Exploring the Curvilinear Relationship between Academic-Industry Collaboration Environment and Innovation Performance: A Multilevel Perspective

Abstract

Academic institutions play a crucial role in knowledge production and driving innovation and economic growth. To enhance their capacity to deliver on these responsibilities, they are increasingly urged to establish academic–industry collaboration (AIC) environments to support research and innovation activities by their faculty and students. Despite the recognized importance of AIC in stimulating innovation performance, there is limited research exploring the cross-level impact of AIC on innovation performance. This study aims to address this research gap by investigating the relationship between the AIC environment and innovation performance while specifically examining the cross-level mediating role of researchers’ technological capability in this relationship. The study used a hierarchical linear model (HLM) approach, drawing on data from 187 researchers at 14 universities and academic institutions in the United Arab Emirates. The study’s results reveal a curvilinear (inverted U-shaped) effect of the AIC environment on innovation performance and a positive correlation between researchers’ technological capability and innovation performance. Furthermore, the AIC environment demonstrates a curvilinear effect on researchers’ technological capability, with researchers’ technological capability partially mediating the relationship between the AIC environment and innovation performance. These findings have significant theoretical and practical implications for policymakers in government agencies and university management seeking to develop evidence-based policies for effective science and technology management that enhance innovation performance.

1. Introduction

Collaboration between academic institutions and industries plays a pivotal role in generating knowledge, driving innovation, and promoting economic growth. The universities in the United Arab Emirates (UAE) have recognized the significance of bringing their students closer to successful entrepreneurs and industry leaders in order to foster innovation and entrepreneurship within student communities, thereby strengthening the national innovation ecosystem [1,2]. As a result, academic–industry collaboration has witnessed positive growth in the UAE over the past decade, with all involved parties, including students, universities, and industries, benefiting significantly. In recent years, a growing number of endeavors have been made to bolster higher education, establish technology parks, and invest in regional research and development (R&D). Collaboration between local universities and the private sector is pivotal to advancing science, technology, and the national innovation agenda. Enhancing academic–industry collaboration (AIC) presents many countries with an opportunity to benefit from AIC, promoting the transformation of scientific and technological achievements, supporting small- and medium-sized enterprises’ innovation, and encouraging entrepreneurship among researchers and students [3]. AIC enables enterprises to gain access to external resources such as talents, new knowledge, and technologies while also expanding academic research directions and ideas and broadening the channels of scientific research funds. For instance, in the UAE, research funding from enterprises and public institutions for national universities has witnessed a significant increase, from USD 31.888 billion in 2010 to USD 68.110 billion in 2022, with a compound annual growth rate of 7.88% [4].

In recent years, partnerships between academic institutions and industries have undergone significant changes, with a growing emphasis on diversity, experimentation, and public visibility. This trend has been accompanied by a call for global policies and laws to promote the commercialization of university science [5]. While industry research is driven by the creation of trade secrets, patents, and exclusive licenses for commercial gain, university research focuses on advancing knowledge and addressing social issues [6,7]. The hierarchical structure of industry research prioritizes confidentiality, intellectual property, and proprietary products, while university research is characterized by a more individualistic organizational structure and transparent priority setting and review processes. The UAE government has implemented various policies to encourage collaboration among academia, industry, and research institutions, including guidance for establishing industry technology innovation strategic alliances [8,9]. The joint guidelines issued in October 2021 by the National Intellectual Property Administration, the Ministry of Education, and the Ministry of Science and Technology aim to promote collaboration and knowledge transfer between academia, industry, and research institutions. These efforts demonstrate a commitment to fostering collaboration, promoting innovation and economic growth, and establishing long-term partnerships that benefit all stakeholders [10,11,12,13,14].

AICs have gained increasing attention from academic research and university bodies due to their strategic importance in promoting innovation and entrepreneurship within university communities. Numerous studies have investigated the relationship between AIC and innovation performance from various perspectives. Bellucci et al. [15] analyzed data from 100 universities and found that an inverted U-shaped relationship exists between the number of academic–industry collaborations and academic innovation performance. Different percentile points showed heterogeneity effects, and there were positive regulatory effects of collaboration breadth and knowledge capacity. Caviggioli et al. [16] examined data from universities in Taiwan and found higher innovation performance in universities with AICs. Using provincial and university statistics data from 2008 to 2017, Caviggioli et al. [17] investigated the relationship between organizational-level academic–industry collaborations and university innovation performance and identified a significant inverted U-shaped relationship. Previous studies have also explored AICs from the perspective of individual researchers. Chai et al. [18] found that there is a positive correlation between funding, knowledge, motivations for university–industry collaboration, and innovation performance among university researchers. Chen et al. [19] conducted a study on Italian scientists in the field of materials science who were globally recognized as highly cited scientists in 2014 and 2022 and discovered that scientists who collaborated with industry had a better academic performance. Similarly, Chen et al. [20] analyzed data from 652 researchers at a university in Denmark and found that researchers who collaborated closely with the industry had higher publication output and citation rates. These studies highlight the importance of AICs in promoting innovation and academic performance and emphasize the need for continued efforts to foster such collaborations.

While previous research has made significant contributions to understanding the connection between AIC and innovation performance, there are still several gaps in the literature that need to be addressed:

• Firstly, previous studies have analyzed the influence of AIC on innovation performance in universities, but the relationship between the two remains complex and not fully comprehended. The existing research has primarily concentrated on either the institutional or individual level. Nonetheless, cross-level research is necessary to investigate the effect of AIC at the institutional level on researchers’ innovation performance within universities.
• Secondly, although some studies have examined the link between AIC and innovation performance, there has been limited investigation into the mediating impact of researchers’ technological capability. Examining the mediating role of technological capability can provide researchers with a more thorough comprehension of how AIC influences innovation performance at the individual level.
• Finally, there is a dearth of research on AIC and innovation performance in universities and academic institutions within the United Arab Emirates (UAE). Given the recent emphasis placed on innovation and technological progress by the UAE, it is imperative to scrutinize the association between AIC and innovation performance within academic settings in the country. The scarcity of scholarly literature on this subject underscores the requirement to explore the impact of AIC on innovation performance in the region.

This study aims to fill the gaps in the existing literature by examining the impact of AIC on innovation performance in academic institutions. Specifically, this study takes a multilevel perspective by investigating the cross-level impact of AIC on innovation performance and examining the role of researchers’ technological capability as a mediator in this relationship, both at the individual and institutional level. The study’s data were sourced from the Dubai Statistics Center (https://www.dsc.gov.ae/en-us/Pages/default.aspx) (accessed on 7 March 2023), which is managed by the Government of Dubai. The academic institutions and universities in the United Arab Emirates (UAE) were specifically analyzed using data from the UAE’s Ministry of Industry and Advanced Technology Statistics (2012–2023), accessed on 7 March 2023. Hierarchical linear modeling (HLM) was used to test the model and hypotheses. The results of this study indicate that the AIC environment has a significant and positive impact on the innovation performance of researchers in universities and academic institutions. Furthermore, the technological capability of researchers partially mediates the relationship between AIC and innovation performance. This research provides theoretical guidance for promoting collaboration between academia and industry and enhancing the innovation performance of researchers in universities and academic institutions.

This study presents a noteworthy contribution to the existing literature on the relationship between AIC and innovation performance in academic institutions. The study’s main contributions are as follows:

• Firstly, the study focuses on the cross-level impact of AIC on innovation performance and the mediating role of technological capability among researchers. The objective of the research is to fill the gap in knowledge by supplementing the scanty research on the collaboration between universities and industries in emerging economies, such as the UAE. The research’s theoretical contribution is its cross-level methodology, which illustrates how the collaborative environment between academia and industry at the organizational level influences the individual level’s innovative performance.
• Secondly, the study investigates the significance of technological capability in augmenting innovation performance within the framework of academic–industry collaboration. The study offers empirical evidence of the mediating role of technological capability in the relationship between AIC and innovation performance.
• Thirdly, the study provides pragmatic guidance for university leaders and policymakers on formulating efficient tactics to foster academic–industry collaboration that bolsters innovation performance. The study presents a model to comprehend the factors that impact the extent of AIC in academic institutions, along with policies that encourage academic researchers to participate actively in AIC to enhance technological capability and innovation performance.

The remainder of this article is structured as follows: Section 2 discusses the research hypothesis, focusing on the direct impact of scientific innovation performance and the cross-level mediating role of technological capabilities. Section 3 outlines the study design, including sample selection, data sources, variable definitions, and model construction. Section 4 conducts empirical analysis through descriptive statistics, correlation analysis, and hypothesis testing. Section 5 discusses the implications of the findings, both theoretically and practically. Section 6 identifies limitations and suggests future research directions. Finally, Section 7 concludes the paper by summarizing the main contributions of the study.

2. Theoretical Analysis and Research Hypotheses

2.1. AIC Climate and Innovation Performance: A Cross-Level Perspective

In the field of psychology research, the term “climate” was first introduced by James, an American psychologist, to describe employees’ perceived work environment as a “psychological climate” and team members’ perceived work environment as an “organizational climate” [21]. Organizational climate is a construct at the institutional level that refers to the collective perceptions of practices, procedures, and behaviors that individuals expect, endorse, and foster within the institution [15]. The prior literature has demonstrated that institutional climate plays a critical role in shaping employee behavior and work performance within the institution. Chen et al. [22] found that researchers’ decision to collaborate with enterprises is influenced by the institutional climate and their cost–benefit assessments of academic–industry collaboration. Delorme [23] conducted a survey of 1428 researchers from nine German universities and discovered that the entrepreneurial orientation and network capabilities of academic chairpersons within the institution are essential components of institutional climate and have a direct impact on collaboration between scientists and non-academic institutions such as enterprises. Additionally, Compagnucci et al. [24] found that employee innovation performance is determined by three factors, namely, employee work engagement and creativity, as well as institutional innovation climate.

Based on relevant concepts and theories of institutional climate, this paper posits that the AIC climate can be defined as the shared perception of individual participation in AIC practices within an institution, which may affect members’ attitudes and behaviors towards AIC [25]. Scholars argue that the AIC climate at the institutional level in higher education institutions has a positive impact on the innovation performance of academic staff. Dias et al. [26] utilized data from 323 researchers at the Taiwan University of Science and Technology to demonstrate that the AIC climate at the institutional level in higher education institutions enhances the innovation performance of academic staff. They suggest that if academic staff work in a university department with close ties to industry, they can leverage resources accumulated from previous collaborations between the institution and industry. Dinu et al. [27] found that if a department within an institution collaborates more with industry, it can facilitate knowledge exchange with industry and obtain financial support from industry, which is conducive to academic innovation among researchers involved in AIC practices.

Nonetheless, while a moderate AIC climate within an institution can offer research personnel more chances to collaborate with industry and acquire support in terms of funding and experimental equipment, a high AIC climate may lead to most research staff within the institution participating in intensive AIC, which may potentially redirect their time and energy [28,29]. Additionally, it could also influence the autonomy of the institution and academic freedom of research staff [22,24,30], ultimately having a negative effect on innovation performance. Therefore, this research proposes that a moderate AIC climate at the institutional level can boost the innovation performance of academic staff, while a high AIC climate may hinder their innovation performance. Based on the above theoretical analysis, the following hypothesis is put forth:

H1. 

The AIC environment at the institutional level and the innovation performance of researchers within the institution have an inverted U-shaped relationship.

2.2. Cross-Level Mediating Effect of Researchers’ Technological Capability

The technological capability of researchers refers to their aptitude in conducting applied research with a focus on technology, which can be demonstrated through activities such as patent licensing and technology transfer. Technological capability is a critical factor in determining the success of collaborations between high-level scientific researchers in universities and enterprises, as well as their scientific innovation performance [31,32,33].

Numerous scholars have emphasized the importance of technological capability in promoting the improvement of innovation performance. For instance, Hong et al. [34], using data from Italian researchers in the field of materials science, found a positive correlation between the number of patents granted to researchers and their research output in terms of academic papers. Similarly, Hou et al. [35], using data from Japanese researchers, discovered a positive linear relationship between the number of patents and publications in journals with low impact factors. Additionally, Hu et al. [36], analyzing patent data in the nanotechnology field from three European countries, found that researchers with patents had more publications and higher citation rates than those without patents. These studies suggest that researchers with strong technological capabilities are likely to have better innovation performance. Thus, the following hypothesis is proposed:

H2. 

There is a significant positive relationship between the technological capability of researchers and their innovation performance.

The collaborative relationship between universities and enterprises can bring significant benefits to both parties. While universities can accelerate the transfer of technological knowledge to enterprises and broaden their research funding channels, enterprises can also gain access to cutting-edge research and technological advancements from academia [37]. This is particularly true for applied research knowledge possessed by enterprises, as university researchers can obtain knowledge from the applied research expertise that enterprises possess. In organizations with a low level of collaboration, increasing the number of researchers participating in academic–industry collaboration can improve the collaboration atmosphere and create more opportunities for researchers to obtain knowledge from enterprises [4,38]. This, in turn, can enhance the technological capabilities of researchers within the organization. However, there is a tipping point where the level of collaboration between universities and enterprises reaches a certain point, and the atmosphere can shift towards high-level collaboration. At this stage, researchers may limit their access to funding and new knowledge, hindering their technological capabilities. Additionally, a high level of collaboration can lead to an increased awareness of confidentiality among members, making researchers less willing to share knowledge for academic or commercial purposes. Such factors can negatively impact the technological capabilities of researchers within the organization [39]. Therefore, it is crucial to maintain an optimal level of the AIC environment that balances the benefits of collaboration with the potential risks. Based on the above theoretical analysis, the following hypothesis is proposed:

H3. 

There is an inverted U-shaped relationship between the AIC environment at the institutional level, viewed from a cross-level perspective, and the technological capability of researchers within the institution.

Viewing AIC from a broader perspective, researchers who engage in such partnerships can derive fresh research ideas from the participating companies. In fact, researchers who possess strong technological capabilities are better equipped to communicate with industry partners and collaborate on the development of new technologies and applications while simultaneously promoting novel academic ideas to enhance their innovation performance [39,40,41]. These researchers have been identified as academic inventors, who play a key role as intermediaries in the transfer of knowledge from universities to enterprises [42,43]. It is important to note that the organizational-level atmosphere of AIC can have a significant impact on the technological capabilities of researchers and, subsequently, on their innovation performance. Therefore, this study posits the following hypothesis:

H4. 

The technological capability of researchers serves as a cross-level mediator in the relationship between the AIC environment and the innovation performance of researchers.

The conceptual model proposed in this article depicts the mediating role of technological capabilities in the relationship between the AIC atmosphere and the innovation performance of university researchers, viewed from a cross-level perspective, as illustrated in Figure 1.

3. Study Design

3.1. Sample Selection and Data Source

The study utilized data obtained from the Dubai Statistics Center’s website (https://www.dsc.gov.ae/en-us/Pages/default.aspx) on 7 March 2023, which is under the governance of the Dubai government. The Ministry of Industry and Advanced Technology Statistics of the UAE was utilized to analyze academic institutions and universities in the United Arab Emirates (UAE). The sample population of the study comprised individual researchers from different universities in the UAE, representing diverse fields, including engineering, science, management, and social humanities.

To meet the HLM requirements, data on the academic–industry collaboration atmosphere at the institutional level were collected from various departments of schools in different universities in the UAE, covering the period from 2013 to 2023. Certain criteria were used to select the data, namely organizational samples with fewer than five individuals were excluded in the first level, and only relevant data from researchers at the main campus of each university were included. Researchers from other campuses were excluded because their data only indicate the school or campus to which they belong, without specific departments and disciplines. This approach ensured that the collected data were representative and relevant to the analysis.

Following the selection process, 187 researchers were included in the final individual-level sample, while 14 academic institutions and universities were included at the institutional level. Based on the methodology described by Moineddin et al. [44], this sample size was considered appropriate for conducting multilevel analysis. Maas et al. [45] suggest that a Level-2 sample size greater than 50 with at least five individual samples per group is necessary to ensure unbiased and accurate estimates of regression coefficients, variances, and standard errors in multilevel analysis. Therefore, the sample size in this study is adequate for multilevel analysis, which enhances the reliability and validity of the study findings.

3.2. Definition of Variables

The present study seeks to examine the association between the academic–industry collaboration (AIC) environment and the innovation performance of academic researchers in higher education institutions, with technological capability as a cross-level mediator. To this end, four main types of variables were utilized. Firstly, the dependent variable was innovation performance, which was measured using the h-index of academic researchers from 2013 to 2023, as suggested by Wang et al. [46]. The h-index is regarded as a comprehensive measure, as it combines both the quantity and quality of an individual’s publications [47,48,49]. Therefore, this study used the h-index to assess the innovation performance of academic researchers.

Secondly, the independent variable was the AIC environment, defined as the collective perception of individuals towards their organization’s participation in AIC practices. The degree of participation in AIC practices at the organizational level was used to measure the AIC environment of higher education institutions [50,51,52]. The proportion of horizontal project funding obtained by all members within an organization to the total research funding was employed as an indicator of the organization’s overall participation in AIC practices. Furthermore, to test hypotheses H1 and H2, the squared term of the AIC environment was included in the model.

Thirdly, the cross-level mediating variable was technological capability, which was measured using the number of patents applied by university researchers from 2010 to 2023, as suggested by previous studies [53,54,55]. The data for journal publications and patent counts were sourced from the UAE National Intellectual Property Administration [3].

Lastly, two control variables at the individual (researcher) level, i.e., educational attainment and professional title, were included to account for their influence on the dependent variable. These variables were highly correlated with the dependent variable, and their inclusion was necessary to control their effects on the study’s outcome. The details of these variables are presented in

Table 1. Description of variables.

Variable		Description	Time Frame
Individual level	Innovation performance	Researcher’s h-index in WoS database	2013–2023
	Technological capabilities	Number of papers published and patents granted to the researcher in the State Intellectual Property Database
	Education level	Dummy variable: Dummy variable: With a value of 1 if the researcher has received a PhD in 2023, 0 otherwise
	Job title	Dummy variable: With a value of 1 if the researcher is a professor in 2023, 0 otherwise	2013–2023
Institutional level	Academic–industry collaboration environment	Measured by the degree of participation of the academic institution as a whole in the practice of academic–industry collaboration, and measured by the proportion of all members of the institution who have received funding for horizontal projects of the total research funding	2013–2023

.

3.3. Model Construction

This study proposes a cross-level model of the influence of the academic–industry collaboration (AIC) environment on the innovation performance of academic researchers. The hierarchical linear and nonlinear modeling (HLM) method was employed to investigate the impact of the AIC environment at the institutional level on the innovation performance of academic researchers, as well as the cross-level mediating effect of the technological capabilities of researchers. The statistical software used for analysis in this study was HLM 7.0.

The cross-level model investigating the effect of the AIC environment on the innovation performance of university researchers consisted of two levels: Level-1 Model and Level-2 Model.

Level-1 Model is represented as follows:

$$\begin{gathered} {Innovation performance}_{ij}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em} \\ ={\beta }_{0j}+{\beta }_{1j}(Education Level)+{\beta }_{2j}(Professional Title)+r_{ij} \end{gathered}$$

(1)

Level-2 Model is represented as follows:

$${\beta }_{0j}={\gamma }_{00}+{\gamma }_{01}(AIC environment)+{\gamma }_{02}(AIC environment squared)+u_{0j}$$

(2)

where Equation (1) is the model at the individual level, while Equation (2) is the model at the institutional level. The variables $r_{ij}$ and $u_{0j}$ represent the residuals at the individual and institutional levels, respectively.

To verify the inverted U-shaped relationship between the organizational-level AIC environment, the technological capabilities of academic researchers, and their innovation performance, this study introduces a cross-level mediating effect model incorporating the technological capabilities variable. However, the traditional three-step testing method is unable to clearly demonstrate the relationship pathway among variables. Therefore, the moderated mediation effect model proposed by Muller et al. [56] was employed to test the research hypothesis. This approach can comprehensively analyze the moderated effects on all the possible pathways in the mediating model. By setting the moderator variable the same as the independent variable, the mediating process of the inverted U-shaped relationship between the independent variable and the dependent variable can be analyzed, as shown in Figure 2.

To examine the mediated moderation model shown in Figure 2, three separate regression equations need to be tested step by step, namely

$$Y={\alpha }_0+{\alpha }_{1}X+{\alpha }_{2}M+{\alpha }_{3}X·M+e_0$$

(3)

$$Z={\beta }_0+{\beta }_{1}X+{\beta }_{2}M+{\beta }_{3}X·M+e_1$$

(4)

$$Y={\gamma }_0+{\gamma }_{1}X+{\gamma }_{2}M+{\gamma }_{3}X·M+{\gamma }_{4}Z+{\gamma }_{5}Z·M+e_2$$

(5)

As described in Equations (3)–(5), testing the mediated moderated effect involves the following three steps:

• In Equation (3), the coefficient ${\alpha }_3$ of the interaction term $X·M$ between the independent variable and the moderating variable is significantly different from zero, indicating that the variable M significantly moderates the relationship between the independent variable X and the dependent variable Y;
• In Equation (4), the coefficient ${\beta }_3$ of the interaction term $X·M$ between the independent variable and the moderating variable is significantly different from zero, indicating that the variable M significantly moderates the relationship between the independent variable X and the mediating variable Z;
• In Equation (5), the coefficient ${\gamma }_4$ of Z is significantly different from zero, while the coefficient γ3 of the interaction term $X·M$ between the independent variable and the moderating variable is either no longer significant or still significant but smaller than α3 in Equation (1). This suggests that the moderating effect of M on the relationship between X and Y occurs on the X~Z path. Additionally, the coefficient γ5 of the interaction term $Z·M$ between the mediating variable and the moderating variable is not significant, indicating that the moderating effect of M does not occur on the Z ~ Y path.

In this study, the dependent variable Y represents the innovation performance of academic researchers, the independent variable X and the moderating variable M are the same variables of the AIC environment, the mediating variable Z is technological capability, $X·M$ represents the squared term of the independent variable of the AIC environment, and $Z·M$ represents the interaction term between technological capability and the AIC environment.

4. Empirical Analysis

4.1. Descriptive Statistics, Correlation Analysis, and Multicollinearity Test

Table 2. Descriptive statistics and correlation analysis of variables.

Variable		Mean	SD	Min	Max	1	2	3	4
Individual level	Research performance	3.5	4.05	0	38	1.000
	Technological capability	6.61	10.623	0	134	0.340 ***	1.000
	Technological capability squared	155.875	794.433	0	17678	0.184 ***	0.820 ***	1.000
	Educational level	0.84	0.38	0	1	0.197 ***	0.139 ***	0.064 **	1.000
	Job title	0.67	0.48	0	1	0.256 ***	0.156 ***	0.095 ***	0.149 ***
Institutional level	Academic–industry collaboration environment	0.452	0.305	0.02	0.20

Note: *** $p<0.01$, ** $p<0.05,$ individual-level sample size was 187; institutional-level sample size was 14.

displays the descriptive statistics, correlation coefficients, and significance levels of the main research variables at both individual and organizational levels, including means, standard deviations, and minimum and maximum values. The following findings can be deduced from Table 2:

• The mean value of technological capability among researchers was 6.61, with a standard deviation of 10.623, indicating a significant variability in technological capability among researchers;
• Education level, professional title, and technological capability were all positively correlated with innovation performance at a significance level of 0.01, with technological capability exhibiting the highest correlation coefficient of 0.340;
• Except for the high correlation between the squared term of technological capability and technological capability (0.820), the correlations between other variables were relatively low;
• According to the correlation analysis, both technological capability and its squared term were significantly and positively correlated with innovation performance at a significance level of 0.01.

It is worth noting that individual-level variables were significantly and positively correlated not only with innovation performance but also with each other, as indicated in Table 2. To investigate the possibility of multicollinearity among independent variables, a VIF test was conducted, and the results are presented in

Table 3. VIF test.

Variable	Technological Capability	Technological Capability Squared	Educational Level	Academic Qualification
VIF	3.60	5.44	1.18	1.24

. The VIF values of all individual-level variables were less than 10, with the highest VIF value being 5.44, indicating no multicollinearity issues among independent variables.

4.2. Hypothesis Testing

The analytical method and software employed in this study was HLM. In conjunction with Equations (1)–(3), the cross-level model was analyzed by following the steps outlined below to test the hypothesis, and the results are presented in

Table 4. Results of the HLM analysis.

Variable	Null Model	Random Model	Intercept Model		Slope Model
	Step 1	Step 2	Step 3	Step 4	Step 5
Intercept term, ${\gamma }_{00}$	3.712 *** (0.318)	0.922 *** (0.293)	1.448 *** (0.332)	6.586 *** (0.529)	1.849 *** (0.313)
Level-1 predictor
Educational attainment, ${\gamma }_{10}$		1.153 *** (0.27)	1.325 *** (0.28)	2.248 *** (0.628)	1.09 *** (0.268)
Title, ${\gamma }_{20}$		1.402 *** (0.308)	1.729 ***(0.326)	2.875 *** (0.57)	1.408 *** (0.315)
Technological capability, ${\gamma }_{30}$		0.128 *** (0.018)			0.137 *** (0.016)
Level-2 Predictors
Academic–industry collaboration environment, ${\gamma }_{01}$			11.673 *** (2.862)	29.829 *** (7.883)	7.856 ** (3.196)
Academic–industry collaboration environment squared, ${\gamma }_{02}$			−12.887 *** (2.748)	−35.079 *** (7.666)	−8.694 *** (3.213)
Technological capability group mean, ${\gamma }_{03}$					0.157 *** (0.068)
Cross-level interaction effect
Technological capabilities × Academic–industry collaboration environment, ${\gamma }_{31}$					0.077 (0.08)
Variance
${\sigma }^2$	13.044	10.729	12.158	95.278	9.749
${\tau }_{00}$	4.265	1.508	2.524	8.04	2.658
${\tau }_{31}$		0.008			0.004
Number of deviations	5428.186	5270.564	5329.024	7323.514	5229.444

Note: *** $p<0.01$, ** $p<0.05$.

.

• Step 1: Null Model

As this study assumes that the research performance of university researchers at the individual level can be predicted by variables at both individual and organizational levels, it is necessary to demonstrate the existence of variation in research performance at both levels. To do so, a null model must be established without any predictor variables, and only when the variance of within-group and between-group is significant ($ICC\left(1\right)>0.12$) [57] can the intercept and slope analysis be performed.

$$\mathrm{Level}-1 \mathrm{Model}: {Innovation performance}_{ij}={\beta }_{0j}+r_{ij}$$

(6)

$$\mathrm{Level}-2 \mathrm{Model}: {\beta }_{0j}={\gamma }_{00}+u_{0j}$$

(7)

In the above model, ${\beta }_{0j}$ represents the average innovation performance of the jth group; ${\gamma }_{00}$ represents the overall average innovation performance; the variance of $r_{ij}$ represents the within-group variance of innovation performance and is denoted as ${\sigma }^2$; and the variance of $u_{0j}$ represents the between-group variance of innovation performance and is denoted as ${\tau }_{00}$.

The total variance of innovation performance = ${\sigma }^2+{\tau }_{00}$, $ICC\left(1\right)$ can be calculated as the percentage of between-group variance using the following formula:

$$ICC\left(1\right)={\tau }_{00}/({\sigma }^2+{\tau }_{00})$$

(8)

From Table 4, it can be observed that the between-group variance of innovation performance was ${\tau }_{00}$= 4.265, and the chi-square test results indicate that this variance was significant: $\chi 2\left(53\right)=321.704$, p < 0.001. The within-group variance of innovation performance (${\sigma }^2$ = 13.044), so $ICC\left(1\right)$ = 0.247. This indicates that 24.7% of the variability in innovation performance among university researchers (17.308 = ${\sigma }^2+{\tau }_{00}$) was due to group differences (4.265), while the other 75.5% of the variability resulted from individual-level factors. According to the criterion suggested by Joyce et al. [57] (greater than 0.12), this represents a strong level of association, and there is no need to calculate the value of $ICC\left(2\right)$. Therefore, it is not appropriate to analyze this with a general regression model, and an HLM model should be used for analysis.

• Step 2: Testing Hypothesis 2 (Random-Parameter Regression Model)

To test Hypothesis 2, in Step 2, a random-parameter regression model was employed by assigning the educational level, job title, and technological capabilities variables of researchers to Level-1, to verify whether Hypothesis 2 holds. At the same time, it was determined whether the innovation performance of different individual researchers has different intercepts and slopes, creating conditions to examine the influence of institutional-level (Level-2) contextual variables. It should be noted that the educational level and job title were control variables in this study, and whether their random variation affects the research performance of researchers was not investigated. Therefore, no random effects (error term $u$) were included in estimating their coefficients at Level-2.

$$\begin{gathered} \mathrm{Level}-1 \mathrm{Model}: {Innovation performance}_{ij}={\beta }_{0j}+{\beta }_{1j}(educational level)+{\beta }_{2j}(job title)+ \\ {\beta }_{3j}(technological capability)+r_{ij} \end{gathered}$$

(9)

$$\begin{gathered} \mathrm{Level}-2 \mathrm{Model}: {\beta }_{0j}={\gamma }_{00}+u_{0j},{\beta }_{1j}={\gamma }_{10}, \\ {\beta }_{2j}={\gamma }_{20}, and {\beta }_{3j}={\gamma }_{30}+u_{3j} \end{gathered}$$

(10)

In the aforementioned model, γ₀₀ represents the mean of the cross-organizational intercepts; ${\gamma }_{30}$ represents the mean of the cross-organizational slopes (used to test Hypothesis 2). The variance of $r_{ij}$ = ${\sigma }^2$ = variance of Level-1 residuals; the variance of $u_{0j}$ = ${\tau }_{00}$= variance of intercepts, while the variance of $u_{3j}$ = ${\tau }_{30}$ = variance of slopes. The parameters γ₀₀ and ${\gamma }_{10}$ are the cross-organizational averages of Level-1 coefficients (i.e., ${\beta }_{0j}$ and ${\beta }_{3j}$), where ${\gamma }_{30}$ indicates the cross-organizational relationship between technological capability and innovation performance, and can therefore be used to test Hypothesis 2. As shown in Table 4, the control variables of education level and job title were both significant at the 0.01 level, and the coefficient of technological capability, ${\gamma }_{30}$ = 0.128, which was positively correlated with innovation performance at the 0.01 level of significance, providing support for Hypothesis 2. In addition, the variance of intercepts, ${\tau }_{00}$, was 1.508 ($\chi 2\left(53\right)=112.907$, p < 0.001), indicating that there may be organizational-level factors at Level-2. Thus, the next step is to test Hypothesis 1.

• Step 3: Testing Hypothesis 1 (Intercept-Only Predictive Model)

Step 3 involved adding the AIC environment as a variable to Level-2 while retaining only the control variables in Level-1. A model was then estimated with the intercept serving as the outcome variable to examine Hypothesis 1.

$$\begin{gathered} \mathrm{Level}-1 \mathrm{Model}: {Innovation performance}_{ij}={\beta }_{0j}+{\beta }_{1j}(educational level)+{\beta }_{2j}(professional title)+ \\ {r}_{ij} \end{gathered}$$

(11)

$$\begin{gathered} \mathrm{Level}-2 \mathrm{Model}: {\beta }_{0j}={\gamma }_{00}+{\gamma }_{01}(AIC environment)+{\gamma }_{02}(square of AIC environment)+u_{3j}, \\ {\beta }_{0j}={\gamma }_{10},{\beta }_{2j}={\gamma }_{20} \end{gathered}$$

(12)

In the above model, ${\gamma }_{00}$ represents the intercept term in Level-2; ${\gamma }_{20}$ represents the effect of the AIC environment on innovation performance (used to test Hypothesis 1); the variance of $r_{ij}$ is ${\sigma }^2$, which is the residual variance of Level-1; the variance of $u_{0j}$ is ${\tau }_{00}$, which is the variance of the intercept. The parameter ${\gamma }_{20}$ represents the estimated relationship between the AIC environment and innovation performance after controlling for educational level and professional title in Level-1. According to Table 4, the value of ${\gamma }_{20}$ was −12.887, with p < 0.01, indicating a nonlinear, inverted U-shaped effect of institutional-level AIC environment on innovation performance. Therefore, Hypothesis 1 is supported.

• Step 4: Testing Hypothesis 3 (Intercept-Only Predictive Model)

Step 4 involved keeping the Level-2 variables constant and estimating a model with the intercept as the outcome variable while using the technological capability of the researcher as the dependent variable in Level-1.

$$\mathrm{Level}-1 \mathrm{Model}: {Technological capabilities}_{ij}={\beta }_{0j}+{\beta }_{1j}(education level)+{\beta }_{2j}(professional title)+r_{ij}$$

(13)

$$\begin{gathered} \mathrm{Level}-2 \mathrm{Model}: {\beta }_{0j}={\gamma }_{00}+{\gamma }_{01}(AIC environment)+{\gamma }_{02}(Square of AIC environment)+u_{3j}, \\ {\beta }_{0j}={\gamma }_{10},{\beta }_{2j}={\gamma }_{20} \end{gathered}$$

(14)

In the aforementioned model, ${\gamma }_{00}$ refers to the intercept term at Level-2, while ${\gamma }_{02}$ represents the effect of the AIC environment on technological capability, which is used to test Hypothesis 3. The variance of $r_{ij}$ is ${\sigma }^2$, which is the variance of the Level-1 residual, while the variance of $u_{0j}$ is ${\tau }_{00}$, which is the variance of the intercept. ${\gamma }_{02}$ represents the estimated relationship between the AIC environment and the technological capability of researchers after controlling for the Level-1 education level and job title variables. According to Table 4, ${\gamma }_{02}$ was −35.079 with p < 0.01, indicating that the influence of the AIC environment at the institutional level on the technological capability of researchers was not linear, but rather had an inverted U-shaped pattern. Therefore, Hypothesis 3 is supported.

• Step 5: testing Hypothesis 4 (the slope-predicting model).

In order to test Hypothesis 4, a model was estimated with the slope as the outcome variable. The Level-2 AIC environment was used as a predictor of the slope coefficient (${\beta }_{3j}$) to determine whether this Level-2 variable can explain the variation in the slope. Additionally, because it involved a cross-level interaction, the technological capability variable in Level-1 was group-centered, while in Level-2, it was used as an independent variable, and the group mean was utilized.

$$\begin{gathered} \mathrm{Level}-1 \mathrm{Model}: {Innovation performance}_{ij}={\beta }_{0j}+{\beta }_{1j}(educational level)+{\beta }_{2j}(professional title)+ \\ {\beta }_{3j}(group-centered technological capability)+r_{ij} \end{gathered}$$

(15)

$$\begin{array}{c}\mathrm{Level}-2 \mathrm{Model}: {\beta }_{0j}={\gamma }_{00}+{\gamma }_{01}(AIC environment)+{\gamma }_{02}(square of AIC environment)+\\ {\gamma }_{03}(group mean of technological capability)+u_{0j}\\ {\beta }_{0j}={\gamma }_{10},{\beta }_{2j}={\gamma }_{20}\\ {\beta }_{3j}={\gamma }_{30}+{\gamma }_{31}\left(AICenvironment\right)+u_{3j}\end{array}$$

(16)

In the above model, ${\gamma }_{00}$ is the intercept of Level-2 (with the intercept of Level-1 model as the dependent variable); ${\gamma }_{01}$ = slope of the AIC environment; ${\gamma }_{02}$ = slope of the square of the AIC environment (used to test Hypothesis 4); ${\gamma }_{30}$ = intercept of Level-2 (with the slope of technological capability in the Level-1 model as the dependent variable, used to test Hypothesis 4); ${\gamma }_{31}$ = slope of Level-2, which represents the moderating effect of the AIC environment on the relationship between technological capability and innovation performance of researchers (used to test Hypothesis 4); the variance of $r_{ij}$ is ${\sigma }^2$, which is the variance of the Level-1 residual; the variance of $u_{0j}$ is ${\tau }_{00}$, which is the variance of the intercept; the variance of $u_{3j}$ is ${\tau }_{30}$, which is the variance of the slope residual.

Hypothesis 4 posits that technological capability serves as a cross-level mediator between institutional-level AIC environment and individual-level innovation performance. As shown in Table 4, the coefficient ${\gamma }_{30}$ for the mean of the Level-1 technological capability group was 0.136 and significant at the 0.01 level, and the coefficient ${\gamma }_{30}$ for the mean of the Level-2 technological capability group was 0.156 and also significant at the 0.01 level. The coefficient ${\gamma }_{02}$ for the square of the AIC environment remained significant at the 0.01 level but with a value of −8.694, which was smaller than the value of ${\gamma }_{02}$ in Step 3 (−12.887). Moreover, the coefficient ${\gamma }_{31}$ for the interaction between technological capability and the AIC environment was not significant, indicating that the moderation effect did not occur on the path from the technological capability to innovation performance among academic researchers. In summary, based on the results of all the models, the institutional-level AIC environment had a curvilinear (inverted U-shaped) relationship with the technological capability of researchers, and this effect was partially mediated by technological capability and influenced the innovation performance of researchers. Thus, Hypothesis 4 is supported.

The results presented in

Table 5. Results of path coefficients and hypotheses testing.

Hypothesis	Relationship	Test Results
H1	The AIC environment at the institutional level and the innovation performance of researchers within the institution have an inverted U-shaped relationship.	Supported
H2	There is a significant positive relationship between the technological capability of researchers and their innovation performance.	Supported
H3	There is an inverted U-shaped relationship between the AIC environment at the institutional level, viewed from a cross-level perspective, and the technological capability of researchers within the institution.	Supported
H4	The technological capability of researchers serves as a cross-level mediator in the relationship between the AIC environment and innovation performance of researchers.	Supported

demonstrate the findings obtained from the testing of hypotheses. Firstly, H1 states that the AIC (Academic-Industry Collaboration) environment at the institutional level and the innovation performance of researchers within the institution exhibit an inverted U-shaped relationship. The analysis provides support for this hypothesis, indicating that there is indeed an inverted U-shaped relationship between these variables. Secondly, H2 proposes a significant positive relationship between the technological capability of researchers and their innovation performance. The results confirm this hypothesis, indicating a significant positive relationship between these variables. Thirdly, H3 suggests an inverted U-shaped relationship between the AIC environment at the institutional level, when considered from a cross-level perspective, and the technological capability of researchers within the institution. The findings support this hypothesis, confirming the existence of an inverted U-shaped relationship. Lastly, H4 posits that the technological capability of researchers serves as a cross-level mediator in the relationship between the AIC environment and the innovation performance of researchers. The results support this hypothesis, providing evidence that the technological capability of researchers indeed acts as a mediator in this relationship.

5. Discussion

This empirical study investigates how the academic–industry collaboration (AIC) environment affects the innovation performance of researchers within academic institutions, and how researchers’ technological capability mediates this relationship from a cross-level perspective. Data were collected from 187 researchers across 14 academic institutions and universities in the United Arab Emirates (UAE) and analyzed using hierarchical linear modeling (HLM) methods and software. The results suggest the following features:

Firstly, there exists an inverted U-shaped pattern in the influence of the academic–industry collaboration atmosphere at the organizational level on the scientific innovation performance of scientific researchers in the organization, as viewed from a cross-level perspective. Despite the importance of academic–industry collaboration, few scholars have examined its impact on the scientific innovation performance of university scientific researchers from a cross-level perspective. This finding is consistent with previous research in the literature (see [29,30,31]). Based on theoretical analysis, this study contends that a low academic–industry collaboration atmosphere within an organization can enhance the scientific innovation performance of scientific researchers in the organization. Conversely, as the academic–industry collaboration atmosphere within the organization strengthens, the positive impact on the scientific innovation performance of scientific researchers will reach a peak. Further elevating the academic–industry collaboration atmosphere beyond this point may impede the improvement in the scientific innovation performance of scientific researchers. Table 4 confirms this hypothesis. Hence, a favorable academic–industry collaboration atmosphere is a valuable resource for scientific researchers in the organization, offering them a platform to acquire new knowledge and opportunities to interact with enterprises, and fostering innovation in both academic and applied fields [32,39,41]. However, an excessive academic–industry collaboration atmosphere may lead to deviation from the research direction and negatively affect the scientific innovation performance of scientific researchers.

Secondly, the technological capability of scientific researchers was found to have a significant positive correlation with their scientific innovation performance. This finding suggests that scientific researchers who possess greater technological skills are better equipped to carry out research projects effectively and efficiently, leading to improved innovation performance. This finding is significant because it challenges the assumption that engaging in applied research may hinder the academic research of scientific researchers [54,55,58]. Instead, the evidence suggests that there is at least some level of complementarity between applied and academic research, where technological skills developed through applied research can actually enhance the performance of academic research. Furthermore, this correlation holds across disciplinary backgrounds, indicating that technological capabilities are important for scientific researchers across different fields. Therefore, organizations that seek to improve their scientific innovation performance may benefit from investing in the development of the technological capabilities of their scientific researchers, regardless of their disciplinary background.

Thirdly, from a cross-level perspective, the academic–industry collaboration atmosphere at the organizational level has an inverted U-shaped pattern on the technological capability of scientific researchers. A moderate level of academic–industry collaboration atmosphere within an organization facilitates scientific researchers in finding cooperative enterprises and acquiring new knowledge or information from enterprises through exchanges with their colleagues. Therefore, a significant positive correlation exists between the academic–industry collaboration atmosphere and the technological capability of scientific researchers [59,60,61]. However, the empirical research findings suggest that a high level of academic–industry collaboration atmosphere at the organizational level has a negative impact on the scientific innovation performance of scientific researchers. One of the reasons for this may be due to confidentiality issues [62,63]. Participating in academic–industry collaboration can limit the exchange of scientific researchers with their colleagues and create confidentiality issues. The academic–industry collaboration atmosphere is a psychological environment. When the academic–industry collaboration atmosphere is low, scientific researchers within the organization do not pay much attention to confidentiality issues while exchanging information with their colleagues. However, when the academic–industry collaboration atmosphere becomes too strong, it is likely that scientific researchers will pay more attention to confidentiality issues while exchanging information with their colleagues, which may impede their scientific innovation performance [3,64].

Finally, it has been found that the technological capability of university researchers plays a partially mediating role in the relationship between the academic–industry collaboration atmosphere at the organizational level and the innovation performance of researchers. In cases where the level of academic–industry collaboration is low, enhancing the atmosphere of collaboration can improve the technological capability of researchers, leading to better innovation performance. However, the influence of the academic–industry collaboration atmosphere on the technological capability of researchers follows an inverted U-shaped relationship, where further improvement beyond a certain point has a negative effect on the technological capability of researchers. This negative effect is mediated by the partial mediation effect of technological capability on the innovation performance of researchers. In other words, the inverted U-shaped relationship between the academic–industry collaboration atmosphere and the technological capability of researchers at the organizational level affects the innovation performance of researchers through the partial mediation effect of technological capability [47,48,65].

6. Implications

The present study contributes to the advancement of theory and research in the field of innovation management, providing insights that have important practical implications for universities, industry, and policymakers.

6.1. Theoretical Implications

The present study provides three noteworthy theoretical contributions to the field of innovation management. Firstly, prior research has concentrated on examining the impact of academic–industry collaboration on researchers’ innovation performance either from an organizational or an individual level perspective. However, this study employs a cross-level (institutional → individual) approach, which enriches our understanding of the complex dynamics between academic–industry collaboration and researchers’ innovation performance. This approach emphasizes the intricate and dynamic interplay between the organizational atmosphere of academic–industry collaboration and the innovation performance of university researchers. Prior research has mainly focused on the impact of academic–industry collaboration from only one level of analysis, neglecting the impact of the other level [46,66]. The cross-level approach used in this study allows researchers to obtain a more comprehensive and nuanced understanding of the mechanisms that link academic–industry collaboration and innovation performance. This approach is particularly pertinent in the current climate, where universities are increasingly partnering with industry to transfer knowledge and stimulate innovation.

Secondly, this study provides valuable insights into the role of technological capability in mediating the relationship between academic–industry collaboration and innovation performance at the organizational level. It highlights the need for universities and industry partners to work together to enhance researchers’ technological capability and maximize the benefits of collaboration. This finding provides a more comprehensive perspective on the impact of academic–industry collaboration on researchers’ innovation performance, as it takes into account the role of technology in the process of innovation. By incorporating technological capability as a mediator, this study shows that academic–industry collaboration has a greater impact on innovation performance when researchers possess the technological capability to take advantage of the knowledge and resources provided by industry partners. Moreover, the mediating role of technological capability highlights the importance of enhancing researchers’ skills and knowledge to drive innovation in academic–industry collaborations. Future research can build on this finding by exploring how universities can best equip their researchers with the technological capability necessary to realize the full potential of academic–industry collaborations.

Lastly, this study provides a new perspective on the mechanisms that underlie the relationship between academic–industry collaboration and innovation performance in the United Arab Emirates (UAE). While prior research has established the positive impact of academic–industry collaboration on innovation performance, this study goes further to explore the specific factors that drive this relationship within the unique context of the UAE [1,3]. The UAE is a country that has placed significant emphasis on developing its knowledge-based economy and fostering innovation in recent years. As such, it offers a valuable setting for investigating the impact of academic–industry collaboration on innovation performance. This study finds that the limited linkage between academia and industry in the UAE provides opportunities for collaboration that can boost student innovation capabilities and entrepreneurship. Furthermore, the study suggests that creating a supportive organizational atmosphere for collaboration can facilitate knowledge exchange and innovation, leading to enhanced innovation performance.

6.2. Practical Implications

This study’s theoretical contributions also have practical implications for policymakers, academic institutions, and industry partners interested in fostering innovation and driving economic growth through academic-industry collaboration.

Firstly, the findings of this study provide guidance for departmental leaders in universities to establish academic-industry collaboration systems that enhance the innovation performance of their departments [7,67]. The research indicates that leaders in universities with limited levels of academic-industry collaboration should formulate appropriate policies and systems to encourage their researchers to actively engage in such collaborations. This is because fostering a collaborative environment not only facilitates the direct enhancement of members’ technological capabilities but also positively impacts researchers’ innovation performance through the mediation of technological capabilities. Conversely, for universities with a high level of academic-industry collaboration, management should develop effective strategies to suppress collaboration intensity among their members to improve researchers’ innovation performance. This not only contributes to the advancement of researchers’ technological capabilities but also improves their overall innovation performance.

Secondly, these findings provide insights that can inform policy and practice in the UAE and other similar contexts seeking to promote innovation through academic-industry collaboration. The study emphasizes the importance of fostering collaboration between academia and industry and highlights the role of supportive institutional environments and technological capability in driving innovation performance. By offering new insights into the mechanisms underlying the relationship between academic-industry collaboration and innovation performance in the UAE, this study contributes to the growing body of literature on innovation management and provides a foundation for future research in this area.

Lastly, this study’s findings have practical implications for policymakers, academic institutions, and industry partners, providing guidance for developing effective academic-industry collaboration systems and strategies that can improve innovation performance. Moreover, the insights gained from this study can inform policymaking and practice in the UAE and other contexts seeking to promote innovation through academic-industry collaboration, contributing to the advancement of the field of innovation management.

7. Limitations and Future Research Perspective

Despite the valuable contributions of this study, some limitations are worth noting. Firstly, the present study was conducted with a sample size of 187 researchers drawn from various academic institutions and universities in the UAE. While this sample size allowed for an in-depth investigation within the specified context, it is important to acknowledge the potential limitations in terms of generalizability to other regions or populations. Given the restricted scope of the sample, caution should be exercised when extrapolating the findings to broader contexts. To mitigate this limitation, it is recommended that future research endeavors consider expanding the sample size by including a more diverse range of participants from universities and regions beyond the UAE. By incorporating a larger and more varied sample, the external validity of the study’s results can be strengthened, facilitating more comprehensive inferences and enhancing the applicability of the findings across diverse settings. Moreover, a more extensive and diverse sample would enable researchers to capture potential variations or nuances in the relationships and dynamics under investigation, offering a more representative understanding of the broader academic community.

Secondly, the data collected for this study covered a relatively short time frame, spanning from 2010 to 2023. This time frame may not fully capture the long-term impact of academic-industry collaboration on innovation performance. Therefore, future studies should extend the research period to track the development of innovation performance over an extended period. Additionally, future studies should consider gathering qualitative data to gain a better understanding of researchers’ experiences and perceptions of academic-industry collaboration and its impact on innovation performance.

Thirdly, this study focused on the role of technological capability as a mediator in the relationship between academic-industry collaboration and innovation performance. However, there are other potential mediators that future research could explore, such as institutional culture, collaboration intensity, and knowledge transfer mechanisms. These factors could all play a significant role in the relationship between academic-industry collaboration and innovation performance. Therefore, future studies should expand the scope of investigation to explore these potential mediators and gain a more comprehensive understanding of the complex dynamics between academic-industry collaboration and innovation performance.

Finally, it should be noted that the data utilized in this study primarily relied on publicly available sources, which did not offer specific details on the job titles or academic backgrounds of industry collaborators. As a result, a comprehensive examination of the impact of managers with academic backgrounds on the performance of Academic-Industry Collaboration (AIC) was not feasible within the scope of this study. Furthermore, the study primarily focused on assessing the influence of AIC collaborations on researchers’ technological capabilities, rather than specifically investigating the role of managers. To address these limitations, future research endeavors could strive to collect more comprehensive and detailed data, including specific information on job titles and the academic backgrounds of industry managers involved in AIC projects. Such an approach would enable a more thorough analysis of the relationship between managers’ academic backgrounds, their role in facilitating collaboration, and the resulting performance of AIC. Additionally, qualitative research methods, such as interviews or surveys, could be employed to obtain insights from both academic and industry stakeholders, shedding light on the influence of managers with academic experience on AIC collaborations.

8. Conclusions

This study investigated the cross-level mediation effect of technological capability on the relationship between the academic–industry collaboration atmosphere and the innovation performance of academic researchers. The results show that academic–industry collaboration atmosphere has a positive effect on the technological capability of academic researchers, which in turn enhances their innovation performance. However, this effect follows an inverted U-shaped curve, indicating that an optimal level of academic–industry collaboration atmosphere is needed to improve technological capability, beyond which it becomes counterproductive and negatively affects innovation performance.

The theoretical contribution of this study lies in its cross-level approach, which highlights the role of technological capability as a mediating variable in the relationship between academic–industry collaboration atmosphere and innovation performance. This study provides practical guidance for department managers to develop academic–industry collaboration policies that can effectively enhance innovation performance. For organizations with a low academic–industry collaboration atmosphere, promoting the active participation of academic researchers in academic–industry collaboration can enhance technological capability and innovation performance. On the other hand, for organizations with a high academic–industry collaboration atmosphere, managers should consider policies to optimize the benefits of academic–industry collaboration for innovation performance.

While this study has some limitations, such as limited sample size and the data being collected from a single university, it provides a foundation for future research in this area. Future studies can expand the scope of investigation by considering other potential mediators, such as organizational culture, collaboration intensity, and knowledge transfer mechanisms, to gain a more comprehensive understanding of the complex dynamics between academic–industry collaboration and innovation performance. Additionally, future studies can extend the research period to track the long-term impact of academic–industry collaboration on innovation performance and include researchers from a diverse range of universities and regions to enhance the external validity of the results.