Positivist,constructivist and critical approaches to international human resource management and some future directions

The roots of international human resource management and its theoretical embeddedness

De Cieri et al. (2007) argue in their review paper that IHRM was atheoretical or mono-theoretical at the beginning and that IHRM was considered as a simple combination of cross-cultural management descriptive theories and HRM literature. This strand of thought can be defined as the first wave of IHRM research. It is in line with Itani’s (2018) review on HRM literature, that until the 1980s, HRM was rather atheoretical and lacked strong ideologies. Early publications on IHRM were rather descriptive and lacked analytical rigour, probably very often limited by accessibility of data, and research results were characterised by ad hoc design and raw data presentation. It showed similarities with international business research in these respects. One of the major themes of IHRM from the beginning was research about expatriation, and this branch of studies grew independently of and beyond IHRM (Doherty, 2013; Harvey and Moeller, 2009).

The second wave of research focused on global coordination (integration) versus local responsiveness (differentiation), that is, multinational enterprise (MNE) staffing policy: ethnocentric, polycentric, regiocentric or geocentric (De Cieri et al., 2007: 286). The second wave is overtly positivist. Itani (2018) analysed six decades of HRM publications, including IHRM, and he concluded that in the 1980s and 1990s, strong positivist ideology was present, and the first publications in critical HRM emerged. Towley (1994), as one of the first critical HRM scholars, introduced a classic Foucauldian analysis of performance management systems. Legge (1995) published a full book on critical HRM, which became influential. However, the book does not include any international dimensions. McKinlay (2002), Pasmore (1999), Steyaert and Janssens (1999), and Nkomo and Ensley (1999) continued the critical branch of HRM, although their work also had no significant international dimensions.

The third wave of IHRM research includes new themes – postcolonial analysis (Boussebaa and Morgan, 2014; Mahadevan, 2015) and gender discrimination in foreign assignments (Adler, 1984, 1987; Altman and Shorthand, 2001; Janssens et al., 2006; Mayrhofer and Scullion, 2002; Metcalfe, 2006; Tung, 2004) – while the dominance of positivism is inevitable and can be identified in perhaps the most influential textbook in the field, that of Dowling et al. (1999). There is a growing number of publications in cross-cultural management, which intentionally apply critical approaches (Mahadevan et al., 2020; Primecz et al., 2016; Romani et al., 2018b, 2020), and these studies can inspire further critical IHRM studies in the future.

The emergence of paradigmatic thinking in organisation studies

IHRM is part of OS and, consequently, paradigms in OS are present in IHRM. Debates about organisational paradigms date back to the late 1960s, when Weick (1969) published The Social Psychology of Organizing, and the early 1970s, when Silverman (1971) published The Theory of Organizations, according to Clegg and Hardy (1996). Burrell and Morgan (1979) organised existing theories and approaches into research paradigms (2x2 matrix), and their taxonomy gave strong drive to the debates about basic assumptions of theories in the field of OS. The two dimensions of the matrix answer two important questions: one about science itself and the other about society. The horizontal dimension defines objective and subjective approaches to science. Burrell and Morgan (1979) distinguish objective and subjective science, based on ontology (the theory of being, namely, if reality is assumed to be subjective or objective). epistemology (theory of knowing: namely if knowledge creation is based on objective or subjective assumptions), human nature (meaning that the relationship between human beings and their environment is voluntarist or determinist), and methodology (whether it is based on subjectivist, such as ideographic, or objectivist, such as nomothetic assumptions). The vertical dimension is about conflict-order debate in societies, where the order side stands for societal status quo, and the conflict side stands for radical (social) change. These two dimensions delineate four research paradigms in organisational studies: functionalist, interpretive, radical humanist, and radical structuralist (Burrell and Morgan, 1979).

The functionalist paradigm builds on objective ontology, epistemology, methodology, and human nature, and includes only theories which do not question the existing societal status quo. Rather, they seek to explain it. The interpretive paradigm is at the subjective end of ontology, epistemology, methodology, and human nature, while it is similar to functionalist theories in terms of its approach to societal status quo. Radical humanism is at the subjective part of ontology, epistemology, methodology, and human nature, and this paradigm questions the existing social status quo, as it wants to change the present societal order. Radical structuralism is objective in ontology, epistemology, methodology, and human nature, while its agenda is also to change the existing social status quo (Burrell and Morgan, 1979).

The debate about the paradigmatic nature of organisation studies intensified in the 1990s (Chia, 1996; Gioia and Pitre, 1990; Hassard, 1991; Jackson and Carter, 1991, 1993; Pfeffer, 1993; Willmott, 1993). Gioia and Pitre (1990) presented the dominance of the functionalist studies visually and argued for equal acceptance of alternative paradigms (see Figure 1). The focus of the debate turned to incommensurability in the late 1990s (Czarniawska, 1998; Donaldson, 1998; Scherer, 1998; Scherer and Steinmann, 1999); after this point, heated debates settled down in OS, and paradigm plurality has since been taken for granted by the majority of organisational researchers (Hassard and Cox, 2013; Hassard and Kelemen, 2002; Primecz et al., 2009, 2015; Romani et al., 2011, 2014; Shepherd and Challenger, 2013; Tsoukas and Knudsen, 2003).

OPEN IN VIEWER

Figure 1.

The dominance of the functionalist paradigm in organisation studies.

Paradigm taxonomies

While paradigm plurality is not questioned in today’s scholarships, there is no full agreement on paradigm taxonomies. Although the Burrell-Morgan matrix initiated the idea of the paradigmatic nature of the discipline, several researchers provided alternative taxonomies. Deetz (1996) criticised the Burrell-Morgan matrix as a closed system and offered his open system, which includes four discourses: traditional, modern, late modern, and postmodern. In his system, the two dimensions deal with similar issues as the dimensions in the Burrell-Morgan matrix: one about the nature of theories in terms of whether they emerge or exist a priori, and the other dimension is about whether society is in consensus or dissensus. The traditional discourses cover interpretive studies, which are emergent, and their approach to society is consensual. The modern discourse covers normative studies, which build on existing (a priori) theories, and it does not question the present social order. Late modern discourses include critical studies, which also build on a priori theories, while they additionally question the existing social order. Postmodern discourses are dialogical studies, which apply deconstruction. They are emergent and they agree that society is in a state of dissensus. Figure 2 provides an overview of the four discourses.

OPEN IN VIEWER

Figure 2.

Four discourses in the Deetz matrix.

Although the Deetz matrix has never become as influential as the Burrell-Morgan matrix, it nonetheless identified the major discourses in the field of organisation studies. Publications on postmodernism were at their peak in the late 1980s and early 1990s (Alvesson and Deetz, 1996; Burrell, 1988, 1996; Chia, 1996; Cooper, 1989; Cooper and Burrell, 1988; Deetz, 1996). Calas and Smircich (1999) closed the debates about postmodernism with their influential article, ‘Past Postmodernism’, in which they argue that although the postmodern discourse had a positive impact on the debate in organisation studies, it is over. This did not, however, stop Tsoukas and Knudsen (2003) from arranging their handbook on organisation theory around four paradigms: namely, the paradigms of positivism, interpretivism, critical, and postmodernism.

Beyond these taxonomies, Denzin and Lincoln’s (1994, 2000, 2005) qualitative research handbook is influential for researchers who do not follow mainstream quantitative studies. The first chapter in each edition is about the paradigmatic taxonomies of social sciences. In the first chapter in the third edition, Guba and Lincoln (2005) define five paradigms: positivism, postpositivism, critical theories et al., constructivism, and the participative paradigm. Taking into consideration all influential taxonomies of research paradigms, I contend that three clearly distinct paradigms can be identified in organisation studies: positivist, constructivist, and critical, as Gephart (2004) has also stated. All other paradigms are either so minor (e.g. the participative paradigm) that it is difficult to designate them as distinct paradigms, or so close to each other because they have similar or the same basic assumptions (e.g. interpretive and constructivist paradigms), whereas other paradigms (e.g. postmodern) are ‘officially’ over (cf. Calas and Smircich, 1999).

Bearing in mind the differences among theories, three paradigms seem to fulfil the status of paradigm in organisation studies: positivist, constructivist, and critical. They have clearly distinctive basic assumptions, numerous publications based on the given paradigms, and there is a critical mass in the research community. These criteria are based on Kuhn’s (1970) description, even though it is known that Burrell and Morgan (1979) and most of their followers used Kuhn’s work only as inspiration, not literally.

Three dominant paradigms: positivist, constructivist, and critical paradigms

A positivist paradigm builds on objective epistemology and ontology, and it does not question the existing societal order. It has clear overlap with Burrell-Morgan’s functionalist paradigm. As an example, GLOBE research investigates how culture influences leadership and management practices in 62 countries (House et al., 2004). Constructivism applies subjective epistemology and ontology, and it is also at the societal consensus end of the scale. It is very similar to Burrell-Morgan’s interpretivism, but as Schwandt (2000) clearly describes, the emphases of the two philosophical approaches differ. Interpretivism’s major concern is Verstehen and explaining (interpreting), while constructivism’s major concern is social construction processes. It is anti-essentialist and focuses on human processes, and, more importantly, it is not about individual sense-making, but rather, the collective generation of meanings or frames of reference (Romani et al., 2011; Schwandt, 2000). As an example, Cunliffe and Karunanayake (2013) present the findings of 3-month ethnographic fieldwork in a tea plantation in Sri Lanka, on how macro- and micro-discourses maintain and construct the identity of the workers. Finally, it is necessary of combining all theories that are essentially concerned with societal change into one critical paradigm. While it is obvious that there are differences between postmodern and critical approaches, and Burrell and Morgan (1979) and their followers make a distinction between radical humanism and radical structuralism, Adler’s (2002) ‘big tent’ approach seems to be widely accepted in Critical Management Studies (CMS), which might include Marxists, post-Marxists, post-modernists, feminists, ecological thinkers, irreductionists, critical-realists, postcolonial researchers, and many others who might have major criticism of the existing social order. This is why it is logical that different or even contradictory forms of epistemology, ontology, and methodology might be present in one paradigm: the critical paradigm. For example, Banerjee (2008) argues that the contemporary organisations, as the result of colonialism, exercise coercive power over ex-colonies, and in this way, many people lose their access to health, livelihoods and communities are destroyed, and eventually vulnerable people die. The overview of three paradigms is illustrated in Figure 3.

OPEN IN VIEWER

Figure 3.

Dominant paradigms in organisational studies.

Positivist paradigm in international human resource management

The main characteristic of positivism is that it takes the natural sciences as its role model (Donaldson, 2003; Pfeffer, 1993). The decisive characteristics of this type of research is that there is an assumption that social phenomenon can be measured with variables, among which there might be mathematical statistical relationships, and the aim of the investigation is to discover if these relationships exist and, if they do exist, whether they are strong or weak. In order to apply this logic to social sciences, many researchers introduce hypotheses, and they develop models about the relationships among dependent and independent variables, such as in Dickmann et al. (2018), García-Cabrera et al. (2018) and Ho et al. (2016). In some cases, proposals are used for the same purpose, such in Vasilaki et al. (2016). The existence of hypotheses is not a proof of positivist research, but it is a strong indicator. I read the hypotheses of the studies, and the hypotheses were checked to ascertain whether the statements were mirrored as causal relationships between (or among) variables developed by the researchers. Models are often involved in positivist studies, and they are frequently visualised by figures with arrows, which symbolize that one variable affect another, and the degree to which it is effected, and the relationship between two variables is typically measured by statistical methods, very often with multivariate statistics building on dependent and independent variables, for example, correlations, factor analysis, and variate analysis, such as in Li and Lee (2015) or Presbitero and Toledano (2018).

The aim and scope of these investigations are often concerned about effectivity, productivity, or performance in general, or in some cases, competitiveness. A model example of these kind of publications is the work of Kornelakis et al. (2017), who developed six hypotheses about the relationship between productivity: and (1) bargaining, (2) presence of trade unions, (3) work councils, (4) performance-related pay, (5) teamwork, and (6) training in post-crisis European workplaces. They analysed data from the European Company Survey of approximately 27,000 establishments. They built a model with dependent and independent variables and applied multivariate statistics, in this case, regression analysis. They provide strong evidence based on their quantitative analysis that productivity is higher where the institutional context is built on neoliberal principles. The study eventually represents a good example of the strong and explicit application of natural science-like epistemology and supporting the existing societal status quo, namely neoliberal capitalism. A similarly good example of a positivist publication, the work of Li and Lee (2015) investigated whether knowledge transfer within MNCs can hurt performance. The authors also developed hypotheses stating clear causal relationships, which are nicely organised into a model (Li and Lee, 2015: 666, Figure 1). Their sample was 1000 randomly selected firms from the China Foreign Enterprises Directory; they developed measures among which hierarchical linear regression analyses were applied, with the result showing that knowledge transfer from the headquarters to the subsidiaries does not always improve productivity, especially when the subsidiary is more entrepreneurial. A causal relationship was assumed among measures (variables) in the study, and large-scale quantitative methods were used to prove this relationship, which is why it can be clearly clustered as a positivist study.

While hypotheses are typically developed for quantitative research, it is possible to develop hypothesis for qualitative research, although it is very rare. I have not found any examples of this during this review process. At the same time, the qualitative research method does not necessarily mean that the study is not positivist. A good example is the grounded theory method developed by Corbin and Strauss (1990), which is often used in a positivist way, while Charmaz (2005) argues that it is possible to apply it in positivist and constructivist approaches, depending on the researchers’ assumptions and practical application of the method. I found several examples in the review of the qualitative method being applied to grounded theory in a positivist way, similarly to Gephart’s (2004) observation. For example, Perera et al. (2018) collected face-to-face, in-depth interviews with expatriates in Malaysia about their perception of and the process of responses to psychological contract breach. While the initial phase of the research was inductive, so the pattern emerged from empirical material, which is in line with the constructivist approach to science, but the researchers applied a model as a theoretical map in Perera et al. (2018: 1459) with a positivist logic. Gephart (2004) notes that data collection and data analysis are equally important, when the characteristics of an article is classified – for example, data collection might be qualitative, but the data analysis is quantitative, and consequently the result is a positivist study. In the work of Perera et al. (2018), they identified phenomena (e.g. attitude, cognition, etc.) among which propositions were drawn up. Eventually, they end up with a model which connected concepts with causal relationships, even though they did not measure them with statistical analysis.

A further example of qualitative research which was categorised as positivist research is the work of McNulty et al. (2009), who investigate the connection between global staffing practices and the variables which influence it. While the authors build their study on in-depth interviews, arguing that it is possible to acquire an in-depth knowledge of complex issues and processes, their overall approach, rhetoric, and analytic style were overtly positivist. Not only is the knowledge production was based on objectivist assumptions in an epistemological sense (Burrell and Morgan, 1979; Guba and Lincoln, 2005; Hassard and Kelemen, 2002), but also, their view of society and organisation leans towards the consensus (status quo) end of the scale, and consequently, the research can be categorised as functionalist (Burrell and Morgan, 1979), which was manifested in a managerialist overtone. As the objective of the research was dominated by the endeavour towards effectivity, return on investment (ROI) was the focus. McNulty et al. (2009) identified environmental, individual and organisational variables which might influence changes in expatriate ROI, and they eventually could conclude that a non-linear causal relationship often exists between strategy, HRM, and firm performance. This kind of basic assumption (objective ontology, epistemology, and human nature and status quo in society) is somewhat common in quantitative research practice. Qualitative methods are often open to alternative paradigms, but not in this case. In this example, the researchers went further with the development of positivist research: they introduced variables and, based on their interviews, calculated their impact on ROI. This step made it inevitable that the research paradigm of the study was positivist.

The last crucial characteristic of a positivist piece of research is that it aims to reach generalisability. Positivist research ideally builds on a full sample (extremely rare) or a representative sample which has identical characteristics to the full sample. While in social science, such as sociology, this endeavour is often expected, in management and organisational studies, it is rather rare. Still, the attempt to have research results which are generalisable for a wider public is often expected and appreciated. The majority of the positivist studies were quantitative, but the aforementioned qualitative studies and similar ones fulfilled similar requirements as the quantitative ones: researchers identified causal relationships between variables, as in natural sciences, and they aimed for generalisable theories, like in Dickmann et al. (2018).

Constructivist paradigm in international human resource management

The number of studies conducted in the constructivist paradigm is rather low compared to positivist works. They build on subjectivist epistemology and ontology. A large part of IHRM research incorporates studies on culture, which is the home context of interpretive and constructivist research, based on Geertz’s (1973) seminal work, which is the starting point of all interpretive and constructivist research. Bearing in mind that interpretive and constructivist paradigms overlap, but they are not the same (Schwandt, 2000), I am using ‘constructivist paradigm’ as an umbrella term for interpretivist and constructivist works. As it is not easy to decide at first sight if a study is constructivist or not, I had to read those papers which applied qualitative methodology more thoroughly. Although qualitative methodology does not indicate a 100% constructivist approach, quantitative methods could be difficult to apply when the researchers follow subjectivist epistemology. The most important feature of the constructivist paradigm is that it investigates and shows the reality construction of the actors, and the investigation remains context-dependent. Large-scale studies do not adhere to this approach, and instead of general knowledge, local reality is emphasised in the constructivist paradigm. The decisive characteristics of the constructivist paradigm are ‘emerging’ themes, such as in Hopkins et al. (2016).

In a few cases, it was easy to identify whether the research was constructivist, when the authors made it explicit in the methodology section, abstract, or introduction that they had built their arguments on the interpretive or constructivist paradigm, quoting seminal theorists of the paradigm, such as Berger and Luckmann (1966), for example, in Kirk (2016), or ‘Heidegger’s interpretive phenomenology methodology to uncover layers of hidden ‘truths’ via the interview medium, as expressed in words and narratives’ (Van Maanen, 2006) in Ang and Tan (2016: 48), and emphasised inductive analysis. The model quality of constructivist research is published by Uygur et al. (2017), who made it explicit that their study is ‘informed by social constructivism’, beyond Berger and Luckmann (1966); they quote the key interpretive concept of Weber’s Verstehen, when they explain ‘explanation and understanding’ in the light of ‘interpretive understanding of social action’ (Uygur et al., 2017: 1217–1218). But these rhetorical forms would not be enough to be categorised as ‘constructivist’, it was rather their actual analysis which built on ‘five emerging themes’ from interviews. Their presentation of findings connected the actors’ own wording with theoretical concepts, and they remained loyal to their qualitative data when they developed their results. Their analysis was insightful and in-depth, and they have not come up with a model which contains variables connected with causal relationships.

The other obvious sign of constructivist research is ethnography. It is, again, not enough that the authors state that their work is ethnographic, because it is possible to analyse ethnographic data collection positivistically, although it is rather rare; it is also possible to conduct critical ethnography, which would imply the critical paradigm, not constructivism. Papers which claimed to be ethnographic were thoroughly scrutinised. For example, Mahadevan (2012) builds on ethnographic research when she presents the research context in great detail, and her analysis involved context-rich insights from the actors’ perspectives, and she eventually draws conclusions from qualitative interpretations of the involved persons, without developing models of variables connected with causal relationships. Similarly, Alberti and Danaj (2017) apply ethnography with participant observation and in-depth interviews. They gain insights from the migrants’ perspective, which they investigate, and they remain faithful to their subjects’ viewpoints when they theorised their findings. They also avoided developing a model with variables connected with causal relationships. It is also worth mentioning the ethnography of Charleston et al. (2018), who built their results on emerging themes of their study, which is an important feature of the constructivist paradigm, where research questions are open, and issues, topics, and phenomena emerge from the unstructured data collected by the researchers.

It is observable that more data do not necessarily lead to a better interpretive and constructivist analysis. Kvale (1996) highlights the problem of 1000-page interview transcripts that he uses as an example of mistakes in qualitative interviewing. He argues that it is the researcher’s task to monitor data saturation, and when new data input (e.g. interview, observation) does not bring further insights, it is necessary to stop data collection, otherwise a redundant and overly large dataset will be the base of analysis. Kvale (1996) made it explicit that a qualitative piece of research usually builds on 5–25 interviews, and this is ideal for in-depth analysis. Some qualitative research in IHRM included more interviews, and these examples led to positivist analysis, for example, in Horak (2017), who collected 44 interviews, although he analysed only 23 interviews for the article. The overall tone and style of analysis remained positivist, the richness of data was not exploited, and the potential of constructivist analysis was lost. It is difficult to evaluate whether the researcher approached the field with a positivistic attitude, and this is why he collected an unnecessarily large number of interviews, or if the large number of interviews made it impossible to undertake an interpretive analysis – but we have to bear in mind that in the case of constructivist analysis, less is more.

Critical paradigm in international human resource management

The number of studies classified as critical was very low compared to constructivist and positivist works. While critical paradigms might have fallen into the subjectivist or objectivist epistemology and ontology, the overall tone of the article was critical, not only in the everyday sense, but it also aimed to criticise existing societies and the world order generally. Although Willmott (2003), the leading scholar of CMS, has stated explicitly that Burrell-Morgan’s radical humanism can be considered as the true legacy of CMS, the term ‘critical paradigm’ is used here in a wider sense, as an umbrella term in line with Adler (2002), Guba and Lincoln (2005), and Gephart (2004). There is a scientific argument to distinguish postmodern and critical paradigms (Deetz, 1996; Primecz et al., 2015; Romani et al., 2018a), but the low number of studies in each of the paradigms and the small differences between them make it reasonable to keep them in one group. The distinction between the two paradigms is instead based on emphasis, and the variety among critical arguments is even greater than the actual difference between a postmodern and an ‘average’ critical study.

The distinctive character of critical IHRM studies is their concern for power inequalities, oppression, exploitation, or injustice. People on the periphery are in the focus, with the explicit or implicit aim of increasing their situation or easing their hardship. One possible topic is the employment of people with a migration background. This topic, however, does not determine if the study will be critical. It is possible to discuss the challenges of migrant integration in the workplace from a merely positivist viewpoint, describing the enabling and interfering forces as variables and developing a model of successful migrant integration into the organisation from managerialist point of view, for example. The same study could be done in a constructivist paradigm, building on the actors’ view of the situation, and remaining neutral on political issues. In contrast with the two previous approaches, a critical approach would problematise the inequalities between powerful and powerless members of the organisation, they would uncover the oppressive mechanism, and they would illuminate the exploitative tendencies. In order to reveal critical articles, I had to analyse more thoroughly the possible manuscripts.

Among the examples, I found that postcolonialism was one of the analytic frames used when critical paradigms were employed. Kamoche et al. (2012), Jackson (2002, 2012), and Jackson et al. (2013) analyse different situations in Africa from postcolonial perspectives. Critical approaches, and especially postcolonial frames, require the historical context to be revealed, like in the work of Al Aris (2014). Other critical paradigm researchers investigate Western hegemony in Russia (Dixon et al., 2014); organisational control, ethics, and power (James, 1994); or othering (Mahadevan and Kilian-Yasin, 2017), applying critical discourse analysis, which is a method relatively often used by critical scholars. Review papers have rarely fallen into the critical paradigm, with one exception being the work of Almond and Menendez (2014), who use ‘critical’ in their title, but not in an everyday sense – they disclose the ideological nature of comparative HRM, as many critical studies aim to show the latent ideology in the seemingly neutral research. In the critical paradigm, it is questioned whether it is possible to conduct any research without ideological assumptions, and this is why they find it important to show the ideological component of positivist and constructivist studies.

Research beyond the three dominant paradigms

Some studies could not be unambiguously classified into one paradigm. In many cases, features of two paradigms were present, most often the positivist and constructivist paradigms. Frequent examples include the application of large-scale questionnaires and series of in-depth interviews, and the researchers followed mixed methods approaches. In line with Burrell and Morgan (1979), there is no research which is ontologically, epistemologically, and methodologically equally subjectivist and objectivist. Other researchers, for example, Gioia and Pitre (1990), have argued, however, that there are transition zones between paradigms. During the review process, I found some articles which could be categorised as being positioned in the transition zones of two paradigms. One example of positivist and interpretive analysis, which is not mixed methods, is that of Yao (2014), who applied exploratory and interpretive approaches in her qualitative studies. She searched for emerging themes, while using the flagship positivist research of Hofstede’s dimensions as analytical frames. Although some steps towards quantification of qualitative data were taken (a positivist feature), the classical interpretive analysis was conducted in a professional way (a constructivist feature). In other cases, critical and constructivist research was combined, where, for example, the researchers started their data collection in a constructivist manner, and they built their findings around their subjects’ sense-making, while the analysis process seemed to be rather critical, for example, in Alberti and Danaj (2017) or Hadjisolomou et al. (2017). Finally, the transition zone between the positivist and critical paradigm resulted a study featuring quantitative data collection with the help of a questionnaire in Portugal and in Mozambique, which ended up in postcolonial analysis (Dibben et al., 2017), which was a somewhat logical analytical frame in the given cultural contexts. All these articles were clearly identifiable with two paradigms, although they were not multi-paradigmatic research (Lewis and Grimes, 1999; Lewis and Kelemen, 2002), because in these cases, the researchers did not keep the two paradigmatic analyses separate, which is a distinctive feature of multiparadigm analysis.

Future research: where shall we go from here?

The overall picture of paradigm plurality in the field of IHRM points to various future directions. One important claim is that it is necessary to increase paradigm reflexivity. As there are rather large proportions of papers which cannot be defined as belonging to one paradigm – as many researchers might not be aware of paradigms and their assumptions or they find it too abstract or impractical to deal with paradigms in their research – they are ‘non-consumers’, to use Hassard and Kelemen’s (2002) term. Some undefinable papers were not the product of ‘non-consumer’ research activities; rather, they were simply the combination of more than one paradigms of studies within one review, often without reflecting on paradigmatic assumptions. I propose that it is desirable to make basic assumptions explicit when research results are presented, especially when more than one paradigms are involved. Finally, it is striking that mainstream positivist research rarely, if ever, deals with abstract questions, such as the philosophy of science and the philosophy of society. Today’s research practice does not require, especially in mainstream studies, scholars to reflect upon the paradigmatic assumptions of their research.

The second claim is connected with the fact that positivist research strongly dominates the IHRM field. Consequently, constructivist and critical research is in the minority. While a large number of papers in positivist research do not bring exciting novelty, and are instead ‘puzzle-solving’, to use Kuhn’s (1970) term, and their contribution is relatively minor to science, non-mainstream research results might generate new insights and even breakthroughs, when compared to mainstream positivist research. Constructivist research is close to actors’ perspectives, and their interpretations and sense-making can renew social science, including IHRM, because, above all, the actors’ perspectives might result in more valuable theories for practice. If we go beyond the constructivist paradigm, we can see that the critical paradigm can bring even more novelty with its unexpected perspectives. CMS, including critical IHRM, is always suspicious of the existing power structures and world order, and such scholars search for more fair organisations and societies. Consequently, the critical paradigm can contribute valuable insights, comprehension, and wisdom to organisational and societal practices. Above all, the critical paradigm enhances critical self-reflexivity.

Moving beyond producing more paradigm-reflexive research and getting involved in non-mainstream studies leads us to the third claim. It is worth experimenting with new paradigmatic directions, such as the participatory paradigm (Denzin and Lincoln, 2005) or any other undiscovered paradigms. The task of the researcher should be to find new paths towards knowledge, and new paradigms might bring about unexpected results. Researchers ought to be encouraged to conduct unconventional studies instead of replicating the same or similar research with a new sample. Besides developing new paradigms, multiparadigm research (Hassard, 1991; Lewis and Grimes, 1999; Lewis and Kelemen, 2002) might be a more feasible way to enhance creativity and novelty, even though it is also challenging at the practical level (Romani and Primecz, 2019). Different forms of multiparadigm research require a strong commitment to paradigm reflexivity, as well as a conscious combination of research processes and results of distinct paradigms (cf. Primecz et al., 2015; Romani et al., 2011; Romani and Primecz, 2019; Schultz and Hatch, 1996). Multiparadigm research is almost entirely absent from the field of IHRM, and this would move the discipline forward.