OGD-driven innovation capacity

OGD is described as ‘making public sector data freely and publicly available for value extraction’ (Kalampokis, Tambouris, & Tarabanis, Reference Kalampokis, Tambouris and Tarabanis2011). Recently, governments have increasingly emphasized the innovation value of OGD and taken a series of measures. For example, the UK has established the Open Data User Group for the promotion of private sector innovation in OGD (Open Data User Group, 2012). And China has held competitions such as the ‘Zhejiang Open Data Innovation and Application Competition’ for the improvement of the innovative application of OGD by individuals, universities, and firms (Zhejiang Open Data Innovation and Application Competition, 2020). These national measures indicate that the government is actively stimulating the innovative application of OGD among users. Simultaneously, firms should also seize the opportunity to overcome multiple obstacles and enhance their capacity to employ OGD for product or service innovation, called OGD-driven innovation capacity.

Most academic research on OGD at the firm level has focused on OGD adoption. For example, Kaasenbrood, Zuiderwijk, Janssen, de Jong, and Bharosa (Reference Kaasenbrood, Zuiderwijk, Janssen, de Jong and Bharosa2015) developed a framework for identifying factors influencing OGD adoption in private organizations. Wang and Lo (Reference Wang and Lo2020) constructed a firm-level model of factors influencing OGD adoption based on a socio-technical perspective. However, these research works only examine the factors driving OGD adoption rather than the factors impacting firms’ OGD-driven innovation capacity. In addition, Jetzek, Avital, and Bjorn-Andersen (Reference Jetzek, Avital and Bjorn-Andersen2014) developed a framework of OGD-driven innovation generation mechanisms to explain how firms stimulate the generation of economic and social value through OGD. This study, however, ignores a vital premise: whether firms have the capacity to drive innovation and achieve economic and social value through OGD. We must explore OGD-driven innovation capacity as a result of the fact that firms will face technical, organizational, and other obstacles after the introduction of OGD as a new innovation resource. The value of OGD-driven innovation cannot be achieved unless firms have enough capacity to cope with these obstacles.

Top leaders’ support and OGD-driven innovation capacity

Top leaders serve as executive sponsors for a given project (Yang, Reference Yang2008). The role of top leaders in the project is mainly responsible for providing precise guidance and allocating appropriate resources to meet the project’s goals (Hsu, Liu, Tsou, & Chen, Reference Hsu, Liu, Tsou and Chen2019; Rodríguez, Pérez, & Gutiérrez, Reference Rodríguez, Pérez and Gutiérrez2008; Rosenbloom, Reference Rosenbloom2000). Therefore, in this study, top leaders’ support can be defined as an act of proactive sponsorship by providing specific guidance and committing appropriate resources to OGD-driven innovation projects (Hsu et al., Reference Hsu, Liu, Tsou and Chen2019; Rodríguez, Pérez, & Gutiérrez, Reference Rodríguez, Pérez and Gutiérrez2008).

The contribution of top leaders’ support to OGD-driven innovation capacity manifests itself in two ways. First, from an economic standpoint, OGD does not have value on its own but rather integrates with other factors of production to generate value (Mei, Reference Mei2018). Top leaders may use their positions and power to deploy the various elements of innovation activities in an efficient manner (Wang & Lo, Reference Wang and Lo2020). This deployment facilitates the integration of OGD with the firm’s current resources, helps to overcome the institutional, human, and technological barriers that OGD-driven innovation activities confront, and improves the firm’s capacity to recognize commercial opportunities and capitalize on innovation. Simply put, top leaders’ support increases the probability of success for OGD-driven innovation by providing the necessary resources.

Second, top leaders’ support is able to strengthen the interaction between employees and OGD through their own influence. A key role of top leaders is to interpret information (Daft & Weick, Reference Daft and Weick1984). When top leaders express support for OGD-driven innovation activities, for example, by communicating their goals and benefits through letters, sensitization sessions, etc., they can shift employees’ attention to OGD-driven innovation (Garcia-Ortega, Lopez-Navarro, & Galan-Cubillo, Reference Garcia-Ortega, Lopez-Navarro and Galan-Cubillo2021; Kaplan, Klebanov, & Sorensen, Reference Kaplan, Klebanov and Sorensen2012). Another key role of top leaders’ support is to provide clarity of goals, which can help eliminate uncertainty about the adoption of new things within the firm and promote innovation (Kaplan, Klebanov, & Sorensen, Reference Kaplan, Klebanov and Sorensen2012; Weiner, Shortell, & Alexander, Reference Weiner, Shortell and Alexander1997). In accordance with Koziol-Nadolna (Reference Koziol-Nadolna2020), the supporting behaviors of leaders (e.g., articulating goal expectations and providing rewards) can effectively stimulate employees’ innovativeness, resulting in a flexible and open innovation firm. In addition, existing literature suggests that top leaders play an important role in promoting organizational learning (Shao, Feng, & Hu, Reference Shao, Feng and Hu2017). After introducing the new knowledge of OGD into the firm, top leaders can enhance employees’ learning and interaction with OGD by providing open data literacy boot camps (Iryna, Åke, & Marijn, Reference Iryna, Åke and Marijn2015).

In conclusion, this study makes the case that the top leaders’ support is, in fact, a proactive leadership approach that qualifies as transformational leadership. By exercising influence, establishing goals and incentives, and other means, this kind of leader may have a direct impact on how the organization views and responds to OGD-driven innovation (Zhou, Wang, Jiang, Li, & Li, Reference Zhou, Wang, Jiang, Li and Li2023). We propose the following hypothesis:

TOE framework as a mediator

OGD-driven innovation, in terms of output results, is the process by which firms use OGD, a new factor of production, to develop new products or services. Multiple variables will undoubtedly impact firms’ OGD-driven innovation capacity during this phase of innovation adoption and deployment. As a result, we require a broad framework to clarify these impacts and their consequences.

Many researchers have acknowledged Tornatzky and Fleischer’s TOE framework as a broad categorization framework for innovation adoption studies (Liang, Wang, Dong, Zhang, & Qi, Reference Liang, Wang, Dong, Zhang and Qi2021; Tornatzky, Fleischer, & Chakrabarti, Reference Tornatzky, Fleischer and Chakrabarti1990; Wang & Lo, Reference Wang and Lo2016; Zhang, Zhao, Zhang, Meng, & Tan, Reference Zhang, Zhao, Zhang, Meng and Tan2017). The TOE framework, which is highly adaptable and extensively relevant, categorizes the factors influencing innovation adoption as technical, organizational, and environmental. By examining these three aspects, we may gain a better understanding of the direct and indirect effects of top leaders’ support on OGD-driven innovation capacity.

Innovation support

According to the theory of organizational behavior, an organization’s survival and growth are dependent on the creativity of its people, and organizations may successfully boost employee creativity by creating a favorable work environment (Amabile & Conti, Reference Amabile and Conti1999). We believe that top leaders may create a good work environment by engaging in innovation-support behaviors such as financial assistance, presentations, and training, which stimulate workers to participate in OGD-related knowledge and skills, hence increasing OGD-driven innovation capacity.

While OGD enters the firm’s production system as an emerging innovation resource, it also brings new knowledge and new risks (Magalhaes & Roseira, Reference Magalhaes and Roseira2020; Zuiderwijk, Helbig, Gil-García, & Janssen, Reference Zuiderwijk, Helbig, Gil-García and Janssen2014). Employees must learn how to use OGD efficiently to achieve innovation in a situation with inconsistent open standards and potential privacy and security risks. Existing literature indicates that top leaders are good at supporting organizational learning and pushing employees to learn new knowledge and experiment with new ideas (Shao, Feng, & Hu, Reference Shao, Feng and Hu2017). Assume that top leaders communicate their support for and even participate in OGD-driven innovation. In such situation, their followers will become more aware of the need of gaining OGD-related knowledge and skills. In addition, top leaders’ leadership styles affect employees’ attitudes toward innovation. Top leaders with transformational leadership styles, for example, promote radical innovations (Zhou et al., Reference Zhou, Wang, Jiang, Li and Li2023). When the outcome of innovative activities is uncertain, transformational leaders’ encouraging and supportive attitudes toward new ideas can boost employees’ confidence in innovative activities and reduce their concerns about potential risks, creating a favorable climate for innovation (Shao, Feng, & Hu, Reference Shao, Feng and Hu2017; Shin & Zhou, Reference Shin and Zhou2003). Finally, we propose the following hypothesis:

Hypothesis 1d: Top leaders’ support positively influences innovation support.

Hypothesis 4: Innovation support positively influences OGD-driven innovation capacity.

Hypothesis 4a: Innovation support plays a mediation role between top leaders’ support and OGD-driven innovation capacity.

Our theoretical model is shown in Fig. 1.

Figure 1.

Theoretical model.

Data collection

Using the Credamo Questionnaire Platform (https://www.credamo.com/home.html) to collect data, we distributed 300 questionnaires to firms that intend or have performed OGD-driven innovation, with all questionnaires returned. After excluding questionnaires with missing or duplicate options, we obtained 236 valid questionnaires, with an effective response rate of 78.7%. The demographics of the samples are shown in Table 1, and the descriptive statistics of firms are shown in Table 2. The questionnaire is divided into two sections: the first section comprises personal and firm-related information, while the second section comprises questions related to each item and all graded on a 7-point Likert scale.

Table 1.

Sample demographics

Table 2.

Firm descriptive statistics

In practice, respondents’ replies may contradict the truth for a variety of reasons, including respondents’ failure to grasp the questions, respondents’ refusal to answer honestly, and respondents’ ambiguity regarding the facts to be replied. As a result, various steps were made in this investigation to assure the quality of the findings. For example, academic experts were widely consulted to revise the questionnaire’s items in order to eliminate the possibility that the items would be difficult to understand or unclearly stated; and it was promised that the information in the questionnaire would only be used for academic research and would not be published to the public in order to gain the respondents’ trust and to fill in the information honestly. Furthermore, we screened respondents using the Credamo Questionnaire Platform’s sample-paying service to guarantee that they were educated about OGD-driven innovation in their firms. The Credamo survey platform has a rigorous and scientific methodology in place to assist this study in inviting qualified target demographics to participate in the survey.

Measures

Data analysis method

The variance-based method (partial least squares structural equation modeling [PLS-SEM]) was employed for data analysis, and SmartPLS software was applied for empirical tests. The PLS-SEM method is appropriate for complicated models with multiple constructs, indicators, and structural paths, as well as for small sample data with non-normal distribution (Hair, Ringle & Sarstedt, Reference Hair, Ringle and Sarstedt2011). Therefore, based on the model in this paper, the justifications for choosing the PLS-SEM method are as follows: (1) the model is reasonably sophisticated (numerous constructs, indicators, and paths); (2) the sample size is comparatively limited (236); and (3) the data are non-normal distributed. Furthermore, in this study, the recommended two-step procedure was applied to assess the measurement and structural models, respectively (Hair et al., Reference Hair, Ringle and Sarstedt2011).

Measurement model

All constructs and indicators in this measurement model were evaluated for reliability and validity. We first evaluated the reliability of each indicator with suggested loadings above 0.7 (Hair, Risher, Sarstedt, & Ringle, Reference Hair, Risher, Sarstedt and Ringle2019). It is shown in Table 3 that all out loadings are above 0.7, suggesting good indicator reliability. And the internal consistency reliability was evaluated on the basis of three criteria, that is Cronbach’s α (CA), ${\rho _A}$, and composite reliability (CR) (Hair et al., Reference Hair, Risher, Sarstedt and Ringle2019). It is shown in Table 4 that the values of CA, ${\rho _A}$, and CR for all constructs are higher than 0.7 (Benitez, Henseler, Castillo, & Schuberth, Reference Benitez, Henseler, Castillo and Schuberth2020), indicating good reliability of the measurement scales.

Table 3.

Outer loadings and VIF

Note. TLS = top leaders’ support; TC = technical competence; OA = organization arrangement; IS = innovation support; OGDIC = OGD-driven innovation capacity; VIF = variance inflation factor.

Bold values are the outer loadings of the constructs.

Table 4.

Reliability and discriminant validity of the constructs

Note. TLS = top leaders’ support; TC = technical competence; OA = organization arrangement; IS = innovation support; OGDIC = OGD-driven innovation capacity.

Bolded values are the square root of the average variance extracted (AVE).

Convergent validity is evaluated by the average variance extracted (AVE) (Fornell & Larcker, Reference Fornell and Larcker1981), and the AVE value above 0.5 is considered that the construct explains no less than 50% of the variance of its indicators (Hair et al., Reference Hair, Ringle and Sarstedt2011). It is shown in Table 4 that all AVE values are above 0.5 and have favorable convergent validity. Discriminant validity, usually evaluated by the Fornell–Larcker criterion, is the degree to which a construct indeed differs from other constructs (Fornell & Larcker, Reference Fornell and Larcker1981), requiring that the square root of AVE should be above the correlation between constructs (Hair et al., Reference Hair, Ringle and Sarstedt2011). Table 4 indicates that the measurement model is in favorable discriminant validity.

In addition, Table 5 shows that a total of 236 respondents participated in this survey, giving the mean and standard deviation, and giving the correlation between all constructs. The results in Table 5 show that there is a good correlation between the various constructs, which is suitable for further hypothesis analysis.

Table 5.

Descriptive statistics and correlation result

Note. M = mean; SD = standard deviation; TLS = top leaders’ support; TC = technical competence; OA = organization arrangement; IS = innovation support; OGDIC = OGD-driven innovation capacity.

Structural model

Following the evaluation of the measurement model, the structural model is examined. PLS-SEM typically tests the relationships among the constructs of the structural model by means of examining the coefficient of determination (${R^2}$), significance and correlation of the path coefficients, and effect sizes (${f^2}$) (Benitez et al., Reference Benitez, Henseler, Castillo and Schuberth2020; Hair et al., Reference Hair, Risher, Sarstedt and Ringle2019). In this stage, we evaluate the structural model to determine whether the hypotheses presented earlier are valid.

Since the relationship between the constructs is a series of regression equations, we need to check multicollinearity to ensure that it does not bias the regression results (Hair et al., Reference Hair, Risher, Sarstedt and Ringle2019). In accordance with Benitez et al. (Reference Benitez, Henseler, Castillo and Schuberth2020), the variance inflation factor was selected to check multicollinearity. It is shown in Table 3 that all variance inflation factor values are under the standard threshold of 3, indicating an absence of multicollinearity.

The evaluation results of the structural model are obtained by running the bootstrapping program with 5,000 iterations of subsamples. Figure 2 indicates the model evaluation results, where the explained variance of endogenous variables (${R^2}$) as well as the standardized path coefficients ($\beta $) of endogenous variables are given. As is illustrated in Fig. 2, top leaders’ support ($\beta $ = 0.21, p < .05), technical competence ($\beta $ = 0.15, p < .05), organizational arrangement ($\beta $ = 0.37, p < .001), and innovation support ($\beta $= 0.22, p < .05) play a significant role in OGD-driven innovation capacity. Consequently, hypotheses H1a, H2, H3, and H4 are supported. In addition, top leaders’ support plays a significant role in technical competence ($\beta $ = 0.65, p < .001), organizational arrangement ($\beta $ = 0.66, p < .001), and innovation support ($\beta $ = 0.63, p < .001), respectively. Thus, hypotheses H1b, H1c, and H1d are supported. Table 6 shows the results of all direct hypotheses.

Figure 2.

Estimated relationships of the structural model.

Table 6.

The results of all direct hypotheses

Note. TLS = top leaders’ support; OGDIC = OGD-driven innovation capacity; TC = technical competence; OA = organizational arrangement; IC = innovation support; SD = standard deviation; LLCI = low-level confidence interval; ULCI = up-level confidence interval.

* p < .05, ***p < .001.

${R^2}$, known as in-sample predictive power, is often used to assess the variance explained of all endogenous constructs (Benitez et al., Reference Benitez, Henseler, Castillo and Schuberth2020; Hair et al., Reference Hair, Risher, Sarstedt and Ringle2019), which is also a measurement of the explanatory power of a model (Hair et al., Reference Hair, Risher, Sarstedt and Ringle2019). On the measurement scale, the ${R^2}$ values of 0.75, 0.50, and 0.25 are regarded as substantial, moderate, and weak levels of explanatory power, respectively (Hair et al., Reference Hair, Ringle and Sarstedt2011; Henseler, Ringle, & Sinkovics, Reference Henseler, Ringle, Sinkovics, Sinkovics and Ghauri2009). It is illustrated in Fig. 2 that our structural model explicates 42% of the variance (${R^2}$) in technical competence, 43% in organizational arrangement, 40% in innovation support, and 68% in OGD-driven innovation capacity, respectively. Consequently, our model has moderate explanatory power.

In addition to evaluating the ${R^2}$ as the explanatory power of all endogenous constructs, the model’s quality should be assessed by the effect sizes (${f^2}$) (Liang et al., Reference Liang, Wang, Dong, Zhang and Qi2021). In case that an exogenous construct is left out from the model, we can use ${f^2}$ to assess the effect of this omitted construct on the ${R^2}$ value of an endogenous construct (Mikalef, Krogstie, Pappas, & Pavlou, Reference Mikalef, Krogstie, Pappas and Pavlou2020). Based on the guideline, the ${f^2}$ values of 0.02, 0.15, and 0.35 refer to small, medium, and large effect sizes, respectively (Hair et al., Reference Hair, Risher, Sarstedt and Ringle2019). ${f^2}$ effect sizes’ results of all exogenous constructs are shown in Table 7. Top leaders’ support (0.06), technical competence (0.03), and innovation support (0.08) have a small effect size on OGD-driven innovation capacity. In contrast, organizational arrangement (0.17) has a medium effect size toward OGD-driven innovation capacity.

Table 7.

The results of ${f^2}$ effect sizes

Mediating effect test

Previous research has frequently used the four-step Sobel test approach to assess for mediating effects (Arpaci, Kesici, & Baloğlu, Reference Arpaci, Kesici and Baloğlu2018; Sobel, Reference Sobel1982). Nevertheless, the Sobel test is inappropriate for analyzing indirect effects (Nitzl, Roldan, & Cepeda, Reference Nitzl, Roldan and Cepeda2016; Preacher & Hayes, Reference Preacher and Hayes2008), and in this study, the Sobel test is unnecessary. As a result that after 5,000 iterations of the bootstrapping procedure, PLS can test the direct and indirect effects of the model (Arpaci, Kesici, & Baloğlu, Reference Arpaci, Kesici and Baloğlu2018; Hair, Hult, Ringle, & Sarstedt, Reference Hair, Hult, Ringle and Sarstedt2017; Preacher & Hayes, Reference Preacher and Hayes2008) and still obtain a high statistical power in the case of small samples (Hair, Sarstedt, Hopkins, & Kuppelwieser, Reference Hair, Sarstedt, Hopkins and Kuppelwieser2014; Henseler, Reference Henseler2010).

The mediating effect test results for this study’s model are shown in Table 8. On the basis of the results in Table 8, we conclude that the direct and indirect effects (mediating paths) of top leaders’ support on OGD-driven innovation capacity are of significance regardless of whether the technical competence, organizational arrangement, or innovation support is introduced. Consequently, we confirm that technical competence, organizational arrangement, and innovation support all partially mediate the relationship between top leaders’ support and OGD-driven innovation capacity, and hypotheses H2a, H3a, and H4a are all supported.

Table 8.

Mediation effect test

Note. TLS = top leaders’ support; OGDIC = OGD-driven innovation capacity; TC = technical competence; OA = organizational arrangement; IS = innovation support; LLCI = low-level confidence interval; ULCI = up-level confidence interval.

* p < .05, ***p < .001.

Theoretical implications

Several theoretical implications are concluded below. First, prior literature has explored a large number of OGD-related topics at the governmental level, such as institutional capacity (Zhao & Fan, Reference Zhao and Fan2018), policy comparison (Bates, Reference Bates2014; Chatfield & Reddick, Reference Chatfield and Reddick2018), adoption (Wang & Lo, Reference Wang and Lo2016), and implementation evaluation (Zuiderwijk & Janssen, Reference Zuiderwijk and Janssen2014). Some scholars have gradually shifted their research from the government level to the firm level, with research topics covering adoption (Kaasenbrood et al., Reference Kaasenbrood, Zuiderwijk, Janssen, de Jong and Bharosa2015; Wang & Lo, Reference Wang and Lo2020), value generation mechanisms (Ahmadi Zeleti, Ojo, & Curry, Reference Ahmadi Zeleti, Ojo and Curry2016; Leviäkangas & Molarius, Reference Leviäkangas and Molarius2020), and realization paths (Jetzek, Avital, & Bjorn-Andersen, Reference Jetzek, Avital and Bjorn-Andersen2014). However, few studies have addressed the capacity of firms to leverage OGD for innovation. This paper shifts the focus of the current study to the firm’s OGD-driven innovation capacity.

Second, this study developed a model and verified its validity through empirical testing to advance the understanding of OGD-driven innovation capacity. Based on the TOE framework, a model has been proposed to identify the intrinsic drivers that influence OGD-driven innovation capacity, which has good explanatory power (67.7%). Unlike the theoretical models commonly used in technological innovation (e.g., Unified Theory of Acceptance and Use of Technology, Technology Acceptance Model, and Task-Technology Fit), this paper provides a different theoretical perspective for researchers of OGD-driven innovation capacity at the firm level.

Furthermore, the findings of this paper make up for and extend the research on firms’ innovation capacity. Previous literature has validated the positive contribution of top leaders’ support to a firm’s innovation capacity on different topics (Guo, Pang, & Li, Reference Guo, Pang and Li2018; Talke, Salomo, & Rost, Reference Talke, Salomo and Rost2010; Yu et al., Reference Yu, Wang, Li and Lin2022). And the findings of this study further validate that top leaders’ support has significant direct and indirect effects on a new innovation capacity such as OGD-driven innovation capacity. In addition, prior literature has examined the impact on innovation by primarily incorporating top leaders’ support into organizational arrangements (Zhao & Fan, Reference Zhao and Fan2021). This study argues that top leaders’ support affects organizational arrangements such as institutional settings and staffing, which in turn has an effect on innovation capacity.

Practical implications

Several practical implications are summarized below. To begin with, top leaders’ support is essential for improving OGD-driven innovation capacity. We encourage top leaders to attach great importance to the innovation value of OGD and optimize the configuration of technology and management within the organization from the top down, in order to prepare for the realization of OGD-driven innovation capacity. Second, the technical competence is essential, although it is not the most critical factor for firms to improve their OGD-driven innovation capacities. Firms are required to keep up with the next generation of information technology to avoid losing competitive advantage continuously. For example, top leaders can enhance the firm’s technical competence by training and developing relevant technological talents and introducing next-generation information technology (e.g., cloud computing). Third, top leaders should realize that verbal expressions of support for OGD-driven innovation are inadequate; they must be followed by action. For example, they can commit resources to establish OGD-driven innovation-related organizations and rationalize staffing arrangements. The organizational arrangement can reflect the firm’s readiness, and a reasonable and adequate organizational arrangement can fully optimize the firm’s workforce and management and enhance the ability of OGD-driven innovation. Finally, the innovation environment within the organization should be given attention. Top leaders should motivate employees to improve their OGD-driven innovation capacities by providing a stable innovation environment (e.g., training, financial support, reward, and punishment system).