Kedia and Bhagat (1988) first coined the term absorptive capacity (ACAP), which then since received considerable attention after the work of Cohen and Levinthal (1990). Cohen and Levinthal define ACAP as a firm's ability to recognize the value of new external knowledge, assimilate it, and apply it to commercial ends (Cohen & Levinthal, 1990: p. 128). Their contribution provided a definition of a construct that continues to evolve today, and is generally accepted as pioneering work. Many studies (e.g., Dyer & Singh, 1998; Lane, Koka, & Pathak, 2006; Van den Bosch, Volberda, & De Boer, 1999; Volberda, Foss, & Lyles, 2010; Zahra & George, 2002) have documented the multidimensional nature of ACAP, but very few (Dyer & Singh, 1998; Lane & Lubatkin, 1998; Lane et al., 2006; Zahra & George, 2002) have revised or expanded on Cohen and Levinthal's definition.

The original definition of ACAP captures a three-stage learning process: recognition, assimilation, and application (Cohen & Levinthal, 1990). However, different alternative approaches to this development process exist (Lane & Lubatkin, 1998; Lane et al., 2006; Szulanski, 1996; Todorova & Durisin, 2007; Van den Bosch et al., 1999; Volberda et al., 2010; Zahra & George, 2002). Today, the definition proposed by Zahra and George (2002) is most widely used, conceptualized, operationalized, and measured. Zahra and George (2002) define ACAP as a set of organizational routines and processes through which firms acquire, assimilate, transform, and exploit knowledge in order to produce a dynamic organizational capability.

Research on knowledge and innovation adoption is studied extensively using the concept of stages in the process of adoption (Ali & Park, 2013). The process is not an instantaneous act, but rather occurs over time and consists of a series of actions (Rogers & Shoemaker, 1971). Based on empirical studies and theoretical backgrounds, researchers have proposed models of different stages in the learning process. However, the present study follows Kim (1976, 1980), using a two-stage conceptualization relevant to the key theme of this study. Rogers (1983) and Zaltman et al. (1973) propose a two-stage model consisting of an initiation stage and an implementation stage, as shown in Table 1. The initiation of learning refers to a process consisting of all activities pertaining to problem perception, information gathering, attitude formation and evaluation, and resource attainment, and leads to a decision to adopt. The implementation of learning is a process consisting of all events and actions that modify both the learning and the organization, including the initial and continued use, until it becomes a routine feature of the organization (Damanpour, 1996). The initiation stage includes sub-stages of knowledge, awareness, evaluation, search, decision, and selection. The implementation stage includes sub-stages of initial implementation, routinization, and stabilization. Based on the two-stage model of Zaltman et al. (1973), Kim (1976, 1980) re-conceptualizes the organizational learning and innovation process model into a single innovative mode and a dual innovative mode. To study ACAP at the organizational level and different stages of the learning process, it may be assumed that the organization has different sub-units to deal with different stages of the learning process.

Structural complexity (or specialization or differentiation) is one of the primary dimensions of organizational design, and refers to the degree of differentiation that exists within an organization (Robbins, 1990). Complexity is likely to improve employees’ skills and capabilities in day-to-day activities because they become specialized in those activities (Pertusa-Ortega, Zaragoza-Sáez, & Claver-Cortes, 2010). Complexity describes the extent and intensity of knowledge in the organization (Kim, 1980), as well as the degree to which functions are distinguished with respect to goals, task orientation, time horizon, and degree of autonomy (Hall, 1962, 1977; Lee & Grover, 2000). Complexity can play an important role facilitating the flow of development, because absorbing new knowledge requires a base of prior knowledge (Cohen & Levinthal, 1990).

Pierce and Delbecq (1977) suggest that differentiation (complexity) within the organization is conducive to the initiation stage. They propose that it is positively related to innovation (in terms of the initiation and implementation stages), but that there is a stronger association in the case of initiation than in implementation. Kim (1980) proposes that in single-mode organizations, complexity has a significant and positive influence on innovation, while in dual-mode organizations, complexity has a positive influence on the initiation stage, but a negative influence on implementation. Kim (1976, 1980) concludes that it is more realistic to consider only the single mode, because it is not easy to change the structure of an ongoing organization, and it is also not realistic to expect members of the organization to change their behavior in accordance with a new structural pattern. In contrast, Damanpour (1996) hypothesizes that the influence of complexity on the initiation stage is not significantly different to its influence on the implementation stage, but concludes that complexity influences the initiation stage less positively than it does the implementation stage. The above discussion leads to the proposition that the two subsets of ACAP (i.e., PACAP and RACAP) require different organizational structures.

Several studies have analyzed the influence of complexity on structural variables, but the relationship between complexity and ACAP in single and dual modes has yet to be investigated. This study analyzes how complexity influences ACAP, along with other dimensions of organizational design.

Thus, in a dual mode, the complexity of the organization has a positive effect on the PACAP, and a potentially negative effect on RACAP. When a single organizational unit is involved throughout the ACAP process, the positive effect of complexity seems to be greater than the negative effect.

The formalization of an organization is a form of control employed by bureaucratic organizations, hierarchical directives, or mechanistic structure, and refers to the degree to which a codified body of formal rules, standard policies, procedures, or behavior prescriptions is developed to govern decisions and work processing (Pierce & Delbecq, 1977; Robbins, 1990). The relevant literature on organization theory supports both a positive impact (Claver-Cortés, Pertusa-Ortega, & Molina-Azorín, 2012; Pertusa-Ortega et al., 2010; Pierce & Delbecq, 1977) and a negative impact (Damanpour, 1991; Kim, 1980; Lee & Choi, 2003) of formalization. One the one hand, those who support the negative impact of formalization on learning argue that strictly following rules and procedures hampers employees from seeking new rules that might lead to an awareness of a performance gap between what the firm is doing and what the employees perceive it should be doing (Kim, 1980; Zaltman et al., 1973). On the other hand, those who support the positive impact of formalization on learning argue that formalization helps to make knowledge more efficient to exploit, easier to apply, and faster to implement (Jansen et al., 2005a). It eliminates the need for communication and coordination among sub-units and creates an organizational memory that helps firms handle routine situations (Van den Bosch et al., 1999).

Several studies have examined the relationship between formalization and the two-stage model of initiation and implementation (Damanpour & Gopalakrishnan, 1998; Kim, 1980; Pierce & Delbecq, 1977; Zaltman et al., 1973). In other words, they examine learning and innovation in the context of the relationship between formalization and learning. Shepard (1967, p. 474) indicates that low formalization may be considered more appropriate for the initiation stage, whereas a higher degree of formalization may be considered more appropriate during the implementation stage. During the initiation phase, the organization needs to be flexible and as open as possible to sources of information and alternative courses of action. During the implementation phase, Shepard (1967) indicates that a singleness of purpose is required. These findings suggest that formalization will be negatively related to initiation, but will have a modest positive impact on adoption and implementation (Damanpour & Gopalakrishnan, 1998; Kim, 1980; Pierce & Delbecq, 1977; Shepard, 1967; Zaltman et al., 1973).

Other studies have examined formalization with regard to PACAP and RACAP (De Araújo, 2010; Jansen et al., 2005a, 2006; Jansen, Van Den Bosch, & Volberda, 2005b). Jansen et al. (2005a) argue that formalization is likely to limit the scope of the effort expended on knowledge acquisition, while it also hinders an individual's assimilation of new external knowledge. Formalization creates an organizational memory, which is essential for transformative and exploitative learning. This involves collecting previously applied solutions to specific problems that can help organizations when reacting to prior knowledge or when matching new knowledge to novel applications. On the other hand, formalization may have a negative impact on the exploratory learning process, where excessive rules may hinder the establishment of informal relations with external partners (De Araújo, 2010). Jansen et al. (2005b, 2006) suggest that formalization enhances exploitative learning and innovation by improving current products, services, and processes, and hypothesize that a higher degree of formalization implies (a) a lower level of exploratory innovation, and (b) a higher level of exploitative innovation.

Considering the above, formalization can improve a firm's ability to assimilate and exploit knowledge. In summary, this study adopts the approach that low formalization facilitates PACAP, and higher formalization is more appropriate for RACAP. In the case of a single-mode organization, low formalization seems have a positive association with ACAP.

Centralization is determined in terms of (1) the allocation of decision-making in the organization, and (2) the hierarchy of authority within an organization (Aiken & Hage, 1971; Kim, 1980; Pierce & Delbecq, 1977). A highly centralized structure hampers the interaction among the organizational members, reduces opportunities for innovation and knowledge solutions and for creating new knowledge (Damanpour, 1991). Kim (1980) points out that a more centralized decision-making structure implies a greater number of channels through which the communication of new ideas and learning must travel. Although the findings of research on the relationship between centralization and ACAP has been positive in some cases (De Araújo, 2010; Liao, Chuang, & To, 2011; Pierce & Delbecq, 1977), in others, the relationship has been negative (Damanpour, 1991; Kim, 1980; Liao et al., 2011; Zheng, Yang, & McLean, 2010). Even though research has found positive results with regard to the relationship between organizational structure and knowledge (Liao et al., 2011; Tsai, 2002), a decentralized organizational structure (i.e., a high degree of participation in decision-making) has often been found to have a significant impact on the knowledge management process (Damanpour, 1991; Jansen et al., 2005b; Pierce & Delbecq, 1977; Zheng et al., 2010). Tsai (2002) argues that greater centralization prevents a unit manager from exercising discretion in dealing with the demands of his or her relevant task environment.

Researchers have also examined formalization using the two-stage model comprising an initiation stage and an implementation stage (Damanpour & Gopalakrishnan, 1998; Kim, 1980; Pierce & Delbecq, 1977; Zaltman et al., 1973). Scholars agree that a low level of centralization is appropriate in the initiation stage of the innovation process, because it reduces the amount of information that decision-making units need to have (Kim, 1976). However, a high level of centralization is suggested for the implementation stage because, here, a more specific level of authority and responsibility is required. Jansen et al. (2006) argue that centralization likely reduces PACAP, because PACAP requires non-routine problem solving and deviation from existing knowledge. In contrast, centralization increases the efficiency of information processing and facilitates RACAP. In summary, we take the approach that centralization has a negative influence on the acquisition and assimilation of the new external knowledge underlying PACAP, and a significantly positive influence on the transformation and exploitation of the new external knowledge underlying RACAP.

This implies that in dual-mode organizations, a low level of centralization seems to be appropriate for PACAP, while a high level of centralization is suitable for RACAP. In the case of single-mode organizations, the negative influence of centralization is more appropriate for ACAP.

Integration is widely recognized as an important structural dimension of an organization (Kim, 1980; Lawrence & Lorsch, 1967; Lee & Grover, 2000; Liao et al., 2011; Miller, 1987). Integration reflects the degree to which the activities of separate actors in the organization can be coordinated through formal coordination mechanisms (Lee & Grover, 2000; Liao et al., 2011) in order to attain the common goals and objectives of the organization (Kim, 1980). Lawrence and Lorsch (1967) suggest that in today's increasingly competitive and dynamic environment, effective organizations have high levels of differentiation and integration. Diversity of occupations, specialization in individual tasks, and horizontal departmentalization according to function are essential to the ACAP process of the organization.

It is also essential to investigate the relationship between integration and the two-stage model, because integration refers to more than just the integration of organizational units involved in the initiation and implementation phases (Kim, 1980). New learning and knowledge is not the outcome of individual effort or that of a single department. Therefore, during the initiation stage, an integrated effort from all relevant departments (e.g., R&D, production, and marketing) is essential. In addition, during the implementation stage, constant coordination among the various sub-departments is needed, and perhaps to a lesser degree, as ideas are adopted and implemented within the existing paradigm of the organization (Kim, 1980). This implies that the process of integration is essential to both PACAP and RACAP, because it enables the organization to collect all previously applied solutions to specific problems, which can help when attempting to react to preceding knowledge or to match new knowledge with novel applications (De Araújo, 2010). For ACAP, the organization needs information from as many sources as possible. In fact, the ACAP process involves the integration of information related to human understanding, equipment (e.g., tools, machinery, building, process technology, technological knowledge, technical skills), and knowledge about marketing, R&D, production, management, finance, human resources, and organizational understanding. For ACAP, an organization requires a differentiated organizational structure with a broad array of formalized managerial positions and specialized tasks, and consequently, liaison personnel to integrate the efforts of the various specialists performing these tasks (Daft & Lengel, 1986; Liao et al., 2011). The above discussion suggests that we can hypothesize that integration is positively related to ACAP in the case of a single mode, and positively related to both PACAP and RACAP in the case of a dual mode.

Table 2 shows a contingency framework that summarizes the relationship between the structural variables and ACAP. Finally, we have the following two propositions:Proposition 1

Intermediate propositions that can be applied to single-mode organizations are as follows:

• 1)

  Complexity relates positively to ACAP.

• 2)

  Formalization relates negatively to ACAP.

• 3)

  Centralization relates negatively to ACAP.

• 4)

  Integration relates positively to ACAP.