papers were related to university activity. In that year, the majority of these papers were part of a special double issue that focused on organizational issues in university-industry technology transfer.

Fig. 2
figure 2

Proportion of published papers with a keyword of University, Universities, or Academic in the Title, by Calendar Year

Full size image

3 Empirical analysis

Empirical models were considered to identify covariates with both the calendar year number of papers with a title using a designated keyword (NumberUses) and with the calendar year proportion of total published papers represented by that metric (ProportionUses). The key variables for the estimation of these two models are defined in Table 2, and descriptive statistics on these variables are presented in Table 3.

Table 2 Definition of calendar year key variables
Full size table
Table 3 Descriptive statistics on calendar year key variables (n = 46)
Full size table

Table 4 shows the least-squares regression results from each model to explain the trend over time in the number of published papers using at least one of the designated keywords (column (1)) and the trend in the proportion of published papers using at least one of the designated keywords (column (2)). For the first model, the total calendar year number of published papers, NumberPublishedPapers, is held constant. The calendar year number of published papers using at least one of the designated keyword appears from Fig. 1 to be related to the total calendar year number of papers. One might expect, based on Fig. 1, that these two metrics are positively correlated. Held constant in both models are three control variables.

Table 4 Least-squares regression results based on the data underlying Figs. 1 and 2 (p values in parentheses)
Full size table

The research environment provided through the Bayh–Dole Act of 1980 is one control variable. The act provided universities right of first refusal for title to federally funded research results, and many scholars have associated the passage of this act with increased research on university patenting activity.Footnote 10 The effect of the Bayh–Dole Act on papers published in the JTT that relate to university and academic behavior, BD, is assumed for estimation purposes to be present for a decade (see Table 2). We hypothesize that the estimated coefficient on BD will be positive.

Given the editorial changes discussed above, post-1996 editorial dynamics are denoted in the models by the variable ED. We offer no hypothesis about the algebraic sign on ED.

Finally, the post-Covid-19 years are also held constant by the variable Covid. In the post-2019 period it became difficult for editors of all journals to find suitable reviewers for submitted papers. It is an empirical issue as to whether that factor affected the topic selection of researchers who, in many instances, were transitioning to online teaching, or the time for a review of submitted papers to be completed. Thus, we offer no hypothesis about the algebraic sign on Covid.

The regression results in column (1) of Table 4 suggest that annual changes in NumberPublishedPapers are positively and significantly related to annual changes in NumberUses. Holding constant NumberPublishedPapers, the estimated coefficient on Trend is positive but it is not significantly different from zero. The regression results in column (2) suggest a mild yet significant trend in the proportion of calendar year published papers that use a designated keyword in the title. None of the control variables are significant in either model.

However, an inspection of the data pattern of NumberUses (see Fig. 1) and ProportionUses (see Fig. 2) over time suggests that neither of the two time series is stationary. To obtain reliable coefficients from a time series regression estimated by ordinary least squares, all variables must be stationary.Footnote 11 Accordingly, we implemented Augmented Dickey-Fuller (ADF) tests for NumberUses, NumberPublishedPapers, and ProportionUses. The ADF tests the null hypothesis that a variable follows a unit root process, and is therefore nonstationary, against the alternative that the series is stationary. The test fails to reject the null hypothesis for each of these three variables.Footnote 12 But, first differencing each variable and re-running the ADF tests shows the differenced series to be stationary. Table 5 shows the least-squares results using first differences (∆) of the variables.

Table 5 Least-squares regression results based on first-differences in the data underlying Figs. 1 and 2 (p values in parentheses)
Full size table

The results in Table 5 show that the estimated coefficient on Trend is again positive and not significantly different from zero. In other words, university related research (as proxied by the three defining key words) has consistently remained as the dominant (i.e., most frequently used keywords) focus of academic and policy research published in the JTT since 1977.Footnote 13

To investigate trends relating to academic entrepreneurship within the context of technology transfer research, we reviewed the 398 published papers that used at least one of the designated keywords (i.e., university, universities, academic) in the title to identify those papers that also used entrepreneur, entrepreneurial, or entrepreneurship as a keyword in the title.Footnote 14 Of the 398 published university papers, 69 additionally used at least one of the entrepreneur-related keywords in the title.

Figure 3 shows by calendar year the number of published university papers and the number of published university papers using entrepreneur, entrepreneurial, or entrepreneurship as a keyword. Both of these time series are nonstationary based on an ADF test. When the first differences of each variable are used in a regression model similar to that in column (1) in Table 5, the estimated coefficient on Trend is again insignificant.

Fig. 3
figure 3

Total number of published university papers and number of published University Papers with a Keyword of Entrepreneur or Entrepreneurship in the Title, by Calendar Year

Full size image

Figure 4 shows the proportion of published university papers using at least one of the keywords of entrepreneur, entrepreneurial, or entrepreneurship, by calendar year. The time series of proportions shown in this figure is stationary. Table 6 shows the least-squares results of the trend in this metric. In column (1), the estimated coefficient on Trend is positive and marginally significant suggesting a change in the scope of technology transfer research from focusing on university activity per se to focusing on academic entrepreneurship. Each year, on average, the proportion of published university papers in the JTT using an entrepreneur-related keyword has increased by about 1 percentage point. Also, from Fig. 4, it is visually clear that the upward trend in academic entrepreneurship as defined therein began in the early 2000s. Albeit that this is a small increase, it nevertheless is suggestive of a trend in the scope of a research area that falls under the umbrella of technology transfer. Also, the regression results show that the Bayh–Dole Act (BD) has a marginally significant impact on the proportion of published university papers addressing academic entrepreneurship topics; however, care should be exercised in generalizing from this finding because it is driven by the 1988 datum as shown in Fig. 4.

Fig. 4
figure 4

Proportion of published University Papers with a Keyword of Entrepreneur or Entrepreneurship in the Title, by Calendar Year

Full size image
Table 6 Least-squares regression results using the proportional data, by calendar year, from Fig. 4 (p values in parentheses)
Full size table

In Fig. 4, the 1988 datum is perhaps an outlier and thus could be omitted. There were only 8 published papers in the sampling population for that calendar year (all other calendar years had double digit numbers of published papers; see Fig. 1) and only 1 paper contained the keyword university. After deleting that observation, the data still show stationarity, and the corresponding regression results are in column (2) of Table 6. The results in column (2) show about the same level of response over trend as in column (1) although the estimated coefficient is more significant, the positive effect of BD is no longer marginally significant, and the negative effect of Covid on the percent of published papers related to academic entrepreneurship is marginally significant. One might reasonably conclude that the Bayh–Dole Act of 1980 has not had a measurable impact on the scope of research activity related to academic activity or more specifically to research published in the JTT.

4 Discussion

Our focus on the scope of technology transfer research emphasizing university activity is motivated by the passage of the Bayh–Dole Act of 1980. Our empirical findings show that the role of universities or the academy in technology transfer has consistently, as revealed by an insignificant coefficient on Trend, been emphasized in the titles of papers published in the JTT, although the extant literature has for the most part been neglected to be explicit in how the term technology transfer has been defined (see Table 1). Our analysis did not reveal that the Bayh–Dole Act has a statistically significant impact on university-related research within the technology transfer literature. However, the research emphasis on academic entrepreneurship appears to have become more emphasized over time (see Table 6), especially since the early 2000s (see Fig. 4).

In an effort to extrapolate from our study of the role of universities in technology transfer research to other research areas of emphasis, we did a word count of all words in all titles of papers published in the JTT (having deleted pronouns and prepositions), and we used our findings to explore several identified research areas. As might be expected, there are a number of frequently used words that are not related to the scope of emphasis of research (e.g., case, event, role, study). From this exploration, we identified the following use of keywords in the titles of published papers based on a sampling population of (n = 1,578). Papers related to innovative activity were published 257 times based on at least one the following words being in the title: innovate, innovation, or innovativeness. Papers related to entrepreneurship were published 153 times based on at least one of the following words being used in the title: entrepreneur, entrepreneurial, or entrepreneurship. Papers related to collaboration were published 79 times based on at least one of the following words being used in the title: collaboration or collaborative. Finally, papers related to commercialization activity were published 49 times based on at least one of the following words being in the title: commercial, commercializing, or commercialization (with standardization to U.S. spelling).

Though our analysis was motivated by interest in academic entrepreneurship and associated keywords embodied within 69 papers (see Fig. 3), we also note that papers generally related to entrepreneurship—reflected by keywords entrepreneur, entrepreneurial, or entrepreneurship—were published 153 times. It may therefore be the case that other keywords and their combination relate to our original research question of interest and may therefore be a topic for future investigation. It is our hope that as artificial intelligence (AI)-based technology becomes more widely used as a research tool, then the findings presented here will be re-examined and expanded upon by scholars.

We also noted infrequent use of the term entrepreneurial university among publications within our analysis. Though the term entrepreneurial university has also been used for more than two decades (e.g., Etzkowitz et al., 2000), the concept frames the emergence of an impact-oriented mission at research universities and the extent to which it relates to traditional teaching and research missions (e.g., Guerrero et al., 2016). Interestingly, scholars have also framed academic entrepreneurship as an integral component of the entrepreneurial university (e.g., Hayter, 2016). This perspective therefore opens the possibility that academic entrepreneurship may not only be framed as a sub-category of technology transfer but also one of the many functions of the entrepreneurial university, the impact of which must be considered within the context of other commercialization functions and university missions. Following Hayter et al. (2020), scholars would therefore do well to avoid myopic conceptualization of academic entrepreneurship and technology transfer, among other areas of related scholarly focus, and instead focus on their emergence and interrelationship.