A Systematic Literature Review into the Adoption,
Diffusion, and Assimilation of Innovations in Information
Systems Research
Master thesis, MSc BA, specialization Change Management University of Groningen, Faculty of Economics and Business
June 22nd, 2015 Arjan Miedema Tuinbouwstraat 104a 9717 JN Groningen +31 (0)648276801 a.miedema.6@student.rug.nl S2577666 Supervisors:
dr. I. Maris-de Bresser & dr. U.Y. Eseryel Co-assessor:
1
Abstract
This research consolidates high-quality literature on the adoption, diffusion, and assimilation of innovations within the context of Information Systems (IS). Starting off with a consideration set of over 2700 academic papers, spanning twenty years, a systematic literature review is chosen as the approach to focus on - and comprehensively synthesize - a scattered field of literature. The synthesis breaks down the selected literature along its conceptual consideration and empirical treatment of innovations as an unproblematic outcome or a problematic process during adoption, diffusion, and assimilation. Following, these distinctions are connected to whether papers study internally conceived inventions and/or innovations adopted from an external source. In the final step of the synthesis, the named conceptual consideration of innovations is connected to a multi-level process framework of adoption, diffusion, and assimilation of innovation in the context of IS. This framework questions the dualistic view which holds „innovation‟ as a process and the diffusion of innovations to be separable. The systematic review provides insights into the state of the literature, which sheds a light on a number of possible avenues for future research.
Keywords: innovation, information systems, innovation adoption, innovation diffusion,
innovation assimilation, systematic literature review
2 A Systematic Literature Review into the Adoption, Diffusion, and Assimilation of
Innovations in Information Systems Research
Introduction
In the 21st century, an ever-changing competitive environment has changed the landscape of the most successful firms from those involved in steel and the automotive industry towards those involved in IT, logistics and financial services (Dess & Picken, 2000). Companies that manage to thrive in the 21st century are those that innovate successfully and manage to be creative, responsive, and flexible (Dess & Picken, 2000). A lack of organizational innovation is named in management literature to connect strongly to the emergence of crises in organizations (Mone, McKinley, & Barker, 1998).
In the context of IS, innovation has been a central concept for at least the past two decades. Recently, a survey by Córdoba, Pilkington and Bernroider (2012) found that in both a European and an American top IS Journal, Davis (1989) on technology acceptance and Rogers (1962) on the diffusion of innovations were among the top five most cited publications. Consequently, reviewing of innovation literature within the field of IS has also been done.
Over the past twenty years, nine literature reviews and meta-analyses on innovation were conducted in AIS basket of eight top IS journals. Mostly, these reviews assume a narrow focus, either on specific types of innovations (e.g. Mueller, Viering, Legner, & Riempp, 2010), or on a specific innovation concept, for example open-innovation alliances (Han, Oh, Im, Chang, Oh, & Pinsonneault, 2012). Some reviews take a slightly broader scope, conveying a history of implementation and innovation research (Lucas Jr., Swanson, & Zmud, 2007), or suggesting possible additions to the positivist paradigm of IT innovation adoption and diffusion research (Fichman, 2004a). Overall, a review of innovation in IS assuming a broad view on the concept is not yet existent, to the best knowledge of the author. Nevertheless, innovation in IS has been named to differ from general innovation on various dimensions (Bajaj, 2000; Swanson, 1994).
3 The general research aim of this study is to provide a first step in composing a comprehensive review of innovation in the context of IS. Specifically, this paper describes an attempt at complementing the work of Crossan and Apaydin (2010) by conducting a synthesis of the diffusion and assimilation aspects of innovation and placing it relative to their framework, as well as an application of their general approach to the context of IS.
A number of research objectives guide the overall process toward the research aim. Firstly, an objective is to collect all aspects of the innovation concept described in high quality IS literature in order to identify adoption, diffusion, and assimilation papers in innovation literature. A second objective is to identify, define, and relate dimensions of the adoption, diffusion, and assimilation of innovations. The final objective of this research is to deliver a structured overview of the adoption, diffusion, and assimilation of innovations in the context of IS, set out to the conceptualization of that innovation as a process or an outcome during the researched trajectory. This final objective provides a nuanced view on the adoption, diffusion, and assimilation of innovations in the context of IS that can be seen as providing a first step towards inclusive comprehension of organizational innovation in IS.
Choosing to take a broad view on innovation inevitably leads to encountering countless applications and definitions of the term (Crossan & Apaydin, 2010). A short selection of definitions demonstrates how scholars have defined innovation in terms that correspond to the contexts they have researched. The first to define innovation was Schumpeter in the 1930‟s, whose concept of innovation, as part of a larger work on economic development, took an economic perspective (Drejer, 2004). Schumpeter‟s innovation covered five areas, which in summary address reaching novel outputs and doing things differently (Drejer, 2004). More recently, Van de Ven (1986) defined innovation as “the development and implementation of new ideas by people who over time engage in transactions with others within an industrial order” (Van de Ven, 1986, p. 590), emphasizing specific intentional actions of individuals. Rogers (2003) on the other hand, identifies three possible entities of innovation: „ideas‟, „practices‟ and „objects‟. These entities solely need to be perceived as new by the unit of adoption for the entity to be seen as an innovation (Rogers, 2003). Assuming a broad perspective, a definition is needed which comprehends the illustrated diversity. Having a similar purpose, Crossan and Apaydin (2010) composed such a definition:
4 of new management systems. It is both a process and an outcome (Crossan & Apaydin, 2010, p. 1155)
This definition covers both invention and adoption, both creative processes and exploitation, it emphasizes adding value, but includes a relative view of the novelty of the innovation, and it crosses multiple levels of analysis (Crossan & Apaydin, 2010). What it explicitly excludes, however, is innovation diffusion, which supposedly takes place after „innovation‟ has occurred (Crossan & Apaydin, 2010). In the remainder of this paper, the definition by Crossan and Apaydin will be embraced. However, the specific purpose of this research is to target and question the exclusion of innovation diffusion from the concept of innovation. Also in line with Crossan and Apaydin (2010), papers that apply a level of analysis beyond that of a single organization are excluded from the synthesis. While a literature review is purposely broad in scope, this confinement encourages a parsimony in fields to which the research is attempting to contribute, improving value for the addressed organizational practitioners and management information systems scholars.
Focussing on the diffusion of innovations requires some flexibility regarding terminology. The three terms „adoption‟, „diffusion‟, and „assimilation‟ are concepts that, as innovation itself, have been defined in numerous ways. All three concepts can be seen as processes, composing multiple stages (Fichman & Kemerer, 1997; Gopalakrishnan & Damanpour, 1997). As described in the results section, these stages themselves are subject of debate as well (e.g. Rogers, 2003). Consequently, definitions of adoption, diffusion, and assimilation often overlap.
5 innovation is continuous (Joshi, Chi, Datta, & Han, 2010), with diffusion trajectories possibly initiating new diffusion trajectories.
Focusing on IS research, a rich body of literature on the adoption, diffusion, and more recently assimilation of innovations has accumulated (Jeyaraj et al., 2006; Lauterbach & Mueller, 2014). This literature has long been dominated by positivist research (Fichman, 2004a). The general assumption amongst positivist research is that a greater quantity of certain determinants for - and characteristics of - innovation inevitably lead to a greater quantity of innovation adopted, which is seen as inherently beneficial (Fichman, 2004a). However, the constructivist paradigm has gained traction in the IS field (Lauterbach & Mueller, 2014). To constructivists, criticism on the technological determinism that implicitly underlies the positivist paradigm is at the roots of its existence (Lauterbach & Mueller, 2014). Assuming comprehensiveness as an objective, both philosophical strands will be included in the review.
By addressing the aim and objectives of this research, the following research question will be answered: how does the adoption, diffusion, and assimilation of innovations in the context of IS relate to innovation as defined by Crossan and Apaydin (2010)?
In the following sections, the methodology used is described, followed by a quantity-oriented description of the initial sampled literature. Then, the literature composing the final sample is synthesized and organized into its identified dimensions. Finally, implications of the findings are discussed, and fruitful possibilities for future research are proposed.
Methodology
6
Table 1
Steps of the systematic review. Adapted from Tranfield et al. (2003, p. 214)
Stage Step
1: Planning 1: Development of a research protocol
2: Execution 2: Identifying initial identification criteria 3: Selection of studies (first coding round) 4: Interrater reliability check
5: Data extraction and monitoring 6: Data synthesis
3: Reporting 7: The report and recommendations
Planning: Development of a Research Protocol
To ensure a systematic approach, a review protocol has been developed ex-ante. Besides providing structure, the review protocol helps decrease researcher bias (Tranfield et al., 2003). The review protocol consists of a coding scheme developed with Microsoft Excel, complemented with a document clarifying the coding. The coding scheme is used to conduct an initial analysis of the studies that are selected. Development of the coding scheme is grounded in earlier systematic reviews and work on systematic reviews in the context of innovation and IS (Crossan & Apaydin, 2010; Crowston, Wei, Howison, & Higgins, 2012; Webster & Watson, 2002). Details of the coding scheme can be found in appendix A. During this review, each paper is coded on some paper descriptives such as year of publication. Papers deemed relevant in the initial review are coded on a number of additional dimensions about methodological features, dimensions relating to the concept of innovation, and dimensions relating to the conceptual relations and results the papers provide. The main construct of the papers is coded by using the Inputs-Mediators-Outputs-Inputs model (IMOI) as used in Crowston et al.(2012). This model is translated to the coding scheme by adding the coding dimensions for Input, Mediator, Moderator, Output and Control Variables. Furthermore, a column themed “remarks” will be utilized to administer emerging themes and information relevant for later data synthesis (Tranfield et al., 2003).
Figure 1
The IMOI model. Adapted from Crowston et al. (2012, p. 7:11)
Execution
Identification of research.
7 selection contains high-quality research. The AIS basket of eight contains the following journals: European Journal of Information Systems, Information Systems Journal, Information Systems Research, Journal of the AIS, Journal of Information Technology, Journal of Management Information Systems, Journal of Strategic IS and MIS Quarterly.
Selection of the studies.
Literature is extracted from EBSCO Host as a primary search engine. However, this search engine does not cover all journals for the entire selected period. Therefore, Palgrave MacMillan, Wiley, AISEL and Science Direct are used to attain the additional literature. The search is limited to a period of 20 years, covering papers published in the period January 1st1995 – December 31st 2014. The keyword search conducted is {“Innovation”, “Innovative” or “Innovate”}, which includes any paper containing at least one of the three keywords. Moreover, the search is confined to “all text”. Table 2 displays the distribution of the 2744 articles that compile the initial consideration set, with search engines used and respective quantities per journal depicted.
Table 2
Distribution of papers among the journals
Journal Number of publications Search engines used European Journal of
Information Systems
501 Palgrave Macmillan
Information Systems Journal 215 EBSCO Host (197), Wiley Online Library (18)
Information Systems Research 218 EBSCO Host Journal of Information Technology 166 EBSCO Host Journal of Management Information Systems 513 EBSCO Host Journal of Strategic Information Systems 333 Science Direct
Journal of the Association for Information Systems
255 EBSCO Host (244), AISEL (11)
MIS Quarterly 543 EBSCO Host
Interrater reliability.
The development of a research protocol, identification of initial identification criteria, selection of the studies and data extraction and monitoring of the initially selected papers were done together with a peer researcher1. After the initial data-extraction and monitoring process concluded with the composition of a concept matrix, the main research split up in two
1The overall research topic, „a systematic review into innovation in IS‟, was initiated by the author and another
8 separate studies. To increase internal validity of the initial coding of the consideration set, the researchers conducted an interrater reliability check by both independently coding an identical set of 50 articles. In 2, or 4% of the cases the researchers judged the article differently on whether it was relevant for inclusion or not. These results are above the threshold of 90 percent that is seen as the minimum acceptable intercoder agreement (Miles & Huberman, 1994). Internal validity was increased further by conducting a second (peer) review specifically targeted at papers about which the researchers were uncertain. Finally, a second round of coding was conducted in which the researchers reassessed each other‟s coding of relevant articles, adjusted errors, and excluded an additional 21 articles from the „relevant in first review‟ set.
In order to resolve differences in viewpoints, assigned codes were compared. During the iterative process of reviewing and discussing, the coding scheme was adjusted with further refinements. Furthermore, researchers added additional codes that provided better content clarity of the papers.
Data extraction and monitoring process.
Carefully extracting and monitoring the data ensures a structured and objective data synthesis (Tranfield et al., 2003). The extraction of the data was done by downloading all the papers and mapping them by journal. After extracting and mapping the papers, the articles were coded. During this process, the articles were scanned on relevancy for this research. The argument used to assess relevance can be described as „the term innovation, innovative(ness) or innovate is directly related to the contribution the paper is attempting to make‟. To achieve this, „innov‟ is entered in the in-document search function. The matches that appear are explored as far as necessary for the researchers to be able to place them in context. When an article is deemed „relevant‟ it is coded by following the coding scheme as described in step 1. Moreover, when an article is deemed „non relevant‟, it is only coded by „article code‟, „authors‟, „year‟ and the reason why it is not relevant. The result of the initial coding process is a dataset2.
Data synthesis.
Data synthesis helps to derive insights from the reviewed literature. Synthesizing existing studies is seen as essential in reaching higher analytic goals and enhancing the generalizability of results (Tranfield et al., 2003). Insights from synthesis include the identification of critical knowledge gaps or common findings in the literature. The data
2
The complete dataset is available at
9 analysis and synthesis has been iterative between first-level coding and second-level or pattern coding (Miles & Huberman, 1994). Pattern finding can be very productive when the dataset is substantial, as is the case for the focal consideration set (Miles & Huberman, 1994).
The initial first-level coding is described above under data extraction and monitoring process and consists of descriptive and interpretive coding (Miles & Huberman, 1994). The first round of pattern coding consisted of categorizing papers on the theory they invoked, the level of analysis they applied, and the consideration of innovation as a process or an outcome. To visualize initial patterns, a concept-matrix was composed (Webster & Watson, 2002). For the purpose of brevity, the concept matrix is not included here. A shortened version can be found in appendix C3. However, from the concept matrix, adoption, diffusion, and assimilation studies were derived as an emergent focus of this research. A second round of first-level coding was then conducted (described in detail under „specific theories‟ and „issues and tensions‟), followed by a second round of pattern coding (e.g., synthesis). Throughout the process, memos were used to interpret and store information about, for example, connections between concepts or other analytical ideas (Corbin & Strauss, 2008). A general format of the memos as applied to the research can be found in appendix D. A chronological description of the reviewing process is depicted in table 3.
Table 3
Types of coding, steps taken and quantities reviewed: a chronological display
Type of coding involved Step in the research Papers involved First-level coding Judging all articles on relevance 2744
descriptive coding of relevant articles 255
Pattern coding Composing the concept matrix 234
Extracting adoption, diffusion, and assimilation (excluding societal level)
84 (70) First-level coding Iterative coding of emergent dimensions 70 Pattern coding Analysing dimensions, performing synthesis 70 Reporting: The Report and Recommendations
In the final phase, reporting and dissemination, it is important that the results of the research are reported in a clear and systematic manner. This is done by following the instructions of Webster and Watson (2002) on writing a literature review. Moreover, the reporting structure of Crossan and Apaydin (2010) is used as an exemplar for structuring the report.
3 Papers included in the concept matrix are available at
https://www.dropbox.com/s/1275new6m3tr2ek/Coding%20scheme%20innovation%20in%20IS%2C%20includ ed%20in%20conceptmatrix.xlsx?dl=0
10
Results
This section provides descriptive characteristics of the papers that compile the consideration set. First, a timeline is shown depicting the number of articles published per year. Then, descriptive characteristics of the papers deemed relevant after the initial review are provided. Finally, an overview of the concepts identified in the iterative initial reviewing phase is composed.
Descriptive Analysis
Figure 2 shows a growth in the number of articles on innovation up to 2007. From there on, the line becomes relatively stable. Comparing the figure to an earlier assessment by Crossan and Apaydin (2010), we see a similar growth in the period 1995 – 2007 as they do. Also in line with Crossan and Apaydin (2010), the growth comes to a sudden stop over 2008. As their coverage ends about halfway through the year 2008, this assessment provides confirmation of the flattened out prominence of innovation in literature. However, generalizing these findings should be treated with caution as this investigation focuses solely on the AIS basket of eight IS journals.
Figure 2
Quantity of papers containing ‘innovation’, ‘innovative’, or ‘innovate’ per year throughout the time frame
Besides relevance, articles were also coded on research methodology. Literature reviews and meta-analyses, which formed 1,7% of the initial consideration set, were excluded from consideration as they aggregate other studies that either are part of the consideration set
11 already or do not match the scope of the consideration set. Three additional possible research methods were identified ex ante: „empirical (theory testing)‟, „empirical (theory building)‟, and „theoretical‟ (Crossan & Apaydin, 2010). Figure 3 depicts the distribution of the 239 relevant articles identified after the initial review and interrater reliability checks. The distribution shows how empirical papers form the vast majority of the papers analysed, 91% in total. Seven percent of the papers were purely theoretical and in two percent of the cases, the researchers were unable to identify the methodology used.
Figure 3
Distribution of papers by research method
Additional coding was done focusing on the level of analysis applied on the relevant articles (figure 4). The categorization proposed by Crowston et al. (2012) is applied, distinguishing the artefact level (3% of the relevant papers), the individual level (23%), the group level (4%), the organizational level (52%) and the societal level (11%). The societal level covers any level of analysis that is beyond one organization, such as inter-organizational (information) systems that are analysed from the perspective of the network adopting the inter-organizational system (Hsu, Lin, & Wang, 2014), or research on the cross-cultural diffusion of innovations (Dewan, Ganley, & Kraemer, 2010).
Organizational-level studies form an absolute majority, followed by individual-level studies at almost a quarter of the total. The prominence of the organizational level is consistent with the findings of Crossan and Apaydin (2010). The prominence of individual-level studies departs from their findings as only 5% of their articles applied an individual level of analysis. This indicates that IS innovation literature is to a greater extent oriented toward the individual than general innovation literature.
50% 41%
7% 2%
12
Figure 4
Distribution of papers by level of analysis
Scoping out the Theoretical Field
In this section, the process and nature of the content analysis will be explained, followed by a summary of the most prominent theories among the selected papers and an elaboration on general characteristics of the theoretical field that inhibit, complicate or enable synthesis of the literature.
Content analysis.
Content analysis of the theoretical field was conducted as part of the two major cycles of first-level and pattern coding (Miles & Huberman, 1994). The first cycle is predominantly described in the methodology above and was anticipated upfront in the research protocol. The second cycle was the result of an emergent choice of the researcher to narrow the analytical scope to papers on the adoption, diffusion, and assimilation of innovations.
Studies primarily interested in the adoption, diffusion, and/or assimilation of innovations compose 37% (N=84) of the relevant articles. The rationale for the limitation to this subset is the notion that diffusion comes after the „innovating‟ has occurred (Crossan & Apaydin, 2010). While Crossan and Apaydin acknowledge that the separation of innovation as a process and innovation as an outcome is possibly problematic (Crossan & Apaydin, 2010; Sood & Tellis, 2005), the same can be found on the separation between innovation as a process and diffusion. The duality of diffusion and innovation provides a simplistic view of reality, ignorant of terms as „adaptation‟ or „improvisation‟ (Foster & Heeks, 2013). Rogers (2003), who contributed strongly to the dualistic view with his seminal work on the diffusion
23%
4%
52% 11%
3% 7%
13 of innovations (Rogers, 1962), introduced the term reinvention in his later editions to describe “the degree to which an innovation is changed or modified by a user in the process of adoption and implementation” (Rogers, 2003, p. 17).
Starting from the above mentioned and other initial observations, the second coding cycle was initiated with another round of first-level coding (Miles & Huberman, 1994). The coding scheme for this coding, which emerged in an iterative fashion, can be found in appendix B. To the outcomes of the second cycle first-level coding, I will turn now.
Specific theories.
A number of theories play a prominent role in adoption, diffusion, and assimilation studies. Naturally, diffusion of innovations theory is among them. Out of the 70 studies that form the final sample of this research, 57 refer to one of Rogers‟ editions of „Diffusion of Innovations‟ (Rogers, 2003). Rogers‟ work is applied at the individual level (16 occasions), the group level (1), the organizational level (30), the artefact level (2) and in multi-level studies (6).
Furthermore, a number of (mostly individual-level) technology acceptance theories are used to study how certain characteristics of the innovation, such as perceived utility or perceived ease of use, affect the acceptance of a technological innovation. The seminal „theory of reasoned action‟ (Fishbein & Ajzen, 1975) is applied two times, as is its later improvement „theory of planned behaviour‟ (Ajzen, 1985). The „technology acceptance model‟ (Davis, 1989) is more popular at ten applications. „Unified theory of acceptance and use of technology‟ (Venkatesh, Morris, Davis, & Davis, 2003) is applied five times, but on four out of five occasions in combination with the technology acceptance model or theory of reasoned action.
At the organizational level, 10 quantitative studies using adoption as a dependent variable apply the contextual elements of technology innovation: the environmental, the technological and the organizational context (Tornatzky & Fleischer, 1990).
14
Table 4
Popular theories in adoption, diffusion, and assimilation research Grand theory Author, specific theory Number of
applications Level of analysis Diffusion of innovations theory Rogers (1962) 5 Organizational (30); individual (16); group (1); artefact (2); multi-level (6) Rogers (1983) 18 Rogers (1995) 30 Rogers (2003) 13 Technology acceptance theories
Fishbein & Ajzen (1975), theory of reasoned action
2 Individual (11); organizational (1) multi-level (1) Ajzen (1985) theory of planned
behaviour
2 Davis (1989) technology
acceptance model
10 Venkatesh et al. (2003) unified theory of acceptance and use of technology
5
TOE framework Tornatzky & Fleischer (1990) 10 Organizational (10) Institutional
influences theory
DiMaggio & Powell (1983) 10 Organizational (6) indi-vidual (2) multi-level (2)
Note. the number of times applied sometimes exceeds the quantities under level of analysis. This occurs because some sampled papers apply more than one specific theory or author within a grand theory.
Composing a framework: issues and tensions.
There are a few general issues with research into the adoption, diffusion, and assimilation of innovations within the IS context. As noted in the introduction, there have been no literature reviews in the consideration set that attempted to provide an overarching framework of the field. This field is characterized by a number of complications that diminish the value of such a framework.
Firstly, adoption, diffusion, and assimilation are each concepts that have been applied to multiple levels of analysis (Lucas Jr. et al., 2007). Moreover, research on the adoption, diffusion, and assimilation of innovations is characterized by assuming stages, with differing determinants and different relations depending on the stage that is investigated (Fichman, 2001). There are two problems with these stages. The jingle-jangle fallacy will be described in detail in the synthesis section. An existential problem, general agreement on stages, is described here.
15 „stages‟ have been themed episodes instead. Even so, research often focuses on a single point in time, analysing one stage. Findings tend to differ significantly across stages (Fichman, 2001). Differences in findings between stages are found in several selected papers. Karahanna, Agarwal and Angst (2006) noted that even though values are by definition seen as relatively stable, new technologies possibly trigger a process of in which dominant value systems are changed. Furthermore, Karahanna, Straub and Chervany (1999) found that pre-adoption beliefs of individuals differ from post-pre-adoption beliefs in IT pre-adoption. Finally, the degree to which an innovation is seen as a success can also vary considerably from stage to stage (Liang, Saraf, Hu, & Xue, 2007). Building on these complications, a comprehensive framework of the innovation diffusion trajectory should at least be level and multi-stage, with special attention to the time-scope of each incorporated determinant influencing the process.
A different issue lies in the generalizability of studies not only across industries, but also between specific types of IS innovations. Types of IS innovations such as inter-organizational systems (Messerschmidt & Hinz, 2013) and enterprise resource planning (Saraf, Liang, Xue, & Hu, 2013) have accumulated significant amounts of research. However, generalizing these findings across these types of innovations is often deemed problematic (Messerschmidt & Hinz, 2013; Saraf et al., 2013). Crossan and Apaydin (2010) find this issue in general innovation literature as well. Their explanation lies in the iterative, path-dependent nature of innovation, which is quite in line with that of Newell et al. (2000) above on the stages of innovation diffusion. From a dualistic, linear viewpoint, these two observations can be treated as separate and unrelated. However, when questioning the separation of innovation as a process and diffusion, it is possible to see how a path dependent, iterative process, which surfaces in particular episodes, complicates the conceptual separation of stages as well as the generalizability of findings from diffusion processes.
16 Summarizing, there are several issues and tensions in literature on the adoption, diffusion, and assimilation of innovations in the context of IS. Partially, these can be attributed to the problematic nature of innovations themselves and the subsequent diffusion and assimilation of innovations. Given the illustrated issues and tensions, it is not deemed feasible to synthesize actual hypothesized relations, as variables are dependent on too many contingencies to be generalized into one generic framework. Instead, the following synthesis will provide a comprehensive overview of the selected literature on a number of emergent dimensions.
Synthesis
In this section, the second round of pattern coding will be reported in detail (Miles & Huberman, 1994). This content analysis will result in a comprehensive synthesis of the selected literature. The section will start with an elaboration on the field for each identified dimension: „the nature of the diffusion trajectory‟ (problematic or unproblematic, in conceptualizations and in the setting of the actual empirical research), „the source of innovation‟ (internal invention or external adoption), „the level of analysis used‟, „the stages of the process that are investigated‟. Assuming the conceptual nature of innovations during the diffusion trajectory as the primary distinction, table 5, figure 5, and figure 6 describe its relation to other dimensions. This synthesis enables a view on the complexity involving composing a comprehensive framework that considers the nuances that should be added. Specifically, figure 6 provides a process framework of the innovation diffusion trajectory in the context of IS set out to the subsequent conceptualization of innovation as an outcome or a process. A basic underlying assumption of the framework is that there are iterative, complex and non-linear aspects to the diffusion trajectory of innovations not susceptible to a sequential model. Therefore, the sequential aspect of the framework only describes the conceptual configuration of stages and the subsequent conceptual view on innovation as a process or an outcome during these stages.
View on Innovation and the Nature of Innovation during the Diffusion Trajectory
17 adopted. However, during the review it appeared that this assumption does not hold for a large section of the literature sample. During the second round of pattern coding, a number of conceptual stances toward the nature of innovation during the diffusion and assimilation process appeared, allowing papers to be placed on a continuum moving from those that view innovation as an unproblematic outcome to those considering innovation a process. All but five papers could be categorized accordingly.
Firstly, an absolute majority of the papers (41) conceptualize innovation as an unproblematic outcome during the diffusion trajectory. Innovations are viewed here as a product which diffuses throughout the target group in a linear manner. All but two technology acceptance model, theory of reasoned action, theory of planned behavior, unified theory of acceptance and use of technology, and technological, organizational, and environmental context papers are in this group. A number of these papers measures its dependent variable in actual use (e.g. Picoto, Bélanger, & Palma-Dos-Reis, 2014). However, the majority focuses on the intention to adopt an innovation. Note that an S-shaped diffusion is not assumed by definition in this group. A number of authors focus on the continued use of innovations after a period of initial use. Discontinuance or rejection can lead innovations that are adopted not to diffuse, or adoptions that seem to diffuse successfully initially to be abandoned in the long term (Cho & Kim, 2001-2002). In addition, the initial decision to adopt can be reversed later on, a frequent occurrence in the case of online services (Parthasarathy & Bhattacherjee, 1998).
Secondly, three papers conceptualize innovations as an unproblematic outcome throughout the research, but acknowledge that this is a simplification of reality. Rai (1995) deliberately chooses a rational approach to his research on the diffusion of CASE innovations, but opts for future researchers to choose a combined rational and political perspective, which would be more appropriate to understand experimentation with CASE innovation by users. Bunker, Kautz, and Nguyen (2007), and Baird, Furukawa, and Raghu (2012) choose a similar approach, and acknowledge in their limitations that this rational approach tells only a part of the story.
18 assimilation of an innovation. For example, a multistage process model that innovations during diffusion, but incorporates usage as a final step, which is defined as the situation where an innovation has become routine (Newell et al., 2000). Finally, some papers conceptualize innovations as unproblematic throughout the scope of their research, but mention a problematic aspect of innovation during diffusion not directly related to their research. For example, Zhu et al. (2006a) focus on how certain determinants affect e-business usage, which then influences e-business impact. During these stages, the innovation is not problematized. However, the authors do mention how the process towards usage, which does not fall within the direct time-scope of the research, is characterized by mutual adaptation of the organization and the innovation (Zhu et al., 2006a).
Finally, there is a group of papers (15) that conceptualize innovations as a process during the stages of the diffusion trajectory their researches cover. This implies that the papers in this category do not see an innovation as finished and stable within their time-scope. What stands out about this group is that 67% of these papers uses either qualitative or both qualitative and quantitative methods, while this number is 35% among the rest of the sample. Moreover, while 7 out of 15 „innovation as a process‟ papers conduct variance research, this number is 48 out of 52 among the rest of the sample. As variance research often measures at discrete points in time, it is of no surprise that the problematic innovation diffusion process, which can only be observed over time, is relatively scarce in this type of research.
By conceptualizing innovation diffusion and assimilation in a problematic manner, researchers are able to grasp innovations that change during these stages, or even iterate between stages in a non-linear manner (Foster & Heeks, 2013). The data gathered allows a simple and direct test of this premise by delineating the actual treatment of innovations during the diffusion trajectory in the selected papers. By doing so, the theoretical conceptualizations are set out against the subsequent empirical observations, when these are present. The results of this test present a coherent picture in which an unproblematic conceptualization of innovation during diffusion trajectories leads to a linear process with a crystallized innovation (see table 5). Furthermore, considering an innovation to some extent as problematic yields two possible consequences for the process. It can be a linear process in which an innovation is modified along the way, or it can be a non-linear process in which an innovation or aspects of it iterate between stages.
19 be explained by a difference in time scope between the focus of the study, on which the conceptual categorization is based, and the observations that are discussed in the paper, which form the input for categorizing the empirical treatment of innovations (e.g. Rai, Brown, & Tang, 2009).
The Source of Innovation
Innovations within an organization can originate from two basic sources. They can either be generated (invented) by the organization or adopted from an external source (Gopalakrishnan & Damanpour, 1997). Organizations can engage in generation, adoption, or in both (Crossan & Apaydin, 2010). In practice, organizations engage in a number of related, partially related or unrelated innovations, of which the first two can be viewed in concert as aggregated innovation (Rai et al., 2009). Gopalakrishnan and Damanpour (1997) explicitly distinguish generation and adoption as they hold that the two are distinct phases, each consisting of multiple stages. It is thus interesting to investigate whether a difference in the conceptualization of innovation exists between studies covering an adoption and studies covering an invention within the field of IS.
Deciding whether a study covers adoptions or inventions requires going beyond the explicit mentions of authors, as these are rarely present. Conveniently, a vast majority of the papers clearly describe their focal innovations. This enables the observer to deduce the answer. As such, it is possible to conclude that within the sample there is a strong dominance of papers that research adoptions. Fifty-five out of the 67 articles that covered any diffusion trajectory stage focused their research on adoptions alone.
20
Table 5
Problematizing innovations conceptually and empirically during the diffusion trajectory and the source of innovation. Empirical: horizontal
Conceptual: vertical
non-linear diffusion and assimilation
alteration possible during diffusion and assimilation
Fixed diffusion and assimilation Innovation as an
unproblematic outcome
Adoption: Rai et al.
(2009)
Adoption: Cavaye (1996); Sharma &
Rai (2015)
Invention: Börjesson et al. (2006)
Adoption: Agarwal & Prasad (1998); Baird & Raghu (2015); Bajaj (2000); Bajwa, Lewis,
Pervan, & Lai (2005); Baskerville & Pries-Heje (1998); Carlo, Lyytinen, & Rose (2011); Carter & Bélanger (2005); Chatterjee, Grewal, & Sambamurthy (2002); Chau & Tam (1997); Cho & Kim (2001-2002); Fichman (2001); Fichman & Kemerer (1999); Gefen & Straub (2000); Karahanna et al. (2006); Karahanna et al. (1999); Lee & Shim (2007); Liang, Xue, Ke, & Wei (2010); Loh & Ong (1998); Lu, Yao, & Yu (2005); Messerschmidt & Hinz (2013); Parthasarathy & Bhattacherjee (1998); Picoto et al. (2014); Premkumar, Ramamurthy, & Crum (1997); Quaddus & Hofmeyer (2007); Rai & Patnayakuni (1996); Reardon & Davidson (2007); Singh & Phelps (2013); Tan & Teo (2000); Teo, Wei, & Benbasat (2003); Teo & Pian (2003); Thong (1999); Van Slyke, Ilie, Lou, & Stafford (2007); Venkatesh & Bala (2012); Wang (2010); Zhu & Kraemer (2005); Zhu, Kraemer, Gurbaxani,& Xu (2006b)
Both: Compeau et al. (2007)
Innovation as an unproblematic outcome, problematic reality recognized
Adoption: Baird et al. (2012); Bunker et al. (2007); Rai (1995)
Innovation as both an outcome and a process
Adoption: Newell et al.
(2000); Swanson & Ramiller (2004);
Both: Cho et al. (2007)
Adoption: Fichman (2004b);
Mangalaraj, Mahapatra, & Nerur (2009); Saraf et al. (2013); Zhu et al. (2006a)
Both: Mustonen-Ollila & Lyytinen
(2004) Innovation as a
problematic process
Adoption: Newell,
Swan, & Robertson (1998); Ramiller & Swanson (2003)
Invention: Bansler et al.
(2000); Baskerville & Pries-Heje (2001); Teo, Srivastava,
Ranganathan, & Loo (2011)
Both: Foster & Heeks
(2013)
Adoption: Ahuja & Thatcher (2005);
Liang et al. (2007); Mishra & Agarwal (2010); Pozzebon, Mackrell, & Nielsen (2014); Vega, Chiasson, & Brown (2008);
Invention: Nidumolu, Goodman,
Vogel, & Danowitz (1996); Yetton, Sharma, & Southon (1999)
Both: Igara (2008); Sherif & Menon
21 Finally, the remaining five papers cover both adoptions and inventions. This implies that either the study explicitly investigates both inventions and adoptions (e.g. Mustonen-Ollila & Lyytinen, 2004), the study does not discuss the origin of the innovation that was studied (e.g. Compeau, Meister, & Higgins, 2007), or the study takes multiple perspectives, including the generator as well as the adopter in the diffusion or assimilation process (e.g. Foster & Heeks, 2013). The view on innovations in these papers can hardly be synthesized, as the grounds for having both a problematic and an unproblematic stance toward innovation differ. However, the blurring of the switch between „innovation‟ and diffusion is present in this group as well, as explicated by Foster and Heeks (2013).
Summarizing the findings of this section and the previous section, table 5 presents a framework of the conceptual and empirical view on the nature of innovation in the sampled IS literature, complemented with a distinction between adoptions, inventions and studies that cover both inventions and adoptions. The table demonstrates how a large group of papers that treats innovations as conceptually and empirically unproblematic focuses on adoptions alone. Adoptions are still an absolute majority amongst adoptions, diffusions or assimilations that are both problematic and unproblematic at six out of eight. However, amongst problematic, innovation as a process trajectories, adoptions form a slight minority at seven out of 15.
Level of Analysis
The distribution of papers on level of analysis within the adoption, diffusion, and assimilation sample is similar to that of the general innovation sample. Two major trajectories can be recognized in the literature sample, based on levels of analysis. The individual-level trajectory (N=23) discusses behavioural occurrences, intentions or decisions of individuals. The organizational-level trajectory (N=45) is more diverse in its approach, researching aggregated behavioural issues, competitive advantage, organizational decision making and a number of other diffusion related organizational-level variables.
22 (1996) paper cannot be categorized in one of the trajectories as is explicitly describes group-level processes and draws on group-group-level theories to compose group-group-level arguments.
Stages of Adoption, diffusion, and assimilation
As the processes of adoption, diffusion, and assimilation of innovations inherently compose stages, all authors within the definitive sample discuss the existence of stages in some form. However, there is not one dominant stage model out there. Instead, authors draw on different stage models and pose diverse definitions of stages. Out of the 70 papers composing the definitive literature sample, 41 papers only define the specific stage(s) they investigate, without explicating the process this stage is, or these stages are a part of. Five papers define their own process, or introduce a new stage model. The remaining 24 papers apply a pre-existing stage model to their research. Illustrative for the lack of agreement on stages is the finding that these 24 papers draw on 13 different stage models. Starting from these observations, it is of no surprise that the use of stages in the literature sample suffers from the jingle-jangle fallacy (Oreg, Vakola, & Armenakis, 2011).
On the one side, different terminology is used to describe the same stages (the jangle fallacy). For example, innovation adoption is defined in multiple ways, amongst which the uptake of an innovation by a certain target group (Cavaye, 1996). At the same time, Igara, (2008) writes a paper about IS innovation uptake, which is again treated as the uptake of an innovation by a specific target group. Though identical in definition, Cavaye (1996) draws solely on adoption literature, while Igara (2008) draws solely on uptake literature.
The other way around, the same words are used to describe different stages (the jingle fallacy). For example, a large group of authors define „adoption‟ as the acquisition of an innovation, explicitly excluding actual use from the concept (e.g., Rogers, 2003). Meanwhile, Zhu et al. (2006b) measure adoption in volume, breadth and depth, of which for example volume refers to “the extent (percentage) to which each of the major value chain activities has been conducted on the open standard IOS” (Zhu et al., 2006b, p. 522), thus incorporating use directly into the concept.
Fully comprehending the interrelations and conceptual differences between stages and the different stage models around would be beyond what is reasonably possible for this research to achieve. As a systematic review, it is nevertheless appropriate for this research to attempt to contribute to progressing comprehension.
23 adoption and use of IT. In order to comprehensively review the entire diffusion trajectory of innovations in the context of IS, the focal literature sample is categorized within the „process framework of IT adoption and use‟, proposed by Lauterbach and Mueller (2014). In the next section, the framework will first be introduced, then it will be applied to the 70 papers that compose the literature sample, and finally it will be connected to innovation as a process and innovation as an outcome. The two latter steps combined are presented as an extended version of the Lauterbach and Mueller (2014) framework.
A Framework of the view on Innovation during the Diffusion Trajectory
As the process framework of IT adoption and use is proposed to be a “first step towards conceptual alignment” (Lauterbach& Mueller, 2014, p. 13), this literature review can thus be seen as taking a second step towards conceptual alignment. In the first place by confirming the applicability of the framework to a specific context, that of literature on innovation in IS, and in the second place by expanding the framework with the conceptual view on innovation during the stages of the diffusion trajectory.
Figure 5
24 Figure 5 depicts an adaptation of the „process framework of IT adoption and use‟ by Lauterbach and Mueller (2014). The terminology and definitions of their stages are applied directly to the literature sample of this research; definitions can be found in appendix E. The adapted framework allows for the categorization of papers on their view of innovation during each stage the researches cover. Furthermore, the framework shows two processes occurring simultaneously, the organizational-level diffusion trajectory, and the individual-level diffusion trajectory. The adapted framework will answer the research question in two steps, which are described here in their logical order.
Placing the sample within the framework.
By assigning each paper within the definitive sample that investigates one or more of the included stages to the respective stage(s) of Lauterbach and Mueller (2014) it addresses, the applicability of the framework is tested in the context of innovation in IS. In total, 64 out of the 70 selected papers can be included into the framework (as depicted in figure 5). This does not imply that these papers choose the same wording as used in the framework to describe their stages. Rather, it implies that these papers define (or when definitions are lacking, treat) their stages in ways corresponding to the definitions used in the framework.
Out of the remaining six papers, three do not investigate any stage. For example, Lyytinen and Rose (2003), in their work on disruptive IS innovations, discuss the implications these innovations have on their trajectory of adoption and use. Their research does not directly cover any stage of the innovation diffusion trajectory. Three papers do use stages, but cannot be placed within the framework of Lauterbach and Mueller (2014). Foster and Heeks (2013) apply an artefact focus, with innovation diffusion being iterative and non-linear. Consequentially, distinguishable stages are absent in this research. Swanson and Ramiller (2004) recognize four processes (comprehension, adoption, implementation and assimilation) which do not have a fixed sequence and can progress simultaneously. Finally, Rai and Patnayakuni (1996) cover the adoption of an innovation, measured in its use. However, the paper applies a group level of analysis, which is not covered by the framework.
In general, the prevalence of research throughout stages differs significantly. At an organizational level, the conversion process predominantly features in multi-stage research. Only one study focuses solely on the conversion process (Baskerville & Pries-Heje, 2001). The other stages are better represented within the sample.
25 studied as part of the multi-stage Pozzebon et al. (2014) research, whereas this research is only assisted by one additional study in the infusion stage (Ahuja & Thatcher, 2005).
Adding the conceptual view on innovation to the framework.
The primary objective of this research is to place the innovation diffusion trajectory relative to innovation as a process and innovation as an outcome in terms of Crossan and Apaydin (2010). The „framework of the conceptual view on innovation in IS during the diffusion trajectory‟ (figure 5) provides a comprehensive overview of the conceptual view on innovations throughout the different stages of organizational and individual-level diffusion trajectory. Starting from the overview of papers with regard to the time-scope their research covers, the research question can thus be answered in detail. In doing so, figure 5 provides an overview of when in the process innovations tend to be problematized, and when innovations are considered an unproblematic outcome. While it is clear from the diverse conceptualization of innovations throughout the trajectory that there is no one conclusive answer available, a few useful observations arise.
In general, virtually all individual-level single-stage papers see the diffusion trajectory of innovations as unproblematic during the specific stage they research. The only exception is formed by Ahuja and Thatcher (2005), who study the „last‟ stage covered by a single-stage individual-level research (infusion). Three out of the five studies that oversee multiple individual-level stages see the trajectory as problematic. At the organizational level, problematizing innovations occurs throughout stages as well (the competitive process being the exception). Mishra and Agarwal (2010), for example, suggest that organizations and innovations need to adapt to each other continuously, regardless of the stage in the innovation process. Overall, the notion that diffusion comes after the innovating has occurred does not hold throughout the individual and organizational-level trajectories.
26
Figure 6
27 In most papers, development of an innovation through time is not the subject of interest. Instead, papers often study correlates between variables at a discrete and defined point in time. This type of measurement is by definition non-susceptible to change, as change is a time-bound phenomenon.
While most „innovation as an unproblematic outcome‟ papers do not discuss the nature of innovations during the diffusion trajectory, three papers within the selection indeed explicitly recognize their rational-linear stance towards innovation diffusion and explicate their choice to purposely simplify the complexity surrounding innovation diffusion for the means of their specific investigation. Not surprisingly, the majority of papers that take a problematic, innovation as a process perspective are those studying multiple stages.
In summary, it appears as though the type of research conducted is a major determinant for the conceptualization of innovation as an outcome or as a process during the diffusion trajectory. Thus, while multi-stage research that does not take a positivist stance might complicate matters beyond what is possible to capture in a framework within reason, there is an invaluable benefit to including these studies for reaching a comprehensive view of the innovation diffusion trajectory in the context of IS (Mangalaraj et al., 2009).
Discussion
In this section, a general discussion of the main findings is provided. Thereafter, three specific directions for future research are proposed. Implications for practitioners are described, followed by the limitations of this research.
28 The primary contribution of this paper and answer to the research question is the addition of the diffusion trajectory to the framework of Crossan and Apaydin (2010). To be precise, by extending the framework of Lauterbach and Mueller (2014) with the process/outcome stance of papers throughout stages, a nuanced placement of the adoption, diffusion, and assimilation of innovations relative to the multi-dimensional framework of organizational innovation by Crossan and Apaydin (2010) is presented. In summary, the framework demonstrates how viewing innovation as an outcome throughout the diffusion trajectory is a simplification that does not hold in reality. „Innovation as an unproblematic outcome‟ research that chooses a discrete point in time to measure, has delivered valuable results. However, to prevent misinterpretation, this school of literature is advised to formulate clearly its philosophical stance and its functional simplifications.
Following the systematic review, researchers taking a multi-stage approach are warranted to match their stance toward innovations to the research approach they intend to adopt. For instance, Karahanna et al. (1999) in their research on pre-adoption and post-adoption (operationalized corresponding to continued usage) beliefs, are not interested in changes to the innovation itself. Rather, their focus is on changes in beliefs of individual adopters (Karahanna et al., 1999). However, Cho et al. (2007) focus specifically on the interaction between groups, and its role in facilitating adoption. As this interaction was in practice partially oriented at the innovation itself, ignoring the possibility for change of the innovation would have muted an important aspect of the adoption process.
Two specific (not mutually exclusive) groups of research are likely to surface „innovation as a problematic process‟ characteristics. First, a constructivist philosophical stance criticizes rational-linear sequential models by its very nature. Second, as a research approach, process research emerged from the systematic review as prone to problematizing innovations during the diffusion trajectory. An explanation for this finding can be found in the inherent multi-stage nature of process research.
29 entanglement of problematizing innovation during the diffusion trajectory with the source of the investigated innovation. Another affordance of systematically identifying dimensions and categorizing papers accordingly is the observation of rich aspects of the field as well as under-studied aspects. These observations allow for the identification of gaps in the literature that can possibly be addressed by future research. To these gaps, I will turn now.
Invention or Adoption of Innovations
Among the selected papers, 83% covers external adoptions alone. Two small groups of papers embrace either both inventions and adoptions, or inventions alone. While internally generated innovations and innovations adopted from an external source are known to have clearly distinguishable phases (Gopalakrishnan & Damanpour, 1997), authors often do not explicate their focus. Wording can be deceiving as well on this notion. Börjesson et al. (2006), for example, speak solely of an internally conceived innovation. However, as their level of analysis is the individual, they write about the adoption of an innovation by organizational members. Furthermore, among papers that study inventions, or both inventions and adoptions, explicit mentions of the chosen focus are virtually absent.
From the above discussion, two gaps emerge. First, the diffusion and assimilation of internally generated innovations is understudied, relative to external adoptions. Second, the distinction between generation and adoption stipulated by Gopalakrishnan and Damanpour (1997) has gone unnoticed in IS literature over the past two decades. Research into the conceptual distinctions between the diffusion and assimilation of innovations originating from these two possible sources of innovation could provide useful insights, informing researchers and practitioners about the generalizability of studies on innovations originating from either source.
Later Stages in the Individual Trajectory
30 (Ahuja & Thatcher, 2005). This paper provides valuable insights into the process of trying to innovate with a new IT and consequentially suggests various avenues for further research.
One of the suggested future research directions of Ahuja and Thatcher (2005) is to conduct process research into factors influencing the infusion process. A similar call was made by Rogers (1995) right at the starting point of the coverage of this research. Rogers (1995) states that innovation research is dominated by variance research, whereas process research could add an understanding of the sequence of events in innovation diffusion processes. Given that only five papers within the selection conducted process research, this call can be prolonged two decades later. Within the context of the adoption, diffusion, and assimilation of innovation in IS, opportunities for process research lie in exploring the necessary conditions for entering and passing through the different individual-level stages as defined by Lauterbach and Mueller (2014). Moreover, as it is recognized that firms and technologies need to adapt to each other continuously (Mishra & Agarwal, 2010) and that individual-level processes significantly shape the use of innovations (Jasperson, Carter, & Zmud, 2005), process research could uncover in detail the degree to which organizational-level conversion processes, use processes and competitive processes are entangled.
The Appropriation, Enactment or Alteration of Innovations after Adoption
31
Implications for Practice
Managers drawing on academic literature to make decisions in the field can benefit from the findings of this research. The main contribution to practitioners is its emphasis on the multi-dimensional complexity surrounding innovation diffusion trajectories in the context of IS. On top of these dimensions, when implementing an innovation, this research provides IS practitioners the awareness that an innovation can possibly be altered at any stage in the diffusion trajectory, perhaps even moving iteratively between stages.
Limitations
This research is prone to a number of limitations, which are recognized by the researcher. Firstly, to ensure high quality data, a choice has been made to confine the search to the „AIS basket of eight‟ top journals. However, high quality IS literature does not conform to strict boundaries. It is likely that high quality, influential IS innovation research has thus been omitted in advance. A follow-up research can solve this problem by „going forward‟ and „going backward‟ (Webster & Watson, 2002). This implies reviewing citations of selected articles and the other way around identifying papers that cite the selected articles (Webster & Watson, 2002).
Secondly, as this research has focused on conceptual dimensions at a rather abstract level, it falls short of linking actual variables into a comprehensive framework. It has been argued before (e.g. Compeau et al., 2007), and is argued again here, that there might be too many situational factors for such a generic framework to be feasible.
32
References
Agarwal, R., & Prasad, J. (1998). A conceptual and operational definition of personal innovativeness in the domain of information technology. Information Systems Journal, 9(2), 204-215.
Ahuja, M., & Thatcher, J. (2005). Moving beyond intentions and toward the theory of trying: effects of work environment and gender on post-adoption information technology use. MIS Quarterly, 29(3), 427-459.
Ajzen, I. (1985). From intentions to actions: a theory of planned behavior. In J. Kuhl, & J. Beckmann, Action Control: From Cognition to Behavior (pp. 11-39). Berlin, Heidelberg, New York: Springer-Verlag.
Aksulu, A., & Wade, M. (2010). A comprehensive review and synthesis of open source research. Journal of the Association for Information Systems, 11(11), 576-656. Baird, A., & Raghu, T. (2015). Associating consumer perceived value with business models
for digital services. European Journal of Information Systems, 24, 4-22.
Baird, A., Furukawa, M., & Raghu, T. (2012). Understanding contingencies associated with the early adoption of customer-facing web portals. Journal of Management
Information Systems, 29(2), 293-324.
Bajaj, A. (2000). A study of senior information systems managers' decision models in adopting new computing architectures. Journal of the Association for Information Systems, 1(1), 4.
Bajwa, D., Lewis, L., Pervan, G., & Lai, V. (2005). The adoption and use of collaboration information technologies: international comparisons. Journal of Information Technology, 20, 130-140.
Bansler, J., Damsgaard, J., Scheepers, R., Havn, E., & Thommesen, J. (2000). Corporate intranet implementation: managing emergent technologies and organizational practices. Journal of the Association for Information Systems, 1(1), 10.
Barrett, M., Heracleous, L., & Walsham, G. (2013). A rhetorical approach to IT diffusion: Reconceptualizing the ideology-framing relationship in computerization movements. MIS Quarterly, 37(1), 201-220.
Baskerville, R., & Pries-Heje, J. (1998). Information technology diffusion: building positive barriers. European Journal of Information Systems, 7, 17-28.
Baskerville, R., & Pries-Heje, J. (2001). A multiple-theory analysis of a diffusion of information technology case. Information Systems Journal, 11, 181-212.
Börjesson, A., Martinsson, F., & Timmerås, M. (2006). Agile improvement practices in software organizations. European Journal of Information Systems, 15, 169-182. Bunker, D., Kautz, K.-H., & Nguyen, A. (2007). Role of value compatibility in IT adoption.
33 Carlo, J., Lyytinen, K., & Rose, G. (2011). Internet computing as a disruptive information
technology innovation: the role of strong order effects. Information Systems Journal, 21, 91-122.
Carter, L., & Bélanger, F. (2005). The utilization of e-government services: citizen trust, innovation and acceptance factors. Information Systems Journal, 15, 5-25.
Cavaye, A. (1996). The implementation of customer oriented inter-organizational systems: an investigation from the sponsor's perspective. European Journal of Information
Systems, 6, 103-119.
Chatterjee, D., Grewal, R., & Sambamurthy, V. (2002). Shaping up for e-commerce: institutional enablers of the organizational assimilation of web technologies. MIS Quarterly, 26(2), 65-89.
Chau, P., & Tam, K. (1997). Factors affecting the adoption of open systems: an exploratory study. MIS Quarterly, 21(1), 1-24.
Cho, I., & Kim, Y.-G. (2001-2002). Critical factors for assimilation of object-oriented programming languages. Journal of Management Information Systems, 18(3), 125-156.
Cho, S., Mathiassen, L., & Robey, D. (2007). Dialectics of resilience: a multi-level analysis of a telehealth innovation. Journal of Information Technology, 22, 24-35.
Compeau, D., Meister, D., & Higgins, C. (2007). From prediction to explanation:
reconceptualizing and extending the perceived characteristics of innovating. Journal of the Association for Information Systems, 8(8), 409-439.
Corbin, J., & Strauss, A. (2008). Basics of qualitative research: techniques and procedures for developing grounded theory (3rded.). Thousand Oaks, CA: Sage Publications. Córdoba, J.-R., Pilkington, A., & Bernroider, E. (2012). Information systems as a discipline
in the making: comparing EJIS and MISQ between 1995 and 2008. European Journal of Information Systems, 21, 479-495.
Crossan, M., & Apaydin, M. (2010). A multi-dimensional framework of organizational innovation: a systematic review of the literature. Journal of Management Studies, 47(6), 1154-1191.
Crowston, K., Wei, K., Howison, J., & Wiggins, A. (2012). Free/Libre open-source software development: what we know and what we do not know. ACM Computing Surveys, 44(2), 35.
Davis, F. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13(3), 319-340.
Dess, G., & Picken, J. (2000). Changing roles: leadership in the 21st century. Organizational Dynamics, 28(3), 18-34.
