Understanding academia-industry collaboration in a developing economic environment: Determinants and implications from Bariloche,Argentina

Abstract

This study examines the determinants of academia-industry collaboration (AIC) in the context of Bariloche, Argentina, and provides valuable insights into collaboration dynamics in a developing economy. Using a quantitative research approach, data was collected through a survey administered to academic researchers in Bariloche. The survey included questions on researcher characteristics, organizational attributes, and disciplinary norms, which were then tested through eight hypotheses related to factors influencing AIC. Data analysis, including descriptive and inferential statistics, revealed significant relationships between independent variables (e.g., prior career experience, organizational affiliation, disciplinary norms) and researchers’ involvement in AIC. The study’s limitations, such as reliance on self-reported data and non-probabilistic sampling, are acknowledged. Nevertheless, the findings contribute to the understanding of AIC in Bariloche and have implications for promoting effective academia-industry partnerships. Future research should consider longitudinal studies and explore additional factors to inform evidence-based policies in developing economies.

Keywords

Academia-industry collaboration Argentina Bariloche determinants developing economy

Introduction

Academia-industry collaboration (AIC) plays a pivotal role in driving innovation and economic growth. Consequently, understanding the determinants of AIC is crucial for policymakers and stakeholders seeking to foster effective collaboration between academic and business organizations (Ankrah and AL-Tabbaa, 2015). This study precisely focuses on examining the determinants that shape AIC in the context of Bariloche, Argentina, a mid-size city renowned as a center of excellence in science and technology activities within Argentina (Kantis et al., 2005; Lugones and Lugones, 2004; Totonelli, 2018).

The literature on AIC highlights the significance of individual researcher characteristics and experiences, organizational attributes, and disciplinary norms in influencing collaboration patterns (Perkmann et al., 2013, 2021). However, as pointed out by Filippetti and Savona (2017), limited research examined these determinants in the specific context of developing economic environments like Bariloche. By addressing this gap, this study contributes to the broader understanding of AIC and provide insights that can inform evidence-based policies in Bariloche.

By focusing on academic researchers as the unit of analysis, this study seeks to capture the determinants that contribute to establishing collaborations with industry partners. Drawing upon the research of Bercovitz and Feldman (2006) and D'Este and Patel (2007), which ask who in academia interacts with industry and why, this research examines whether the determinants identified as relevant by these authors for explaining AIC in developed economies hold true in developing economic contexts.

This study adopts a quantitative approach, examining individual characteristics, organizational attributes, and disciplinary norms as potential determinants of AIC in Bariloche. By considering the entire city of Bariloche as the scope of investigation, this research distinguishes itself from previous studies that focused on specific sectors, universities, or single collaborations. Acknowledge for its numerous knowledge-based organizations, Bariloche hosts a longstanding web of academia and industry interactions dating back to the 1950s (Carro, 2022; Mariscotti, 2016).

The study utilizes quantitative data obtained from an original survey conducted in Bariloche, distributed to academic researchers listed as authors in publications with at least one affiliation to a Bariloche organization. A total of 135 valid responses were collected for the analysis and logistic regression models are employed to analyze the data. The findings reveal similarities between developed and developing contexts, indicating that individual and disciplinary variables are more influential than organizational factors in establishing inter-institutional linkages.

The findings of this study contribute to the understanding of AIC in developing economic environments and provide insights that can offer valuable evidence for policy making. By identifying the determinants that shape AIC in Bariloche, policymakers can design targeted strategies to foster effective collaboration between academia and industry, leading to enhanced innovation, economic growth, and sustainable development.

The remainder of the paper is organized as follows: the subsequent second section reviews the existing literature on determinants of AIC in both developed and developing economic environments, formulating research hypotheses for comparison. The third section then describes the data collection process and the methodology used for data interpretation. The fourth presents the main empirical findings, followed by their discussion in the fifth section. Finally, the sixth and last section concludes the article and suggests future research and policy implementations.

Background and research hypotheses

The literature on determinants of AIC is extensive and examines various factors influencing collaboration between academia and industry. These factors include individual researchers’ characteristics, organizational attributes, and disciplinary norms (Giuliani et al., 2010; Perkmann et al., 2021). Individual characteristics include demographic traits and prior career experience, while organizational features relate to affiliation, research group size, and peer effects. Disciplinary norms mainly refer to differences among research fields.

At the individual research level, variables for researcher characteristics have been differentiated into two types: demographic (e.g., gender and age) and prior career experience (e.g., seniority, teaching and experience outside academia) (Bercovitz and Feldman, 2006; D’Este and Patel, 2007; Giuliani et al., 2010). As indicated below, many of these variables have been found to play a significant role in shaping AIC.

Organizational attributes have also been explored in relation to AIC (Boardman, 2009; Giuliani et al., 2010; Perkmann et al., 2021). Although the role of organizational embeddedness remains inconclusive, the context of developing economies necessitates a deeper understanding of how organizational factors influence collaboration. Previous research suggested that institutional affiliation may affect AIC outcomes, with studies in other developing economic contexts indicating its significance (Giuliani et al., 2010).

Disciplinary norms have been recognized as influential factors in shaping AIC patterns, underscoring the pivotal role of knowledge fields in the possibility of establishing collaborations (Abreu and Grinevich, 2013; Bekkers and Bodas Freitas, 2008; Cohen et al., 2002; Lawson et al., 2019). Applied fields, such as Engineering and Technology or Medical Sciences, often exhibit a higher propensity for collaboration with industry due to their specialized knowledge and technical skills (Bekkers and Bodas Freitas, 2008; D’Este and Patel, 2007). However, it is important to consider the influence of disciplinary norms in specific developing economic contexts, such as Bariloche, to better understand collaboration dynamics among scholars in different knowledge fields.

Drawing from these insights, this study aims to examine the determinants of AIC in the developing economic environment of Bariloche, Argentina. By considering individual characteristics, organizational attributes, and disciplinary norms that shape AIC in Bariloche, we seek to contribute to the existing literature and provide insights that inform evidence-based policies.

Research hypotheses

By reviewing the relevant literature, this study proposes eight research hypotheses that examine key determinants of AIC at the individual, organizational, and disciplinary levels:

1. H1: Gender

Female researchers are less likely to engage in AIC than their male counterparts.

2. H2: Age Differences

Older academics are more likely to interact with industry.

3. H3: Seniority

Researchers with higher seniority are more likely to collaborate with industry.

4. H4: Years of Experience

More years of experience make researchers more likely to collaborate with industry.

5. H5: Teaching Activities

Researchers who teach are less likely to collaborate with industry.

6. H6: Non-Academic Work Experience

Academics with previous industry work are more prone to collaborate with industry.

7. H7: Affiliation

University researchers are more likely to collaborate with industry.

8. H8: Research Disciplines

Applied disciplines are more prone to have industry partners.

The following subsections elaborate on the rationale behind each hypothesis, providing insights into how the literature guided hypothesis postulation in order to capture similarities and differences between developed and developing contexts.

Individual determinants

H1 – Gender

Previous studies present conflicting findings regarding the relationship between gender and AIC. Some studies suggest that women are less engaged with industry than men (Abreu and Grinevich, 2013, 2017; Tartari and Salter, 2015). However, others find no significant differences (Blind et al., 2018; Meng, 2016). Moreover, certain studies even suggest that women are more active in AIC (Giuliani et al., 2010). Considering the existing gender parity gap in developing economies, we propose the hypothesis that female researchers are less likely to be involved in academic collaborations than male researchers.

H2 – Age differences

The relationship between age and AIC is complex and has yielded contradictory evidence. Some studies suggest a positive correlation between age and collaboration with industry, indicating that older researchers have advantages such as access to contacts and greater time availability (Boardman and Ponomariov, 2009). Conversely, younger scholars may be more inclined to collaborate with industry to enhance their reputation and establish their careers (Bercovitz and Feldman, 2007; D’Este and Patel, 2007; Giuliani et al., 2010). We propose that, overall, there is a positive correlation between age and industry collaborations, as more experienced researchers have accumulated networking opportunities and prior collaborative experiences.

H3 – Seniority

Researchers’ prior-career experience, particularly their seniority, has been found to positively correlate with AIC (D’Este and Patel, 2007; Boardman and Ponomariov, 2009; Ding and Choi, 2011; Tartari and Breschi, 2012; Abreu and Grinevich, 2013). This relationship suggests that more experienced researchers inspire confidence in companies and possess larger networks that facilitate access to industry partners. Therefore, we hypothesize that researchers with higher seniority are more likely to collaborate with industry partners.

H4 – Years of experience

Similarly, the number of years a researcher has spent in academia has been found to positively influence AIC (Schuelke-Leech, 2013; Van Rijnsoever et al., 2008). Increased experience in academia is associated with larger networks and a higher likelihood of establishing collaborations with industry (Giuliani et al., 2010). Therefore, we propose that more years of experience in academia make researchers more likely to engage in AIC.

H5 – Teaching activities

The impact of teaching activities on research engagement in academia-industry is less studied and inconclusive (Bianchini et al., 2016; Hughes et al., 2016; Lin and Bozeman, 2006). While teaching activities may expand research networks and provide access to new collaborators, they may also compete for time and reduce the likelihood of industry collaborations. We postulate that the substitution effect will predominate, as researchers involved in teaching activities may have less time available for industry collaborations. Thus, we propose that researchers who teach are less likely to collaborate with industry compared to researchers who do not.

H6 – Non-academic work experience

The non-academic work experience of researchers has been found to increase likelihood of establishing collaborations with non-academic agents, including industry partners (Abreu and Grinevich, 2013; Gulbrandsen and Thune, 2017; Tartari et al., 2012). This positive relationship suggests that experience outside academia reduces barriers for industry collaborations and enhances researchers access to industrial partners. We hypothesize that academics with previous work experience in the private sector are more inclined to engage in collaborations with industry than those without such experience.

Organizational determinants

H7 – Affiliation

Researchers’ organizational affiliation, specifically whether they are affiliated with universities or other types of academic organizations, may influence their collaboration with industry. In developed countries, this variable has been generally found not significant in explaining differences between researchers participation in industrial partnerships (Bekkers and Bodas Freitas, 2008; D’Este and Patel, 2007; Perkmann et al., 2013, 2021). Nonetheless, previous research in developing countries suggests that researchers affiliated with universities are more likely to collaborate with industry compared to those affiliated with other types of academic organizations (Giuliani et al., 2010). This finding may be attributed to the perceived advantages of universities in terms of peer effects and the perception of detachment from government inefficiencies. Therefore, we propose that researchers affiliated with universities are more likely to collaborate with industry partners.

Disciplinary norms

H8 – Research disciplines

Different research disciplines exhibit varying levels of engagement in AIC. Applied research disciplines, such as engineering, chemistry, and biotechnology, are generally associated with higher rates of collaboration with industry compared to disciplines focused on basic research or social sciences and humanities (Abreu and Grinevich, 2013; Bozeman, 2000; Lee, 1996; Owen-Smith and Powell, 2001). We expect this trend to be consistent in the developing economic environment considered in this study. Therefore, we propose that scholars conducting research in applied disciplines are more likely to have industry partners compared to those involved in basic research, social sciences, and humanities.

These hypotheses establish a framework for examining AIC determinants from the researcher’s perspective in developing countries, contributing to the existing literature. Conducting empirical research to test these hypotheses can inform the specificities and challenges of collaboration in these contexts.

Methodology

This section presents the methodology employed in the study, outlining the data collection process, sampling strategy, and the statistical analysis method.

Study context: The Argentinean city of Bariloche

Bariloche was selected as the ideal location for distributing the survey aimed at analyzing determinants of collaboration for several compelling reasons. Firstly, the city boasts both an extensive network of academic organizations and a high ratio of researchers per 1000 inhabitants.¹ Additionally, it serves as the geographical headquarters of INVAP,² a prominent technological firm renowned throughout Latin America. This dynamic environment establishes Bariloche as a vibrant intellectual hub (Britto and Lugones, 2020; Kantis et al., 2005; Lugones and Lugones, 2004; Totonelli, 2018), enhancing the likelihood of interaction and collaboration between academic organizations and innovative companies.

Secondly, by focusing on a single location like Bariloche, the study can ensure greater validity of results when compared to similar contexts. This is particularly significant as prior research conducted in developed countries is more likely to be comparable with Bariloche’s research landscape than with municipalities lacking academic research institutions or innovative firms. Thus, Bariloche provides an ideal setting for making meaningful comparisons with more developed environments.

Furthermore, the selection of a unique location such as Bariloche allows for a greater level of local applicability of the study’s findings, enabling a clearer understanding of differences among influential factors while keeping contextual variables constant. In light of these considerations, the decision to distribute the survey among Bariloche’s researchers is justified and promised valuable insights into the determinants of collaboration.

Data collection, sampling, and limitations

Data for the study were collected through an original online survey distributed among Bariloche’s researchers. The initial target population consisted of researchers with at least one publication indexed in Scopus or Web of Science and affiliated with San Carlos de Bariloche. Initially, 4442 unique email addresses were extracted from these databases, with two email failures and 982 bounced emails due to potential errors or discontinued addresses. This resulted in a final sample of 3458 email addresses.

Due to the inability to differentiate between local researchers and external collaborators, the decision was made to distribute the survey to the entire sampling frame. Although this approach may limit the generalizability of the results (Fricker, 2016), it ensured a significant number of responses and minimized biases in respondent selection. To capture the viewpoints of local researchers not represented in the indexed databases, anonymous survey links were also shared among local peers and promoted by communication managers of academic organizations in Bariloche. This supplementary approach yielded an additional 35 responses, resulting in a total of 180 recorded responses after 1 month. Out of the total responses, only 135 were complete and valid, forming the basis for analysis. While the sample size may seem modest, it aligns with the context of Bariloche’s academic community (Civitaresi and Colino, 2019; Quiroga et al., 2020; Totonelli, 2018), suggesting that the findings might serve as valuable insights for implementing evidence-based policy changes in Bariloche.

Though not perfect, this methodology presents an inexpensive and feasible alternative that allows for a general overview of AIC determinants in developing countries. This would otherwise be very difficult or impossible, as there are usually no reliable records available in these contexts. Limiting the analysis to a particular geographical location allows reaching a relatively large number of cases to approach a certain degree of representativeness.

Independent variables

This study examines various independent variables to test the postulated hypotheses related to AIC determinants, explaining the existence of collaboration with industry partners. The measurement and contrast of each variable are described below, furthermore the survey questions associated with each variable are included in Appendix A.

Individual variables

• Gender (gen): This binary variable represents gender of the researchers, where 0 indicates male and one represents female.

• Age (age): The age of the researchers was calculated based on the year of birth reported in the survey. This variable captures the researcher’s age at the time of data collection.

• Seniority (senior): This variable categorizes the researcher’s position in their institution into three levels: “Formation/recognized”, “Established”, and “Leading”. It is a polytomous and ordinal variable.

• Academic Experience (yearExp): This binary variable captures the researchers’ experience in academia. Respondents were asked about their previous years of experience in the academic sector, which encompassed multiple levels. To maintain degrees of freedom and ensure a sufficient sample size, a threshold of 20 years was chosen as it divided the sample of respondents into nearly equal halves. Researchers with 20 or more years of experience were assigned a value of 1, indicating extensive academic experience, while those with less than 20 years were assigned a value of 0.

• Teaching (teach): This binary variable indicates whether the researcher is involved in teaching activities. A value of one denotes engagement in teaching, while 0 indicates no teaching involvement. The variable was constructed based on an original survey question that asked about the average number of hours of classes taught per week.

• Industry Experience (indExp): This binary variable captures any previous work experience in the industry sector since obtaining the researcher’s last degree. A value of one represents experience in the industry, while 0 indicates no such experience.

Organizational and discipline variables

• Organization Type (uni and pro): These binary and nominal variables classify the researchers’ affiliations. The variable “uni” takes the value one if the researcher is affiliated with a university, and 0 otherwise. Similarly, the variable “pro” is set to one only for members of public research organizations. These separate variables were used to account for cases where researchers had concurrent affiliation to both types.

• Research Disciplines (rField): The researchers self-reported their research fields based on the categorizations provided in the OECD Frascati Manual’s scheme of ‘Fields of science and technology’. However, due to the specialized nature of academic production in Bariloche, the analysis focused on the knowledge fields that are characteristic of the region. Based on a frequency threshold of 10 observations, the categories considered for analysis were biological sciences, physics, engineering and technology, earth and related environmental sciences, other natural sciences, and other.

Logistic regression models

To assess the hypotheses and determine the effect of the chosen determinants on researchers’ likelihood of establishing collaborations with industry partners, logistic regression models were employed (Agresti, 2015; Hosmer et al., 2013). Models were estimated using R software (version 4.0.2). The dependent variable (collab) is binary, indicating whether a researcher has collaborated with industry partners at least once during their academic career. The models include the relevant independent variables necessary to test the proposed hypotheses.

The full econometric model can be expressed as:

l o g i t (c o l l a b) = β_0 + β_1 g e n + β_2 a g e + β_3 s e n i o r + β_4 y e a r E x p + β_5 t e a c h + β_6 i n d E x p + β_7 u n i + β_8 p r o + β_9 r F i e l d + ε

In the equation above, logit (collab) represents the natural logarithm of the odds ratio of collaboration with industry. The β coefficients denote the estimated effects of the independent variables on the log odds ratio, and ε represents the error term. Notably, the explanatory variable senior has been treated as an ordered factor, accounting for its inherent order and progression. Polynomial trends, including linear and quadratic trends, have been incorporated into the model to capture potential non-linear relationships between the senior variable and the outcome. This approach enables a more comprehensive understanding of this variable and enhances the interpretation of the results.

The methodology employed in this study provides an inexpensive and feasible alternative for understanding AIC determinants in developing countries. Despite the limitations associated with the non-probabilistic nature of the sampling approach, the comprehensive survey distribution and the inclusion of anonymous links helped achieve a relatively large sample size of surveyed academic researchers, thereby ensuring a representative inclusion of local researchers. The selection of Bariloche as a case study location allows for meaningful comparisons with developed economic environments. The subsequent analysis using logistic regression models provides insights into the determinants influencing AIC in the region.

Results

To ensure accurate estimates and avoid problems caused by collinearity, we conducted an evaluation of the relationship between independent variables. When two independent variables are highly correlated, it is advisable to include only one of them in the estimated models (Ranganathan et al., 2017). The correlation matrix analyzing statistical associations among explanatory variables is included in Appendix B.

As expected, several variables related to researchers’ experiences in academia were found to be interrelated. Involvement in teaching activities, years of experience in academia, and seniority showed significant dependencies on each other. This observation can be attributed to the fact that more experience enables access to higher-ranking roles that often involve teaching responsibilities. These variables also exhibited a positive correlation with age. To address collinearity issues in our analysis, we included them separately in the estimated models. Consequently, the partial models yielded more robust results compared to the full model.

Descriptive statistics

This section provides a brief overview of the main characteristics of researchers in Bariloche and their scientific production. Table 1 summarizes the key features of the explanatory variables for both collaborators and non-collaborators. Notably, a majority of researchers reported having collaborated with industry, which we hypothesize higher than expected for two reasons. Firstly, collaborators might have had a greater interest in responding to the survey. Secondly, a significant portion of AICs cannot be accounted for by other means. Nonetheless, there is no specific reason to assume that the determinants captured for the two groups and the differences between them do not characterize average behavior.

Table 1.

Descriptive statistics.

Variable	Overall, N = 135^a	No, N = 52^a	Yes, N = 83^a	p-value^b
Age	48.01 (11.66)	43.90 (10.84)	50.59 (11.47)	0.002
Gender				0.13
Male	76 (56%)	25 (48%)	51 (61%)
Female	59 (44%)	27 (52%)	32 (39%)
Residence in Bariloche	114 (84%)	47 (90%)	67 (81%)	0.13
Seniority				0.024
Formation/recognized	28 (21%)	13 (25%)	15 (18%)
Established	51 (38%)	25 (48%)	26 (31%)
Leading	56 (41%)	14 (27%)	42 (51%)
20 or more year experience	59 (44%)	14 (27%)	45 (54%)	0.002
Industry experience	34 (25%)	7 (13%)	27 (33%)	0.013
Teaching	79 (59%)	23 (44%)	56 (67%)	0.008
University affiliation	66 (49%)	26 (50%)	40 (48%)	0.8
PRI affiliation	87 (64%)	36 (69%)	51 (61%)	0.4
Grouped research fields				0.028
Engineering and technology	16 (12%)	3 (5.8%)	13 (16%)
Other	23 (17%)	4 (7.7%)	19 (23%)
Biological sciences	51 (38%)	25 (48%)	26 (31%)
Other natural sciences	14 (10%)	4 (7.7%)	10 (12%)
Earth and related environmental sciences	14 (10%)	8 (15%)	6 (7.2%)
Physical sciences	17 (13%)	8 (15%)	9 (11%)

^aMean (SD); n (%).

^bWilcoxon rank sum test; Pearson’s Chi-squared test.

Table 1 demonstrates a balanced proportion of male and female researchers, with an average age of 48.01 (11.66). As stated above, regarding years of experience in the academic sector, the sample can be divided into two comparable groups based on a threshold of 20 years of experience, with 44% of researchers above and 56% below this threshold. Teaching activities were more prevalent than non-teaching activities, and the distribution among academic positions groups was not significantly skewed. However, the presence of researchers in training, such as PhD candidates and postdoctoral fellows, was less frequent, as belonging to this group increased the likelihood of not having published in indexed journals. Finally, in terms of organizational and research-field features, researchers were primarily affiliated with public research institutes and predominantly belonged to the fields of biological sciences and physics.

Regression results

In this section, we present the results obtained through logistic regression models, which estimate the probability of researchers collaborating with industry (collab) and identify the influencing factors discussed in the literature. The main results are summarized in Table 2, with the first and last columns representing the null model and the full model, respectively, for comparison. Between these extremes, models 2, 3, 4, and 5 include solely individual variables, while models 6, 7, 8, and 9 incorporate organizational and discipline variables. We included four models in each group to avoid the inclusion of interdependent explanatory variables, such as those accounting for age, years of experience in academia, seniority, and teaching. Models 2 and 6 include age, models 3 and 7 include years of experience in academia, models 4 and 8 include seniority, and models 5 and 9 evaluate the effects of teaching.

Table 2.

Estimated parameters of logistic regression models.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)
(Intercept)	0.468**	−2.276*	0.290	0.673	0.310	−1.835	0.811	1.130	0.387	−0.413
(Intercept)	(0.177)	(1.169)	(0.610)	(0.574)	(0.627)	(1.455)	(1.010)	(1.019)	(1.103)	(1.823)
genFemale		−0.204	−0.300	−0.168	−0.352	−0.124	−0.242	−0.047	−0.343	−0.188
genFemale		(0.390)	(0.385)	(0.392)	(0.380)	(0.433)	(0.429)	(0.441)	(0.422)	(0.450)
Age		0.061**				0.063**				0.024
Age		(0.019)				(0.021)				(0.034)
indExpYes		1.368**	1.188*	1.311*	1.081*	1.219*	1.045*	1.247*	0.967*	1.199*
indExpYes		(0.510)	(0.490)	(0.516)	(0.484)	(0.540)	(0.525)	(0.552)	(0.518)	(0.563)
residenceBchYes		−0.367	−0.491	−0.584	−0.487	−0.878	−0.919	−1.122	−1.077	−0.194
residenceBchYes		(0.585)	(0.578)	(0.574)	(0.576)	(0.919)	(0.896)	(0.906)	(0.913)	(0.722)
yearExpYes			1.174**				1.284**			0.577
yearExpYes			(0.401)				(0.446)			(0.719)
senior.L				0.886*				1.127*		0.509
senior.L				(0.391)				(0.445)		(0.563)
senior.Q				0.290				0.332		0.074
senior.Q				(0.322)				(0.364)		(0.405)
teachYes					0.885*				0.844*	0.476
teachYes					(0.383)				(0.458)	(0.483)
uniYes						1.067	1.083	1.260	1.354	−0.026
uniYes						(1.085)	(1.088)	(1.120)	(1.084)	(0.492)
proYes						0.965	0.898	1.020	1.242	0.152
proYes						(0.995)	(0.998)	(1.034)	(1.026)	(0.519)
rFieldOther						0.117	0.120	0.265	0.581	0.234
rFieldOther						(0.909)	(0.903)	(0.938)	(0.922)	(0.933)
rFieldBiological sciences						−1.285*	−1.348*	−1.267	−0.874	−1.202
rFieldBiological sciences						(0.762)	(0.749)	(0.787)	(0.751)	(0.801)
rFieldOther natural sciences						−0.617	−0.674	−0.779	−0.323	−0.589
rFieldOther natural sciences						(0.979)	(0.967)	(0.988)	(0.954)	(0.995)
rFieldEarth and related environmental sciences						−1.613*	−1.639*	−1.903*	−1.024	−1.454
rFieldEarth and related environmental sciences						(0.934)	(0.911)	(0.948)	(0.938)	(0.988)
rFieldPhysical sciences						−1.574*	−1.719*	−1.790*	−1.191	−1.749*
rFieldPhysical sciences						(0.899)	(0.901)	(0.934)	(0.855)	(0.927)
uniYes:proYes						−1.327	−1.201	−1.341	−1.686
uniYes:proYes						(1.171)	(1.171)	(1.195)	(1.176)
N	135	135	135	135	135	135	135	135	135	135
logL	−89.98	−79.42	−80.47	−81.30	−82.31	−73.74	−74.33	−74.30	−77.07	−72.82
AIC	181.97	168.84	170.94	174.60	174.62	173.47	174.66	176.60	180.15	177.64
BIC	184.87	183.36	185.46	192.04	189.15	211.24	212.43	217.28	217.91	224.13

***p < .001; **p < .01; *p < .1.

Regarding researchers’ demographic features, being female consistently displayed a negative sign in almost every model, although it did not reach statistical significance. On the other hand, age had a positive and significant effect in models that did not include other explanatory variables significantly correlated with age. These variables primarily pertain to researchers’ careers in academia.

Among the variables related to researchers’ careers in academia, individual analysis indicated that greater years of experience in academia, higher seniority (in its linear trend – senior. L), and involvement in teaching activities had a positive and significant impact on the likelihood of collaborating with industry, echoing the findings for age. However, when these variables were considered together in the full model, the positive effect also attenuated, highlighting the potential collinearity issues. The non-significance of the quadratic senior trend (senior.Q) indicated that the non-linear relationship between the ‘senior’ levels and the odds of collaboration is not supported by the data.

The variable representing experience in the industry sector consistently displayed a positive and significant effect in all models, including the full model. Among the organizational variables, neither of the two dummy variables representing affiliation with universities or public research organizations (PROs), nor their combination, showed a significant relationship with collaborating with industry partners. However, the variable measuring research disciplines indicated that researchers in the engineering and technology interacted significantly more with industry compared to researchers in the field of biological sciences, earth and environmental sciences, and physics.

Discussion

The findings of this study provide insights into the determinants of AIC in the developing economic context of Bariloche. Our results confirm certain similarities with more developed contexts, while also suggesting a few context-specific attributes. It is important to note that this study has some limitations, such as its focus on a specific geographic location, which may limit the generalizability of the findings to other regions. Future studies should aim to expand the scope and incorporate multiple data sources to enhance the external validity of the results.

The assessment of the postulated hypotheses is summarized in Table 3, but a more explicit prose contrasting previous findings is worthwhile. In terms of individual determinants, indicating researchers’ characteristics, our study reveals that prior work experience plays a crucial role in shaping AIC in Bariloche, while demographic characteristics do not show a significant influence.

Table 3.

Assessment of hypotheses on AIC determinants.

Hypothesis	Result	Notes on interpretation
H1: Female researchers are less likely to be involved in AIC.	Not supported/Inconclusive	Despite positive sign of male, gender was not a significant predictor of AIC.
H2: Older academics are more likely to interact with industry.	Partially supported	Age had a positive and significant effect on AIC propensity. Yet, significance mitigates when controlling by previous academic experience.
H3: Researchers with higher seniority are more likely to collaborate.	Supported	Seniority increased likelihood of AIC.
H4: More years of experience make researchers more likely to collaborate.	Supported	Experience was a significant predictor of AIC until controlled for seniority.
H5: Researchers who teach are less likely to collaborate.	Not supported	Teaching activities increased AIC likelihood until controlled for experience.
H6: Academics with previous industry work are more prone to engage.	Supported	Previous non-academic experience significantly increased AIC propensity. Non-academic work experience confirms to be the most robust determinant across all models
H7: University researchers are more likely to collaborate.	Not supported	Researcher affiliation was not a significant predictor of AIC.
H8: Applied disciplines are more prone to have industry partners.	Supported	Researchers in engineering and technology fields showed higher AIC rates.

Specifically, researchers with previous experience in the industry sector are more likely to engage in collaboration. Non-academic work experience emerged as the most robust predictor across all models. This finding is consistent with prior research, indicating that practical experience outside academia equips researchers with valuable skills and networks that facilitate engagement with industrial partners (Hughes et al., 2016; Tartari et al., 2012). Given the limitations of institutional support that often characterize developing economic contexts, the relative importance of this variable deserves further investigation, as it suggests that it may be particularly important in these settings.

Furthermore, the study demonstrates that research experience and seniority within the academic hierarchy positively influence the likelihood of collaborating with industry. In line with previous studies (Abreu and Grinevich, 2013; Lawson et al., 2019), longer careers in academia and higher seniority are both associated with a higher propensity for AIC. The persistent significance of these variables suggests that longer careers in academia are more likely to foster collaboration with industry. These findings also reinforce the hypothesis that researchers’ connections and positions within academia are more relevant for AIC than simply the passage of time.

The positive correlation between teaching and collaborating with industry indicates an interesting effect that networks and experienced gained from teaching may have. It is plausible that scholars involved in teaching activities are subject to greater exposure and therefore have access to more diversified contacts that may lead to potential partners. Being unclear even in developed country environments (Bianchini et al., 2016; Hughes et al., 2016), the causes, and effects of teaching on AIC still require further exploration.

In contrast to expectations, gender does not emerge as a significant differentiating factor in AIC propensity. This finding aligns with some studies conducted in specific STEM disciplines in the US and Germany (Blind et al., 2018; Meng, 2016), although it diverges from findings in other contexts (Giuliani et al., 2010). Thus, the influence of gender on AIC remains inconclusive and may vary across different settings.

In terms of organizational attributes, affiliation did not emerge as a differentiating factor for AIC, contrasting findings from other developing economic contexts (Giuliani et al., 2010). This suggests that country-specific factors may contribute to differences in AIC across different types of organizations. Additionally, the absence of significant peer effects or the presence of similar collaboration propensities across organizations in Bariloche may also explain the findings.

Lastly, our study confirms that disciplinary norms influence collaboration patterns among Bariloche’s scholars. Researchers in applied fields, such as Engineering and Technology, exhibit a higher likelihood of collaborating with industry partners. This observation aligns with expectations, as applied fields often possess specialized knowledge and technical skills that are in high demand from industry partners, suggesting that these fields might present some professional standards that distinguish them from others.

In line with Bekkers and Bodas Freitas (2008) and D’Este and Patel (2007), overall, the results highlight the significant influence of individual and research field determinants on AIC in Bariloche, with individual researchers’ trajectories and disciplinary backgrounds assuming greater importance than organizational embeddedness. However, given the plausible effect of Bariloche’s specific characteristics, it is essential to approach the results with caution.

Finally, it is worth noting some limitations of this study. Firstly, the research was conducted in a specific geographic location, which may limit the generalizability of the findings to other regions. Therefore, future studies should aim to expand the scope and incorporate multiple data sources to enhance the external validity of the results. Additionally, the study’s cross-sectional design restricts the ability to capture the systemic and longitudinal dynamics of AIC in developing countries fully. Longitudinal research efforts could provide insights into the evolution of AIC and its implications for economic development.

Conclusion

This study highlights on the determinants of AIC in the developing economic context of Bariloche, Argentina. By examining researchers’ individual characteristics, organizational attributes, and disciplinary norms, it provides valuable insights into the dynamics of AIC in this specific region.

The findings highlight the crucial role of researcher experiences and disciplinary inclination in fostering collaborations between academia and industry. Researchers with prior work experience in the industry sector demonstrated a higher propensity for engagement, emphasizing the value of practical knowledge and networks gained outside academia. Additionally, research seniority within the academic hierarchy and involvement in teaching activities were identified as influential determinants, suggesting the importance of both expertise and networking opportunities provided by teaching.

Importantly, the study reveals that individual variables have a stronger influence on AIC than organizational ones, which challenges the notion suggested by Giuliani et al. (2010) that organizational embeddedness, particularly affiliation with universities, is a primary driver of collaboration. These findings have implications for evidence-based policies in developing economies, emphasizing the need to tailor support mechanisms to the specific characteristics and trajectories of researchers.

The study also underscores the significance of disciplinary norms, within applied fields such as Engineering and Technology exhibiting a higher likelihood of collaboration with industry partners. This highlights the demand for specialized knowledge and technical skills from industry, further emphasizing the importance of interdisciplinary collaboration and knowledge transfer.

In summary, the findings emphasize the significance of individual researcher experiences and their disciplinary inclination in fostering collaborations with industry. More concretely, the study highlights the importance of considering researcher experiences, seniority, disciplinary inclination, and the specific characteristics and trajectories of researchers when designing support mechanisms to foster AIC. These insights can inform policymakers in developing targeted strategies to promote effective AIC in Bariloche and similar contexts.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Universitat Oberta de Catalunya.

ORCID iD

Ian Hulskamp

Notes

Appendix

Appendix A.

Table of survey questions and translations.

Associated variable	Original Spanish question	Translated question	Type of created variable
Collab	Desde su rol como académica/o, ¿colaboró alguna vez con empresas o productores? Considere todo tipo de acuerdos formales e informales (remunerados y no remunerados), por ejemplo, desarrollo de proyectos comunes, cualquier tipo de asesoramiento o intercambio de información profesional	From your role as an academic, have you ever collaborated with companies or producers? Consider all types of formal and informal agreements (paid and unpaid), such as joint project development, any kind of advice or professional information exchange.	Binomial categorical
ResidenceBCH	¿Reside o residió en la ciudad de San Carlos de Bariloche?	Do you currently reside or have you resided in the city of San Carlos de Bariloche?	Binomial categorical
Age	Año de nacimiento	Year of birth	Continuous
gen	Indique su género	Indicate your gender	Binomial categorical
Senior	¿Cuál es su cargo de investigador/a en la institución a la que pertenece? Si pertenece a más de una institución marque únicamente el cargo de mayor jerarquía	What is your research position at the institution you belong to? If you belong to more than one institution, please indicate only the highest-ranking position.	Ordinal multinomial categorical
yearExp	¿Cuántos años de experiencia lleva en el campo de la investigación?	How many years of experience do you have in the field of research?	Binomial categorical
Teach	Además de sus tareas como investigador/a, ¿cuántas horas de clase dicta por semana?	In addition to your research duties, how many hours of classes do you teach per week?	Binomial categorical
indExp	¿Trabajó en una alguna empresa por fuera del ámbito académico desde que finalizó su última formación universitaria (grado, máster o doctorado)?	Have you worked for any company outside the academic field since completing your last university education (undergraduate, master’s, or doctoral)?	Binomial categorical
uni	¿A cuál de las siguientes instituciones de Bariloche está afiliada/o? Si ya no se encuentra afiliada/o marque aquella que fuera su institución principal.	Which of the following institutions in Bariloche are you affiliated with? If you are no longer affiliated, please mark the institution that was your main affiliation.	Binomial categorical
pro	¿A cuál de las siguientes instituciones de Bariloche está afiliada/o? Si ya no se encuentra afiliada/o marque aquella que fuera su institución principal.	Which of the following institutions in Bariloche are you affiliated with? If you are no longer affiliated, please mark the institution that was your main affiliation.	Binomial categorical
rField	¿Cuál es su campo de investigación principal?	What is your main field of research?	Nominal multinomial categorical

Appendix B.

correlation matrix and statistical associations among explanatory variables.

	Age	gen	Profile	yearExp	Teach	indExp	uni	pro	rField
Age	1.00	−0.18	0.64	0.77	0.26	−0.08	0.10	−0.24	0.04
gen		1.00	0.22	0.10	0.06	0.12	0.00	0.05	0.27
Senior			1.00	0.65	0.31	0.25	0.11	0.21	0.40
yearExp				1.00	0.21	0.01	0.00	0.17	0.23
Teach					1.00	0.02	0.15	0.26	0.38
indExp						1.00	0.11	0.06	0.25
uni							1.00	0.31	0.24
pro								1.00	0.18
rField									1.00

Note. For the numeric variable ‘age’, Pearson correlation coefficients were computed using ANOVA. Categorical variables (‘gen’, ‘profile’, ‘yearExp’, ‘teach’, ‘indExp’, ‘uni’, ‘pro’, ‘rField’) were converted to numeric format, enabling the calculation of Pearson correlation coefficients. Categorical-categorical associations were measured using the square root of Cramer’s V, with Fisher’s exact test employed for small sample sizes.

References

Abreu

Grinevich

(2013) The nature of academic entrepreneurship in the UK: widening the focus on entrepreneurial activities. Research Policy 42(2): 408–422.

Abreu

Grinevich

(2017) Gender patterns in academic entrepreneurship. The Journal of Technology Transfer 42(4): 763–794.

Agresti

(2015) Foundations of Linear and Generalized Linear Models. Hoboken, NJ: John Wiley & Sons.

Ankrah

AL-Tabbaa

(2015) Universities–industry collaboration: a systematic review. Scandinavian Journal of Management 31(3): 387–408.

Bekkers

Bodas Freitas

(2008) Analysing knowledge transfer channels between universities and industry: to what degree do sectors also matter? Research Policy 37(10: 1837–1853.

Bercovitz

Feldman

(2006) Entpreprenerial universities and technology transfer: a conceptual framework for understanding knowledge-based economic development. The Journal of Technology Transfer 31(1): 175–188.

Bercovitz

Feldman

(2007) Academic entrepreneurs: organizational change at the individual level. Organization Science 19(1): 69–89.

Bianchini

Lissoni

Pezzoni

, et al. (2016) The economics of research, consulting, and teaching quality: theory and evidence from a technical university. Economics of Innovation and New Technology 25(1): 668–691.

Blind

Pohlisch

(2018) Publishing, patenting, and standardization: motives and barriers of scientists. Research Policy 47(7): 1185–1197.

10.

Boardman

(2009) Government centrality to university–industry interactions: university research centers and the industry involvement of academic researchers. Research Policy 38(10): 1505–1516.

11.

Boardman

Ponomariov

(2009) University researchers working with private companies. Technovation 29(2): 142–153.

12.

Bozeman

(2000) Technology transfer and public policy: a review of research and theory. Research Policy 29(4): 627–655.

13.

Britto

Lugones

(2020) Bases y Determinantes para una Colaboración Exitosa Entre Ciencia y Producción. Buenos Aires: CIECTI. Epub ahead of print 2020.

14.

Carro

(2022) Vinculación y transferencia tecnológica en el Centro Atómico Bariloche de la CNEA en el período 2006-2015. Ciencia, Docencia y Tecnología 33(66): 4.

15.

Civitaresi

Colino

(2019) Turismo, transformaciones territoriales y resiliencia: Bariloche como evidencia de una ciudad turística intermedia argentina. Revista de Estudios Latinoamericanos sobre Reducción del Riesgo de Desastres REDER 3(1): 41–52.

16.

Cohen

Nelson

Walsh

(2002) Links and impacts: the influence of public research on industrial R&D. Management Science 48(1): 1–23.

17.

Ding

Choi

(2011) Divergent paths to commercial science: a comparison of scientists’ founding and advising activities. Research Policy 40(1): : 69–80.

18.

D’Este

Patel

(2007) University–industry linkages in the UK: what are the factors underlying the variety of interactions with industry? Research Policy 36(9): 1295–1313.

19.

Filippetti

Savona

(2017) University–industry linkages and academic engagements: individual behaviours and firms’ barriers. Introduction to the special section. The Journal of Technology Transfer 42(4): 719–729.

20.

Fricker

Ronald D.

(2016) Sampling methods for online surveys. The SAGE handbook of online research methods 162–183.

21.

Giuliani

Morrison

Pietrobelli

, et al. (2010) Who are the researchers that are collaborating with industry? An analysis of the wine sectors in Chile, South Africa and Italy. Research Policy 39(6): 748–761.

22.

Gulbrandsen

Thune

(2017) The effects of non-academic work experience on external interaction and research performance. The Journal of Technology Transfer 42(4): 795–813.

23.

Hosmer

Lemeshow

Sturdivant

(2013) Applied Logistic Regression. Hoboken, NJ: John Wiley & Sons.

24.

Hughes

Lawson

Kitson

, et al. (2016) The Changing State of Knowledge Exchange: UK Academic Interactions with External Organisations 2005-2015. London: National Centre for Universities and Business. Epub ahead of print 2016.

25.

Kantis

Federico

Drucaroff

, et al (2005) Clusters y Nuevos Polos Emprendedores Intensivos en Conocimiento en Argentina. Buenos Aires: Universidad Nacional de General Sarmiento. Epub ahead of print 2005.

26.

Lawson

Salter

Hughes

, et al. (2019) Citizens of somewhere: examining the geography of foreign and native-born academics’ engagement with external actors. Research Policy 48(3): 759–774.

27.

Lee

(1996) ‘Technology transfer’ and the research university: a search for the boundaries of university-industry collaboration. Research Policy 25(6): 843–863.

28.

Lin

M-W

Bozeman

(2006) Researchers’ industry experience and productivity in university–industry research centers: a “scientific and technical human capital” explanation. The Journal of Technology Transfer 31(2): 269–290.

29.

Lugones

(2004) Bariloche y su grupo de empresas intensivas en conocimiento: Realidades y Perspectivas. Buenos Aires: Documento de trabajo Centro Redes, No 17. Epub ahead of print 2004.

30.

Mariscotti

MAJ

(2016) El secreto atómico de Huemul: Crónica del origen de la energía atómica en Argentina. Lenguaje claro Editora.

31.

Meng

(2016) Collaboration patterns and patenting: exploring gender distinctions. Research Policy 45(1): 56–67.

32.

Owen-Smith

Powell

(2001) Careers and contradictions: faculty responses to the transformation of knowledge and its uses in the life sciences. The Transformation of Work. Bingley: Emerald Group Publishing Limited.

33.

Perkmann

Tartari

McKelvey

, et al. (2013) Academic engagement and commercialisation: a review of the literature on university–industry relations. Research Policy 42(2): 423–442.

34.

Perkmann

Salandra

Tartari

, et al. (2021) Academic engagement: a review of the literature 2011-2019. Research Policy 50(1): 104114.

35.

Quiroga

Carro

Lugones

, et al. (2020) Vinculación y transferencia tecnológica en pequeñas empresas. El caso de la levadura salvaje patagónica en la producción de cerveza artesanal. Pymes, Innovación y Desarrollo 8(1): 112–131.

36.

Ranganathan

Pramesh

Aggarwal

(2017) Common pitfalls in statistical analysis: logistic regression. Perspectives in Clinical Research 8(3): 148–151.

37.

Schuelke-Leech

B-A

(2013) Resources and research: an empirical study of the influence of departmental research resources on individual STEM researchers involvement with industry. Research Policy 42(9): 1667–1678.

38.

Tartari

Breschi

(2012) Set them free: scientists’ evaluations of the benefits and costs of university–industry research collaboration. Industrial and Corporate Change 21(5): 1117–1147.

39.

Tartari

Salter

(2015) The engagement gap:. Research Policy 44(6): 1176–1191.

40.

Tartari

Salter

D’Este

(2012) Crossing the Rubicon: exploring the factors that shape academics’ perceptions of the barriers to working with industry. Cambridge Journal of Economics 36(3): 655–677.

41.

Totonelli

(2018) Bariloche, nieve y ciencia: algunas consideraciones respecto a la conformación del Cluster CTI. Santiago de Chile: XII Jornadas Latinoamericanas. Epub ahead of print 20 July 2018.

42.

van Rijnsoever

Hessels

Vandeberg

RLJ

(2008) A resource-based view on the interactions of university researchers. Research Policy 37(8): 1255–1266.