Article Review

Locate a peer-reviewed article that discusses research in an organization(see attached).  Topic ideas include the role of research, development of research studies, integration of different types of methods, or the communication of research results within an organization. Note that these are ideas; please expand within the parameters of research used within organizations. Respond to the following questions/topics: 

1. Summarize the article with an eye on the author’s main point. 

2. How does this article contribute to contemporary thinking about research? 

3. How does this article illustrate the importance of using research to make decisions within an organization? 

4. How can information in this article be applied to your field?

5. What is your opinion on the topic of this article? 

Your APA-formatted response must be a minimum of three pages in length (not including the title page and the reference page). Your sources must be peer-reviewed. All sources used must be referenced; paraphrased and quoted material must have accompanying citations. 

Q Academy of Management Review 2017, Vol. 42, No. 4, 577–595.

2016 Decade Award Invited Article


GARY JOHNS Concordia University and University of British Columbia

This is a reflection on my 2006 article, “The Essential Impact of Context on Organiza- tional Behavior,” which received the 2016 Academy of Management Review Decade Award. I review some studies supportingmy earlier contention that the impact of context has been underappreciated in management research and then recount the genesis of the article, particularly emphasizing the capacity of context to explain anomalous and counterintuitive research findings. I offer conjectures as to why the article has been cited and present evidence that contextual appreciation is increasing in management, and that this is part of a general trend in the social and behavioral sciences. I discuss some newer theories and measures of context and consider the desirable properties of theo- ries that incorporate context. Finally, I argue that it is not easy to control away context, that context is about similarities as well as differences and about change as well as stability, and that variables and relationships vary in their sensitivity to context.

The purpose of this article is to offer some re- flections on “The Essential Impact of Context on Organizational Behavior” (Johns, 2006), the re- cipient of the 2016 AMR Decade Award. I was surprised to learn that the context article had re- ceived the award, because its contribution to theory is less traditional in approach than that of most AMR articles, and one reviewer disliked the manuscript in all of its iterations. Furthermore, the subject of the article is not my main scholarly preoccupation. However, I had visited the topic of context earlier (Johns, 1991, 1993, 2001), and I thought I had something more to say on the mat- ter, even though that initial 1991 effort was first rejected by AMR, to my distinct displeasure!


The basic premise of the 2006 article was that the impact of context on organizational behavior is underrecognized and underappreciated. I rather broadly defined context as situational or environmental constraints and opportunities that have the functional capacity to affect the occur- rence and meaning of organizational behavior. Contextual stimuli can be located at, above, or below a focal level of analysis and can operate as main or moderator effects. I offered several ways

of thinking about context, including context as salient situational features, situational strength, cross-level effects, configurations of stimuli, en- vironmental events, situational shapers of meaning, and a fairly constant ambient back- ground factor. The point of the article was not that context had never been studied before. Rather, it was that it should be incorporatedmoremindfully and systematically into our research. In the article I argued that it is helpful to think

about context as operating at a broader, more gen- eral, more distal level (omnibus context) and a nar- rower, more specific, more proximal level (discrete context), with the latter usually roughly nested un- der the former and serving as a mediator of more distal effects. I suggested that the journalistic im- peratives to report who, what, when, where, and why exemplify omnibus context. Drawing on social and environmental psychology, I asserted that dis- crete context comprised task, social, and physical stimuli. I proposed that more important aspects of these context dimensions are theoretically perva- sive and operate at multiple levels of analysis. For instance, the task variables uncertainty, autonomy, and interdependence appear in a variety of man- agement theories and have been applied at levels ranging from individuals to industries. Despite this, there are fewer theories of such variables (Whetten, 2009) and even fewer literature reviews focusing on such variables than one might expect. And if they are not explicitly modeled and measured, they are

I thank Blake Ashforth, Aparna Joshi, and Sharon Parker for comments that improved this article.

577 Copyright of the Academy of Management, all rights reserved. Contents may not be copied, emailed, posted to a listserv, or otherwise transmitted without the copyright holder’s express written permission. Users may print, download, or email articles for individual use only.

often simply ignored, despite their proven influence on organizational behavior.

In what I consider to be the core of the article, I provided a number of examples of how context affects organizational behavior and related re- search results by restricting range, determining base rates, reversing causality, reversing signs, prompting curvilinearity, and tipping relation- ships. A key argument was that many anomalous research findings can be explained when context is taken into consideration. Unrecognized, context effects can threaten internal validity, challenge external validity, and limit the application of management research. Recognized, such effects can identify boundary conditions for theories, of- fer opportunities to enlarge the scope of theories, and facilitate research application.

The article concluded with some ways to con- textualize research (see also Rousseau & Fried, 2001) and its reportage, along with relevant ex- amples. Indesign terms, I encouraged cross-level, comparative, and qualitative research, as well as the study of processes and events. I also encour- aged the provision of more qualitative data in conjunction with otherwise quantitative designs, particularly to facilitate future meta-analyses. In terms of measurement and analysis, I placed special emphasis on the choice of dependent variables, with “more and varied” being the pre- scription. The general idea of this is to identify variables that are differentially susceptible to contextual opportunities and constraints. This prescription contrasts with disciplinary conven- tions to dwell on a rather limited range of de- pendent variables (as in strategy) or to study a plethora of seemingly similar dependent vari- ables one by one, in isolation (as in organizational behavior). I discouraged the cavalier designation of contextual features as control variables, a point I will reemphasize in this article. Finally, I encouraged better reporting of context, particu- larly in the omnibus domains of who or what is studied and when, where, and why the research is conducted.


During the review process, I was fortunate that no one asked for proof that context had been neglected in the organizational sciences, since I did not really have any data to this effect. Since then, however, some empirical evidence has emerged for the contention. Gorgievski and

Stephan (2016) examined 142 articles concerning the psychology of entrepreneurship published between 2000 and 2015. Only eight articles ex- plored context at any level of analysis. A compi- lation of 373 articles on leadership published between 1990 and 2005 concluded that only 16 percent “took into account the organizational context to at least a moderate extent” (Porter & McLaughlin, 2006: 561), and a review of 52 articles concerning knowledge sharing concluded that 31 were “context free” and only 7 were decidedly context aware (Sergeeva & Andreeva, 2016). Cronin, Weingart, and Todorovic (2011) enu-

merated the articles on workgroups and teams published in six prominentmanagement journals in 2010. Only 16 percent of the group-level con- structs studied were contextual (e.g., resource levels, presence of intergroup competition), and over half of these were employed as control vari- ables rather than employed substantively. Simi- larly, a review of team diversity research concluded “overwhelmingly . . . contextual vari- ableswere considered as control variables rather than directly incorporated into study hypotheses” (Joshi&Roh, 2007: 29).Maloney, Bresman, Zellmer- Bruhn, and Beaver (2016) rated the richness of the contextual descriptions included in 271 teams articles published between 2004 and 2013. Only 25 percent provided richdescriptions that addressed the social, physical, organizational, and in- dustrial nature of the research settings, while 29 percent provided no contextual information at all. The remaining articles simply provided the geo- graphic location of the research site. Given all of this, it is fair to conclude that “much extant research treats groups as closed systems” (Kouchaki, Okhuysen, Waller, & Tajeddin, 2012: 171). Although these findings present a rather neg-

ative picture, it must be remembered that they are necessarily looking backward rather than for- ward. As things progress, Iwill present additional evidence that changes have occurred in the ap- preciation of context in and beyond the manage- ment discipline. These changes should be reflected in such reviews in the future.


Over the years, as a reviewer of manuscripts, I had been struck by the frequency with which re- searchers attempted to test hypotheses that could almost certainly not be confirmed given the

578 OctoberAcademy of Management Review

apparent distribution of at least one of the vari- ables in question. Technically, this is a problemof restriction of range, in which there is not enough variance in a variable to detect an underlying true-score association with another variable. In essence, researchers are studying things that don’t exist, because they have sampled from an inappropriate context. It is tempting to write this off as a methodological error and something that should have been resolved before the data were even collected (see Aguinis & Vandenberg, 2014). However, I began to discern that constraints on behavior (which lead either to direct restriction of range or to an inappropriate range of response to address the question at hand) were equally a property of substance, not just methodological choice, and were often inherent in the context chosen for a study (Johns, 1991). Hence, trying to validate the existence of a need hierarchy using respondents from a single occupation is probably a methodological misstep, but the collected data will still accurately reflect the reality of occupa- tional constraints and opportunities for that sample, such as they are. The bottom line here is that the interplayamong theory, data, andmethod cannot be avoided (Van Maanen, Sørensen, & Mitchell, 2007).

At about this same time, I read a book chapter that cited a survey indicating awork absenteeism rate of 14 percent in Italy and 1 percent in Swit- zerland, Italy’s next-door neighbor (Steers & Rhodes, 1984). Subsequently, I read a meta- analysis of the relationship between job perfor- mance and turnover that recorded turnover rates ranging from 3 percent to 106 percent for the an- alyzed samples (McEvoy & Cascio, 1987). These data suggested strong respective effects for na- tional and organizational context. As an absen- teeism researcher struggling to account for an additional percentage point of variance in individual-level absence, I found such effects most impressive, and I knew that the Swiss were not thatmuchhealthier than Italians. This insight, of course, required a shift in level of analysis, as an appreciation of context often does, with the caution that collective rates of behavior are not reflecting the same pool of variance as individual behavior. Nevertheless, I saw the value of being attentive to levels effects (Hackman, 2003) some- what before it became fashionable, and this set the stage for thinking about context. One concrete outcome of this was the development and re- finement of the absence culture construct

(Nicholson & Johns, 1985), which specifies how individual employee absenteeism levels are markedly determined by the social context, in- cluding workgroup peer behavior, occupational norms, social class dynamics, andmanifestations of national culture. This construct complemented the prevailing individual-level ethos that absen- tees were ill, unethical, or job dissatisfied. During this period, a number of surveys

appeared in the literature documenting the fail- ure of organizations to adopt various scientifi- cally validated human resources (HR) practices. It seemed to me that this paradox could be explained in part by applying the massive liter- ature on innovation to the context in which such practices are actually implemented. In particular, I argued that scholars tend to frame valid HR practices as technical innovations while man- agers frame them as administrative innovations (Johns, 1993). Practices associated with the latter are more likely to be seen as matters of manage- rial style and taste, offering more freedom of choiceandmoreopenness tononexpert influence. All of this induces uncertainty, allowing contex- tual factors such as institutional forces, govern- ment regulation, organizational politics, and management fads and fashions to overwhelm technical merit as a determinant of HR practice adoption (cf. Birkinshaw, Hamel, &Mol, 2008). The overarching point was that not understanding the context in which organizational innovations are enacted contributes to the so-called relevance gap or science-practice gap for management research. Perhaps the most impactful factor that shaped

the genesis of the Decade Awardmanuscript was my collection, over the years, of a number of arti- cles revealing (sometimes extremely) counterin- tuitive results that appeared to be attributable to contextual factors. These results included dra- matic sign reversals, reversals of causality, and reversals of conventional scholarly wisdom that were understandable when put in context (Johns, 2001). Thus, workers can feel more controlled un- der self-management than under bureaucratic control (Barker, 1993), convenience stores with friendlier personnel can have lower sales (Sutton & Rafaeli, 1988), employees can choose to work off-site when they don’t really want to do so (Rockmann & Pratt, 2015), and more domain ex- perts on corporate boards can increase the chan- ces of organizational failure (Almandoz & Tilcsik, 2016).

2017 579Johns

Similar to the investigative journalism dictum to “follow the money,” such anomalies and para- doxes are signals to look for the operation of context, and they often constitute “empirical mysteries” that can serve as a basis for theorizing (Alvesson & Kärreman, 2007), in the spirit of the recently established Academy of Management Discoveries. Consider the supply chain position paradox (Schmidt, Foerstl, & Schaltenbrand, 2017). Schmidt and colleagues correctly pre- dicted that green supply chain practices would facilitate firm market and financial performance, and that such practices would be most common lower in the supply chain. However, they in- correctly anticipated that the link between green practices and performance would be stronger lower in the chain; in fact, upstream firms profited more from going green, despite their general ret- icence to do so. Applying an explicitly contextual frame, the authors were able to theorize an ex- planation for the paradox based on factors such as stakeholder attention, green practice maturity, and economies of scope.

Sign reversals are among the most common research anomalies. Weyman and Clarke (2003) studied the degree of perceived danger of various questionable working practices among English underground coal mining personnel, finding that those who were physically and experientially closer to the actual work (e.g., miners, as opposed to managers) were most likely to view the prac- tices as risky. It is easy enough to concoct a theory to explain these results, but it is much harder to simultaneously incorporate the results of Östberg (1980), whose study Weyman and Clarke were replicating. Studying Swedish forestry personnel, Östberg found a distance-danger gradient pre- cisely the opposite—those closer to the work found the various questionableworking practices less risky! Although the reason for this particular sign reversal is not clear, careful analysis of the context will often yield explanatory dividends. For instance, Bresman and Zellmer-Bruhn (2013) found that more organizational structure gener- ally hurt the external learning of self-managed pharmaceutical teams but facilitated learning when the teams themselves lacked structure. Similarly, in a recent meta-analysis concerning genderdifferences innegotiation, the researchers concluded that while, on average, men obtained better economic outcomes thanwomen, this effect was actually reversed to favor women when the negotiation contextwas role congruent forwomen

(i.e., negotiating for another person, bargaining range specified; Mazei et al., 2015). This finding supports Yoder and Khan’s (2003) contention that contexts themselves are gendered and that established gender effects can be greatly modi- fied by context. All three of these examples of sign reversals

suggest how an appreciation of context can de- marcate theoretical boundary conditions con- cerning underlying processes (Busse, Kach, & Wagner, in press), but they simultaneously invite us to incorporate context into our theories (Bamberger, 2008; Maloney et al., 2016), thus en- larging their scope. For a thorough discussion of such sign reversals, see Cavarretta, Trinchera, Choi, and Hannah (2016). These rather dramatic opposite effects should

not obscure the fact that context often operates in a more nuanced way. For instance, Latham and Erez had to conduct a meticulous series of repli- cations to uncover how subtle contextual varia- tions led one of these researchers to conclude that participation in goal setting enhanced goal com- mitment while the other researcher did not (see Latham, Erez, & Locke, 1988). Sometimes, salient situational features have little impact because their effects are countervailed by other features that comprise context. In other cases, apparently minor changes in context have big effects be- cause of tipping mechanisms (Johns, 2006). For example, Konrad, Kramer, and Urkut (2008) de- scribed how the dynamic of corporate boards changes when the number of women directors reaches or exceeds the critical mass of three.


The AMR editor asks the authors of Decade Award articles to reflect on the reasons these ar- ticles have been cited. In the spirit of the subject matter at hand, one reason is, well . . . context! That is, the 2006 article appeared at an opportune time, corresponding to a trend away from uni- versalism and toward a more nuanced and con- tingent view of natural and social phenomena. In addition, this trend has occurred across disci- plines, underlining the fact that context is rele- vant to a wide range of scholarly pursuits. In his bookCosmic Jackpot:WhyOurUniverse Is

Just Right for Life, the physicist Paul Davies pro- poses that “the laws of physicsmight be just local by-laws” (2007: 166). By thishemeans that the laws

580 OctoberAcademy of Management Review

of physics might vary over the universe or over time, rather than applying universally. Many other disciplines and areas of study outside of management have been experiencing calls for greater attention to context, including soft- ware engineering and development (Clarke & O’Connor, 2012), information systems (Davison & Martinsons, 2016; Hong, Chan, Thong, Chasalow, & Dhillon, 2014; Venkatesh, Thong, & Xu, 2016), information and communication technology (Senarathne Tennakoon, da Silveira, & Taras, 2013), supply chain management (Schmidt et al., 2017), gender studies (Yoder & Kahn, 2003), com- munity psychology (Calvard, 2015), social psy- chology (Reis, 2008), personality psychology (Rauthmann, Sherman, & Funder, 2015; Rentfrow, 2010), positive psychology (McNulty & Fincham, 2012), intelligence (Sternberg, 2004), cognition (Smith & Semin, 2004), memory and aging (Hess, 2005), immunology (Morey, Boggero, Scott, & Segerstrom, 2015), health (Short & Mollborn, 2015; Weibe, Helgeson, & Berg, 2016), safety (Rosness, Blakstad, Forseth, Dahle, & Wiig, 2012), criminol- ogy (Smith, Torstensson, & Johansson, 2001), and public administration (O’Toole & Meier, 2015). Special sections and issues of Social Science & Medicine (Placing Health in Context, November 2007), American Psychologist (Geography and Psychology, September 2010), European Journal of Personality (European Personality Reviews, May/ June 2015), Journal of Personality (Contextualized Identities, December 2007), Human-Computer In- teraction (Context-Aware Computing, issues 2–4, 2001), and Journal of Information Technology (De- bates and Perspectives, September 2016) signal the thrust of this phenomenon. Given all this, it is not surprising that the management discipline has also exhibited increased interest in context, a point that will be documented below. To its credit, our discipline, in fact, served as a bell- wether in this regard, with prescient attention de- voted to issues concerning context (Cappelli & Sherer, 1991;Hattrup& Jackson,1996;Heath&Sitkin, 2001; Mowday & Sutton, 1993; Rousseau & Fried, 2001).Muchof thisworkmakes thepoint that such attention should indeed be a core competence of the management discipline.