1 s2.0 S2096248717300012 Main

International Journal of Innovation Studies 1 (2017) 107e133
Contents lists available at ScienceDirect
International Journal of Innovation Studies

journal homepage: http://www.keaipublishing.com/en/journals/in-
ternational-journal-of-innovation-studies
Innovation attributes and managers' decisions about the

adoption of innovations in organizations: A meta-analytical
review
Gianluca Vagnani*, Loredana Volpe
Department of Management of Sapienza, University of Rome, Castro Laurenziano 9, Roma 00191, Italy
a r t i c l e i n f o a b s t r a c t
Article history: The adoption of innovations has emerged as one of the dominant research topic in the
Received 28 July 2016 management of innovations in organizations, though investigations often yield mixed
Received in revised form 26 April 2017 results. To help managers and researchers improve the effectiveness of their efforts, the
Accepted 20 June 2017
authors employed meta-analysis integrated by structural equation modeling to analyze the
Available online 24 October 2017
associations between the attributes of innovations, behavioral preferences of managers
and organizations' innovation adoption decisions in a mediated-moderated framework.
Keywords:
Our findings offer evidences that attributes of innovations influence behavioral prefer-
Attributes of innovations
Managers' behavioral preferences
ences of managers and, in turn, adoption decisions in organizations. We also observe that
Organizations' innovation adoption the context in which the adoption decision takes place and the research settings employed
decisions by scholars matter. Finally, the theoretical contribution and practical implications of our
Meta-analysis meta-analytical results are discussed.
Mediation-moderation © 2017 Publishing Services by Elsevier B.V. on behalf of KeAi Communications Co., Ltd. This
is an open access article under the CC BY-NC-ND license (http://creativecommons.org/
licenses/by-nc-nd/4.0/).
1. Introduction
Innovations are essential for firms to build core competences and create sustainable competitive advantage (Porter, 2005;
Reed & DeFillippi, 1990; Rogers, 1983; Schumpeter, 1934). At the organizational level, Damanpour and Gopalakrishnan (1998)
observed that a new idea, behavior, practice or product (i.e., innovation) may come to organizations in two distinct ways, i.e.
innovations may be either generated or adopted. When generated, an innovation is initiated and developed in organizations;
when adopted, it is generated elsewhere than in the adopting organization (Angle & Van de Ven, 1989; Rogers, 1983; Zaltman,
Duncan, & Holbek, 1973). And, as observed by March and Simon (1958), most innovations in organizations result from
borrowing rather than invention. This proposition is supported by the wide-ranging research on the adoption of innova-
tionsdparticularly, on the identification of the factors that facilitate organizations to adopt a new technologydacross several
levels of analysis that have been produced within several disciplines in management sciences over the decades (for a review
see Crossan & Apaydin, 2010; Wolfe, 1994).
Since the elaboration of the innovation diffusion theory (Rogers, 1983), a large number of studies have considered how
attributes of innovations influence the adoption decisions taken by managers in organizations. However, results from
* Corresponding author.
E-mail addresses: [email protected] (G. Vagnani), [email protected] (L. Volpe).
Peer review under responsibility of China Science Publishing and Media Ltd.
https://doi.org/10.1016/j.ijis.2017.10.001
2096-2487/© 2017 Publishing Services by Elsevier B.V. on behalf of KeAi Communications Co., Ltd. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
108 G. Vagnani, L. Volpe / International Journal of Innovation Studies 1 (2017) 107e133
different studies addressing the same topic often appear to conflict and often really do conflict. Downs and Mohr (1976)
suggested that attributes of innovations are interdependent with characteristics of organizations. Therefore, it is unlikely
to find the same relationship between a given attribute of an innovation and adoption decisions across a large array of or-
ganizations. Wolfe (1994) noted that the most consistent result of innovation research is that results are inconsistent. Other
scholars have further noted that extant studies exhibited somehow conflicting results and variations in magnitude, statistical
significance, and direction of the observed relationships between the attributes of innovations and adoption decisions (see
also Bruque & Moyano, 2007; Premkumar, 2003). In particular, several classes of conflicts can be identified. One concerns the
relationship between perceived benefits and adoption decisions. Consider, for example, Kurnia, Choudrie, Mahbubur, and
Alzougool (2015) who found that the perceived benefits are associated negatively to the adoption decision. Conversely,
Premkumar (2003) found a significant and positive association between the same variables.
Similar contradictory results can be found on the relationship between compatibility and adoption decisions. Some studies
have reported that compatibility has a positive relationship with the decision to adopt an innovation (Gupta, Seetharaman, &
Raj, 2013; Luqman & Abdullah, 2011), while others have provided evidence of a significant, negative relationship between
compatibility and the innovation adoption decision (Wanyoike, Mukulu, & Waititu, 2012). Consequently, it has been pointed
out that several decades of empirical studies into the adoption of innovations by organizations have failed to provide clear and
consistent findings (Keupp, Palmie , & Gassmann, 2012; Tidd, 2001).
Hence comes the first research question underlying this paper which aims at understanding how practitioners and re-
searchers identify the attributes of innovations which are more critical to managerial decisions about the adoption of in-
novations by organizations. The systematization of the work to date on innovation research and the identification of areas for
future enquiries represent an important task for future studies (Anderson, Poto cnik, & Zhou, 2014). Therefore, we employ
meta-analysis to integrate empirical findings on the effects of specific attributes of innovations (i.e., relative advantage,
compatibility, complexity, observability, and trialability) on managerial decisions on the adoption of innovations in organi-
zations. Note that the use of meta-analysis allows us to combine data collected from companies with different characteristics
and operating in various industries and countries, which would be infeasible in typical on-field research.
Moreover, we conceive the effects of attributes of innovations (i.e., relative advantage, compatibility, and complexity) on
adoption decisions as mediated by one or more behavioral preferences of managers represented by attitude, subjective norm,
and perceived behavioral control. In conceptualizing such mediation relationship, we move from a number of studies in
extant literature. Pierce and Delbecq (1977) conceive organization innovation adoption behavior as a decision being made by
the appropriate manager providing mandate and resources for the change in the organization. Kimberly and Evanisko (1981)
provide insights that decision makers' preferences are keys to explain innovation adoption decisions in organizations.
Fishbein and Ajzen (1975, 1985; 1987; 1991) seminal works on the theory of planned behavior identified and modeled the
effects of decision maker preferences on adoption decisions. Taylor and Todd (1995a) developed a decomposed theory of
planned behavior that linked some of the attributes of innovationsdas elaborated within the innovation adoption theorydto
the behavioral preferences of decision makers. Considering these studies, we therefore offer a first attempt to integrate the
innovation diffusion theory with the theory of planned behavior. In the field study of adoption of innovations by organiza-
tions, efforts to integrate theories would probably constitute one of the most important contributions to the progress of
research activities (Downs & Mohr, 1976).
If one considers that existing literature on managerial decisions to adopt innovations in organizations offers a
wide range of antecedents for the considered mediators, and researchers disagree about which one best influences
these decisions. For example, Rogers (1983) suggests that attitudes of organizational decision makers are likely to
mediate the relationships between all the five attributes of innovations and the adoption decision. Other studies based on
the technology acceptance model (Davis, 1989) suggested instead that only a subset of attributes of innovations (i.e.,
relative advantage and complexity) are mediated by attitudes of decision makers. Other studies suggest that not only
attitudes but also the decision maker's subjective norms and perceived behavior control mediate the innovation
attributes-innovation adoption linkages (Taylor & Todd, 1995b). Furthermore, while different studies suggests that the
effects of attributes of innovations are fully mediated by behavioral preferences of decision makers (e.g., Harrison,
Mykytyn Jr & Riemenschneider, 1997; Riemenschneider, Harrison, & Mykytyn, 2003), others consider that there is a
direct effect of the attributes of innovations on manager decisions about the adoption of innovations in organizations
(Tornatzky & Klein, 1982). Some other studies consider managers' behavioral preferences as independent variables (as
the attributes of innovations) that are likely to produce a direct effect on the adoption decision (Kimberly & Evanisko,
1981; Premkumar & Roberts, 1999; Thong & Yap, 1995). Finally, researchers have also offered evidence that the impli-
cations of managers behavioral preferences on the adoption decision may be influenced by contexts in which such de-
cision takes place (e.g., see Damanpour, 1996) as well as by research settings employed by scholars (e.g., see Tornatzky &
Klein, 1982).
The different mediators and the various antecedents, moderators, and consequences of such mediators pointed out in
literature thus trigger our second research question on how decision makers' behavioral preferences contribute to mediate
the attributes of innovations-adoption decisions relationship in organizations.
Although narrative and quantitative review of the literature on the attributes of innovations-adoption decisions
linkage are available (e.g., Anderson et al., 2014; Jeyaraj, Rottman, & Lacity, 2006; Kapoor, Dwivedi, & Williams, 2014a;
Kapoor, Dwivedi, & Williams, 2014b; King & He, 2006; Tornatzky & Klein, 1982; Weigel, Hazen, Cegielski, & Hall,
2014), an empirical review that fully addresses the attributes of innovation-decision makers behavioral preferences-
G. Vagnani, L. Volpe / International Journal of Innovation Studies 1 (2017) 107e133 109
Fig. 1. Mediator-moderator meta-analytic framework.
adoption decisions relationships in organizations has yet to be conducted. This study attempts to fill this void by pro-
posing a meta-analytic framework (see Fig. 1), which tests hypotheses based on a sample of 165 unique studies published
in two decades of research on the antecedents of innovation adoption decisions in organizations. In doing so, this paper
builds on previous literature and takes it further by offering insight into a number of related issues, which follow from
our two main research questions mentioned above: (1) how do variations in attributes of innovations account for var-
iations in managers' decisions about the adoption of innovations in organizations? (2) Which decision makers' behavioral
preferences are mostly affected by variations in attributes of innovations? (3) How do variations in decision makers'
behavioral preferences account for variations in innovation adoption decisions in organizations? (4) How do variations in
contexts of adoption and research settings influence the linkages between behavioral preferences of decision makers and
adoption decisions in organizations?
Our results provide answers to each of these issues. Firstly, in terms of the role of the attributes of innovations, we found
consistent evidence that relative advantage, compatibility, and complexity are likely to promote greater stimulus on the
decision to adopt innovations in organizations. In addition, our results offer evidence that the correlation coefficient of the
trialability-adoption association has the lowest magnitude among the attributes of innovations. Secondly, in terms of the
effect of decision makers' behavioral preferences, a key finding is that the behavioral preferences of decision makers mediate
(although partially) the attributes of innovation-adoption decision linkages. Thirdly, for what concerns the influence of the
context of adoption, we observed that both the hierarchical position of the decision maker and the research settings
employed by scholars both matter for empirical predictions. In particular, we found evidence that measuring adoption as
intention or as expectation as well as measuring dependent and independent variables with the same/different format in-
fluence the magnitude of empirical estimates.
We expect that, by advancing our understanding on each of the issues mentioned above, our results help organizations
develop new technologies that even other organizations are likely to successfully adopt and, at the same time, provide re-
searchers with insights into ways to build more precise, more comprehensive, and more rigorous models about the adoption
of innovations in organizations.
2. Conceptual framework
In this study, we developed a meta-analytic framework that encompasses the potential mediating mechanisms and the
moderating variables that may potentially shape the innovation attributes-innovation adoption relationships in organizations
(see Fig. 1). In the following paragraphs, we present our theoretical background as divided into different sections, each
addressing our two main research questions and their related issues as explained in the introduction to this work. We first
review the literature on the innovation attributes-innovation adoption relationship and then investigate the mediators of
such relationship. Finally, we introduce their potential moderators.
2.1. Main relationships between attributes of innovations and managerial decisions about the adoption of innovations in
organizations
The basic theoretical perspective addressing the main effects of attributes of innovation on adoption decisions in orga-
nizations is the innovation diffusion theory (Rogers, 1983). This perspective has inspired most of the empirical studies on
innovation adoption (for a recent review see Kapoor et al., 2014a; Kapoor et al., 2014b) and it has represented a building block
for a number of widely acknowledged frameworks in literature such as the technology, organization, environment framework
(Tornatzky & Klein, 1982), the technology acceptance model (Davis, 1989), five contextual factors model (Kwon & Zmud,
1987), and competitive effects model (Robertson & Gatignon, 1986).
In the innovation diffusion theory, the adoption decision is conceived as a choice to make full use of an innovation as the
best course of action available in an organization (Rogers, 1983, p. 20). In defining the adoption decision, the innovation
diffusion theory considers a decision maker that is able to assess the attributes and therefore the expected net benefits of
innovations and according to such assessment she makes a choice on whether to adopt innovations or not. As in expectancy-
value models (Fishbein & Ajzen, 1975), the idea is that decision makers are more likely to adopt innovations that are slightly
uncertain and are expected to produce outcomes with a positive valence (see also Abrahamson, 1991; Taylor & Todd, 1995a).
In the innovation diffusion theory, the attributes of innovations that are likely to influence the adoption decisions are
relative advantage, compatibility, complexity, observability, and trialability (Rogers, 1983, p. 211). Specifically, relative
advantage refers to the degree to which an innovation provides desirable consequences for the adopter compared to other
available alternatives, providing the decision makers with insight of its net benefits, which in turn will favor the innovation
adoption decision (e.g., Premkumar, 2003). Compatibility refers the degree to which an innovation consistently fits with
existing business processes, values, past experiences, and needs. An innovation that is more compatible is also less uncertain
to the decision maker (Rogers, 1983, p. 223). Moreover, greater levels of compatibility increase the expected net benefits of an
innovation since the innovation will require the decision maker fewer efforts to integrate the new technology with the
already deployed ones. Therefore, greater levels of compatibility of an innovation are expected to influence positively
adoption decisions in organizations (e.g., Jeon, Han, & Lee, 2006). Concerning complexity, a new technology that is intricate
and difficult to master requires decision makers to invest greater amount of resources to understand it and, subsequently, to
catch up with its adoption, which in turn has negative impacts on the assessment of the net benefits of the innovation.
Moreover, an innovation that is more complex is more uncertain to the decision maker. Therefore, greater levels of complexity
are expected to influence negatively adoption decisions in organizations (Thong, 1999). Trialability and observability offer the
decision maker the possibility to assess the net benefits of the innovation more effectively. Observability may help decision
maker assess the positive features of an innovation. In addition, greater levels of observability may offer the decision maker
greater opportunities to learn about the new technology, which increases its net benefits and, in turn, favors adoption de-
cisions (e.g., Hashem & Tann, 2007). Finally, the availability of the innovation on a trial basis allows decision makers to
experiment with the new technology, exploring in advance its potentialities, and even to identify and resolve ex-ante
problems associated to the innovation before committing to fully adopting it. These possibilities increase the net benefits
offered by innovations to a decision maker. Thus, greater levels of trialability are expected to influence positively the adoption
decision (e.g., Ramdani, Kawalek, & Lorenzo, 2009). From the discussion, it follows that:
Hypothesis 1. Greater levels of relative advantage, observability, observability, and trialability (complexity) are likely to have
positive (negative) effects on innovation adoption decisions in organizations.
2.2. Mediators of the relationship between attributes of innovations and managerial decisions about the adoption of innovations in
organizations
If attributes of an innovation have an effect on the decision to adopt a new technology, this effect should be mediated by
mechanisms that affect this decision. First, adoption decisions in organizations require an intervention of a decision maker.
Second, as noted by March and Simon (1958), a manager brings her own preferences to the decision making context and these
preferences are likely to exert a meaningful influence on organizational behaviors (Staw, 1991), and hence on innovation
adoption behavior. As observed by Rogers (1983), not only the assessment of the attributes of innovations but also the
preferences of decision makers are key for an adoption decision to occur (see also Baldridge & Burnham, 1975; Hage & Dewar,
1973; Kimberly & Evanisko, 1981; Miller, Manfred, & Toulouse, 1982).
In representing the behavioral preferences of managers (i.e., attitude, norm, and perceived behavioral control), we referred
to the theory of planned behavior (Ajzen, 1991). This behavioral theory is derived from social psychology and represents an
extension of the expectancy theory (Vroom, 1864) that has been generally employed in understanding different behaviors
(Armitage & Conner, 2001; Sutton, 1998). Extant studies have employed the considered theory to explain important orga-
nizational decisions (e.g., Cordano & Frieze, 2000), and, more specifically, the adoption of innovations in organizations (e.g.,
Harrison, Mykytyn, & Riemenschneider, 1997; Marcati, Guido, & Peluso, 2008; Riemenschneider et al., 2003). Both the
innovation diffusion theory and the theory of planned behavior focus on decision makers, yet on two distinct but comple-
mentary aspects: the assessment of the attributes of innovations on the one hand, and the behavioral preferences of decision
makers on the other hand. By jointly considering both these aspects in our framework, we expect to offer further explanatory
power on the decision to adopt innovations in organizations.
In our framework, the assessment of attributes of innovations are likely to influence organization decision makers'
behavioral preferences and these preferences create, in turn, opportunities and pressures for adopting innovations. Specif-
ically, we integrated the innovation diffusion theory (Rogers, 1983) and the theory of planned behavior in its decomposed
form (Oliver & Bearden, 1985; Taylor & Todd, 1995a) and considered the relationship between the attributes of innovations-
behavioral preferences of managers and adoption decision. It must be noted that the theory of planned behavior in its
decomposed form considers all the attributes of an innovation but observability and trialability (see also Davis, 1989; Davis,
Bagozzi, & Warshaw, 1989). Moreover, by relating the adoption decision to both attributes of innovations and preferences of
managers, we introduced a behavioral element. According to the latter, an innovation can be adopted or rejected in orga-
nizations not only due to its expected net benefits but, for example, because of pressures stemming from the environment or
because of perceived lack of control over an innovation which are necessary to adopt the new technology.
The first potential mediating variable is attitude, which refers to individual's positive or negative evaluation of an action to
be taken (Ajzen, 1987; Fishbein & Ajzen, 1975). Concerning the relationship between the assessments of the attributes of
innovations (i.e., relative advantage, compatibility, and complexity) and decision-makers’ attitudes, the considered attributes
are similar to the attitudinal beliefsdwhich typically identify the advantages and disadvantages associated with performing a
behavior (Shimp & Kavas, 1984)dused in studies based on the theory of planned behavior, where they are considered as
antecedents of the attitudes (Ajzen, 1991). Specifically, attitudes are formed by associating a behavior to a certain outcome.
Consistently with the expectancy-value model, the attributes that come to be linked to the behavior can be then valued
positively or negatively. Because of these evaluations, decision makers tend to develop a favorable or unfavorable attitude
toward the behavior (Ajzen, 1987; Fishbein & Ajzen, 1975). Thus, some attributes are more likely to favor the development of a
favorable attitude toward innovation adoption than others are. In the first case, can be included innovations characterized by
such desirable consequences as increases in customer satisfaction, reduction in costs, enhancement of the relationships with
customers and suppliers, and various economies associated to their compatibility with organization's existing technology and
to their ease of use. In the second case can instead be encompassed those innovations characterized by more limited (greater)
levels of relative advantage and compatibility (complexity) (Taylor & Todd, 1995a). Concerning the implications of attitudes
on the adoption decision, in the theories of reasoned actions and of propositional control, if a decision maker holds the belief
that a given behavior will lead to a certain event and such event is positive, the decision maker's adoption of the behavior
should increase (Dulany, 1961; Fishbein, 1967). Finally, in the theory of planned behavior it has been considered that decision
makers tend to adopt behaviors toward which they have positive attitudes. At the same time, other different variables (e.g.,
attributes of innovations) are likely to influence the choices of decision-makers only via their effects on their attitude (Ajzen,
1987, 1991; Rogers, 1983). It then follows that the more favorable the attitude of a decision maker toward a new technology,
the stronger should be the tendency of the decision maker to adopt the innovation. Moreover, when attitude is accounted for,
the strength of the attributes of innovations-adoption decisions relationships in organization are likely to be weakened The
discussion leads to the following hypothesis.
Hypothesis 2. Attitude is positively related to relative advantage and compatibility and negatively related to complexity and
it mediates the relative advantage-, compatibility-, and complexity-innovation adoption decisions relationships in
organizations.
The second mediator is subjective norms (or simply norms) which reflect one's agreement (or disagreement) with
statements that refer to the preferences of others (e.g., like owners, competitors, suppliers, customers) about performing the
decision under examination (Ajzen, 1987; Fishbein & Ajzen, 1975). Concerning the relationship between attributes of in-
novations (i.e., relative advantage, compatibility, and complexity) and norms, Ryan (1982) offered a premise based on
Fishbein and Ajzen (1975) that the perception of the attitudinal beliefs that stem from assessments of attributes of in-
novations may affect not only attitudes but other types of behavioral preferences (like norms) through secondary mecha-
nisms. Shimp and Kavas (1984) elaborated on these mechanisms by introducing the notion of false consensus (Oliver &
Bearden, 1985; Ross, 1977; Taylor & Todd, 1995a). Developed within studies of behaviors of individuals, the mechanism of
“false consensus” was observed in all social environments in which some behaviors are important to recognize but often
difficult to determine (Cross & Brodt, 2001; Flynn & Wiltermuth, 2010; Gross & Miller, 1997). Specifically, the false consensus
mechanism involves that decision makers tend to overestimate the importance of personal and dispositional factors relative
to external pressures and by doing so they tend to infer broad personal disposition and expect consistency in behavior or
outcomes across widely disparate situations and contexts. A decision maker's strong held perception about the relative
advantage, the compatibility, or the complexity of a given innovation may then consider that these perceptions are common
and widely legitimate, and that others would not only agree with these beliefs, but would state them in prescriptive manner.
In this vein, false consensus resembles to an “anchoring and adjustment” process, whereby decision makers anchor on their
own preferences and inadequately adjust for ways in which they tend to vary from others (Davis, Hoch, & Ragsdale, 1986).
Hence, come the implications of subjective norms for the adoption decision. Even if the decision maker has a tendency to
produce a distorted interpretation of the social environment by overestimating the support of others to its own perceptions of
an innovation's relative advantages, compatibility, and complexity (Taylor & Todd, 1995a), these erroneous norms are likely to
influence people's decisions via compliance mechanisms. Moreover, in the theory of planned behavior, norms capture all the
pressures exerted by others on a decision-maker decision to adopt a behavior (Ajzen, 1991; Fishbein & Ajzen, 1975). In this
vein, in cases in which the decision maker believes that other people support the adoption of an innovation (e.g. a new
technology) she may be sufficiently motivated to comply with this idea, which, in turn, is likely to influence her decision to
adopt the innovation. Consequently, norms are likely to influence a decision-maker decision to adopt an innovation and, once
this behavioral preference is considered, the strength of the association between the attributes of innovations and adoption
decisions is likely to be weakened. From the discussion, it follows that:
Hypothesis 3. Norm is positively (negatively) related to relative advantage and compatibility (complexity) and it mediates
the relative advantage-, compatibility-, and complexity-innovation adoption decisions relationships in organizations.
The third mediator considered in this paper is perceived behavioral control which reflects beliefs regarding access to the
resources and opportunities needed to perform a behavior (Ajzen & Driver, 1992; Ajzen, 1991). Concerning the relationship
between attributes of innovations and perceived behavioral control, it has been observed that the greater the amount of
resources and opportunities a decision maker perceives to possess, and the fewer obstacles and impediments she anticipates,
the greater should be her perceived control over the behavior (Ajzen, 1991). The adoption of innovations that are highly
complex requires the decision maker to overcome different obstacles and invest greater levels of resources for its under-
standing (Triandis, 1979) and, more relevantly, to sustain the processes that are required to use and assimilate the new
technology within the organization once adopted (Attewell, 1992). Moreover, complex technologies are often composed of
many interacting parts that need to be configured with high reliability, and even small variations in performance can induce
high risks of failures (Perrow, 1994). These risky characteristics of complex innovations increase the possibility of a decision
maker to perceive the innovation as subjectively threatening (Bandura, 1977). Accordingly, greater levels of complexity of an
innovation may induce the perception of the inadequacy of available resources as well as of more extended obstacles and
impediments associated to the adoption of the new technology which, in turn, reduces the perceived control upon the
innovation (Taylor & Todd, 1995a). In terms of the implications of the perceived behavioral control on the adoption decision,
greater levels of perceived control on innovations are associated to greater levels of confidence that the decision maker will
bring the adoption to a successful conclusion. In addition, high perceived control on innovations reduces uncertainties and
risks associated to innovation adoption, which also favors the adoption decision. Moreover, a decision-maker decision to
adopt a behavior is influenced only by her perceived behavioral control about such behavior and its complexity is likely to
influence only her perceived behavioral control belief (Ajzen, 1985). Consequently, when perceived behavioral control is
considered, the complexity-adoption decision relationship is likely to be weakened. From the discussion, it follows that:
Hypothesis 4. Perceived behavioral control is negatively related to complexity and it mediates the complexity-innovation
adoption decisions relationship in organizations.
2.3. Potential moderators of the relationship between behavioral preferences of decision makers and the adoption of innovations in
organizations
The model we conceptualized herein can be referred to many different contexts in which an adoption decision may take
place (Huber, 1990; Mowday & Sutton, 1993). Therefore, we introduced a set of post hoc analyses that seek to explain the
moderating effect on the mediators-adoption linkages. These moderators are introduced in order to account for the contexts
in which the adoption decision takes place as well as the research settings employed by scholars in measuring constructs
under examination.
The organizational contexts of the adoption decision. The behavioral preferences of managers can find expression in an
adoption behavior only if the decision maker can actually influence the behavior, i.e., if the person can decide as whether to
perform or not perform the behavior (Ajzen, 1991). In organizations, it is well acknowledged that organizational decision
maker can influence the adoption behavior, and researchers recognize that such influence vary from large to small (Bass &
Stogdill, 1990). Consequently, when the influence of a decision maker on the organizational behaviors is large, innovation
adoption behavior can be expected to occur in the organization if the decision maker develops appropriate attitudes, norms,
and behavioral control beliefs on the innovation. By contrast, if such influence is small, the decision maker's behavioral
preferences may play a more limited role in shaping the final adoption decision. In particular, in the following of this study, we
consider two determinants that may affect the influence exerted by the decision maker on the organization: a) the hierar-
chical position of the decision maker and b) the size of an organization (Ettlie, 1983; Miller & Friesen, 1982).
The hierarchical position of the decision maker represents a key variable in studies of organizational behavior (Staw, 1991).
Because of their positions, top managers are less subject to rules and procedure and their decisions often entail a good degree
of discretion (Baldridge & Burnham, 1975). As one descends the organizational levels, rules, procedure, and supervision
processes take instead prominence, and thus the influence of any single individual on organizational behaviors is likely to
decrease (Staw, 1991). Consequently, top managers may take actions that reflect their own preferences more easily than other
decision makers characterized by less degrees of autonomy. In addition, top managers have direct and personal contacts with
most levels of management whereby they shape key organizational behaviors (Kets de Vries & Miller, 1984). Consequently, we
should expect the hierarchical position of decision makers to moderate the effects of behavioral preferences on innovation
adoption decision in organizations.
Hypothesis 5. The effect of attitude, norm, and perceived behavioral control on innovation adoption decisions in organi-
zations will be stronger when the decision maker is a top manager rather than when the decision maker is a lower-level
manager.
Organizational size represents one of the most important variables in micro-organizational studies (Kimberly, 1976). In
organizations of more limited size, individuals make most of the key decisions themselves (Chen & Hambrick, 1995), directly
influence other managers, and tightly control and channel operations (Mowday & Sutton, 1993). For example, Miller and
Droge (1986) observed that the individual decision maker preferences are likely to be a critical and perhaps tremendously
important element in small organizations, in which the impact of the leader can be very direct and pervasive (Premkumar,
2003; Unsworth, Sawang, Murray, & Sorbello, 2009; Welsh & White, 1981). In contrast, large organizations may have pro-
cedures, routines, style, and bureaucratic momentum of their own which can more easily absorb or resist leaders' initiatives.
In addition, extant studies in strategic management posit that larger firms tend to be associated with many and varying
stakeholders with somehow conflicting interests and preferences (Fombrun & Shanley, 1990) and are more likely to be under
regulatory and public scrutiny (Bloom & Kotler, 1975). Consequently, in organizations of larger size, individual behaviors are
less under the control of a single decision maker (Thompson, 1967). These arguments suggest that organizational size will
moderate the effects of behavioral preferences on innovation adoption decision in organizations.
zations will be stronger in small organizations than in larger ones.
The research settings of the adoption decision. In organizations the adoption decision involves different, contingent
choices (Eveland, 1979). Accordingly, the key issue is just which one of these choices is in fact the crucial onedi.e., the one
correctly identifying the moment at which the organization shifted from the category of non-adopter to the category of
adopter. In their research settings, scholars tended to first identify a decision (or a few) which they believe to be critical in
some sense, and used the occurrence of that decision as an indication of adoption. Second, they searched for documentary
evidence in organizations or specified critical decision makers and asked them about organization's adoption (Becker, 1970).
It is commonly understood that selecting one measure for the adoption decision instead of another as well as measuring the
adoption decision according to self-reported or objective data matter for empirical results. For example, data from decision
maker is easy to collect, but may be affected by the choice of people to supply it. To what extent, for example, does a top
manager know about what her organization is really doing with the innovation (Eveland, 1979)? We therefore consider in our
study additional moderators (i.e., measures of the adoption decision and presence of common method variance in self-
reported data) that have a methodological nature and capture some of the key elements of research settings employed by
scholars in their studies.
In extant empirical studies, the dependent variable has been measured either as a future expectation or as a current
intention or as a behavioral action. Yet, these different measures of the same dependent variable (i.e., adoption) matter
for empirical estimates. First, in assessing the behavioral preferences-adoption decisions linkage, literature has intro-
duced the distinction between intentions and expectations. Whereas intention involves making a behavioral commit-
ment to perform (or not perform) an action, expectation is one's estimated likelihood of performing the action, whether
or not a commitment has been made (Warshaw & Davis, 1985). In forming expectations, decision makers are likely to
consider not only their current behavioral preferences but also their future beliefs as well as beliefs toward alternative
actions or outcomes. Therefore, current behavioral preferences may play a less relevant role in predicting expectations
rather than intentions. In addition, expectations involve considerations of the potential impediments/barriers that may
prevent the behavior from being implemented (Warshaw & Davis, 1985) which may not be under the volitional control of
the decision maker (Saltzer, 1981), and therefore render the decision maker's current behavioral preferences less relevant
for the adoption decision (for an early review see Sheppard, Hartwick, & Warshaw, 1988). The above arguments suggest
the following hypothesis.
zations will be stronger when decision makers are asked to indicate their present intentions than when they are asked to
estimate their future expectations.
Moreover, literature has also introduced the distinction between the behavior (the adoption decision) and intention/
expectation toward the behavior (the intention to adopt). In the view of the theory of planned behavior it has been conceived
that attitudes and norm are associated to intention but not to adoption while perceived behavioral control has an effect on
both intention and adoption (Ajzen, 1985, 1991; Fishbein & Ajzen, 1975). Consequently, the measure of the adoption decision
moderates the attitude and norm-adoption linkages. The above arguments suggest the following hypothesis.
Hypothesis 8. The effect of attitude and norm on innovation adoption decisions in organizations will be stronger when
decision makers are asked to indicate their intentions than when they are asked to estimate the effective adoption of the
innovation.
The common method variance refers to cases in which variance is associated to the measurement method rather than to
constructs that measures represent (Campbell & Fiske, 1959). There are studies that suggested how the common method
variance bias affects estimates of researches based on self-reported data (Crampton & Wagner, 1994). For example, extant
literature provided evidence that attitudes and norms are strongly associated with self-reported everyday behaviors but did
not correlate with objective evidence of the same behavior (Armitage & Conner, 1999, 2001). The effect of the common
method variance bias on estimates is likely to vary not only across studies based on objective data but also within studies
based on self-report data. For example, in the context of the adoption of an information system, the considered bias, induced
by research settings employed by scholars, increases the magnitude of the association between dependent and independent
variables (Malhotra, Kim, & Patil, 2006). The above arguments suggest the following hypothesis.
zations will be stronger in studies that are highly affected by the common method bias than in studies that are affected by the
same bias to a more limited extent.
2.4. Empirical studies on the attributes of innovations-adoption decisions linkages in organizations
Empirical studies provided evidences on the relative advantage-, compatibility-, complexity, and observability-innovation
adoption decisions associations that are consistent with the theoretical predictions. However, despite these consistent pat-
terns, a few studies offered contrary evidences. Several “classes” of conflicting evidence can be identified. One concerns the
effects of perceived relative advantage on innovation adoption decisions. Indeed, some studies reported how greater levels of
perceived benefits about an innovation are likely to reduce the propensity of organizations to adopt such new technology
(Kurnia et al., 2015). Another class addresses the effects of complexity. Although complexity discourages innovation adoption,
there are studies that provided evidences for a positive, significant association between complexity and innovation adoption
(Messerschmidt & Hinz, 2013; Seyal & Rahman, 2003). A further class of conflicting evidence deals with trialability. Studies on
the trialability-innovation adoption linkage are extremely mixed, with some works that offer evidences of positive, significant
associations (Ramdani, Chevers, & Williams, 2013) and others which demonstrate negative, significant associations
(Premkumar, Ramamurthy, & Crum, 1997).
Concerning the mediating variables, empirical studies instead offered consistent evidences on the attitude, norm, and
behavioral control-innovation adoption associations. There is just one study that reported a slightly negative association
between norm and innovation adoption in organizations (Hsu, Ray, & Li-Hsieh, 2014). Regarding the attributes of innovations-
mediating variables associations, empirical studies have shown consistent pattern on the relative advantage-, compatibility,
and complexity-attitude associations. Such consistent patterns are observable for relative advantage- and compatibility-norm
associations. However, although studies provided evidences of negative associations between complexity and norm, one
work offered empirical evidence of a positive, significant association (Tashkandi & Al-Jabri, 2015).
We must note that although extant empirical studies on the main effect of the attributes of innovations on innovation
adoption decisions in organizations abounds, we found only a few studies that tested the mediation effects of the preferences
of managers on the attributes of innovations-adoption decisions linkage in organizations (Harrison et al., 1997).
A number of previous meta-analyses and reviews on the innovation adoption have focused on many important issues,
including (a) factors affecting information technology usage (Mahmood, Hall, & Swanberg, 2001); (b) drivers of intention and
adoption in consumers' behavior (Arts, Frambach, & Bijmolt, 2011); (c) the determinants of information systems success
(DeLone & McLean, 2003; Sabherwal, Jeyaraj, & Chowa, 2006); (d) the performance implications of the adoption of in-
novations in organizations (Rosenbusch, Brinckmann, & Bausch, 2011); (e) the role of organizational characteristics and the
adoption of IT technologies in organizations (Hameed, Counsell, & Swift, 2012); (f) the technological, organizational and
individual factors that influence the adoption and diffusion of IT-based innovations by individuals and organizations (Jeyaraj
et al., 2006); (g) the effect of Rogers' attributes of innovations on the adoption decisions for both individuals and organizations
(Arts et al., 2011; Kapoor et al., 2014b; Tornatzky & Klein, 1982); (h) the determinant of the acceptance of technologies by
individuals (King & He, 2006; Schepers & Wetzels, 2007); (i) the applications of models of diffusion of innovations (Sultan,
Farley, & Lehmann, 1990); (j) the performance implications of the adoption of innovations (Rosenbusch et al., 2011).
One contribution of our paper therefore already lies in focusing on the adoption of innovations in organizations while
integrating the potential mediators/moderators of the attributes of innovations-adoption decisions relationship. To the best of
our knowledge, no previous meta-analysis has offered an integration of the potential mediators/moderators on the attributes
of innovations-adoption decisions relationship in organizations. One exception is the meta-analysis by Weigel et al. (2014)
where researchers combined diffusion of innovations and the theory of planned behavior to develop a model of innova-
tion adoption-behavior. Yet, in their analysis, researchers combined studies on the adoption of innovation by consumers as
well as by managers in organizations. Moreover, they offer only empirical evidences of the primary direct effects of ante-
cedents derived from the diffusion of innovations theory and from the theory of planned behavior on innovation adoption
decisions for individuals and for organizations.
3. Research method
3.1. Database development and inclusion criteria
In selecting the most important studies that explored the relationship between an adoption decision and its correlates, we
followed some general rules set out in the literature (e.g., see Chen, Damanpour, & Reilly, 2010). First, we searched relevant
bibliographic databases, namely ABI/INFORM Global, Business Source Premier, JStore, ProQuest, IEEExplore, Science Direct,
Scopus, and ISI Web of Science. In order to identify relevant journal articles, conference papers and dissertation works
published in English, we performed the search based on the following Boolean expression: (“organization”) AND (“adoption”)
AND (“innovation”), where “organization” includes, for example, such keywords as business, enterprise, company,
organization. The term “adoption” includes, for example, adoption, intention to adopt, adopters, non-adopters, whereas the
term “innovation” includes such terms as innovation, new technology, information system, cost accounting, payroll, enter-
prise resource planning, computer-assisted manufacturing, e-commerce, e-business, and customer relationship
management.
We integrated the first database search with thorough examination of primary journals in relevant research fields, such as
innovation management, strategic management, technology and R&D, organizational behavior, entrepreneurship and mar-
keting. In order to mitigate the publication bias (Begg, 1994), we used a bottom-up search approach and considered existing
meta-analyses and seminal papers that have dealt with relationships under study, and carefully examined their reference
lists. Additionally, we analyzed cross-references in the acquired reports, sent requests for working papers and forthcoming
publications. Specifically, we used two-way ‘snowballing’, backward-tracing all references reported in previously identified
articles and forward-tracing all articles that cited these articles using Google Scholar and ISI Web of Knowledge.
Following Damanpour (1991), each collected study was scrutinized and included into the meta-analysis if only it met the
following eligibility criteria. First, we focused on studies that deal with the adoption of innovations. Reports focusing spe-
cifically on either awareness or post-adoption aspects (e.g., “infusion,” or “satisfaction with the innovation,” or “impact of the
innovation of firms' performance”) were excluded. Although we acknowledge the importance of post-adoption aspects
(Angle & Van de Ven, 1989; Damanpour & Wischnevsky, 2006; Rogers, 1983; Zaltman et al., 1973), it has been observed that
mixing innovation stages increases the risk of finding inconsistent and contradictory research results because the direction of
the influence of some determinants on innovation is also contingent on the decision being considered (Wolfe, 1994). In
particular, adoption and post-adoption stages imply different conceptualization of the decision to be taken in organizations
(for an early review of post-adoption studies see DeLone & McLean, 1992, 2003). Because of the conceptual differences in the
dependent variable between adoption and post-adoption studies, the nature of the considered correlates between pairs of
variables differs, too. And, even if one can observe a partial overlapping between the attributes of an innovation emphasized
in adoption and post-adoption studies (e.g., see Larsen, 2003), these aspects may express dissimilar underlying conceptu-
alizations. Consequently, aggregating under the same label variables that express very different constructs may induce an
excess of heterogeneity into our study and at the same time may render the interpretation of results more difficult.
Second, we conducted our analysis at an organizational, rather than an individual or organizational population level. As
observed by Tornatzky and Klein (1982) research should study innovations that will be adopted by managers of organizations,
not by individuals operating alone since it is not logical to attempt to generalize from the individual adoption process to the
organizational innovation process as the two processes may in fact be quite different. Moreover, it must be noted that such
concepts and variables as organizational size and decision makers' hierarchical position do apply to organizations but not to
individuals.
Third, to be eligible, studies had to include at least one of the attributes of innovations that is expected to correlate with the
decision to adopt innovations. Fourth, we checked for the presence in the study of information about the adopters, and
excluded reports that presented no sourced information whatsoever about the sample of studied organizations. Since the
concepts of attitude, norm and perceived behavioral control have been referred to individual decision makers and their
extension to teams or group decision-making has not been developed, we controlled that studies measured the behavioral
preferences of managers with respect to an individual decision maker. In case of multiple respondents, we controlled, for each
sample organizations, that these preferences were derived individually for each member belonging to an organization
(Harrison et al., 1997). Fifth, in order to be included in our analysis, we verified that each study was published in a scholarly
book, PhD thesis, or journal. Finally, we required a study to include appropriate statistics. In that, we only synthesized studies
that provided information to calculate the correlation between considered constructs of Table 1 (Hunter & Schmidt, 2004).
Table 1
Description of key constructs.
Key Construct Definitions Sample measures

Attributes of innovations
1. Relative advantage (RA) The degree to which an innovation is perceived as being better than Relative advantage (Gu, Cao, & Duan, 2012;
the idea it supersedes (Rogers, 1983, p. 213) Jarrett, 2003; Seyal & Rahman, 2003)
Perceived usefulness (Daryanto, Khan, Matlay,
& Chakrabarti, 2013; Gamal Aboelmaged, 2010;
Heyder, Theuvsen, & Hollmann-Hespos, 2012)
Perceived benefits (Ghobakhloo & Tang, 2013;
Kurnia et al., 2015; Quaddus & Hofmeyer, 2007)
2. Compatibility (CO) The degree to which an innovation is perceived as consistent with Compatibility (He, Duan, Fu, & Li, 2006;
the existing values, past experiences, and needs of potential Henderson, Sheetz, & Trinkle, 2012; Lin & Ho,
adopters (Rogers, 1983, p. 226). 2011)
Technological compatibility (Uzoka & Ndzinge,
2009; Weng & Lin, 2011)
3. Complexity (CX) The degree to which an innovation is perceived as relatively difficult Complexity (Gu et al., 2012; Hsing Wu et al.,
to understand and use (Rogers, 1983, p. 231). 2013; Hung, Hung, Tsai, & Jiang, 2010)
Ease of use (reverse coded) (Arpaci, 2013;
(continued on next page)
Table 1 (continued )
Key Construct Definitions Sample measures

Jarrett, 2003; Li, Troutt, Brandyberry, & Wang,
2011)
4. Observability (OB) The degree to which the results of an innovation are visible to others Observability (Hussin et al., 2008; Leung, 2005;
(Rogers, 1983, p. 232). Sophonthummapharn, 2009)
Demonstrability (Askarany, Brierley, & Yazdifar,
2012; Plouffe, Vandenbosch, & Hulland, 2001)
5.Trialability (TR) The degree to which an innovation may be experimented with on a Trialability (Hsu & Lin, 2015; Limthongchai &
limited basis (Rogers, 1983, p. 231). Speece, 2003; Ramdani et al., 2013)
Preferences of managers
6. Attitude (A) Refers to the degree to which an individual has a favorable or Attitude (Gamal Aboelmaged, 2010; Harrison
unfavorable evaluation or appraisal of the behavior under et al., 1997; Riemenschneider et al., 2003)
examination (Ajzen, 1991, p. 188). Top management attitude (Lin, Lin, Roan, & Yeh,
2012; Thompson, Sijie, & Kee-hung, 2009)
7. Norm (N) Reflects perceived social pressure to perform or not to perform the Subjective norm (Harrison et al., 1997;
behavior under examination (Ajzen, 1991, p. 188). Riemenschneider et al., 2003; Xu & Quaddus,
2012)
External pressures (Alam & Noor, 2009; Hossain
& Quaddus, 2015; Tung & Rieck, 2005)
8. Behavioral control (BC) Refers to the availability of requisite opportunities and resources Perceived behavioral control (Harrison et al.,
(e.g., time, money, skills, cooperation of others) to staff the behavior 1997; Hsu et al., 2014; Riemenschneider et al.,
under examination (Ajzen, 1991, p. 182). 2003)
Self-efficacy (Jackson, 2008; Lee & Larsen, 2009;
Segaar, Bolman, Willemsen, & de Vries, 2006)
Facilitating conditions (Leung, 2005;
Messerschmidt & Hinz, 2013)
Innovation adoption
9. Adoption (AD) It is a decision to make full use of an innovation as the best course of Adoption (Jeon et al., 2006; Leung, 2005; Thong,
action available (Rogers, 1983, p. 20) 1999)
Intention to adopt (Daryanto et al., 2013;
Martinez-Garcıa, Dorward, & Rehman, 2013;
Quaddus & Hofmeyer, 2007)
Expectation to adopt (Hussin et al., 2008; Lee &
Larsen, 2009; Marcati et al., 2008)
Although unintentional omissions might have occurred, after these comprehensive search steps we ended up identifying a
total of 986 correlations from 165 different manuscripts to yield a combined sample sized of 31,626. A list of the articles used
in our empirical meta-analysis is available as an electronic companion from the journal webpage.
3.2. Variables coded
In considering the literature on the decision to adopt innovations in organizations, we observed many constructs with
similar definitions that operate under diverse aliases and constructs with comparable names but diverse operationalization.
Thus, we used a single construct definition to code existing research. In doing so, we included only those aspects analyzed
extensively across multiple studies and which fit conceptually (see Table 1). In terms of the followed procedure, four experts
coded the variables considered in Table 1. Two experts evaluated all of the studies for all variables excluding the attributes of
innovations. A third expert coded the variables for the attributes of innovations and coded correlations. Following this, a
fourth expert re-evaluated a random sample of 20% of all of the manuscripts. The initial average inter-rater agreement for the
experts was 96%. Occasional disagreements were resolved by repeated discussion between the coders until reaching
consensus, according to a procedure originally suggested by (Bullock & Svyantek, 1985)) and already employed in previous
meta-analytic works (Bauer, Bodner, Erdogan, Truxillo, & Tucker, 2007; Chen et al., 2010; Damanpour, 1991).
Attributes of innovations. Extant studies observed that the relative advantage concept is akin to the concept of perceived
usefulness (Moore & Benbasat, 1991) and in studies that consider both their correlation coefficient is often very high.
Therefore, we included perceived usefulness under the relative advantage label. Moreover, some studies referred to the
variable of ease of use instead of the variable complexity (Moore & Benbasat, 1991). Because of the differences in the expected
sign of correlation coefficients between complexity and ease of use, on the one hand, and the decision to adopt, on the other
hand, we reported results for ease of use under the complexity label but recoded the sign of the associated correlation co-
efficients. In addition, it must be observed that some studies refer to the concept of demonstrability. Since the concept of
demonstrability is strongly related to the construct of observability (Moore & Benbasat, 1991), in this meta-analysis we
included under the label observability those reports that referred to demonstrability as an attribute that is correlated to the
decision to adopt a new technology.
Behavioral preferences of decision makers. We evaluated attitudes by referring to the evaluative propertydi.e., how
positively or negatively a decision maker feels toward the innovation to adopt. Measures of the evaluative property consider
how much favorable versus unfavorable, or supportive versus antagonistic decision makers are toward the innovation
(Ostrom, Bond, Knosnick, & Sedikides, 1994). Scholars sometimes derived a decision maker's attitude by weighting her belief
of the potential outcomes associated to the adoption of the innovation by its subjective evaluation of the effect of these
outcomes (e.g., see Marcati et al., 2008; Nasco, Toledo, & Mykytyn, 2008). Finally, there are some studies that measured
attitude according to the decision maker's outcome beliefs and omitted to correct such beliefs by the decision maker's
outcomes evaluation (Thong, 1999). In order to attenuate the possibility that the variable attitude may enter the mediation
analysis as a simply reflection of attributes of innovations (Fiedler, Schott, & Meiser, 2011), we excluded such operationali-
zation from the current meta-analysis. Following Fishbein and Ajzen (1975), norms are affected by a set of salient beliefs about
the normative prescriptions of specific actors, weighted by the incentive to fulfill with each of those actors. Norms are
typically measured by global, subjective measures as well as according to belief-based measures. In this meta-analysis we
combined both (e.g., see Riemenschneider & McKinney, 2001/2002). Finally, the perceived behavioral control construct
encompasses two components. The first component includes such aspects as the self-efficacy (Ajzen, 1991), that is an in-
dividual's self-confidence in his/her ability to perform a behavior (Bandura, 1977). The second component includes the
facilitating conditions and the resources needed to engage in a behavior (Triandis, 1979). This might include access to the
time, money and other specialized resources required to engage in a behavior (e.g., skills of employees, experience of the
employees, availability of complementary resources).
Adoption decision. The considered construct commonly included assessments on subjective base of whether the
decision maker has adopted or not adopted an innovation (Pontikakis, Lin, & Demirbas, 2006) or a set of correlated
technologies (e.g., Thong, 1999) or not. Other studies used as a dependent variable a measure of intentions which
assessed the intent or willingness to adopt an innovation in the future (e.g., Harrison et al., 1997). Other else studies
considered the adoption as a process and ask the decision maker about the stage of adoption at which the organization is
currently engaged (e.g., Oliveira, Thomas, & Espadanal, 2014). We also found few studies that employed objective
measures of adoption. As observed, we excluded studies dealing specifically with post-adoption aspects. We did not
consider also studies in which organizations had adopted the innovation and were considered it for further adoption (e.g.,
Lefebvre, Lefebvre, & Harvey, 1996). We also ruled out studies in which the innovation is related to the benefits offered to
the firm (e.g., Wu & Chuang, 2010), studies dealing with the decision maker's satisfaction about the innovation (e.g.,
Thong, 2001), and works that deal with the adoption decision, yet they measured the adoption decision in ways are
difficult to distinguish this decision from either the attributes of innovations or the managers' behavioral preferences
(e.g., Gupta et al., 2013).
Moderators. We labeled a study as predominantly small when the number of employees was less than the median
number of employees in the sampled organizations; conversely, we labeled the study as predominantly large. We also
analyzed the percentage of respondents represented by top managers (senior executives/directors or higher positions as
defined in Hambrick and Mason (1984) and we labeled a study as predominantly higher positions when the percentage of
top managers in the sample was greater than the median percentage of top managers in the sampled organizations.
Conversely, we labeled the study as predominantly lower positions. We considered the different measures used by
scholars for the dependent variable and distinguished between adoption, intention to adopt, and expectation to adopt an
innovation (see Table 1). As observed by Mitchell (1985), common method variance may influence estimates when
dependent and independent variables are measured with the same format. Accordingly, we distinguished studies that
measure the dependent and independent variables with the same/different format. Note that there are alternatives ways
to control for common method variance (i.e., measures based on self-report vs. Objective data, respondents for dependent
and independent variables are the same or different), although infeasible for this study because of constraints from
available data.
3.3. Univariate analyses
Following Chen et al. (2010) and Damanpour (1991), we based our meta-analysis procedure on three fundamental
steps: (1) the main effect testing; (2) the mediation analysis; (3) the moderator analysis. Concerning the effect size, we
relied on correlation coefficients. It must be observed that, in most cases, correlation coefficients were the primary
statistics in our meta-analysis, but a few coefficients derived from studies in which logistic regression models or
discriminant analysis were applied. Although no universally accepted method for handling such data in a meta-analysis is
reported, we retained these studies and used available formulas to calculate effect size estimates from findings presented
in a logistic regression and discriminant analysis formats (Fern & Monroe, 1996; Lau, Sigelman, Heldman, & Babbitt, 1999;
Lipsey & Wilson, 2001). We also tested the robustness of our results by excluding findings that originate from studies that
used either logistic regression or discriminant analysis and observed consistent results. Finally, when studies drawn from
the same sample of observations were published several times, we primarily considered works published on academic
journals, when available.
On the collected data, we calculated the mean correlation on the relationship between the decision to adopt
an innovation and its correlates across studies, weighted by sample size. Next, the variance among correlation (observed
variance), the variance due to sampling error and residual variance were determined. Following Hunter and Schmidt
(2004) we also calculated corrected means and residual variances. A 95% confidence interval around mean correlation
was computed (based on the mean correlation, the number of the correlations, and the variance due to sampling error)
to check if this interval included the zero. Similarly to previous meta-analyses, we made no adjustments for range re-
striction, which was hardly assessable in many of the selected studies on organizations' innovation adoption decision
(Camiso n-Zornoza, Lapiedra-Alcamí, Segarra-Cipre s, & Boronat-Navarro, 2004; Chen et al., 2010; Damanpour, 1991;
Gooding & Wagner, 1985). We also checked the salience of each association by considering the comparison with
the conventional medium effect size of the population according to 0.30 rule (Cohen & Cohen, 1983). We additionally
addressed the publication bias by computing, for each considered associations, the classical file drawer N (Rosenthal,
1979) and evaluated publication and eligibility biases with the normal quantile plot method (Wang & Bushman, 1999).
Concerning the mediation analysis, the magnitude of effects was evaluated according to multiple methods. First, we
computed a partial correlation (corrected for unreliability) for each attributes of innovations-adoption decisions re-
lationships after controlling for the considered mediators. We inferred the presence of a mediating effect when the
original correlation coefficient between attributes of innovations and the decision to adopt a new technology (direct
effect) was reduced once the influence of the potential mediator was controlled for (Blalock, 1961). In particular, at the
one hand, if the correlation coefficient associated to the direct effect is not significantly divergent from zero, then it
suggests that the mediator accounts completely for the relationships between attributes of innovations and the decision
to adopt. On the other hand, if the partial correlation is identical to the original correlation, then it indicates that there is
no mediation effect. If the original correlation is higher than the partial correlation and the significance test of the partial
correlation indicates that it is different from zero, then one can conclude that there is a partial intervening effect of the
mediator (Gajendran & Harrison, 2007). Second, following Sobel (1982) we calculated the total unstandardized indirect
effect (i.e., estimated by the product of correlation coefficients associated to attributes of innovations-mediators and
mediators-adoption decisions paths). The indirect effect is then divided by its estimated standard error, which
approximately follows a standard normal distribution. The estimated corrected standard errors were calculated according
to the multivariate delta method introduced by Bobko and Rieck (1980). The suggested method has been employed in
different studies (Cheung, 2009) and it has the favorable characteristic that allows one to obtain the standard errors
directly from correlation coefficients. Finally, following Viswesvaran and Ones (1995), we used as sample size the har-
monic mean of the sample sizes for each of the involved meta-analytic correlations.
We introduced multiple tests to evaluate the need for moderation analysis. The first test refers to the so-called “75%
rule” (Gooding & Wagner, 1985) according to which if the variance caused by the sampling error does not account for 75
percent of the observed variance, then we cannot accept the hypothesis of homogeneity among empirical correlations
and had to look for other moderating variables that affect the correlations. The second test refers to the Chi-square
homogeneity test (Hunter & Schmidt, 2004). Specifically, a non-significant Chi-square statistic suggested there was no
true variation across studies. When moderation analysis is appropriate, we followed Hunter and Schmidt (2004) sub-
group analysis. A moderator could produce its effects on estimates in two ways: (1) the expected variance will be lower in
the subgroup than in the entire population, and (2) the corrected mean correlations will differ from one group to another.
For capturing the two effects we conducted a t-approximation test of the difference between the mean correlations of the
associations across different subgroups (Wagner & Gooding, 1987, p. 249). Note that the focus of moderating analysis is
on the differences between different groups of organizations and studies. In the context of our study, the relative
difference between the characteristics of organizations/studies is more relevant than the absolute characteristics of
each organization/study considered. For example, in considering organization size, the categories of small and large
relates to a relative concept rather than to an absolute one. In other words, organizations that differ greatly in size will
reveal differences in adoption behavior conceivably not of the extent of differences between absolutely large and small
organizations, but significant variations however. Finally, some recent studies suggested for robustness to apply two
techniques for moderation analysis. In that, we employed a meta-regression procedure with random effects (Lipsey &
Wilson, 2001).
3.4. Structural equation modeling
In addition to the pairwise analyses, we aggregated sampled studies to test the causal model implicit in Fig. 1. This
multivariate technique has the benefits of considering all linkages simultaneously and has been recommended by a number of
scholars (Colquitt, LePine, & Noe, 2000; Shadish, 1996; Viswesvaran & Ones, 1995).
We followed a procedure based on two stages. In the first stage, the application of this technique needed to comply to
the following prerequisites: (i) primary study must report either the covariance matrix or the variances of the variables
included in their analysis and (ii) the effects (i.e., correlation coefficients) must be available between each construct in the
model and all other constructs (Brown & Peterson, 1993). In analyzing data, we followed two procedures. Firstly, we
meta-analyzed the covariance matrix, then converting the results to structural relationships. Secondly, we converted
covariance matrix for each individual study in path coefficients, then meta-analyzing the obtained results (King &
He, 2006). Since the two procedures tended to produce only trivial differences, we reported only results
obtained from the first procedure. For the analysis, we needed to evaluate an overall sample size and employed the
harmonic mean to measure the matrix sample size (Colquitt et al., 2000; Viswesvaran & Ones, 1995). Sample sizes are
reported in Table 3.
In the second stage, the intercorrelation matrix was then analyzed in a random effect model using maximum likelihood
estimator available in the statistical package STATA 13 (see also Montazemi & Qahri-Saremi, 2015). We estimated the model
Table 2
Average Reliability-Adjusted Intercorrelations Among Constructs in meta-analytic model.
1 2 3 4 5 6 7 8 9
1. Relative advantage [0.87]
SD of correlations e
N. of correlations e
Sample size e
2. Compatibility 0.47* [0.86]
SD of correlations 0.04 e
N. of correlations 32 e
Sample size 6034 e
3. Complexity -0.37* -0.35* [0.85]
SD of correlations 0.05 0.08 e
N. of correlations 51 26 e
Sample size 10,187 4914 e
4. Observability 0.39* 0.41* -0.31* [0.80]
SD of correlations 0.02 0.03 0.04 e
N. of correlations 12 10 10 e
Sample size 2297 1958 1958 e
5. Trialability 0.29* 0.30* -0.21* 0.42* [0.85]
SD of correlations 0.01 0.03 0.04 0.02 e
N. of correlations 11 10 11 9 e
Sample size 2290 2157 2290 1719 e
6. Attitude 0.50* 0.45* -0.38* 0.39*a 0.27* [0.89]
SD of correlations 0.04 0.07 0.04 0.01 0.01 e
N. of correlations 20 6 12 1 3 e
Sample size 3955 1433 2503 95 666 e
7. Norms 0.41* 0.30* -0.26* 0.42* 0.22*a 0.56* [0.83]
SD of correlations 0.04 0.02 0.04 0.01 0.01 0.02 e
N. of correlations 47 12 19 4 3 13 e
Sample size 7991 1920 3467 793 453 2605 e
8. Behavioral control 0.35* 0.36* -0.39* 0.36* 0.18*a 0.42* 0.37* [0.83]
SD of correlations 0.03 0.03 0.03 0.03 0.01 0.04 0.03 e
N. of correlations 40 15 20 2 3 9 29 e
Sample size 7300 3051 3898 313 751 1550 4670 e
9. Adoption 0.39* 0.36* -0.33* 0.30* 0.11* 0.54* 0.38* 0.40* [0.89]
SD of correlations 0.04 0.03 0.04 0.02 0.02 0.03 0.04 0.03 e
N. of correlations 153 68 90 22 20 33 86 74 e
Sample size 30,414 14,301 19,685 4605 4328 6559 15,819 14,273 e
Notes: Entries on the diagonal in brackets are weighted mean Cronbach's alpha coefficients.
*p < 0.05.
a
Means that the coefficient is cross-situationally consistent according to the Chi-squared homogeneity test.
fit by measuring the root mean square error of approximation (RMSEA), the comparative fit index (CFI), and the standard chi-
square statistic (Bentler, 1990). Note that given the constraints posed by data, we limited the use of the structural equation
model to the analysis of main and mediation effects.
4. Results
In this section, we provide the results of our analyses structured around our focal research questions. Specifically, we
describe the sample of considered studies and provide evidence of the correlations between the considered constructs. Then,
we introduce mediation and moderation analyses and we propose additional analyses to test the robustness of our results.
4.1. Description of the sampled studies and correlations between constructs
Concerning the characteristics of sampled studies, retrieved manuscripts were published between 1995 and 2015. Most of
them were gathered around 2009. In terms of continents, 46% were collected in Asia, 28% in North America, 10% in Europe, 6%
in Africa and 2% in Australia. In terms of firms' characteristics, about 44% were operating in non-service industries. In terms of
technologies, about 67% of studies considered such technical innovations as CAD, CAM systems, flexible manufacturing
systems, ERP, EDI, e-commerce and only 5% considered multiple technologies for adoption. Finally, in terms of adoption,
almost 31% of the sample organizations already adopted an innovation at the time of the study.
The average number of the sample firms is 197 and the standard deviation is 143. About 86% of respondents were top
managers. Concerning the characteristics of organizations, their median number of employees was 100. Concerning the
research method employed by scholars in their works, the average response rate is 31%. Moreover, of the reports, 25% used a
random sample, 28% used personal interview for collecting data, and 27% addressed the potential differences between re-
spondents with non-respondents. For what concerns the measurement procedures, the cross-sectional design was the
method of choice.
Moreover, in measuring the adoption decision, 41% (24%) of studies measured the adoption decision with a single item
(with more than 3 items). We also observed that 85% of studies reported Cronbach's alpha and 73% referred to past studies in
building their measures of studies 57% adapted measures of constructs. Content, convergent and discriminant validities were
reported in 57%, 77%, and 72% of available reports, respectively. Dependent and independent variables were measured with
different scale in 45% of reports. Finally, the adoption of cross-validation or holdout samples as well as tests for endogeneity
between dependent and independent variables were nearly non-existent.
Given the available pairwise correlation coefficients, we then determined the average-adjusted intercorrelation among all
constructs in our proposed framework. Results are presented in Table 2.
From results reported in Table 2, we observed that although the associations between variables concerning both the
attributes of innovations and decision makers' behavioral preferences are generally high, no correlation is above the rec-
ommended level of 0.65 (Tabachnick & Fidell, 1996) and, thus, problems induced by a lack of discriminant validity are not
likely to bias our data. In addition, we calculated the variance inflation factor (VIF) for multicollinearity (Montazemi &
Qahri-Saremi, 2015), and observed that for all antecedent pairs included in our analysis the VIF ranges from 1.03 to 1.73.
Since VIFs are less than the level of 1.87, the assumption of independence of constructs is not violated by the data (Larsson,
1993).
Moreover, the correlation matrix gives some first indications about the relationships between the constructs: the
reliability-corrected relationships between the attributes of innovations and the adoption decision range between 0.39 and
0.33 while those between the behavioral preferences of managers and the adoption decision range between 0.54 and 0.38.
4.2. Which attributes of innovations are most effective for triggering managerial decisions about the adoption of innovations in
organizations?
Table 3 shows the results for meta-analysis on the attributes of innovations-innovation adoption decision linkages.
Given the results reported in Table 3, we noted that the attributes of innovations tend to influence the adoption with
a sufficient consistency and directionality across different studies, organizations, contexts and research settings. More-
over, results appear to be robust with regard to the number of null studies needed to render the observed effects zero
(mean file-drawer N is 51,443). Consequently, the attributes of innovation matter for the adoption decision in
organizations.
Our results further suggested that the attributes of innovation influence the adoption decision with different signs and
salience. In terms of signs, relative advantage, compatibility, observability, and trialability tend to favor the adoption
decision while complexity hamper such decision. According to the 0.30 rule, attributes of innovations that turn out to
be salient determinants of organizations' innovation adoption decision are relative advantage, compatibility,
and complexity (see also Tornatzky & Klein, 1982). We further calculated the combined direct effect of the salient at-
tributes of innovations on the adoption decision as estimated by structural equation model. The path coefficient is
positive, greater than the 0.30 rule, and highly significant. This result suggests synergic, combined effects between the
three salient attributes of an innovation in triggering an adoption decision. Overall, results provide evidence that confirm
Hypothesis 1.
4.3. Which decision makers' behavioral preferences are mostly affected by variations in attributes of innovations?
Table 3 shows also the results for meta-analysis on the attributes of innovations-decision makers' behavioral preferences
linkages. The average of the absolute sample-weighted reliability-adjusted correlations among attributes of innovations and
mediators is 0.39. All paths from antecedents to mediators are supported in both the pairwise analyses and the structural
equation model. All of the findings appear to be robust with regard to the number of null studies needed to render the
observed effects zero (mean file-drawer N is 4333).
Different attributes of innovations have also differential influences on decision maker-level variables. Relative advantage
has the largest positive impact on the mediators of all antecedents, in support of the relevance of this attribute in forming a
favorable attitude in the decision makers. The significant impact of attributes of innovations on norm indicates that the
mechanism of false consensus in decision makers matter. Yet, the explained variance for the considered associations suggests
also that the mechanism of false consensus does not monopolize the process of norms formation. Complexity has instead a
negative impact on all the behavioral preferences of decision makers, suggesting that greater levels of complexity have a
negative influence on attitude, norm and perceived control perceptions of the innovation to be adopted. We also evaluated
the combined effects of the attributes of innovations on the considered mediators. The path coefficient is positive, greater
than the 0.30 rule, and highly significant and indicates joint effects of the attributes of innovations on decision maker's
behavioral preferences.
4.4. How do variations in decision makers' behavioral preferences account for variations in innovation adoption decisions in
organizations?
We now consider the back-half part of the model, the mediators-adoption decisions relationships. From Table 3 the
average sample-weighted reliability-adjusted correlation among mediators and adoption decisions is equal to 0.44. All paths
Table 3
Descriptive statistics and influences of Attributes of innovation and of decision maker-level mediators on adoption decisions.a
Number of Total sample Mean correlation Weighted mean Corrected Observed variance T-value (d.f.) Explained 95% credibi- File Drawer X2 homogeneity
raw effects size coefficients correlation mean correlation [z-values] variance lity interval N (Using test (d.f.)
G. Vagnani, L. Volpe / International Journal of Innovation Studies 1 (2017) 107e133

coefficients [Beta] correlation coefficients [S.E.] (in %) Two-Tailed
coefficients Test)
Attributes of innovations→Adoption
RA/AD 153 30,414 0.34* 0.34* 0.39* 0.04 22.59 (152) 0.16 0.42 184,144 1205.23 (152)
CO/AD 68 14,301 0.31* 0.31* 0.36* 0.03 14.07 (67) 0.16 0.42 29,941 524.23 (67)
CX/AD 90 19,685 -0.26* -0.28* -0.33* 0.04 14.24 (89) 0.15 -0.43 40,600 757.31 (89)
OB/AD 22 4605 0.25* 0.25* 0.30* 0.02 7.87 (21) 0.22 0.33 2009 110.33 (21)
TR/AD 20 4328 0.14* 0.10* 0.11* 0.02 3.25 (19) 0.27 0.23 518 74.90 (19)
Attributesb/Adoption e 1365 e [0.92*] e [0.03] [28.75] e e e e
Attributes of innovations→Mediators
RA/A 20 3955 0.47* 0.44* 0.50* 0.04 9.41 (19) 0.15 0.49 6292 215.84 (19)
CO/A 6 1433 0.42* 0.38* 0.45* 0.07 3.48 (5) 0.07 0.71 531 124.48 (5)
CX/A 12 2503 -0.35* -0.34* -0.38* 0.04 5.70 (11) 0.12 0.46 1269 120.61 (11)
RA/N 47 7991 0.36* 0.35* 0.41* 0.04 12.11 (46) 0.16 0.47 17,638 363.58 (46)
CO/N 12 1920 0.24* 0.25* 0.30* 0.02 5.54 (11) 0.25 0.33 439 51.17 (11)
CX/N 19 3467 -0.18* -0.22* -0.26* 0.04 4.66 (18) 0.13 0.48 1099 156.26 (18)
CX/BC 20 3898 -0.32* -0.34* -0.39* 0.03 7.86 (19) 0.16 -0.46 3065 158.70 (19)
Attributesb/Mediators e 1365 e [0.88*] e [0.02] [29.84] e e e e
Mediators→Adoption
A/AD 33 6559 0.49* 0.48* 0.54* 0.03 15.12 (32) 0.19 0.40 19,183 286.28 (32)
N/AD 86 15,819 0.34* 0.32* 0.38* 0.04 14.07 (85) 0.14 0.51 46,896 775.23 (85)
BC/A 74 14,273 0.34* 0.35* 0.40* 0.03 17.05 (73) 0.19 0.39 40,930 488.10 (73)
Mediators/Adoption – 1365 e [0.94*] e [0.02] [34.44] e e e e
Note: harmonic mean of total sample size, and standardized beta coefficients are reported for structural equation model. All Ns greatly exceed the critical N calculated as Ncrit ¼ 5xN þ 10 (Rosenthal, 1991). The
model associated to the Attributes-Adoption linkages has the following indexes c2(4) ¼ 16.49, RMSEA ¼ 0.04, p < 0.01, CFI ¼ 0.99. The model associated to the Attributes-Mediator linkages has the following
indexes c2(8) ¼ 186.86 RMSEA ¼ 0.12, p < 0.01, CFI ¼ 0.89. The model associated to the Mediators-Adoption linkages has the following indexes: c2(4) ¼ 38.96, RMSEA ¼ 0.08, p < 0.01, CFI ¼ 0.98. RA is relative
advantage, CO is compatibility, CX is complexity, OB is observability, TR is trialability, A is attitude, N is norm, BC is behavioral control, AD is adoption.
*p < 0.05.
a
Means that the correlation coefficient is cross-situationally consistent according to the Chi-squared homogeneity test.
b
Only salient variables (i.e., relative advantage, compatibility, and complexity) are included under the label Attributes.
121
Table 4
Relationships between Attributes of innovation and adoption decision after controlling for behavioral preferences of decision makers.
Attitude Norm Behavioral control
Corrected mean partial Total unstandardized Corrected mean partial Total unstandardized Corrected mean Total unstandardized
correlation coefficients mediation effect (SE) correlation coefficients mediation effect (SE) partial correlation mediation effect (SE)
coefficients
RA/AD 0.16*(6846) 0.23*(0.01) 0.29*(13,560) 0.11*(0.01) e e
CO/AD 0.15*(3260) 0.22*(0.01) 0.27*(4587) 0.09*(0.01) e e
CX/AD -0.14*(4977) -0.19*(0.01) -0.25*(7454) -0.08*(0.01) -0.19*(8706) -0.13*(0.01)
Note. Note that the significance of partial correlations was evaluated against the harmonic mean of sample sizes (in parentheses). RA is relative advantage,
CO is compatibility, CX is complexity, AD is adoption.
*p < 0.05.
from mediators to adoption decisions are supported in the pairwise analyses as well as in structural equation model. All of the
findings appear to be robust with regard to the number of null studies needed to render the observed effects zero (mean file-
drawer N is 35,669).
In general, decision makers' behavioral preferences have a wide range of effectiveness for the adoption of innovation in
organizations. The combined effect on the adoption decision, as captured by structural equation model, is positive, well
greater than the 0.30 rule and highly significant. This result suggests that decision makers' behavioral preferences may
produce synergistic effects. The chances of an innovation to be adopted tend to increase when the new technology is able to
simultaneously stimulate favorable attitude, norm and perceived behavioral control in the decision maker. Concerning the
different mediators, they tend to produce varying influences on the adoption decision. Attitude has the greatest influence on
the adoption decision. Norm and behavioral control have a more attenuated although highly significant effects. This implies
that stimulating a favorable attitude in a decision maker toward the innovation represent a key determinant for its adoption
in organizations. We further evaluated the strengths of the effects in the attributes of innovations-adoption decision linkages
that are likely to be mediated by decision makers' behavioral preferences. Results for the hypothesized relationships are
reported in Table 4.
In particular, the analysis provides evidence that if one controls for attitude, norm, and perceived behavioral control in the
attributes of innovations-adoption decisions relationships, the resulting corrected partial correlation coefficients will be
lower than the un-mediated correlations between the considered variables. Moreover, we evaluate the strength and sig-
nificance of the mediations effects by test developed by Sobel (1982) and observed that the total standardized mediation
effects for the considered variables are significant (see Table 4). The pairwise analysis then offers evidence that support
Hypotheses 2, 3, And 4. Note that the magnitude of corrected partial correlation coefficients is lower than the magnitude of
coefficients associated to unmediated effects, yet their significance level were still high for all the considered associations (see
Table 4). These results provide evidence of a partial intervening effect of the considered mediators on the attributes of
innovations-adoption decisions chain in organizations.
As robustness, we referred again to the structural equation model. Specifically, we run a first structural equation
model that included decision maker's behavioral preferences as mediators of the attributes of innovations-adoption
decisions association in organizations (model 1). We also run a second model that included the direct and the media-
tion effects between attributes of innovations and adoption decision (model 2). The fit statistics of Model 1 are c2
(8) ¼ 433.49, RSMEA ¼ 0.20, p < 0.01, CFI ¼ 0.77 and those of model 2 are c2 (5) ¼ 381.42, RSMEA ¼ 0.24, p < 0.01;
CFI ¼ 0.75. These models allowed us to calculate direct, indirect, and total effects (Brown, 1997). The direct effects are the
effects of the attributes of innovations on the adoption decision unmediated by other variables. The indirect effects were
computed as the product of the paths from the attributes of innovation to the behavioral preferences of manger and from
the behavioral preferences of managers to the adoption decision. These effects for relative advantage (Beta ¼ 0.13,
p < 0.05), compatibility (Beta ¼ 0.20, p < 0.05) and complexity (Beta ¼ 0.09, p < 0.05) were all significant, further
supporting hypotheses 2, 3, and 4. We also calculated the direct effects of attributes of innovation on the adoption
decision controlling for decision makers behavioral preferences, for relative advantage (Beta ¼ 0.13, p < 0.05),
compatibility (Beta ¼ 0.05, p < 0.05), and complexity (Beta ¼ 0.09, p < 0.05). We observed that these coefficients are
much lower than the corresponded unmediated coefficients for relative advantage (Beta ¼ 0.21, p < 0.05), compatibility
(Beta ¼ 0.17, p < 0.05), and complexity (Beta ¼ 0.15, p < 0.05). On the one hand, these results confirm the mediating role
of decision makers' behavioral preferences on the attributes of innovations-adoption decisions linkages in organizations.
On the other hand, the mediated coefficients for the considered attributes of innovations are still significant. This result is
consistent with the analysis of partial correlation coefficients and suggests that the considered mediators have in our
model a partial intervening effect.
4.5. How do variations in contexts of adoption and research settings influence the linkages between behavioral preferences of
decision makers and innovation adoption decisions in organizations?
In introducing the moderation analysis, we observed that the explained variance of associations between mediators and
the adoption decision is well below the threshold of 75% and the homogeneity test is always significant for all the considered
Table 5a
Influence of organization contexts on decision Maker's behavioral preferences-adoption decision linkage.
Hierarchical position Organizational size
Higher Lower t-test Beta (t-value) Larger Smaller t-test Beta (t-value)
Mediators/Adoption decision
A/AD 0.57 (15) 0.50 (9)a 0.87 0.32 0.62 (10) 0.58 (14) 0.36 0.09
3205 1863 (1.53) 1695 2737 (0.74)
N/AD 0.46 (28) 0.40 (32) 0.52 0.01 0.43 (36) 0.38 (28) 0.74 0.02
4372 5444 (0.04) 5721 5608 (0.50)
BC/AD 0.43 (34) 0.30 (18) 2.66* 0.28* 0.41 (25) 0.40 (26) 0.42 0.01
6053 4038 (2.69) 5460 4891 (0.01)
Note: Number of correlations is reported in parentheses. Sample size is reported under correlation coefficients. Differences in sample sizes depend on
availability of data for moderators. A is attitude, N is norm, BC is behavioral control, AD is adoption.
*p < 0.05.
a
Table 5b
Influence of research settings on decision Maker's behavioral preferences-adoption decision linkage.
Adoption measures Adopion measures Measures of dependent/Indipendent
Intention Expectation t-test Beta Intention/ Adoption t-test Beta (t-value) Same Different t-test Beta
(t-value) expectation format format (t-value)
Mediators/Adoption decision
A/AD 0.71 (8) 0.45a(8) 2.55* 0.57* 0.53 (12) 0.55 (20) 0.29 -0.09 0.61 (21) 0.46a(12) 2.34* 0.31*
1213 1506 (2.65) 2850 3576 (0.58) 3671 2888 (1.98)
N/AD 0.55a(13) 0.34 (24) 2.79* 0.33* 0.35 (33) 0.38 (39) 0.65 -0.16 0.43 (45) 0.32 (40) 2.36* 0.31*
1876 3485 (2.36) 7379 6781 (1.42) 8238 7432 (2.69)
BC/AD 0.42a(4) 0.38 (26) 0.29 0.17 0.41 (34) 0.40 (29) 0.04 -0.09 0.44 (32) 0.39 (42) 1.20 0.14
544 4839 (0.79) 7240 5140 (0.74) 5795 8478 (1.21)
Note: Number of correlations is reported in parentheses. Sample size is reported under correlation coefficients. Differences in sample sizes depend on
availability of data for moderators. A is attitude, N is norm, BC is behavioral control, AD is adoption.
*p < 0.05.
a
associations between mediators and innovation adoption decisions in organizations. Outcomes from the considered tests
motivate an investigation of the potential sources for the reported differences of effect size. Results are reported in Table 5a
and b.
The premise that behavioral control beliefs have a greater impact on the adoption decision when one considers top
managers was supported. The same effect was not found for attitudes and norms. Therefore, Hypothesis 5 is partially
confirmed by the data. The moderating effect of the organization size on decision maker's behavioral preferences-adoption
associations was found not significant: the t-test and the beta coefficients calculated with a meta-analytical regression model
are never significant. Consequently, our data does not provide consistent evidence that support Hypothesis 6. As robustness,
we test the moderating effects of organizational size by looking at different thresholds for the size of organizations. Spe-
cifically, we followed the OECD and considered the thresholds of both 250 and 500 employees for the size of organizations.
From the analysis, we observed consistent results with our main findings.
Considering the research settings, the magnitude of the attitude and norm-adoption associations is greater when the
adoption decision is measured as an intention rather than as an expectation. Therefore, we found partial support for
Hypothesis 7. We do not find any confirming evidence for Hypothesis 8. Attitudes and norm have significant implications not
only for the intention to adopt an innovation but also for the adoption decision. Finally, we found evidence that the common
method variance may inflate correlation coefficients, particularly for the attitude and norm-adoption linkages. Therefore,
Hypothesis 9 is partially confirmed by the available data. It must be noted, on the one hand, that, when significant, the
observed differences concerned the magnitude of the correlation, while the direction and significance level are consistent
with our main results. On the other hand, non-significant results must be considered cautiously and one should avoid
concluding on the limited relevance of the moderator since its non-significance may be related to the power of the test.
4.6. Additional analyses
As for robustness, we evaluated the possibility that a mediating variable, as it was measured in the study, may only be a
spurious one. Thus, we run a regression with random effects (Lipsey & Wilson, 2001) for each mediator on the variables
capturing other measured characteristics of our sampled studies (i.e., country, technology, industry, and year). From the
analysis, we observed that none of the considered variables was associated significantly to the attitude (min p-value ¼ 0.21;
max p-value ¼ 0.96; R-squared ¼ 0.23), norm (min p-value ¼ 0.10; max p-value ¼ 0.90; R-squared ¼ 0.18), and perceived
behavioral control (min p-value ¼ 0.06; max p-value ¼ 0.72; R-squared ¼ 0.26).
We also provided a further assessment of publication and eligibility biases by running a quantile-plot analysis, here
omitted but available on request from the corresponding author, in which the empirical values of considered associations
were contrasted with the expected values under the assumption of normality. From the analysis, we observed that the sample
of effect sizes for the considered associations gathered around the diagonal and generally felt within the 95% CIs of the
normality line. Therefore, we can conclude that publication and eligibility biases is limited in our study (Wang & Bushman,
1999). We also excluded extreme outliers from the sample of observation and observed that our findings remained unaltered,
which confirms the robustness of our results. We finally analyzed associations for the specific sample measures as reported in
Table 1 and found again that our main results hold. Finally, we conducted a one-sample removed analysis to report the in-
fluence of each individual sample on our results (Borenstein, Hedges, Higgins, & Rothstein, 2009). From the analysis, we
observed that our main findings hold constant.
5. Discussion
No single study has analyzed the linkages within Fig. 1 at once. Meta-analysis combined with path modeling permitted us
to conduct such a study. We found consistent evidence that attributes of innovations play a key role in the adoption of in-
novations in organizations, and relative advantage, compatibility, and complexity are those attributes that are likely to
promote greater stimulus on the decision to adopt innovations in organizations. Yet, we found in our sampled studies a work
in which greater levels of relative advantage are significantly associated (p ¼ 0.02) to more limited levels of adoption (Kurnia
et al., 2015). Authors suggested that the finding is likely to be attributed to the lack of understanding within the surveyed
organizations of the benefits of the new technologies as well as to the presence of barriers that favor the formation of negative
perceptions, which outweigh the positive impressions about the impact of the new technologies. For an alternative expla-
nation, we referred to studies developed within the protection motivation theory for individuals. We proposed that, in cases
in which the level of efficacy of a behavior is low or there is uncertainty that the behavior will result in the desired outcome,
the perceptions of lower (even negative) benefits may be more effective than perceptions of greater positive benefits in
stimulating the decision maker to adopt the behavior. In addition, decision makers are more likely to involve in extended
search for a behavior that is less certain to lead to the expected outcome than for a behavior that is more likely to lead to the
expected outcome (Block & Keller, 1995). This result calls for future studies considering the mechanisms and the conditions
that may explain a reverse relationship between relative advantage and adoption of innovations in organizations. For
instance, among such conditions, a further explanation may derive from analyzing the process by which organizations adopt
innovations by resorting to external sources. On this point, scholars have usually distinguished between pecuniary and non-
pecuniary inflows (Dahlander & Gann, 2010). There might be significant pecuniary costs involved in adopting innovation from
external sources, which usually takes place through contractual and licensing agreements (West & Bogers, 2014). The
pecuniary cost of innovation adoption may be seen as a proxy for the innovation perceived competitive advantage. However, a
number of scholars in this vein have demonstrated that the strength of the intellectual property protection regime, as re-
flected by the “price” of the innovation, may end up deterring innovation adoption from external providers because of the fear
that the adopting organization might experience reduced value capture and decreasing profitability, especially when the cost
of innovation acquisition exceeds the incremental value creation associated to it (Faems, de Visser & van Looy, 2010). Thus,
future research might look at the role of the price variable in the innovation adoption process and more generally on the
performance net benefits of innovation adoption.
Incidentally, we must observe that out of Rogers (1983) five attributes of innovations, observability and trialability were
the least used in empirical studies. Some of the reasons behind their more limited utilization are related in our sampled
studies to the specificity of the adopted technologies (e.g., Hashem & Tann, 2007), to the findings proposed by the study of
Tornatzky and Klein (1982) as well as to prior research based on the technology, organization, environment framework
(Agarwal & Prasad, 1998; Hashem & Tann, 2007; Hsing Wu, Kao, & Lin, 2013; Thong, 1999). Other studies referred to the
technological acceptance model in which only usefulness and ease of use (which relates to the concepts of relative advantage
and complexity, respectively) are considered (Davis, 1989) and to the decomposed theory of planned behavior where only
relative advantage, compatibility, and complexity are discussed as potential variables affecting the behavioral preferences of
decision makers (Taylor & Todd, 1995a). However, we provided evidence that both observability and trialability matter for the
adoption of innovations in organizations.
Our results offer evidence that the correlation coefficient of the trialability-adoption association has the lowest magnitude
among the attributes of innovations. This result may be influenced by available empirical estimates as well as by the
conceptualization/operationalization of trialability. Concerning the empirical estimates, in our sample we found three studies
in which greater levels of trialability were associated in a significant way (p < 0.05) to more limited levels of adoption (Aubert,
Schroeder, & Grimaudo, 2012; Hussin, Nor, & Suhaimi, 2008; Ramdani et al., 2009). Authors have explained this result by
considering that the negative trialability-adoption association may be related to the wrong impression of non-adopters about
the possibility to carry out tests of the new technology that, being complex and composed of different parts, is inherently
difficult to try. In the considered studies, non-adopters were considered to perceive the innovation to be easier to try than it
really was (Aubert et al., 2012). We may attempt to propose an alternative explanation for this result. Instead framing the
considered association within the expectancy-value model, one can consider the cognitive reassurance theory and observe
that when expectations of efficacy are high (e.g., the perceived trialability of technology is high) the subjects use this opinion
as a basis for their behavior, without further evaluations. However, when the expectations of efficacy are low (e.g., the
perceived trialability of the technology is low) people seek reassurance, which may influence their attitudes toward the
behavior and, in turn, the decision to adopt it (Gleicher & Petty, 1992). Therefore, we could consider the possibility that
decision makers that perceived the innovation as difficult to try were involved in understanding and experimenting with the
innovation to greater extents, which in turn favored their decision to adopt the innovation. Understanding mechanisms in
which the observed negative association between trialability and adoption prevails needs to be addressed by future research.
In the same vein, research should compare the predicting validity of the expectancy-value model with the cognitive reas-
surance theory or with alternative theories across different contexts and technologies in understanding the role of trialability
in innovation adoption decisions in organizations.
Concerning the conceptualization/operationalization of the construct of trialability, Banerjee, Wei, and Ma (2012)
observed that, if one considers contexts that are highly uncertain and conceives trialability not just as a belief-based factor
but as an active post-intent recursive process of experimentation focused on accumulating information and experience on the
innovation, then trialability may become an important determinant of the adoption decision. For an adoption decision to
occur, it is important not only the initial perception of the trialability of an innovation but also the effective intention of the
decision maker to try it. Considering the negative association between trialability and adoption, we may also envisage the
possibility that perceived greater levels of trialability may reduce the probability of a decision maker to try the innovation,
which in turn may negatively influence the adoption decision. Therefore, further elaborating on the concepts of observability
and trialability as well as investigating their implication across different organizations, contexts and technologies will
represent an important task for future research.
We found strong support for the attributes of innovations-behavioral preferences of decision makers relationships as well
as for the behavioral preferences of decision makers-adoption decisions linkages in organizations. A key finding is that the
behavioral preferences of decision makers mediate (although partially) the attributes of innovation-adoption decisions
linkages in organizations. Therefore, to the extent that future studies are interested in understanding the mechanisms
involved in the adoption decision, the considered constructs of attitudes, norm, and perceived behavioral control should be
included.
We also observed that contexts in which adoption takes place (i.e., the hierarchical position of the decision maker) as well
as the research settings employed by scholars both matter for empirical predictions. We found evidence that measuring
adoption as intention or as an expectation as well as measuring dependent and independent variables with the same/
different format influence the magnitude of empirical estimates, yet direction and significant level hold constant. Future
studies, particularly those interested in making precise estimates of the attributes of innovations-adoption decisions in or-
ganizations should make sure to address the problem of measuring the adoption construct, considering alternatives that are
both consistent with the conceptual definitions and address the common method variance bias (e.g., by combining data from
multiple, different sources or by using multiple respondents).
5.1. Our findings in the context of past meta-analytical works
Our study relates to a number of previous meta-analyses. Concerning the main effect of attributes of innovations on the
adoption of innovations, our study relates to the work of Arts et al. (2011) that deals with the adoption of innovations by
consumers. It is interesting to note that the influence of the attributes of innovations on the decision to adopt was greater
when the adopting unit is an organization rather than when the adopter was a consumer. In addition, considering consumers'
adoption of innovations, correlation coefficients associated to the attributes of complexity, observability, and trialability were
close to zero and generally not significant. The comparison suggested that, while organizations generally tend to care about all
the attributes of an innovation, consumers generally tend to care less about such attributes and in a selective manner.
It must be further noted that those meta-analysis that combined adoptions by organizations and by consumer (Kapoor
et al., 2014b), reported coefficients associated to the attributes of an innovation that were generally lower than ours and
greater than those reported by Arts et al. (2011) for consumers. In that, both our study and the meta-analysis of Arts et al.
(2011) offered a picture of the upper and lower boundaries within which the mean correlation coefficients associated to
the attributes of relative advantage (0.39 .19), compatibility (0.36 .21), complexity (0.33;-0.01), observability (0.30 .05),
trialability (0.11 .03) tend to be distributed.
Our study focused on the adoption decision and, thus, our results relate with studies on post-adoption, particularly with
those reviews that examined the performance outcomes of innovation adoption in organizations (Rosenbusch et al., 2011).
Although results from the two studies cannot be combined since samples are different, both reviews address an important
research need e i.e. understanding antecedent innovation performance chaindand suggest a call for future studies aimed at
developing an integrated model that accounts for the relationships among attributes of innovations, adoption decisions, and
performance implications of innovations in organizations.
5.2. Limitations of this meta-analysis
Prior to concluding, there are several limitations of this study to discuss. These limitations concern the correlational nature
of the results, the validity of self-reports of innovation adoption decisions, the difficulty of analyzing more complex in-
teractions in the proposed framework as well as the generalizability of its conclusions to the sample of studies and to the
population of potential studies on the topic. Firstly, our study has a correlational nature. The specific composition of the
sample of organizations, the preferences of particular researchers, and other un-controlling variables can co-vary with other
characteristics of the studies or the methods being used. Fortunately, however, this limitation is mitigated by the use of
mediational analysis, multiple measures and methods. Another limitation refers to the potential inaccuracy of the managers'
self-reported behavior. However, various factors influence the accuracy of self-reports, such as the length of the time interval
and the social context of the adoption as well as the order in which participants answer questions (March & Sutton, 1997). The
accuracy of self-reports may also differ across groups. For example, if the innovation conveys a particular status to the adopter,
respondents may tend to report that their firm has adopted the innovation even in the absence of an adoption decision. Given
these possibilities, future studies may combine self-report measures of the adoption decision with objective obtained, for
example, from archival data or from third parties (e.g., suppliers of the new technology).
An additional limitation concerns the associations between attributes of innovations and behavioral preferences of
managers. Because of the available conceptualization, we excluded observability and trialability. Yet, we expect future studies
to consider how the attributes of observability and trialability may influence the behavioral preferences of managers and,
consequently, stimulate the adoption decision. Mediating processes and moderating variables may interact among each other
and such interactions may generate further effects on the attributes of innovations-adoption decisions relationships in or-
ganizations. In addition, second order relationships could be also envisaged. Finally, by synthesizing the largest number of
studies on the adoption decision that provide source information, the findings from our meta-analysis are probably the most
generalizable to date. To complement these findings, future research may offer a sufficiently large number of effect sizes to
estimate the population variance and establish the tenability of our conclusions in the broader universe of all possible studies.
5.3. Implications for practice
This study has also implications for practice. Without knowing the characteristics of the potential adopters, business
executives interested in the diffusion of an innovation should design the new technology in a way that at least offers greater
benefits compared with already available alternatives and, at the same time, such innovation should be compatible, easy to
use, and its benefits observable.
Business executives must also recognize that the influences of attributes of innovation (particularly relative advantage,
compatibility, complexity) on the decision to adopt innovation are mediated by a greater extent by influences on attitude,
norms, and behavioral control perceptions of the decision maker. This means that the attributes of innovation need to
stimulate positive attitudes, norm and perceived behavioral control over the new technology for decision makers, particularly
those of top managers. Our results suggest that targeted approaches may improve the diffusion of innovations. Because
attributes of innovations tend to operate through different mediators that influence the adoption decision differently, a
business executive who recognizes that the decision maker of adopting organizations may lack the necessary resources to
spot the innovation should design new technologies whose complexity is very limited. Managers that targets firms that are
highly connected with other organizations (i.e., like in business ecosystems) could improve the diffusion of innovations by
investing resources for increasing the perception of the relative advantage of the innovation not only by the target adopting
organizations but also by partners with which these target organizations are associated.
Business executives should also consider the differential effects of the attributes of innovations on the behavioral pref-
erences of decision makers and, in turn, on the adoption decision. Consider for example that while relative advantage has a
significant positive effect on decision makers' attitude and norm, complexity has a negative effect on attitude, norm, and
perceived behavioral control of the innovation. Therefore, one can increase the diffusion of an innovation by increasing its
relative advantage or by reducing its complexity. From our structural equation model, we also observed that the positive
effect of relative advantage on decision maker's behavioral preferences (Beta ¼ 0.58, p < 0.05) is larger in magnitude than the
negative effect of complexity on these behavioral preferences (Beta ¼ 0.27, p < 0.05). Thus, when there are trade-offs be-
tween relative advantage and complexity in the design of a new technologies, investing resources for providing innovations
with greater relative advantages may turn to be beneficial in terms of its diffusion, even if such increases in the advantages of
the innovation will be achieved at the expenses of moderate increases in the levels of complexity. In a different perspective,
however, if an organization wants to deter the adoption of an innovation by other organizations (e.g., to protect a firm
competitive advantage), business executives could increase the complexity of the innovation and reduce its relative
advantage (and also its compatibility) for other organizations and favor that this perception will be shared by most of the
players of an industry.
5.4. Future directions for research
Our findings provide evidence that something can be gained by integrating the innovation diffusion theory and the theory
of planned behavior. Such integration offers a richer but more articulated explanation of the adoption decision in organization
as well as of its determinants. Specifically, we conceived and provided empirical evidence that decision makers will
contribute to the adoption decision not just by simply forming their own interpretation of the attributes of innovation but by
developing a more complex set of behavioral preferences which, in turn, will influence the adoption decision. The important
role of decision maker in the adoption decision was also envisaged by Rogers who suggested that in the adoption process an
individual is expected to pass from the knowledge of an innovation to the formation of an attitude toward the new technology
and once formed such attitudes are keys for explaining a decision to adopt an innovation (Rogers, 1983, p. 36). We add to
Roger's initial insight that not only attitude but other behavioral preferences (i.e., norms and perceived behavioral control)
matter for the adoption decision.
Our results further suggest the importance of scholars to continue to analyze decomposition and crossover effects in the
theory of planned behavior (Taylor & Todd, 1995a). Future research should also incorporate as antecedents of decision makers'
behavioral preferences the attributes of observability and trialability. To the best of our knowledge, a theoretical discussion of
nature (e.g., direct correlates or moderating variables of the attributes of innovation-behavioral preferences linkage), ex-
pected association (i.e., which behavioral preferences is influenced), and direction of the effects of these two attributes of
innovations on decision makers' behavioral preferences is required.
Moreover, direct effects of attributes of innovations on the adoption decision are still significant even if we controlled for
the behavioral preferences of the decision maker. The presence of these direct effects may originate from considering the
adoption decision in an organization. For example, an organization decision maker may have a very limited attitude toward
the innovation but she will pursue its adoption in order to achieve various benefits and reward for herself or for the orga-
nization, with the latter benefits and reward that are extrinsic to the use of the technology by the decision maker herself
(Davis et al., 1989). Therefore, studies on the adoption of innovations in organizations should elaborate on the distinction
between behavioral preferences that are considered by decision makers as intrinsic or extrinsic to the adoption of an
innovation and evaluating how these distinct preferences are likely to mediate the attributes of innovation-adoption de-
cisions relationship in organizations.
It seems useful to clarify that the adoption decision may be affected by the function assumed by the decision maker (i.e.,
entrepreneur or manager). Although data do not allow us to introduce such distinction, conceptually conceiving and
empirically separating adoption decisions taken by entrepreneurs from those decisions taken by managers may be relevant
for understanding the adoption of innovations by organizations. On the one hand, entrepreneurial decisions, more than
managerial ones, tend to be affected by specific biases and heuristics, which can lead to perceive the competitive environment
differently and thereby influence innovation adoption decisions, such as overconfidence and representativeness (i.e. the
willingness to rely on small, nonrandom samples in decision making, see e.g. Busenitz and Barney (1997)). In general, it is
commonly acknowledged within strategic management literature that entrepreneurs do think and behave differently than
managers especially in large organizations (Busenitz & Barney, 1997). On the other hand, the role of managers as the “ar-
chitects of the innovation imperative” is also widely acknowledged in literature (Cooper, 1998, p. 493). Put simply, scholars
have offered evidences that it is managers who help the organizations to develop all the pre-requisites in terms of resources
that precede innovation adoption. And, once developed, such resources are likely to facilitate the adoption of innovations in
organizations. Therefore, the process of resources development by an organization represent an important aspect to be
considered in understanding the adoption of innovations in organizations.
In understanding the adoption of innovation in organizations, research should also elaborate and even extend the concept
of behavioral preferences of decision makers. In that, areas of study that appear to hold promise include such aspects as
decision maker's awareness of the attributes of an innovation and trust in the providers/vendors of the innovation. Note that
awareness has been extensively considered within the awareness-motivation-skills and capabilities framework for under-
standing adoption decision of various behaviors (Dutton & Jackson, 1987; Kiesler & Sproull, 1982; Lant, Milliken, & Batra,
1992) and trust has been considered in studies of innovation adoption as antecedent of the decision maker's attitude
(Pavlou & Fygenson, 2006).
Research should also consider the possibility to take a multidimensional perspective on the behavioral preferences of
managers and investigate potential interdependencies both among the behavioral preferences of an organization decision
maker and across the preferences of different organization decision makers, identify relations that are potentially comple-
mentary or substitutable, and evaluating how these interactive preferences mediate the attributes of innovations-adoption
linkages. At the same time, given that the decision maker matter in the adoption decision, we envisage the possibility to
integrate the innovation adoption and the behavioral economics literature and explicitly study the effects of heuristics and
biases of decision makers on the adoption decision.
Moreover, future research could employ such multidimensional perspective even to tackle the combinative effects of
different attributes of innovation on different categories (or types) of innovation. Research on innovation adoption has
provided a myriad of categorizations, by relating innovation for example to product, process, disruptive, architectural and
“jugaad” innovation (Radjou, Prabhu & Ahuja, 2012), that all attempt to shed light on the Schumpeterian process of creative
destruction. Among the most popular innovation categories, for instance, Damanpour (1991) has identified pair of types of
innovation (i.e. administrative and technical, product and process, radical and incremental) and demonstrated that the
adoption of each type does not relate equally to the same predictor variables. Thus, an interesting future avenue of research
would be to address how the behavioral preferences of decision makers mediate the relationship between attributes of
innovation and adoption decisions depending on the specific type of innovation being considered. Additional tests of our
theoretical argument and mediation effects across different types of innovation should add to the conclusiveness of our
findings.
Our study provides also evidence of the presence of heterogeneity across nearly all linkages, even after we accounted for
the moderators. Therefore, theoretically addressing and empirically analyzing the implications of moderating variables on the
antecedents-adoption decisions linkages in organizations represent an important task for future research. Concerning
moderating variables associated to the contexts of the innovation adoption, past experience of decision maker (Cordano &
Frieze, 2000) and espoused national cultural values (Mueller, Rosenbusch, & Bausch, 2013; Srite & Karahanna, 2006)
represent important aspects to be considered in future studies. The moderating effect of the organization size on decision
maker's behavioral preferences-adoption associations was instead found not significant in this study. However, it must be
noted that studies adopting a different operationalization of organizational size, such as those focusing on Small and
Medium-Sized Enterprises (SMEs) as defined by the European Commission (i.e. businesses which employ up to 500 people in
Europe, see OECD (1998)), have found a significant relationship between the perceived direct and indirect benefits of
innovation and adoption decisions depending on the size of the organization (see e.g. Scupola (2003) study on the adoption of
e-commerce by SMEs in the South of Italy). Therefore, it would be interesting to address, in future research, as whether and
how different conceptualizations of organizational size and thresholds may differently affect the decision maker's behavioral
preferences-adoption decisions linkages. In addition, the non-significant moderating effect of organizational size found in
this paper stimulates further research on the antecedents of such effect. A well-known explanation in literature, for instance,
looks at the different endowment of financial resources between small and large-sized organizations as one common
explanation of diverging innovation adoption behaviors. For instance, in their study, Van de Vrande, De, Jong, Vanhaverbeke,
and De Rochemont (2009) conclude that innovation adoption in SMEs is hampered by lack of financial resource, whereas
larger size organizations not only are more prone to adopt innovation but also reserve structural funds to finance innovation
exploitation and exploration once adopting. In light of these results, our non-significant moderating effect of size should thus
encourage scholars to analyze in greater depth organizational size as an important contingent factor, which is able to in-
fluence trends toward the adoption of innovations. Finally, one could also explore the relative advantage e adoption decisions
linkage in the view of the “innovator's dilemma” idea and including additional moderating variables both at the firm (e.g.,
age) and at the industry (e.g., turbulence, dynamicity and heterogeneity) levels. This is an important issue to be addressed in
future studies.
Concerning the overall research settings employed by scholars, this review suggests that still different studies do not meet
some of the “ideal” conditions suggested by Tornatzky and Klein (1982). We then suggest future studies to continue to
embrace a predictive instead of a retrospective stance, use research approaches and measures that are reliable, replicable and
permit some degree of statistical power, consider longitudinal design and introduce cross sample validation. Scholars should
also carefully spot the issues of endogeneity and chronology. Scholars should design their works and consider their findings in
a way that considers that (a) the decision to adopt innovations may induce the decision maker to associate certain attributes
to the new technology (i.e., the adoption decision determines the perception of the attributes of innovations); (b) there may
be different time lags between the time at which a decision maker is exposed to an innovation and the time at which a
decision is taken (i.e., the adoption decision becomes contingent upon the period the interview is conducted); (c) in some
organizations, the decision makers always changes (i.e., the decision maker that decides to adopt the innovation may be
different from the individual who evaluate the attributes of the innovation).
Extant studies suggest that the innovation diffusion theory is affected by pro-innovation biases (Jeyaraj et al., 2006). The
presence of these biases makes difficult for scholars to address such challenging questions as when and how do organizations
adopt “inefficient” innovations? When and how do organizations reject “efficient” innovations? In addressing these chal-
lenging questions, in an early study Abrahamson (1991) considered, among other conditions, how influences exerted by other
organizations on the decision maker may represent a potential mechanisms for understanding the adoption of inefficient
innovations (or the rejection of efficient ones). The integration of the innovation diffusion theory with the theory of planned
behavior may offer a complementary perspective on the adoption of inefficient innovations. Specifically, our model not only
incorporates the possibility that an organization may adopt inefficient innovations (e.g., an innovation with very limited
relative advantages and/or highly complex), via decision makers' norms formation, but also offers some directions on the
conditions for these adoption decisions to occur. Specifically, our model suggests that organizations are likely to adopt
inefficient innovations when the magnitude of effects produced by external pressures on decision makers' norm will be
greater than the magnitude of both the direct effects of attributes of innovations on attitudes as well as of the crossover effects
of these attributes on the decision maker's norms. An empirical estimate of conditions according to which effects of external
pressures tend to dominate the effects of the attributes of innovations in the adoption of inefficient innovations represent an
important issue to be discussed in future studies.
References
Abrahamson, E. (1991). Managerial fads and fashions: The diffusion and refection of innovations. Academy of Management Review, 16(3), 586e612.
Agarwal, R., & Prasad, J. (1998). The antecedents and consequents of user perceptions in information technology adoption. Decision Support Systems, 22(1),
15e29.
Ajzen, I. (1985). From intentions to actions: A theory of planned behavior. In J. Kuhl, & J. Beckmann (Eds.), Action-control: From cognition to behavior (pp.
11e39). Heidelberg: Springer.
Ajzen, I. (1987). Attitudes, traits, and actions: Dispositional prediction of behavior in personality and social psychology. In L. Berkowitz (Ed.), Advances in
experimental social psychology. New York: Academic Press.
Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50(2), 179e211.
Ajzen, I., & Driver, B. L. (1992). Contingent value measurement: On the nature and meaning of willingess to pay. Journal of Consumer Psychology, 1(4),
297e316.
Alam, S. S., & Noor, M. K. M. (2009). ICT adoption in small and medium enterprises: An empirical evidence of service sectors in Malaysia. International
Journal of Business and Management, 4(2), 112e125.
Anderson, N., Poto cnik, K., & Zhou, J. (2014). Innovation and creativity in organizations a state-of-the-science review, prospective commentary, and guiding
framework. Journal of Management, 40(5), 1297e1333.
Angle, H. L., & Van de Ven, A. H. (1989). Suggestions for managing the innovation journey. In A. H. Van de Ven, H. L. Angle, & M. S. Poole (Eds.), Research on
the management of innovation: The Minnesota studies: Da trovare. New York: Oxford University Press.
Armitage, C. J., & Conner, M. (1999). Distinguishing perceptions of control from self-efficacy: Predicting consumption of a low-fat diet using the theory of
planned behavior. Journal of Applied Social Psychology, 29(1), 72e90.
Armitage, C. J., & Conner, M. (2001). Efficacy of the theory of planned behaviour: A meta-analytic review. British Journal of Social Psychology, 40(4), 471e499.
Arpaci, I. (2013). Organizational adoption of mobile communication technologies. Ankara: Middle East Technical University.
Arts, J. W. C., Frambach, R. T., & Bijmolt, T. H. A. (2011). Generalizations on consumer innovation adoption: A meta-analysis on drivers of intention and
behavior. International Journal of Research in Marketing, 28(2), 134e144.
Askarany, D., Brierley, J. A., & Yazdifar, H. (2012). The effect of innovation characteristics on activity-based costing adoption. International Journal of
Managerial and Financial Accounting, 4(3), 291e313.
Attewell, P. (1992). Technology diffusion and organizational learning: The case of business computing. Organization Science, 3(1), 1e19.
Aubert, B. A., Schroeder, A., & Grimaudo, J. (2012). IT as enabler of sustainable farming: An empirical analysis of farmers' adoption decision of precision
agriculture technology. Decision Support Systems, 54(1), 510e520.
Baldridge, J. V., & Burnham, R. A. (1975). Organizational innovation: Individual, organizational, and environmental impacts. Administrative Science Quarterly,
20(2), 165e176.
Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84(2), 191e215.
Banerjee, P., Wei, K. K., & Ma, L. C. (2012). Role of trialability in B2B e-business adoption: Theoretical insights from two case studies. Behaviour & Information
Technology, 31(9), 815e827.
Bass, B. M., & Stogdill, R. M. (1990). Handbook of leadership. New York: Free Press.
Bauer, T. N., Bodner, T., Erdogan, B., Truxillo, D. M., & Tucker, J. S. (2007). Newcomer adjustment during organizational socialization: A meta-analytic review
of antecedents, outcomes, and methods. Journal of Applied Psychology, 92(3), 707.
Becker, M. H. (1970). Sociometric location and innovativeness: Reformulation and extension of the diffusion model. American Sociological Review, 35(2),
267e282.
Begg, C. B. (1994). Publication bias. In H. Cooper, & L. V. Hedges (Eds.), The handbook of research synthesis (pp. 399e409). New York: Russel Sage Foundation.
Bentler, P. M. (1990). Comparative fit indexes in structural models. Psychological Bulletin, 107(2), 238e246.
Blalock, H. (1961). Causal inferences in nonexperimental research. Chapel Hill: University of North Carolina Press.
Block, L. G., & Keller, P. A. (1995). When to accentuate the negative: The effects of perceived efficacy and message framing on intentions to perform a health-
related behavior. Journal of Marketing Research, 32(2), 192e203.
Bloom, P. N., & Kotler, P. (1975). Strategies for high market-share companies. Harvard Business Review, 53(6), 63e72.
Bobko, P., & Rieck, A. (1980). Large sample estimators for standard errors of functions of correlation coefficients. Applied Psychological Measurement, 4(3),
385e398.
Borenstein, M., Hedges, L. V., Higgins, J. P., & Rothstein, H. R. (2009). Introduction to meta-analysis. West Sussex, UK: Wiley.
Brown, R. L. (1997). Assessing specific mediational effects in complex theoretical models. Structural Equation Modeling: A Multidisciplinary Journal, 4(2),
142e156.
Brown, S. P., & Peterson, R. A. (1993). Antecedents and consequences of salesperson job satisfaction: Meta-analysis and assessment of causal effects. Journal
of Marketing Research, 30(1), 63e77.
Bruque, S., & Moyano, J. (2007). Organisational determinants of information technology adoption and implementation in SMEs: The case of family and
cooperative firms. Technovation, 27(5), 241e253.
Bullock, R. J., & Svyantek, D. J. (1985). Analyzing meta-analysis: Potential problems, an unsuccessful replication, and evaluation criteria. Journal of Applied
Psychology, 70(1), 108e115.
Busenitz, L. W., & Barney, J. B. (1997). Differences between entrepreneurs and managers in large organizations: Biases and heuristics in strategic decision-
making. Journal of Business Venturing, 12(1), 9e30.
Camiso n-Zornoza, C., Lapiedra-Alcamí, R., Segarra-Cipre s, M., & Boronat-Navarro, M. (2004). A meta-analysis of innovation and organizational size. Or-
ganization Studies, 25(3), 331e361.
Campbell, D. T., & Fiske, D. W. (1959). Convergent and discriminant validation by the multitrait-multimethod matrix. Psychological Bulletin, 56(2), 81e105.
Chen, J., Damanpour, F., & Reilly, R. R. (2010). Understanding antecedents of new product development speed: A meta-analysis. Journal of Operations
Management, 28(1), 17e33.
Chen, M.-J., & Hambrick, D. C. (1995). Speed, stealth, and selective attack: How small firms differ from large firms in competitive behavior. Academy of
Management Journal, 38(2), 453e482.
Cheung, M. W. L. (2009). Comparison of methods for constructing confidence intervals of standardized indirect effects. Behavior Research Methods, 41(2),
425e438.
Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioral sciences (Second ed.). Hillsdale, NJ: Lawrence Erlbaum As-
sociates, Inc.
Colquitt, J. A., LePine, J. A., & Noe, R. A. (2000). Toward an integrative theory of training motivation: A meta-analytic path analysis of 20 years of research.
Journal of Applied Psychology, 85(5), 678.
Cooper, J. R. (1998). A multidimensional approach to the adoption of innovation. Management Decision, 36(8), 493e502.
Cordano, M., & Frieze, I. H. (2000). Pollution reduction preferences of U.S. environmental managers: Applying Ajzen's theory of planned behavior. Academy
of Management Journal, 43(4), 627e641.
Crampton, S. M., & Wagner, J. A. (1994). Percept-percept inflation in microorganizational research: An investigation of prevalence and effect. Journal of
Applied Psychology, 79(1), 67e76.
Crossan, M. M., & Apaydin, M. (2010). A multi-dimensional framework of organizational innovation: A systematic review of the literature. Journal of
Management Studies, 47(6), 1154e1191.
Cross, R. L., & Brodt, S. E. (2001). How assumptions of consensus undermine decision making. MIT Sloan Management Review, 42(2), 86e94.
Dahlander, L., & Gann, M. D. (2010). How open is innovation. Research Policy, 39(6), 699e709.
Damanpour, F. (1991). Organizational innovation: A meta-analysis of effects of determinants and moderators. Academy of Management Journal, 34(3),
555e590.
Damanpour, F. (1996). Organizational complexity and innovation: Developing and testing multiple contingency models. Management Science, 42(5),
693e716.
Damanpour, F., & Gopalakrishnan, S. (1998). Theories of organizational structure and innovation adoption: The role of environmental change. Journal of
Engineering and Technology Management, 15(1), 1e24.
Damanpour, F., & Wischnevsky, D. J. (2006). Research on innovation in organizations: Distinguishing innovation-generating from innovation-adopting
organizations. Journal of Engineering and Technology Management, 23(4), 269e291.
Daryanto, A., Khan, H., Matlay, H., & Chakrabarti, R. (2013). Adoption of country-specific business websites. The case of UK small businesses entering the
Chinese market. Journal of Small Business and Enterprise Development, 20(3), 650e660.
Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13(4), 319e340.
Davis, F. D., Bagozzi, R. P., & Warshaw, P. R. (1989). User acceptance of computer technology: A comparison of two theoretical models. Management Science,
35(8), 982e1003.
Davis, H. L., Hoch, S. J., & Ragsdale, E. E. (1986). An anchoring and adjustment model of spousal predictions. Journal of Consumer Research, 13(1),
25e37.
DeLone, W. H., & McLean, E. R. (1992). Information systems success: The quest for the dependent variable. Information Systems Research, 3(1), 60e95.
DeLone, W. H., & McLean, E. R. (2003). The DeLone and McLean model of information systems success: A ten-year update. Journal of Management Infor-
mation Systems, 19(4), 9e30.
Downs, J. G. W., & Mohr, L. B. (1976). Conceptual issues in the study of innovation. Administrative Science Quarterly, 21(4), 700e714.
Dulany, D. E., Jr. (1961). Hypotheses and habits in verbal 'operant conditioning.'. Journal of Abnormal and Social Psychology, 63(2), 251e263.
Dutton, J. E., & Jackson, S. E. (1987). Categorizing strategic issues: Links to organizational action. Academy of Management Review, 12(1), 76e90.
Ettlie, J. E. (1983). Organizational policy and innovation among suppliers to the food processing sector. Academy of Management Journal, 26(1), 27e44.
Eveland, J. D. (1979). Issues in using the concept of “adoption of innovations”. Journal of Technology Transfer, 4(1), 1e13.
Faems, D. M., de Visser, P. A., & van Looy, B. (2010). Technology alliance portfolios and financial performance: Value-enhancing and cost-increasing effects of
open innovation. Journal of Product Innovation Management, 27(6), 785e796.
Fern, E. F., & Monroe, K. B. (1996). Effect-size estimates: Issues and problems in interpretation. Journal of Consumer Research, 23(2), 89e105.
Fiedler, K., Schott, M., & Meiser, T. (2011). What mediation analysis can (not) do. Journal of Experimental Social Psychology, 47(6), 1231e1236.
Fishbein, M. (1967). Attitude and the prediction of behavior. In M. Fishbein (Ed.), Readings in attitude theory and measurement (pp. 477e492). New York: John
Wiley & Sons.
Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention and behavior: An introduction to theory and research. Reading, MA: Addison-Wesley.
Flynn, F. J., & Wiltermuth, S. S. (2010). Who's with me? False consensus, brokerage, and ethical decision making in organizations. Academy of Management
Journal, 53(5), 1074e1089.
Fombrun, C., & Shanley, M. (1990). What's in a name? Reputation building and corporate strategy. Academy of Management Journal, 33(2), 233e258.
Gajendran, R. S., & Harrison, D. A. (2007). The good, the bad, and the unknown about telecommuting: Meta-analysis of psychological mediators and in-
dividual consequences. Journal of Applied Psychology, 92(6), 1524e1541.
Gamal Aboelmaged, M. (2010). Predicting e-procurement adoption in a developing country: An empirical integration of technology acceptance model and
theory of planned behaviour. Industrial Management & Data Systems, 110(3), 392e414.
Ghobakhloo, M., & Tang, S. H. (2013). The role of owner/manager in adoption of electronic commerce in small businesses: The case of developing countries.
Journal of Small Business and Enterprise Development, 20(4), 754e787.
Gleicher, F., & Petty, R. E. (1992). Expectations of reassurance influence the nature of fear-stimulated attitude change. Journal of Experimental Social Psy-
chology, 28(1), 86e100.
Gooding, R. Z., & Wagner, J. A. (1985). A meta-analytic review of the relationship between size and performance. Administrative Science Quarterly, 30(3),
462e481.
Gross, S. R., & Miller, N. (1997). The 'golden section' and bias in perceptions of social consensus. Personality & Social Psychology Review, 1(3), 241e272.
Gu, V. C., Cao, Q., & Duan, W. (2012). Unified Modeling Language (UML) IT adoptionda holistic model of organizational capabilities perspective. Decision
Support Systems, 54(1), 257e269.
Gupta, P., Seetharaman, A., & Raj, J. R. (2013). The usage and adoption of cloud computing by small and medium businesses. International Journal of In-
formation Management, 33(5), 861e874.
Hage, J., & Dewar, R. (1973). Elite values versus organizational structure in predicting innovation. Administrative Science Quarterly, 18(3), 279e290.
Hambrick, D. C., & Mason, P. A. (1984). Upper echelons: The organization as a reflection of its top managers. Academy of Management Review, 9(2),
193e206.
Hameed, M. A., Counsell, S., & Swift, S. (2012). A meta-analysis of relationships between organizational characteristics and IT innovation adoption in or-
ganizations. Information & Management, 49(5), 218e232.
Harrison, D. A., Mykytyn, P. P., Jr., & Riemenschneider, C. K. (1997). Executive decisions about adoption of information technology in small business: Theory
and empirical tests. Information Systems Research, 8(2), 171e195.
Hashem, G., & Tann, J. (2007). The adoption of ISO 9000 standards within the Egyptian context: A diffusion of innovation approach. Total Quality Man-
agement & Business Excellence, 18(6), 631e652.
He, Q., Duan, Y., Fu, Z., & Li, D. (2006). An innovation adoption study of online e-payment in Chinese companies. Journal of Electronic Commerce in Orga-
nizations, 4(1), 48e69.
Henderson, D., Sheetz, S. D., & Trinkle, B. S. (2012). The determinants of inter-organizational and internal in-house adoption of XBRL: A structural equation
model. International Journal of Accounting Information Systems, 13(2), 109e140.
Heyder, M., Theuvsen, L., & Hollmann-Hespos, T. (2012). Investments in tracking and tracing systems in the food industry: A PLS analysis. Food Policy, 37(1),
102e113.
Hossain, M. A., & Quaddus, M. (2015). Radio frequency identification (RFID) adoption: A cross-sectional comparison of voluntary and mandatory contexts.
Information Systems Frontiers, 17(5), 1057e1076.
Hsing Wu, C., Kao, S.-C., & Lin, H.-H. (2013). Acceptance of enterprise blog for service industry. Internet Research, 23(3), 260e297.
Hsu, C.-L., & Lin, J.-C. (2015). Factors affecting the adoption of cloud services in enterprises. Information Systems and e-Business Management, First online: 21
December, 2015, 1e32.
Hsu, P.-F., Ray, S., & Li-Hsieh, Y.-Y. (2014). Examining cloud computing adoption intention, pricing mechanism, and deployment model. International Journal
of Information Management, 34(4), 474e488.
Huber, J. (1990). Macro-micro links in gender stratification: 1989 presidential address. American Sociological Review, 55(1), 1e10.
Hung, S.-Y., Hung, W.-H., Tsai, C.-A., & Jiang, S.-C. (2010). Critical factors of hospital adoption on CRM system: Organizational and information system
perspectives. Decision Support Systems, 48(4), 592e603.
Hunter, J. E., & Schmidt, F. L. (2004). Methods of meta-analysis: Correcting error and bias in research findings. New York: Sage.
Hussin, H., Nor, R. M., & Suhaimi, M. A. (2008). Perceived attributes of e-commerce and the adoption decision: The case of Malaysian SMEs. Journal Teknologi
Maklumat & Multimedia, 5, 107e125.
Jackson, E. L. (2008). Behavioural determinants of the adoption of forward contracts by Western Australian wool producers. Curtin: Curtin University of
Technology.
Jarrett, S. M. (2003). Factors affecting the adoption of e-business in the aerospace industry. Nova: Nova Southeastern University.
Jeon, B. N., Han, K. S., & Lee, M. J. (2006). Determining factors for the adoption of e-business: The case of SMEs in Korea. Applied Economics, 38(16),
1905e1916.
Jeyaraj, A., Rottman, J. W., & Lacity, M. C. (2006). A review of the predictors, linkages, and biases in IT innovation adoption research. Journal of Information
Technology, 21(1), 1e23.
Kapoor, K. K., Dwivedi, Y. K., & Williams, M. D. (2014a). Innovation adoption attributes: A review and synthesis of research findings. European Journal of
Innovation Management, 17(3), 327e348.
Kapoor, K. K., Dwivedi, Y. K., & Williams, M. D. (2014b). Rogers' innovation adoption attributes: A systematic review and synthesis of existing research.
Information Systems Management, 31(1), 74e91.
Kets de Vries, M., & Miller, D. (1984). The neurotic organization. San Francisco: Jossey-Bass.
Keupp, M. M., Palmie , M., & Gassmann, O. (2012). The strategic management of innovation: A systematic review and paths for future research. International
Journal of Management Reviews, 14(4), 367e390.
Kiesler, S., & Sproull, L. (1982). Managerial response to changing environments: Perspectives on problem sensing from social cognition. Administrative
Science Quarterly, 27(4), 548e570.
Kimberly, J. R. (1976). Organizational size and the structuralist perspective: A review, critique, and proposal. Administrative Science Quarterly, 21(4), 571e597.
Kimberly, J. R., & Evanisko, M. R. (1981). Organizational innovation: The influence of individual, organizational, and contextual factors on hospital adoption
of technological and administrative innovations. Academy of Management Journal, 24(4), 689e713.
King, W. R., & He, J. (2006). A meta-analysis of the technology acceptance model. Information & Management, 43(6), 740e755.
Kurnia, S., Choudrie, J., Mahbubur, R. M., & Alzougool, B. (2015). E-commerce technology adoption: A Malaysian grocery SME retail sector study. Journal of
Business Research, 68(9), 1906e1918.
Kwon, T., & Zmud, R. (1987). Unifying the fragmented models of information systems implementation. In R. Hirschheim, & R. J. Boland (Eds.), Critical issues
in information systems research (pp. 227e262). New York: Wiley.
Lant, T. K., Milliken, F. J., & Batra, B. (1992). The role of managerial learning and interpretation in strategic persistence and reorientation: An empirical
exploration. Strategic Management Journal, 13(8), 585e608.
Larsen, K. R. T. (2003). A taxonomy of antecedents of information systems success: Variable analysis studies. Journal of Management Information Systems,
20(2), 169e246.
Larsson, R. (1993). Case survey methodology: Quantitative analysis of patterns across case studies. Academy of Management Journal, 36(6), 1515e1546.
Lau, R. R., Sigelman, L., Heldman, C., & Babbitt, P. (1999). The effects of negative political advertisements: A meta-analytic assessment. American Political
Science Review, 93(4), 851e875.
Lee, Y., & Larsen, K. R. T. (2009). Threat or coping appraisal: Determinants of SMB executives' decision to adopt anti-malware software. European Journal of
Information Systems, 18(2), 177e187.
Lefebvre, L. A., Lefebvre, E., & Harvey, J. (1996). Intangible assets as determinants of advanced manufacturing technology adoption in SME’ s: Toward an
evolutionary model. IEEE Transactions on Engineering Management, 43(3), 307e322.
Leung, M. T. (2005). Determinants of the adoption of technological innovations in organizations. Hong Kong: Hong Kong Polythechnic University.
Limthongchai, P., & Speece, M. (2003). The effect of perceived characteristics of innovation on e-commerce adoption by SMEs in Thailand. In Proceedings of
the Seventh International Conference on Global Business and Economic Development, Bangkok, Thailand.
Lin, C.-Y., & Ho, Y.-H. (2011). Determinants of green practice adoption for logistics companies in China. Journal of Business Ethics, 98(1), 67e83.
Lin, C.-H., Lin, I.-C., Roan, J.-S., & Yeh, J.-S. (2012). Critical factors influencing hospitals' adoption of HL7 version 2 standards: An empirical investigation.
Journal of Medical Systems, 36(3), 1183e1192.
Lipsey, M. W., & Wilson, D. B. (2001). Practical meta-analysis. Thousand Oaks, CA: SAGE Publications.
Li, X., Troutt, M. D., Brandyberry, A., & Wang, T. (2011). Decision factors for the adoption and continued use of online direct sales channels among SMEs.
Journal of the Association for Information Systems, 12(1), 1e31.
Luqman, A., & Abdullah, N. K. (2011). E-business adoption amongst SMEs: A structural equation modeling approach. Journal of Internet Banking and
Commerce, 16(2), 99e112.
Mahmood, M. A., Hall, L., & Swanberg, D. L. (2001). Factors affecting information technology usage: A meta-analysis of the empirical literature. Journal of
Organizational Computing & Electronic Commerce, 11(2), 107e130.
Malhotra, N. K., Kim, S. S., & Patil, A. (2006). Common method variance in is research: A comparison of alternative approaches and a reanalysis of past
research. Management Science, 52(12), 1865e1883.
Marcati, A., Guido, G., & Peluso, A. M. (2008). The role of SME entrepreneurs' innovativeness and personality in the adoption of innovations. Research Policy,
37(9), 1579e1590.
March, J. G., & Simon, H. A. (1958). Organizations. New York: Wiley.
March, J. G., & Sutton, R. I. (1997). Organizational performance as dependent variable. Organization Science, 8, 698e706.
Martinez-Garcıa, C. G., Dorward, P., & Rehman, T. (2013). Factors influencing adoption of improved grassland management by small-scale dairy farmers in
central Mexico and the implications for future research on small holder adoption in developing countries. Livestock Science, 152(2e3), 228e238.
Messerschmidt, C. M., & Hinz, O. (2013). Explaining the adoption of grid computing: An integrated institutional theory and organizational capability
approach. The Journal of Strategic Information Systems, 22(2), 137e156.
Miller, D., & Droge, C. (1986). Psychological and traditional determinants of structure. Administrative Science Quarterly, 31(4), 539e560.
Miller, D., & Friesen, P. H. (1982). Innovation in conservative and entrepreneurial firms: Two models of strategic momentum. Strategic Management Journal,
3(1), 1e25.
Miller, D., Manfred, F. R. K. D. V., & Toulouse, J.-M. (1982). Top executive locus of control and its relationship to strategy-making, structure, and environment.
Academy of Management Journal, 25(2), 237e253.
Mitchell, T. R. (1985). An evaluation of the validity of correlational research conducted in organizations. Academy of Management Review, 10(2), 192e205.
Montazemi, A. R., & Qahri-Saremi, H. (2015). Factors affecting adoption of online banking: A meta-analytic structural equation modeling study. Information
& Management, 52(2), 210e226.
Moore, G. C., & Benbasat, I. (1991). Development of an instrument to measure the perceptions of adopting an information technology innovation. Infor-
mation Systems Research, 2(3), 192e222.
Mowday, R. T., & Sutton, R. I. (1993). Organizational behavior: Linking individuals and groups to organizational contexts. Annual Review of Psychology, 44(1),
195.
Mueller, V., Rosenbusch, N., & Bausch, A. (2013). Success patterns of exploratory and exploitative innovation: A meta-analysis of the influence of insti-
tutional factors. Journal of Management, 39(6), 1606e1636.
Nasco, S. A., Toledo, E. G., & Mykytyn, P. P. (2008). Predicting electronic commerce adoption in Chilean SMEs. Journal of Business Research, 61(6), 697e705.
OECD. (1998). SMEs and e-commerce, ministerial conference on e-commerce. Ottawa: OECD.
Oliveira, T., Thomas, M., & Espadanal, M. (2014). Assessing the determinants of cloud computing adoption: An analysis of the manufacturing and services
sectors. Information & Management, 51(5), 497e510.
Oliver, R. L., & Bearden, W. O. (1985). Crossover effects in the theory of reasoned action: A moderating influence attempt. Journal of Consumer Research, 12(3),
324e340.
Ostrom, T. M., Bond, C. F., Knosnick, J. A., & Sedikides, C. (1994). Attitude scale: How we measure the unmeasurable. In S. Shavitt, & T. C. Brock (Eds.),
Persuasion: Psychological insights and perspectives. Boston, MA: Allyn and Bacon.
Pavlou, P. A., & Fygenson, M. (2006). Understanding and predicting electronic commerce adoption: An extension of the theory of planned behavior. MIS
Quarterly, 30(1), 115e143.
Perrow, C. (1994). Accidents in high-risk systems. Technology Studies, 1(1), 1e20.
Pierce, J. L., & Delbecq, A. L. (1977). Organization structure, individual attitudes and innovation. Academy of Management Review, 2(1), 27e37.
Plouffe, C. R., Vandenbosch, M., & Hulland, J. S. (2001). Intermediating technologies and multi-group adoption: A comparison of consumer and merchant
adoption intentions toward a new electronic payment system. Journal of Product Innovation Management, 18(2), 65e81.
Pontikakis, D., Lin, Y., & Demirbas, D. (2006). History matters in Greece: The adoption of Internet-enabled computers by small and medium sized enter-
prises. Information Economics and Policy, 18(3), 332e358.
Porter, B. E. (2005). Time and implementing change. British Journal of Educational Technology, 36(6), 1063e1065.
Premkumar, G. (2003). A meta-analysis of research on information technology implementation in small business. Journal of Organizational Computing &
Electronic Commerce, 13(2), 91e121.
Premkumar, G., Ramamurthy, K., & Crum, M. (1997). Determinants of EDI adoption in the transportation industry. European Journal of Information Systems,
6(2), 107e121.
Premkumar, G., & Roberts, M. (1999). Adoption of new information technologies in rural small businesses. Omega, 27(4), 467e484.
Quaddus, M., & Hofmeyer, G. (2007). An investigation into the factors influencing the adoption of B2B trading exchanges in small businesses. European
Journal of Information Systems, 16(3), 202e215.
Radjou, N., Prabhu, J., & Ahuja, S. (2012). Jugaad innovation: Think frugal, be flexible, generate breakthrough growth. San Francisco: Jossey-Bass.
Ramdani, B., Chevers, D., & Williams, D. A. (2013). SMEs' adoption of enterprise applications: A technology-organisation-environment model. Journal of
Small Business and Enterprise Development, 20(4), 735e753.
Ramdani, B., Kawalek, P., & Lorenzo, O. (2009). Predicting SMEs' adoption of enterprise systems. Journal of Enterprise Information Management, 22(1/2),
10e24.
Reed, R., & DeFillippi, R. J. (1990). Causal ambiguity, barriers to imitation, and sustainable competitive advantage. Academy of Management Review, 15(1),
88e102.
Riemenschneider, C. K., Harrison, D. A., & Mykytyn, P. P. (2003). Understanding it adoption decisions in small business: Integrating current theories. In-
formation & Management, 40(4), 269e285.
Riemenschneider, C. K., & McKinney, V. R. (2001/2002). Assessing belief differences in small business adopters and non-adopters of web-based e-commerce.
Journal of Computer Information Systems, 42(2), 101e107.
Robertson, T. S., & Gatignon, H. (1986). Competitive effects on technology diffusion. Journal of Marketing, 50(3), 1e12.
Rogers, E. M. (1983). Diffusion of innovations. New York: The Free Press.
Rosenbusch, N., Brinckmann, J., & Bausch, A. (2011). Is innovation always beneficial? A meta-analysis of the relationship between innovation and per-
formance in SMEs. Journal of Business Venturing, 26(4), 441e457.
Rosenthal, R. (1979). The “file drawer problem” and tolerance for null results. Psychological Bulletin, 86(3), 638e641.
Rosenthal, R. (1991). Meta-analytical procedures for social research. Newbury Park, CA: Sage.
Ross, L. (1977). The intuitive psychologist and his short-comings. In L. Berkowitz (Ed.), Advances in experimental social psychology (pp. 173e220). New York:
Academic Press.
Ryan, M. J. (1982). Behavioral intention formation: The interdependency of attitudinal and social influence variables. Journal of Consumer Research, 9(3),
263e278.
Sabherwal, R., Jeyaraj, A., & Chowa, C. (2006). Information system success: Individual and organizational determinants. Management Science, 52(12),
1849e1864.
Saltzer, E. B. (1981). Cognitive moderators of the relationship between behavioral intentions and behavior. Journal of Personality and Social Psychology, 41(2),
260e271.
Schepers, J., & Wetzels, M. (2007). A meta-analysis of the technology acceptance model: Investigating subjective norm and moderation effects. Information
& Management, 44(1), 90e103.
Schumpeter, J. A. (1934). The theory of economic development: An inquiry into profits, capital, credit, interest, and the business cycle. Cambridge, Mass: Harvard
University Press.
Scupola, A. (2003). The adoption of Internet commerce by SMEs in the south of Italy: An environmental, technological and organizational perspective.
Journal of Global Information Technology Management, 6(1), 52e71.
Segaar, D., Bolman, C., Willemsen, M. C., & de Vries, H. (2006). Determinants of adoption of cognitive behavioral interventions in a hospital setting: Example
of a minimal-contact smoking cessation intervention for cardiology wards. Patient Education and Counseling, 61(2), 262e271.
Seyal, A. H., & Rahman, M. N. A. (2003). A preliminary investigation of e-commerce adoption in small & medium enterprises in Brunei. Journal of Global
Information Technology Management, 6(2), 6e26.
Shadish, W. R. (1996). Meta-analysis and the exploration of causal mediating processes: A primer of examples, methods, and issues. Psychological Methods,
1(1), 47e65.
Sheppard, B. H., Hartwick, J., & Warshaw, P. R. (1988). The theory of reasoned action: A meta-analysis of past research with recommendations for modi-
fications and future research. Journal of Consumer Research, 15(3), 325e343.
Shimp, T. A., & Kavas, A. (1984). The theory of reasoned action applied to coupon usage. Journal of Consumer Research, 11(3), 795e809.
Sobel, M. E. (1982). Asymptotic intervals for indirect effects in structural equations models. In S. Leinhart (Ed.), Sociological methodology (pp. 290e312). San
Francisco: Jossey-Bass.
Sophonthummapharn, K. (2009). The adoption of techno-relationship innovations. A framework for electronic customer relationship management. Mar-
keting Intelligence & Planning, 27(3), 380e412.
Srite, M., & Karahanna, E. (2006). The role of espoused national cultural values in technology acceptance. MIS Quarterly, 30(3), 679e704.
Staw, B. M. (1991). Dressing up like an organization: When psychological theories can explain organizational action. Journal of Management, 17(4), 805.
Sultan, F., Farley, J. U., & Lehmann, D. R. (1990). A meta-analysis of applications of diffusion models. Journal of Marketing Research, 27(1), 70e77.
Sutton, S. (1998). Predicting and explaining intentions and behavior: How well are we doing? Journal of Applied Social Psychology, 28(15), 1317e1338.
Tabachnick, B. G., & Fidell, L. S. (1996). Using multivariate statistics. New York: Harper & Collins.
Tashkandi, A. N., & Al-Jabri, I. M. (2015). Cloud computing adoption by higher education institutions in Saudi Arabia: An exploratory study. Cluster
Computing, 18(4), 1527e1537.
Taylor, S., & Todd, P. (1995a). Decomposition and crossover effects in the theory of planned behavior: A study of consumer adoption intentions. International
Journal of Research in Marketing, 12(2), 137e155.
Taylor, S., & Todd, P. A. (1995b). Understanding information technology usage: A test of competing models. Information Systems Research, 6(2), 144e176.
Thompson, J. D. (1967). Organizations in action. New York: McGraw-Hill.
Thompson, S. H. T., Sijie, L., & Kee-hung, L. (2009). Adopters and non-adopters of e-procurement in Singapore: An empirical study. Omega, 37(5),
972e987.
Thong, J. Y. (1999). An integrated model of information systems adoption in small businesses. Journal of Management Information Systems, 15(4), 187e214.
Thong, J. Y. (2001). Resource constraints and information systems implementation in Singaporean small businesses. Omega, 29(2), 143e156.
Thong, J. Y., & Yap, C.-S. (1995). CEO characteristics, organizational characteristics and information technology adoption in small businesses. Omega, 23(4),
429e442.
Tidd, J. (2001). Innovation management in context: Environment, organization and performance. International Journal of Management Reviews, 3(3).
Tornatzky, L. G., & Klein, K. J. (1982). Innovation characteristics and innovation adoption-implementation: A meta-analysis of findings. IEEE Transactions on
Engineering Management, 29(1), 28e43.
Triandis, H. C. (1979). Values, attitudes, and interpersonal behavior. In Nebraska symposium on Motivation: Beliefs, attitudes, and values (pp. 195e259).
Lincoln, NE: University of Nebraska Press.
Tung, L. L., & Rieck, O. (2005). Adoption of electronic government services among business organizations in Singapore. Journal of Strategic Information
Systems, 14(4), 417e440.
Unsworth, K. L., Sawang, S., Murray, J., & Sorbello, T. (2009). Developing an integrative model for understanding innovation adoption. Academy of Man-
agement Proceedings, 1e6.
Uzoka, F.-M. E., & Ndzinge, T. (2009). Empirical analysis of biometric technology adoption and acceptance in Botswana. Journal of Systems and Software,
82(9), 1550e1564.
Van de Vrande, V., De Jong, J. P., Vanhaverbeke, W., & De Rochemont, M. (2009). Open innovation in SMEs: Trends, motives and management challenges.
Technovation, 29(6), 423e437.
Viswesvaran, C., & Ones, D. S. (1995). Theory testing: Combining psychometric meta-analysis and strucutural equations modeling. Personnel Psychology,
48(4), 865.
Vroom, V. H. (1864). Work and motivation. New York: John Wiley & Sons.
Wagner, J. A., & Gooding, R. Z. (1987). Effects of societal trends on participation research. Administrative Science Quarterly, 32(3), 241e262.
Wang, M. C., & Bushman, B. J. (1999). Integrating results through meta-analytic review using SAS software. Cary, NC: SAS Institute Inc.
Wanyoike, D. M., Mukulu, E., & Waititu, A. G. (2012). ICT Attributes as determinants of e-commerce adoption by formal small enterprises in urban Kenya.
International Journal of Business and Social Science, 3(23), 65e74.
Warshaw, P. R., & Davis, F. D. (1985). Disentangling behavioral intention and behavioral expectation. Journal of Experimental Social Psychology, 21(3),
213e228.
Weigel, F. K., Hazen, B. T., Cegielski, C. G., & Hall, D. J. (2014). Diffusion of innovations and the theory of planned behavior in information systems research: A
meta-analysis. Communications of the Association for Information Systems, 34(1), 31e43.
Welsh, J. A., & White, J. F. (1981). A small business is not a little big business. Harvard Business Review, 59(4), 18e27.
Weng, M.-H., & Lin, C.-Y. (2011). Determinants of green innovation adoption for small and medium-size enterprises (SMES). African Journal of Business
Management, 5(22), 9154e9163.
West, J., & Bogers, M. (2014). Leveraging external sources of innovation: A review of research on open innovation. Journal of Product Innovation Management,
31(4), 814e831.
Wolfe, R. A. (1994). Organizational innovation: Review, critique and suggested research. Journal of Management Studies, 31(3), 405e431.
Wu, I.-L., & Chuang, C.-H. (2010). Examining the diffusion of electronic supply chain management with external antecedents and firm performance: A multi-
stage analysis. Decision Support Systems, 50(1), 103e115.
Xu, J., & Quaddus, M. (2012). Examining a model of knowledge management systems adoption and diffusion: A partial least square approach. Knowledge-
based Systems, 27(1), 18e28.
Zaltman, G., Duncan, R., & Holbek, J. (1973). Innovations and organizations. New York: John Wiley & Sons, Inc.

1 s2.0 S2096248717300012 Main

Uploaded by

Copyright:

Available Formats

1 s2.0 S2096248717300012 Main

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1 s2.0 S2096248717300012 Main

Uploaded by

Copyright:

Available Formats

International Journal of Innovation Studies 1 (2017) 107e133

Contents lists available at ScienceDirect

International Journal of Innovation Studies

Innovation attributes and managers' decisions about the

Fig. 1. Mediator-moderator meta-analytic framework.

2.4. Empirical studies on the attributes of innovations-adoption decisions linkages in organizations

3.1. Database development and inclusion criteria

Key Construct Deﬁnitions Sample measures

Key Construct Deﬁnitions Sample measures

3.2. Variables coded

3.3. Univariate analyses

3.4. Structural equation modeling

4.1. Description of the sampled studies and correlations between constructs

G. Vagnani, L. Volpe / International Journal of Innovation Studies 1 (2017) 107e133

Attitude Norm Behavioral control

Hierarchical position Organizational size

Adoption measures Adopion measures Measures of dependent/Indipendent

4.6. Additional analyses

5.1. Our ﬁndings in the context of past meta-analytical works

5.2. Limitations of this meta-analysis

5.3. Implications for practice

5.4. Future directions for research

You might also like