Knowledge Synthesis for Scientific Management: Practical Integration for Complexity Versus Scientific Fragmentation for Simplicity

A Brief History of the Scientific Management Research

Historically, knowledge about management was generated inductively, based on insights from practitioners and concerned with concrete, real management problems in “the world of practice” (see Taylor, 1911). Before business schools were created, management knowledge was generated primarily outside of the scholarly community; developed by businessmen; disseminated in non-refereed reports, books, and popular press publications; and taught in vocational and trade schools. This kind of craft knowledge suffers from what Herbert Simon (1967) called a “dangerous illusion that good business teaching consists of ‘telling the boys how I did it’” (p. 7). To foster academic teaching in business, many universities developed collegiate business schools in the late 19th (e.g., Wharton, Haas [Berkeley], St. Gallen) and the beginning of the 20th century (e.g., Harvard, Kellogg, Turk [Dartmouth]). After World War II, in response to the growing number of questions regarding the quality of incidences, cases, and descriptions of existing practices as the primary knowledge sources, a major mission of management education and scholarship has been to introduce more scientific and discipline-based theories, methodologies, and quality scientific results into the management (and the broader business) field. These efforts have played a significant role in building a rigorous scientific climate for business research, as we explain below.

Although management is partially an art and practice based on emergent, intuitive, and craft knowledge (Mintzberg, 2004; Simon, 1967), a strong demand was developed in academia for creating scientific and peer-reviewed publication outlets after World War II, led by the Association to Advance Collegiate Schools of Business (AACSB; 2019) and later reinforced by Gordon and Howell (1959) and Pierson (1959). This movement was reinforced (or initiated) by the skepticism of nonbusiness academic peers about the business school’s academic worthiness, with a significant focus on making a unique theoretical contribution (Hambrick, 2007; Porter & McKibbin, 1988). In this context, many peer-reviewed scientific journals were created, all of which require that every article make a unique contribution to theory (Hambrick, 2007). ²

Scientific Journals in Management Studies

The business knowledge ecosystem can be identified primarily as a collection of three interconnected communities: teaching and education institutions, the academy of scholars, and the world of managers and practitioners, as depicted in Figure 1. First, teaching and education institutions’ end goal is to help students achieve learning objectives through instruction and mentoring. They are responsible for developing educational programs to prepare students to become impactful and responsible citizens. Second, the academy of scholars often overlaps with teaching and education institutions because both are often hosted within the universities or colleges but differ in their end goals. The end goal of scholars, as Simon (1967) suggested, is to create knowledge, through means of research, driven primarily by their curiosity and identification of research needs in their field. This community could be hosted outside the teaching and education institutions (e.g., Center for Creative Leadership, McKinsey Knowledge Center; Microsoft Research; IBM Research) and often has its own communities (e.g., Academy of Management [AOM1]; Institute for Operations Research and the Management Sciences [INFORMS]), which are independent of their members’ institutional affiliations. The third is the world of practice. This is the source of many students (e.g., MBAs and executive programs), and ultimately, most students at both undergraduate and graduate levels are destined for practice after graduation. In this community, as Simon (1967) suggested, the end goal of management is to achieve specific performance metrics through practice.

OPEN IN VIEWER

Figure 1.

Scientific management journals and knowledge flows in business knowledge ecosystem.

These three communities are interrelated because one’s output is an input into the other. For instance, knowledge is a major output for the academy of scholars, but is an input into teaching activities; student learning is the output of education, but is an input as human capital for the world of practice; performance is the output of the world of practice, but the underlying drivers and processes are an input into research as observation/testing fields, research data, and occasionally a source of research funding. Facilitated especially by libraries and databases (e.g., Web of Science, Google Scholar, and other media to find information and knowledge), all communities have access to the same pool of knowledge.

In this view of the science-practice ecosystem, scientific journals seek to motivate management faculty to focus on rigorous research based on theory and evidence and to ensure that the research is consonant with developed knowledge from foundational disciplines such as economics, psychology, behavioral sciences, sociology, political science, and statistics (Agarwal & Hoetker, 2007). In this way, these journals are central to a vast knowledge network that connects cutting-edge knowledge in established disciplines (e.g., refereed discipline publications), management research, teaching, and management applications composing practice. We depict this knowledge network in Figure 1, in which the arrows show knowledge flows. In the figure, the two gray circles (“Management Journals” and “Collegiate Business Schools”), the white circle (“Libraries/databases”), and the thick solid arrows between them show the knowledge flows related to business schools, that is, from foundational disciplines (e.g., refereed discipline publications) into management journals and then into a common pool of libraries and databases, from which collegiate business schools draw knowledge for teaching and application problems.

Certainly, the three communities do not always follow the linear arrangements described above. There are additional linkages between knowledge and practice—the two ends of management knowledge ecosystem, some of which bypass the journals. For instance, many scholars, including some of the field’s most influential minds (e.g., Michael Porter and Clayton Christensen), write popular books for general audiences or serve as consultants for business clients. But to create rigorous and generalizable knowledge, their works are highly integrated with journal publications.

Between Management and Foundational Disciplines

The first disruption of connections is between management knowledge and foundational disciplines. We examined the hypotheses (and propositions) in and citations (i.e., reference lists) of papers published in two sample volumes (1990 vs. 2014) of the nine journals noted earlier (see a discussion of the method and review in Supplement 2). The findings are mixed (see Supplement 2 Figure 1). On one hand, as a proxy for knowledge inflows, citations drawn from source journals by articles in these nine management journals that came from nonmanagement/nonbusiness disciplines slightly increased (38% in 1990 vs. 39.5% in 2014). On the other hand, about 66% of these citations were to articles published more than 10 years earlier (i.e., older than 2004), translating into only 9.07% of references in nine journals that were drawn from relatively current (less than 10 years old) knowledge from these source journals outside of management/business. ³

Although the classic works in many social sciences are important, there are areas in which we need to constantly channel knowledge synthesis (e.g., reviews, meta-analysis) from these disciplines to ensure that some of the premises in our management theories are constantly updated by more developed theories and rigorous techniques from these disciplines. Foundational disciplines have a longer history of scientific development and often have developed a complex network of branches, each of which is highly specialized and hosts its own peer-reviewed journals, as well as more knowledge synthesis efforts (e.g., reviews, meta-analysis). ⁴ Built on a longer history, a larger number of concurrent specialized branches, and a larger number of PhDs and researchers, scholars in foundational disciplines may have more established tools and experience more competitive intensity to publish new knowledge.

We need to balance the emphasis on consolidation within the management field (Agarwal & Hoetker, 2007) and timely integration of knowledge synthesis (e.g., reviews, meta-analytical evidence) from foundational disciplines (Daft & Lewin, 1990, p. 1). On one hand, after decades of knowledge development, research in the management field has increased in sophistication and quality (Agarwal & Hoetker, 2007). Thus, it has begun building and consolidating its own knowledge base, which becomes increasingly distinct from the foundational disciplines but yet has many of its roots in these disciplines (Daft & Lewin, 1990, p. 1). And, with the enhanced knowledge development in management, work published in the areas relevant for management may be less foundational than in prior years. ⁵ On the other hand, regularly integrating the knowledge consolidations (e.g., reviews, meta-analytical evidence) from the foundation disciplines can help us address some of the unquestioned premises of management theories. ⁶ In addition, not examining the recent knowledge syntheses in foundational disciplines may cause management scholars to overlook theoretical extensions and conflicts and, thereby, perpetuate outdated and even bad theory from these disciplines. ⁷

Within Management Studies

The second breakdown of the connections is among different knowledge domains within the management field. An important cause for this interruption is the pressures from scientific journals for theoretical (often reduced and disciplinary) coherence in each paper published, rather than a more holistic understanding of the management concerns regarding the topic on which the research focused. Litchfield (1956) suggested in the first issue of Administrative Science Quarterly (ASQ) that

most of the new thought has come from the fields of mathematics, engineering, anthropology, sociology, or some [forms] of the emerging behavioral sciences [. . . but] these additions to our knowledge have been concerned with selected parts of administration and not with the whole. (p. 4)

Prioritizing theories as a major contribution (Hambrick, 2007), often grounded in a specialized and disciplinary approach, and trying to understand complex social challenges such as management are to reduce an irreducible complex whole into more coherent, partial, and yet biased pieces (e.g., the micro–macro divide) (Aguinis, Boyd, Pierce, & Short, 2011; Hitt, Beamish, Jackson, & Mathieu, 2007).

As an unintended consequence of rewarding theoretical coherence and common sense by editors, reviewers, and contributors, our leading journals tend to display biases toward specific types (and levels) of analysis and disciplines over the others (Aguinis et al., 2011). For individual papers, scholars often find it easier to ensure coherence by drawing logics from a single discipline and avoiding inherent conflicts between some disciplines (and thus avoiding reviewers from conflicting theoretical camps) (see a summary of our citation analysis in Supplement 2).

However, the practice of management across all levels of analysis can only be sufficiently understood by integrating simultaneous multiple disciplinary foci, including those that draw on conflicting assumptions of human behaviors and social contexts (Richerson & Boyd, 2005). A relevant example regards the assumptions underlying corporate governance relationships. The neo-economic approach largely draws on agency theory, which assumes self-interest and moral hazard to analyze the managerial behaviors (Ghoshal, 2005), whereas many scholars following sociological and psychological approaches gravitate toward a model that assumes subordinates to be “collectivists, pro-organizational, and trustworthy” (Davis, Schoorman, & Donaldson, 1997, p. 20). Because humans are inherently both economic and social (Lindenberg, 1990; Wrong, 1961; Wynn, 2008), managers need to entertain both economic interests and social relationships. As a result, neither of the above views alone is likely to provide a full understanding of the corporate governance context and outcomes.

Also related to disciplinary silos is the potential lack of neutral views of the research subjects, because researchers’ logics and tools are bounded by discipline/theory-specific values and assumptions. In organizations, general management practice needs to align diverse and pluralistic interests of multiple constituencies, several of which may have conflicting values, logics, rationales, and/or perspectives (Mitchell, Weaver, Agle, Bailey, & Carlson, 2016). But to have a coherent logic in their research, researchers are often trained to simplify their focus by discipline or to have a particular theoretical focus within that discipline, and to avoid such pluralism ensuring that it is relevant to and coherent with their own training. Yet, other views may be salient to address the research question (Campbell, 2004). ⁸ Bounded values and assumptions provide the contexts within which to understand, apply, and interpret a theory (and a discipline). Without a common framework to specify, compare, and synthesize these bounded values and assumptions of alternative theories underlying a complex management issue, there is a risk of overgeneralizating a theory outside its contexts (Edwards & Berry, 2010). ⁹

(Re)Inventing Technologies

First, the lack of (or the difficulty in) knowledge integration is a generic problem in perhaps all knowledge fields. One fundamental challenge is the increasing volume and complexity of information and the relatively limited research capacity of humans. A research project typically starts with a thorough literature review, and for any research subject, the set of existing studies can be quite large. ¹⁰ One may argue that we do not need to read the entire set of papers to capture the essence, if the goal is to understand the main story in each paper about cause-and-effect relationships (or of processes). The time for capturing the essence can still be daunting. ¹¹ In reality, no researchers are likely to take so many years to thoroughly review these studies, but instead will limit their focus to a shorter list of particular journals in certain discipline(s) while ignoring other areas (including foundational disciplines) or subordinating them to a secondary role as needed. As a result, new publications remain fragmented and only distantly connected to the more holistic picture.

We suggest that there is a need to shift our paradigm from “human agents” as researchers to human–machine cooperation with “intelligent assistants” (Detlor, 2004; Hitt, 1998). We do not recommend replacing human researchers with machines, but introducing or developing “intelligent” assistants for human researchers to more rapidly and precisely identify and organize scientific literature in management studies. For artificial intelligence (AI), such as IBM Watson, it will require considerably less time, involving much more comprehensive work, such as connecting similar concepts, analyzing logics, generating new hypotheses, and problem solving. ¹² Several machine reading techniques have been developed in the domains of natural sciences to speed up knowledge discovery and integration from scientific papers, such as the Big Mechanism program funded by Defense Advanced Research Projects Agency (DARPA), seeking to automate the extraction and integration of causal relations from scientific papers as well as hypothesis development (Cohen, 2015). This program has yielded several promising machine reading tools such as Reach (Valenzuela-Escárcega et al., 2018), which is being commercialized by lum.ai to automate the machine reading of scientific papers for actionable knowledge discoveries. These technologies, however, have not been sufficiently integrated into or utilized by management studies.

Relatedly, these technologies, once introduced into management studies, can be customized to assist editorial work as well. Editors are the key players in shaping a knowledge field by designing the agenda, organizing the community of editors and reviewers, and cultivating the ways in which scholars do research (Davis, 2014). One important judgment they regularly make is whether they should certify a particular paper for publication, using the criteria and standards of the journal. As illustrated above, “intelligent assistants” can help a human editor (and a reviewer) more quickly conduct a “fair hearing” based on a thorough analysis on all of the prior publications and the relevant relationships. In reality, it is very difficult for a human editor (or reviewer) to do this, similar to the challenges experienced by the researchers. The judgment often is confined within a particular field or a given set of major publications. ¹³

We need to enrich our publications using advanced text mining techniques. For example, we need a shared ontology of constructs, labels, descriptions, and examples of expressions, combined with text analytics algorithms, through which scholars across fields can translate their own academic languages. One practical challenge is that advancing these types of technology would require close collaboration between management scholars and information technology (IT) experts, who typically do not target the same academic peers, audience, or outlets. We notice a few initiatives (although at early stages) trying to address this challenge. One notable example is theorizeit.org, which aims to integrate the behavioral science constructs, assisted by natural language processing (NLP) algorithms. This project has developed a shared search engine to group similar constructs and descriptions. However, the application of it into a broad range of management disciplines remains untested.

A more proactive way is to redesign the format of our papers and journal submissions to be more machine-friendly. In addition to regular manuscripts, contributors can submit a three-dimensional (3D), visualized causal map (3D model) of variables demonstrating causal relations and processes; the model can be reviewed by both human editors and computers and prepared for comparisons and integration with other 3D models (submitted by other researchers). The descriptions of all variables and the relationships between them should be explicit, logical, and clear enough to enable machine coding and programming. Importantly, this approach does not restrict the model to only quantitative studies; rather, conceptual and qualitative studies, which offer constructs as well as their logic relations, can and should be included. In Figure 2, we demonstrate an example of such 3D models for integrating two submissions, each of which follows a different theoretical logic (denoted as logic red and logic blue). ¹⁴

OPEN IN VIEWER

Figure 2.

Three-dimensional submissions and integration.

(Re)Inventing Outlets

Another constraint relates to knowledge and its outlets. Only a few outlets are of high status based on their contribution to novel scientific knowledge in the field. If we did more review syntheses, it might matter less where something was published as long as it was peer reviewed by quality researchers. There is a lack of diversity in the outlets for impactful scholarly publications and knowledge sharing, which discourages efforts focused on replications, synthesis, and tests for generalizations in different samples and contexts. Thus, with brand names, financial resources, and public media ranking, elite schools/universities attract the brightest PhD students and faculty to focus on publishing in these scientific journals. Meanwhile, all PhD students and faculty in academia compete to excel in this process to join the elite institutions as their ultimate career goal. All other business schools experience peer pressure and try to imitate these elite institutions, which reduce their innovations and diversity (Glick, 2008). This systematic pressure reinforces the conformity in the entire academic community to scientific work, repelling other types of scholarship.

While there is no need for more journals, books, or publishing firms, or even more non-peer-reviewed publications, different media can be developed. All traditional media suffer from the problems experienced by human readers (editors and reviewers), such as lack of capabilities for processing a complete review of all information (as discussed above), gatekeeping of belief systems, and biased attention to “counterintuitive” findings. Rather, more creative thought is needed about what other forms of outlets could facilitate the daunting work of knowledge accumulation and synthesis.

One potential mode is a knowledge synthesis portal, which has been widely used in the industry (e.g., the IBM Global Services K Portal) and in a few scientific communities (e.g., CancerMA for synthesizing results from independent cancer trials). A knowledge portal is an IT-enabled, typically Web-based platform for “gathering, organizing, analyzing, creating, and synthesizing information and expertise” (Mack, Ravin, & Byrd, 2001, p. 925). It combines “content” (e.g., information access to data and documents), “communication” (e.g., channels for conversations, negotiations on collective interpretations), and “coordination” (e.g., workflows and routines to support cooperative work actions) into a single point (Detlor, 2004, p. 12). Unlike traditional outlets (e.g., journals, books), this Web-based platform is not limited by page size, page limit, formatting (e.g., two-dimensional visuals), or total space. A potential starting point could be the design of a Wikipedia-styled community, in which experts can regularly make major or incremental improvements in the definition and description of a concept or construct to enhance its specificity and explicitness, reducing the probability of different interpretations. This also requires scholars to consolidate the different definitions, jargon, and terminologies for the same constructs and concepts into a general language of taxonomic terms across disciplines. Following that, the portal could be accompanied by a search engine on which people can find the cause-and-effect relationships (enabled by the aforementioned 3D submissions) between any two specific concepts or constructs, based on theoretical logics and empirical evidence. This saves time for researchers avoiding duplication of each other’s efforts in reviewing earlier publications in the area, helping them to more quickly find and organize the current knowledge.

A working prototype is metaBUS.org, which hosts, organizes, and meta-analyzes correlates from empirical studies in select journals in applied psychology (and related micro disciplines) (Bosco et al., 2020). We encourage a broader integration of knowledge from both theories and empirical evidence—which would include conceptual, qualitative, and quantitative studies. Specifically, when integrating theories into meta-frameworks, we need to include theoretical propositions derived from qualitative research such as those based on intensive case studies; similarly, when meta-analyzing quantitative papers, we need to incorporate qualitative research to enrich the interpretation of empirical findings, thereby providing more valuable contributions to knowledge.

More provocatively, an open portal would be useful that not only integrates published information but also, as a journal equivalent, accepts submissions of new manuscripts. To enable knowledge accumulation, submissions can be accompanied with cause-and-effect principles and, if empirically tested, findings in the 3D model we mentioned. To enhance research transparency, reproducibility, and replicability of empirical studies (Aguinis, Ramani, & Alabduljader, 2018), this outlet may also mandate submissions of raw data, data processing, and programming. This practice can help to ensure rigorous experimental (or synthetic matching) designs to ensure causal inference—a limitation with most meta-analytic approaches based on correlations such as metaBUS.org. Combined with AI, the portal can integrate disparate academic findings into a holistic knowledge network. It could be designed to group similar concepts and constructs (based on highly explicit descriptions) and to connect these concepts and constructs using cause-and-effect relations, and finally to quantify all of these relationships using meta-analysis. This shared platform could be continuously updated, navigating the search and development of knowledge as well as integrating all new submissions that are accepted.

Similar to knowledge portals in workplaces that connect knowledge workers, knowledge portals in our academic systems could serve as an interface among editors, researchers, and others engaging in teaching and application activities. For editors, this tool can assist researchers in conducting a thorough literature review and inform editors of the knowledge gaps that exist, where the findings are conflicting and weak and where the theories and logics are unclear or contradictory. In this way, editors can more rapidly analyze and evaluate the potential value of papers (e.g., based on the research questions addressed) and determine their priority for publication. This tool can also help researchers to identify their contributions. For teaching and application activities, this portal would provide a means to synthesize all of the evidence and the underlying logics and reasons to solve a practical problem.

Gradually, new community tools could be developed to enrich this portal. We propose two ideas for such tools. First, one can link this portal with internal knowledge portals of workplaces (e.g., companies) to allow theories and evidence from sciences to be integrated with real-time big data (both quantitative and qualitative). It helps alleviate one difficulty in replications, that is, the challenge of accessing a large sample of data from multiple independent sources to run replications. A portal linking the scholarly community and workplaces could create timely feedback loops between the models suggested by the scholarly portal and the unique data from multiple workplace contexts held by the community of practice. It helps to speed up the knowledge synthesis from both deductive and inductive approaches. For deductive research (e.g., conceptual modeling followed by hypothesis testing), these feedback loops help to more quickly identify the relevant real-time data, and test and retest the validity of conceptual models in real-time contexts. For inductive research (e.g., propositions from qualitative data; pattern recognitions from quantitative data), these feedback loops connect propositions and patterns from workplaces to domain experts in academia, who may help interpret them using well-established theories and logics, thus enabling analytics that can be explained effectively. Certainly, these feedback loops are two-ways and can be repeated to reach the optimal point, where conceptual models and real-time data can consistently support each other.

Second, decision-making tools can be developed from this shared platform, which can assist managers in making rapid decisions based on logic in practice (which we refer to as “informed intuition”). As Simon (1996) suggested, professional schools (e.g., business schools) need to work on the science of the artificial such as decision tools. Given very limited time and resources for decision making, managers often rely on heuristic frameworks or “rules of thumb” to identify solutions for complex problems (Bettis, 2017). Many of these tools are ready-to-use, but may lack rigor such as specific measures for all of the variables and timely updates when new theories and evidence arise. Visualized tools for conceptualizing and analyzing problems can be developed based on the knowledge platform as managers’ interfaces, through which they apply rigorous theories and big data to identify solutions. These tools can save time and resources for managers to search for logic shortcuts. Again, we suggest the platform should not only host and meta-analyze quantitative findings, but also integrate both deductive theoretical works such as conceptual papers and inductive theoretical works such as qualitative studies, because they offer rich interpretations of empirical findings.

(Re)Inventing Incentives

Importantly, neither of the two initiatives discussed above will easily emerge and develop unless the business school (and the university) can create incentives for them. This is a basic problem derived from knowledge fragmentation. Currently, business schools and, also the job market, PhD programs overemphasize top journal publications or scholarly impact measured as citations (mostly within a closed academic specialty) as the primary and most critical criterion for academic performance appraisals (Aguinis, Shapiro, Antonacopoulou, & Cummings, 2014; Aguinis, Suárez-González, Lannelongue, & Joo, 2012; Shapiro & Kirkman, 2018).

As discussed above, our top management journals are structured to motivate new discoveries or new perspectives (i.e., state-of-the-art knowledge of a field) that are built on disciplines and specializations, rather than to engage in integrative efforts that contribute to a more holistic knowledge structure. To tenure and promote only people who focus on new discoveries and new perspectives thus undervalues others who seek to integrate new knowledge into a holistic framework, the suitable outlets for which are usually books, edited volumes, special issues in traditionally less impactful journals, or integrated course curricula. AOM Annals, International Journal of Management Reviews (IJMR), Annual Reviews, review issues at Journal of Management (JOM), Journal of Organizational Behavior (JOB), and, most recently, Journal of International Business Studies (JIBS) and Journal of World Business (JWB) are welcome efforts. Harvard Business School’s tenure and promotion practices based on integrated curriculum development are an exception. Outside academia, Campbell Collaboration (campbellcollaboration.org/), which has recently launched a Business and Management library for systematic reviews, is another good example. However, because publications in these outlets receive considerably less credit for academic performance appraisals, some of them may struggle to obtain high-quality submissions.

Changes in the approaches used in business schools (and in PhD programs) and their hosting universities are needed to produce greater advances in our knowledge stocks and structure. More fundamentally, we also perhaps need to challenge two well-established cultures in a top scientific journal-centric system. The first is “publish or perish” as the driver of scholarly efficiency without any concerns for broader social impact. This means we need to develop new measures of publishability by rewarding knowledge breadth (as opposed to a sole focus on depth), novelty in frameworks (rather than only in contents), and a holistic (as opposed to only a coherent) explanation of a complex phenomenon. These changes should encourage junior scholars to invest time in developing knowledge synthesis/accumulation skills and contribute to review issues. These new measures are also important for the university, which would benefit from solution-based research that is often interdisciplinary in nature and has social impact for its stakeholders (e.g., see Arizona State University Office of Knowledge Enterprise Development).

The second is the established notion of peer review by a small group of specialists. There is need to broaden the measures of scholarly impact to include the extent to which it contributes to solutions regarding a major social challenge (concerning a broader group of stakeholders). Aguinis et al. (2012) found evidence that scholarly impact measured as citations within the AOM was different from the impact generated by research outside the Academy. Aguinis, Ramani, Alabduljader, Bailey, and Lee (2019) found weak correlation between research cited most frequently in journals and that most frequently cited in textbooks. We may extend their methodology to measure scholarly impact on solving a major social challenge, regardless of disciplines or citing outlets. For instance, one can measure the relative weight (e.g., incremental R²) of a scholar or a scholarly contribution in explaining the variation of a specific dependent variable (e.g., organizational effectiveness) as a proxy for solving the social challenge in a meta-analysis.

Ironically, several of the most influential scholars chose not to conform to the norms and went beyond any personal concerns for the tenure and promotion cycle (Finch et al., 2017; Finch, Deephouse, O’Reilly, Massie, & Hillenbrand, 2016). In fact, they are often best known for their integrated frameworks of dispersed theories published by other scholars to provide a holistic explanation of grand social challenges, and such works are typically published in outlets that are accessible to nonspecialists. ¹⁵ Relatedly, the knowledge portals discussed above can enable more effective and impactful knowledge integration and can potentially serve as a new host for integrative scholarship.

In addition to performance appraisals, we also need to proactively create and scale a new market for knowledge synthesis workers, starting with integrative PhD programs focusing on knowledge synthesis to solve complex social challenges, industry-engaged research programs to make timely prescriptions based on knowledge synthesis, and more diversified tenure/promotion tracks based on new performance appraisal criteria mentioned earlier. In terms of budgeting, the National Science Foundation (NSF) Industry-University Cooperative Research Centers (I/UCRC) Program offers a good model, providing phased initial funding for up to 15 years, aimed at growing the number and financial contributions of industry members to make centers fully independent afterward. For instance, business schools can initiate university–industry partnerships with data analytics firms, which can provide not only financial support but also specialized access to unique data resources from inside the organizations and firms. One motivation for such collaboration is that more holistic analytic and logic algorithms from integrative scholarship, hosted and shared by new technologies and outlets, help analytics professionals to analyze and monetize their big data resources. The results and feedback from running these algorithms using big data can be hosted, constantly updated, and accumulated on the IT-based platform recommended earlier, whenever new big data become available from the organizations.

Limitations of Scientific Knowledge Synthesis

Limited Incentives of a Broader Community

More fundamentally, encouraging collaborative efforts in paradigm development in this untheorized/untested domain requires more radical rethinking of both the incentives for the industry community to collaborate with academia. One major constraint of the incentives for industry community collaboration with academia is the accessibility, quality, and comparability of context-specific data resources in the workplaces. Our discussions of the collaboration on the knowledge synthesis platform are based on a major assumption that industry members are willing to share their context-specific (often sensitive) data, make efforts to audit the data for academic research-ready quality, and format the data in a timely manner so that they can be compared with and harmonized with data from other sources (e.g., publicly available data, context-specific/sensitive data from other industry members). We realize that some industry members may make their data accessible to researchers on the platform. We suggest that for effective collaborations the data descriptions (not necessarily the sensitive raw data) and the managerial recommendations from synthesized and/or meta-analyzed multi-theoretical frameworks be made available so that the two groups (e.g., data owners, framework authors) can enter private discussions for meaningful retest opportunities. In the future, collaborative opportunities can be posted on the digital knowledge platform we recommended for industry and academic members.

Although we celebrate the scientific movement that began in the 1950s, led by elite scientific journals and AACSB policies, we still need to address the tensions between ideologies (scientific vs. intuitive/heuristic knowledge) and power structures (top journal-centric scholarship vs. other types of scholarship) for the business school incentive systems. There are positive signs of potential change represented by the calls for more collaborative relationships among scientific researchers and practitioners, which may eventually lead to more appreciation of the diversity in ideologies and scholarship. For instance, in business and management areas, Responsible Research in Business & Management (RRBM) network (rrbm.network) has already attracted approximately 1,000 endorsers (mostly scholars). As another example, The Declaration on Research Assessment (DORA; sfdora.org), developed in 2012, represents an initiative targeting a broader array of disciplines. It seeks to change research assessment from journal-based metrics to research quality. To scale the community, we suggest that these networks can also collaborate or even merge with the communities of other online academic networks such as Social Science Research Network (SSRN), Academia, and ResearchGate, and open source platforms such as GitHub for software developers to constantly translate synthesized knowledge into business analytics software.