Mixed Methods in Business, Management and Accounting Research: an Experimental Design in the Entrepreneurship Domain

All research strategies suffer from some weaknesses. Combining qualitative and quantitative approaches might offset some flaws. Recently a growing methodological trend arose from the acknowledgement that mixed methods may increase the value of several research projects. With this background, the paper analyses trends of mixed methods in the business, management and accounting area during the last twenty years. The analysis highlights domains where mixed methods are not frequently used, such as entrepreneurship, and presents ongoing research based on mixed methods on the Italian community enterprises, a set of collective entrepreneurial initiatives working for sustainable regeneration in their territories. These enterprises are elusive since they are neither easily identifiable nor extractable from databases. A fixed and sequential mixed method approach turns out to be effective for investigating these evolving enterprises. The work is relevant for novices to mixed methods research and provides meaningful insights to analyze a type of organisation that is very important in depleted contexts.


Introduction
Starting a scientific study implies several research decisions or design dilemmas to face. Indeed, there is not a set of choices that ensures maximum results, as there are goals that are difficult to match (for instance, the generalizability of results and realism of contexts). Therefore, it can be argued, as stated by McGrath (1981, p. 179), that "all research strategies and methods are seriously flawed". Nevertheless, the methodological triangulation may mitigate some limitations (Dezin, 2009;Turner et al., 2017), as one method can compensate for the weaknesses of another (Jick, 1979), and support the growth of understanding of a phenomenon and the advance of the knowledge frontier.
However, it is worth saying that mixed methods are not a solution to all problems. Echoing the words of Johnson and Onwuegbuzie (2004, p. 17), it is important to remember that: "In some situations the qualitative approach will be more appropriate; in other situations, the quantitative approach will be more appropriate. In many situations, researchers can put together insights and procedures from both approaches to produce a superior product".
In other words, although some weaknesses cannot be completely overcome, when the combination of different methods in a study is performed effectively, new avenues for knowledge creation arise (Hurmerinta-Peltomäki et al., 2006). This is a reason why mixed methods have achieved a certain success in several fields, such as organisational science (Molina-Azorin et al., 2017), to the extent that they have been recognised as "the third major research approach or research paradigm" (Johnson et al., 2007, p. 112). However, there are still fields or areas of research where they are taking hold more slowly. In this perspective, the study addresses the following question: do mixed methods represent an emerging methodology in business, management and accounting research? In this vein, to promote their use, the paper illustrates the design of ongoing research concerning Italian community enterprises, which adopts the mixed methods approach to address some specific research complexities.
To this end, the paper is organised as follows. The next section offers background information on the mixed methods approach and a brief description of the community enterprise concept together with some elements characterizing Italian community enterprises. The following section depicts the methodology adopted to analyse the mixed methods research in the investigated areas. In addition, this section provides some information on the methodological features concerning the application case. Then, the paper describes the results of the analysis and presents how the mixed methods approach has been fruitfully used in the application case. Finally, the discussion section proposes some reflections on the literature trends, and the conclusions section illustrates some future issues and limitations of the work.
Mixed methods: literature backgrounds For many years, researchers have integrated qualitative and quantitative methodologies, but until the 1980s there was no conceptualisation of this way of doing research (Bryman, 1988). Subsequently, mixed methods gained more recognition and were frequently adopted, at least in certain research areas, such as education, health science, sociology (Denscombe, 2008;Molina-Azorin et al., 2017).
For those new to the subject, it is relevant to distinguish between multi-methods and mixed methods, since the former approach concerns the use of either multiple quantitative methods or multiple qualitative methods in the same study, while the latter involves -as indicated in the research methods map proposed by Sage Publications 1 -to mix that crosses the quantitative and qualitative boundary in inquiries by combining two or more methods within given research.
However, adopting mixed methods is not so easy and may present obstacles because, as Mason (2006) pointed out, it requires thinking outside the box. In practice, researchers may face perceptual and epistemological obstacles (Tunarosa and Glynn, 2017). From this point of view, as highlighted in Howe (1988), who endorses the compatibility thesis, some authors (see for instance Smith, 1983;Smith and Heshusius, 1986) think that the compatibility between qualitative and quantitative methods is unlikely. Indeed, these methodological approaches have different goals and use different techniques. Theory testing is usually the primary objective of a quantitative study (Dezin and Lincoln, 2005) and, in this perspective, it analyses hypotheses deductively, operationalizes (independent and dependent) variables, gathers data by trying to reach a representative sample and demonstrates statistical significance. Whereas, theory building is typically the goal of a qualitative study. In this type of study, the adopted approach is inductive and the analysis is developed on archival data or data collected through interviews, case studies and other field research methods. In addition, there are divergences from an epistemological point of view, as quantitative studies tend to refer to the positivist paradigm, whereas qualitative studies embrace the interpretivist paradigm; therefore, these studies have discordant postures and this implies differences in the way of knowing and in research questions. These features represented obstacles to methods integration, even if many authors considered it positively (Cresweel et al., 2003).
In the first attempts, the combination of two different types of methods resulted from a hierarchy of them, as authors considered one type of method as an extra part with a completing function. For example, several quantitative studies employed qualitative analyses to provide detailed information about the context of reference. This practice cannot be considered as an effective integration if anything a collage of methods, so much so that the subordination of the qualitative method to the quantitative method has met with some criticism (see, for instance, Jick, 1979).
As noted in the introduction, some situations and motivations suggest the adoption of mixed methods. Green et al. (1989) provided a non-exhaustive list of non-mutually exclusive cases aiming at the following purposes: the identification of convergences and for confirmatory goals (triangulation), the use of the results of one method for the description and interpretation of the results achieved through the other (complementarity), the identification of contradictions to refine or even to reframe theory (initiation), the use of the results of one method for the determination of the elements characterising the second (development), the use of different methods to increase the level of detail and/or the breadth of the analysis (expansion) and, the use of different methods to reveal the different aspects of a given phenomenon (diversity).
To maximize the potential benefits of mixed methods, suitable choices have to be made in the research design phase. In particular, they concern not only the selection of the type of technique employed (case study, survey, focus group, simulation, etc.) but also the weight assumed in the combination (qualitative dominant, equal status, quantitative dominant) (Johonson et al., 2007) and the application order (concurrent or sequential). Consequently, those who opt for this type of methodological approach are called upon to construct a specific pathway to reduce complexity and facilitate the rearrangement of information. To systematize these choices, Figure 1 shows the mixed-methods design matrix presented in Johnson and Onwuegbuzie (2004).  (Grafton et al., 2011). Concerning the latter, even if mixed methods are potentially relevant to different strands of research, they have only been used in some of them, like in management accounting studies (as they are more prone to qualitative methods). In this case, however, the degree of integration between the methods was sometimes low, as the qualitative methods served to complement contributions (Tucker and Hoque, 2017).
Community enterprises: peculiarities hindering analysis and gaps in the Italian case Community enterprises are usually bottom-up collective entrepreneurial initiatives aiming to attain economic, social and environmental goals for the well-being of the community of reference (Johannisson, 1990;Peredo and Chrisman, 2006;Somerville and McElwee, 2011). They usually grow in depleted contexts, namely in territories characterised by depopulation, low-income levels, lack of services and job opportunities, trying to counter and reverse these negative trends. For this reason, community enterprises are initiatives to be analysed carefully, not so much for the economic dimension they achieve, but for the regenerative capacity, they show in depleted contexts. In addition, if one considers that these enterprises also operate in urban areas (not necessarily marginalised areas), then their significance and the plurality of their implications become evident. As a consequence, several studies and reports have been published on this topic, but some gaps persist.
For example, on the subject of entrepreneurship, very few studies have been conducted in recent years. Nevertheless, some meaningful pieces of evidence arose in the work of Buratti et al. (2022), which showed that these enterprises tend to have a humane entrepreneurial posture (Kim et al., 2018;Parente et al., 2018). Similarly, except for a few countries (where specific plans have been launched to promote and support this type of enterprise), there are not many analyses of the phenomenon on a national scale and almost no studies or comparisons on an international scale. Suffice it to say that even literature reviews are scarce, the last one was published about eight years ago (Pierre et al., 2014). In light of this, the following application case is derived from an ongoing research project aiming to represent Italian community enterprises through the adoption of mixed methods.
In the Italian case, community enterprises mainly take the form of cooperatives (Buratti et al., 2022) and in the various regions of Italy there are regulations governing community cooperatives (CCs), but at the national level, there is no law recognising them as a type of enterprise. Therefore, it can be argued that the label community (or cooperative) enterprise may represent status and not a recognised type of company (Alleanza delle Cooperative Italiane, 2018). This is relevant from the researcher point of view, because CCs are elusive organisations as they are difficult to detect in databases as community enterprises. Hence the unavailability of a national picture of the phenomenon and the need to design an analysis to face this gap. In the light of these difficulties, research based on a mixed-methods approach proved to be a cost-effective option. Therefore, the paper describes the application case both to show how a mixed-methods approach can address certain research complexities and to promote this method in fields where it has been little used.

Methodology
To analyze mixed methods literature trends in the investigated domains, the work collected data from the Scopus database. To this end, following the Scopus classification, the analysis selected the subject areas of business, management and accounting (BMA) and, as a term of comparison, social sciences. Concerning these areas, the paper considered English-language articles published between 2002 and 2021 that had the term mixed methods in their title or keywords. To identify these documents, the search was carried out at the end of January 2022 through specific queries on the Scopus database.
Then, to highlight (as far as possible) the most frequent subfields and the quality of journals that published the selected articles, the analysis gathered data from the Scimago database. In particular, based on the SCImago Journal Rank (SJR) and journal categorization, the paper shows the quartile 1 in which the publication outlet was ranked in 2020 and the thematic categories of the BMA area with which the journal is linked.
The application case concerns the Italian context where, since it is difficult to identify CCs, there is no up-to-date national picture of this type of enterprise. Following Creswell and Plano Clark (2011), the application case adopted a fixed mixed-methods approach, as this choice was made at the research design stage and not afterwards. In addition, considering the mixed methods design matrix (Figure 1), the adopted approach can be placed in quadrant 2, as it places qualitative and quantitative methods on the same level and applies them sequentially. The order of application is in line with Morse's suggestion (1991) for research on topics that cannot be considered mature and/or where there is a need for more knowledge. However, if one considers the articles that have the term mixed methods in their keywords, one notices that the gap is narrowing -although not by much -and this might suggest that researchers in the area of BMA are starting to use this methodology more.  Table  1

): International Journal of Production Research, Journal of Cleaner Production, International Journal of Production Economics, Evaluation and Program Planning and, Journal of the Operational Research Society.
Overall, the journals publishing mixed methods articles are of a good standard, as they are in the top two quartiles in most cases (see Table 1). In addition, to provide an overview of the research strands to which the selected articles can be linked, the thematic categorisation of the BMA area was considered. In this respect, there are few journals related to accounting; whereas journals related to the categories strategy and management, business and international management and marketing show a greater interest in mixed methods research. Again, a certain degree of openness emerges for the category related to management information systems. The selected articles deal with a variety of topics, among the most frequent ones are analyses concerning social interaction, innovation and environment, while papers concerning corporate social responsibility, intellectual capital, entrepreneurship or other meaningful topics are fewer in number. Given this, to promote the adoption of mixed methods in studies concerning entrepreneurship, this paper describes a pilot research project on community entrepreneurship in Italy.

Application case
The pilot research project was designed to answer several questions such as a) when did CCs originate and in which areas of Italy did they become more widespread? b) What are their distinctive features? c) What is the type of entrepreneurship that characterises them? d) What activities do they carry out? e) On average, how much wealth do they produce and how many employees do they have? Due to space constraints, it is not possible to present the results of an extensive research project in this contribution. Therefore, considering the aim of the work, the paper merely illustrates the adopted approach to answer the above-mentioned questions effectively and efficiently.
The research on Italian CCs started with the conduct of seven case studies selected considering different reference contexts (rural/urban) and life cycle phases. In particular, through face-to-face interviews, information was collected on different aspects, including the birth of the CC, its mission, strategy and business model, governance, etc. This qualitative analysis allowed the researchers to detect the basic characteristics and typically distinctive features of an Italian CC. This represented a key result because it allowed the researchers to recognise just under 250 CCs in the Italian Register of Cooperatives, which gathers all active Italian cooperatives (at the beginning of 2022 their number will exceed 110,000). Subsequently, the email contacts of these CCs were searched and this task triggered the design and administration of a questionnaire. The second part of the research made it possible to depict the phenomenon on a national scale, both by using archival analysis on economic data of the sample, and by delving into various topics through the questionnaire, which included not only the themes of the face-to-face interviews but also new questions concerning, for example, the factors that triggered the entrepreneurial initiative, the obstacles met (for example, in terms of education and training), the relationships with the context of reference, the ways to communicate the results to the stakeholders, etc. The results obtained are more than satisfactory since, considering that a valid email contact was not available for all CCs, the number of responses exceeded 24% of the sample. Thanks to all the data collected, the research project is now completing the analysis of CCs in Italy and preparing the related report.

Discussion
The literature trends revealed some interesting elements, e.g. focusing on the articles in which mixed methods are considered as characterizing elements (namely those with the term in the title). In this perspective, if one compares the number of articles in the selected subject areas to the respective total production of English-language articles in the Scopus database, it emerges that: mixed methods have a limited impact since in the years 2002, 2010 and 2021 they usually did not exceed 0.3% of the total scientific production; considering these years, researchers from the social sciences area have contributed more to mixed methods research than BMA researchers. Indeed, although starting from a lower percentage of articles (in 2002, in the social sciences area it was 0.007% of the total scientific production, whereas in the BMA area it was 0.023%), in 2021 the percentage has increased since in the social sciences area it is close to 0.3% and in the BMA area, it is just over 0.166%.
As for the BMA area, it is clear from Table 1  This might be influenced by branches of accounting, such as financial accounting, which are less predisposed to mixed methods, as they are traditionally linked to quantitative methods. In general, however, as Cameron (2011) pointed out, several barriers might have influenced the trends shown, those being philosophical (deriving from incompatibility of paradigms), cultural (prejudices against mixed methods), psychological (more confidence in a specific methodological approach), practical (preference towards routines and methods already known). Some of these obstacles are also increasing in importance because of the pressure on researchers to be productive (publish or perish). This is not good. Indeed, as Marchi (2021) noted, hyper-specialisation and the tendency to avoid certain types of research (probably the least profitable) do not benefit the growth of the scholar.
To promote mixed methods, it might be useful if widely recognized journals showed more interest in mixed methods articles. Indeed, although the ranking of the journals in the various categories of the BMA area is appreciable ( As for the application case, it was useful for the author to describe this pilot experience, not only to promote mixed methods in the domains where they were least used, but also to highlight the chance of achieving multiple purposes with this methodological approach. Indeed, since mixed methods can have non-mutually exclusive purposes, the presented experimental design is a good example of the way mixed methods can achieve different purposes. In particular, the research project described achieves the aims of: development, since the interviews supported the design and administration of the questionnaire; expansion, since the response rate and the diversity of the analyzed topics contribute to expanding the knowledge of CCs in Italy, and complementarity, as the results of qualitative and quantitative analyses and secondary data (archival analysis on economic data) were combined to produce a realistic picture of Italian CCs.

Conclusions
The paper presented the development of the literature on mixed methods in the last twenty years in social sciences and BMA areas. The analysis showed that in the latter area this methodological approach is not very widespread and proposed some determinants. However, these answers are not definitive, since it would be necessary to carry out a structured literature review. This represents a limitation of the work, but it should also be noticed that a structured literature review is a complex task, even if limited to a specific domain such as accounting, as noted by Grafton et al. (2011). This is certainly a future issue, but along with this, it is hoped that there will be an increase in the adoption of mixed methods in different research fields such as in analyses concerning social enterprises, corporate social responsibility, intellectual capital, integrated reporting, business model or entrepreneurship. To promote research in this direction, the design of a research project adopting mixed methods was briefly presented. By effectively integrating the methods adopted, it benefited from the strengths of each method and filled a gap in the current state of knowledge of community entrepreneurship in Italy.