Introduction

Qualitative research is supported by qualitative data and theoretical inductions. It enables underlying mechanisms and processes to be more clearly understood (Bouncken et al., 2021). It allows scholars to investigate the origins of phenomena, why they occur, and how they affect those involved (Williams & Moser, 2019). It has played an important role in the development of theories of new phenomena (Van Burg et al., 2020). Qualitative research methods help scholars explain phenomena that are not quantifiable. They are flexible, based on human behavior, and allow the symbolic dimensions and social meaning of different topics to be examined in one study (Mohajan, 2018).

Despite the relevance and advantages of qualitative methods, most scientific publications feature articles that use quantitative approaches (Bluhm et al., 2011). This includes the study of international business, where the impact of the use of qualitative methods on article citation is not clear. In general terms, qualitative methods have been used much less than quantitative research, a trend that continues (Doz, 2011). The proportion of use of these respective methods has been estimated to be 80% quantitative and 20% qualitative (Bansal et al., 2018).

More specifically, research using qualitative methods in international business can develop new finding about several understudied topics, such as the potential contribution of bi- and multi-cultural actors in context, the current actions of multinationals in the international arena, and the current problems of daily life facing people and companies carrying out international operations (Doz, 2011).

Therefore, the main objective of the present study is to determine whether the use of qualitative methods is a predictor of the impact of articles that use them from the perspective of the number of citations they receive. The study of this topic is important for two reasons. First, due to the less frequent use of qualitative methods in research on international business, the analysis of citations can contribute to researchers considering phenomena using qualitative techniques. Second, qualitative research is an important source of knowledge, but there is an erroneous perception among top journals that papers using qualitative methods lack scientific rigor and generate less impact in the academic community, so the present research will analyze this perception. The research question was, Do qualitative methods have more impact than others? To address it, 925 articles in the most prestigious journal in the area of international business, the Journal of International Business Studies (JIBS), were analyzed (Liesch et al., 2011).

The evaluation of the impact of scientific articles helps to advance science (Zhang & Wu, 2021). The determinants of citation have been addressed in the literature (Confraria et al., 2017; Molina-Azorín, 2012; Urlings et al., 2020). In general, results have shown that the use of certain research methods influences the rate of citation but that the type of country of origin of articles has no effect. The present study focuses on the reasons why articles using qualitative methods in the field of international business are cited. It encourages the use of qualitative methods because these can advance knowledge of the subject (Birkishaw et al., 2011).

The study is structured as follows. The academic literature relating to qualitative research methods and article citation is reviewed. The methodology used to collect the information needed to achieve the study’s main objective is then discussed. Third, the results from the statistical procedures are presented, and finally, several conclusions are drawn.

Literature review

Qualitative research methods

From a pragmatic point of view, qualitative research involves textual rather than numeric data (Punch, 2013). The main qualitative research methods include logic; ethnography; discourse analysis; case studies; interviews; observation; grounded theory; biography; comparative methods; focus groups; literary criticism; historical research; and content analysis (Cibangu, 2012). Each of these methods have particular characteristics that make them useful for investigating non-quantifiable phenomena.

Such methods—as has been touched on—have several advantages. First, they present in detail the feelings, opinions, and experiences of members of the sample. Second, they describe the human experience in a certain context. Third, they can be used to understand people’s behaviors. Fourth, they are flexible because they can be quickly modified if necessary (Maxwell, 2012; Rahman, 2017; Tsushima, 2015). Fifth, they can capture stages in the evolution of a given phenomenon (Langley, 2011).

When an investigation uses qualitative methods, the observations contained within the data can introduce generalizable knowledge (Banzal et al., 2018). These types of methods are suitable when the phenomenon and its context have been under-researched (Nadkarni et al., 2018). In addition, the researcher who applies qualitative methods becomes one with the phenomenon to be studied, which can be helpful in international business because the contexts and cultures can be so diverse (Birkinshaw et al., 2011).

The most beneficial aspect of qualitative research lies in its capacity to construct theory from three perspectives. First, it can provide more accurate descriptions of a phenomenon because the researcher is more immersed in it. This is not possible with quantitative approaches. Second, it can test theories by comparing the various findings in a specific field (Doz, 2011). Third, it can present opportunities for a greater understanding of phenomena and their contexts by introducing new variables and identifying new relationships (Charmaz, 2014; Shufutinsky, 2020).

Article impact and citation

Article citation is the most commonly used indicator to measure the impact of a scientific article (Lashitew et al., 2021; Lu et al., 2017). Lipetz (1965) was one of the first to recognise the significance of article citation. He identified four clusters: (a) the original scientific contribution or intent of the paper cited; (b) the contributions of the citing paper other than its original scientific contribution; (c) the identification of relationships between the citing paper and the cited paper; and (d) the scientific contribution of the cited paper to the citing paper. Another important element to consider is the field to which an article belongs because this can affect the rate of citation (Confraria et al., 2017). Comparisons between fields result in significant biases. This may be due to the dynamics of the field of study, for example, whether it is expanding or contracting, and the average number of references per article (Seglen, 1998).

According to Repiso et al. (2020) there are three essential factors for the citation of an article: accessibility (or the degree of difficulty of access, e.g., was it published in an open-access journal?), dissemination (i.e., the communicative capacity of the journal in which it has been published). and scientific authority (which may be related to the prestige of the author[s]). It is important to mention that article citation is not standard in the different disciplines of knowledge, and it may be used differently even in related areas (for example management and international business; Vieira & Gomes, 2010).

Several studies have addressed article citation in areas related to management. Figg et al. (2006) sought to uncover a relationship between the citation rate and the degree of collaboration. The results indicated that there was a relationship between the number of authors and number of times the article was cited. Jang (2021) examined whether the level of the publication was related to citation and again concluded that the number of authors was connected to the number of citations. Finally, Ferreira (2018) mapped the field of management using bibliographic coupling and co-citation to identify the most important articles, journals, and authors (and their networks of collaboration).

According to the above authors, international business can be used successfully to study the subject of article citation. However, the specific influence of qualitative methods has not been given a great deal of attention, so it is hoped that the present study will fill this gap in the literature.

Methodology

The methodology adopted for the study was proposed by Molina-Azorín (2012), though in the present case it is being used to investigate the impact of qualitative studies on international business, as has been noted. The JIBS was selected because it has for many years been the world leader in the field (Inkpen, 2001; Nielsen et al., 2020; Rialp et al., 2019). In addition, it does not publish articles related to other areas of management, which made it easier to achieve the objectives of the study. The Journal of International Business Studies (JIBS) has been a key reference in the field of international business since the 1970s and in 2021 was ranked in the first quartile in Journal Citation Reports (JCR), the most important index in scientific publications, with an impact factor of 11.382 (Impactfactorjournal, 2022). The use of articles published in JIBS as the data source indicates that the database focused only on international business.

To identify qualitative studies, a total of 925 articles from the period 2000 to 2020 were computed. Although the JIBS has been published since 1970, this period (21 years in total) was chosen because we consider it to be the time when its articles became more visible (mainly as a result of the development of online platforms). Research notes, editors’ commentaries, perspective articles, introductions, and dissertations were excluded. Only those articles that were classified on the JIBS platform as “Articles” and “Original articles” were considered. These empirical contributions were chosen because we wanted to analyze articles that contained methodology sections. Twenty-five articles from the original 925 were excluded because they did not meet the criteria, so the final sample was 900 (Table 1). We divided the articles into qualitative, quantitative, and mixed categories. We considered an article to be quantitative when it used databases and statistical procedures to obtain results. Likewise, to identify qualitative articles, we reviewed the methodology section to determine if the article includes a method consider qualitative such as the case study, content analysis, storytelling, literature review, and other methods that do not use statistical procedures.

Table 1 Types of articles in the sample (2000 − 2020)
Full size table

Table 1 shows the results of the analysis. There were 733 quantitative, 162 qualitative, and five mixed-method articles. The imbalance between the quantitative and the qualitative reflected the general trend in the use of the former in management studies.

Variables and statistical procedure

A regression analysis was performed to achieve the central objective of the present study. The dependent variable was article impact, which was measured with the citations of each article found on the JIBS website, which in turn were taken from the Journal of Citation Reports (JCR), one of the most important and reliable citation indexes (Haba-Osca et al., 2019). Given that article citations are affected by other factors, other variables were considered (Aguinis et al., 2020; Van Burg et al., 2020; Vieira & Gomes, 2010).

The first variable was article access. This corresponded to the number of times an interested party reviewed at least the abstract of the article. The data were obtained from the web page of each article within the JIBS. The second variable was article age. To measure this, a scale was designed where the number 21 was assigned to articles published in 2000, 20 to those published in 2001, and so on; those published in 2020 were assigned the number 1. The third variable was article length, which was measured according to the number of pages in the article. Whether an article has been published in a regular or in a special issue can affect citation, so a dummy variable was created to measure this. Articles published in regular issues were assigned a 0, and articles published in special issues were assigned a 1. Finally, the number of authors variable was operationalized by assigning a number according to the number of authors. For example, an article with two authors was assigned the number 2, and an article with three authors was assigned the number 3.

An article was classified as qualitative if the data were processed using a method such as content analysis, case studies, ethnography, interviews, and grounded theory. Finally, an article using qualitative and quantitative techniques was placed in the mixed methods category. A dummy variable was created to measure the qualitative method variable; a qualitative article was assigned the number 1 and a quantitative and a mixed-method article the number 0.

Results

The database employed in the study was constructed using data from the JIBS for the period 2000 − 2020. Nine hundred articles were analyzed according to the characteristics illustrated in Fig. 1. A dummy variable was used for articles that used qualitative methods. The following model was proposed:

Fig. 1
figure 1

Theoretical model

Full size image
$${y}_{i}={AC}_{i}^{{\alpha }_{0}}{AA}_{i}^{{\alpha }_{1}}{AL}_{i}^{{\alpha }_{2}}{NA}_{i}^{{\alpha }_{3}}{e}^{{\alpha }_{4}{SE}_{i} +{ \alpha }_{5}{D}_{i }+ {\alpha }_{6}{MM}_{i} + {u}_{i}}$$
(1)

where \({y}_{i}\) is the endogenous variable (article citation), \({AC}_{i}\) the number of times the article was accessed (article access), \({AA}_{i}\) the article age, \({AL}_{i}\) the article length, \({NA}_{i}\) the number of authors, \({SE}_{i}\) special issue, \({D}_{i}\) the dummy variable representing articles that used quantitative and mixed methods, \({MM}_{i}\) mixed methods, and \({u}_{i}\) the stochastic disturbance with zero mean and constant variance.

The model of Eq. 1 is proposed, first, to carry out the homogenization of the database, because the combination of level variables and dummies would generate estimation problems, such as heteroskedasticity, autocorrelation, or structural breakage. Second, the publication of a scientific article in a journal does not depend on its individual characteristics, but on a combination of variables as they are introduced in the model—hence its multiplicative form. Finally, although the qualitative response regression models are adjusted to the phenomenon analyzed, the probability that an article is cited or not was not the target information, according to the methods used to test their hypotheses, but rather the level of dependency of the characteristics proposed in the model of Eq. 1 for an article to be cited.

The maximum verisimilitude method was used to estimate Eq. 1. It was represented by the following equation:

$$L=\prod_{i = 1}^{N}\frac{1}{\sqrt{2\pi {\sigma }^{2}}}{e}^{- \frac{{\left({y}_{i} - \mu \right)}^{2}}{2{\sigma }^{2}}}$$
(2)

where \(L\) is the function of maximum verisimilitude, \({y}_{i}\) the study variable (in this case the number of citations, \({\sigma }^{2}\) the variance, and \(\mu\) the mean. If we suppose that is possible to estimate the mean and variance of Eq. 2, logarithms must be applied:

$$\mathrm{ln}L=-\frac{N}{2}\mathrm{ln}2\pi -\frac{N}{2}\mathrm{ln}{\sigma }^{2}-\sum_{i = 1}^{N}-\frac{{\left({y}_{i}-\mu \right)}^{2}}{2{\sigma }^{2}}$$
(3)

where \(\mathrm{ln}L\) is the natural logarithm of the function of maximum verisimilitude, \({y}_{i}\) the variable of study, \({\sigma }^{2}\) and \(\mu\) the variance and the mean required for estimation. Exogenous variables have to be introduced; hence, we suppose that \(\mu\) can be rewritten as \(X\upbeta\).

$$\mathrm{ln}L=-\frac{N}{2}\mathrm{ln}2\pi -\frac{N}{2}\mathrm{ln}{\sigma }^{2}-\frac{1}{2{\sigma }^{2}}{\left(y-X\upbeta \right)}^{^{\prime}}\left(y-X\upbeta \right)$$
(4)

where \(\mathrm{ln}L\) is the logarithm of the function of maximum verisimilitude, \(y\) the vector that has the endogenous variable, \({\sigma }^{2}\) the variance, \(X\) the information matrix that contains the exogenous variables and vector \(\upbeta\) the unknown parameters. By maximizing Eq. 4, the value of the parameters of Eq. 1 can be obtained.

Table 2 shows that, on average, each article was cited around 18 times, with a standard deviation of 36 and a bias of 5.05, so most of the articles published in the JIBS were cited more than the mean. The kurtosis for citation was 127.08, which indicates that the Gaussian bell shape was of the leptokurtic type. The rest of the statistics for the variables can be explained in the same way. The correlation matrix indicates that the endogenous variable had a positive relationship with article access and age (statistically significant at 95% confidence). The null hypothesis was rejected for the other variables. The other data in the correlation matrix were introduced to discover the relationships between exogenous variables and to avoid a possibly spurious regression.

Table 2 Basic statistics and correlation matrix
Full size table

Table 3 shows the different regressions that were carried out on Eq. 1 using maximum likelihood. Only those parameters associated with the variables that contained two asterisks were statistically significant at 95% confidence; hence, the number of citations for an article depended positively on the number of times it was accessed, its age, and whether it featured in a special issue. The figure depended negatively on article length and whether quantitative methods were used. For example, if the number of times the article was accessed increased by one unit, keeping all else constant, we would suppose that the number of citations would increase by 0.5958. However, if the length of the article increased by one unit, the number of citations would decrease by 0.8802. The other variables could be explained in the same way.

Table 3 Regression results
Full size table

On the whole, the exogenous variables explained 70.77% of the number of citations. The Durbin − Watson statistic rejected the null hypothesis of autocorrelation. In Model 2, the variables that were statistically insignificant in the previous model were eliminated, and as can be seen, the value of the parameters associated with the variables grew marginally and kept its sign. In Model 3, Eq. 1 was estimated by changing the dummy variable from quantitative methods to qualitative methods (i.e., quantitative articles for qualitative articles) to measure the effect on the number of citations. As can be observed, there was no change, since this variable was statistically insignificant at 95% confidence. It was concluded that Model 2 was the best fit.

Table 3 shows three different estimates of the model of Eq. 1 using the maximum likelihood method. For model 1, Article Access, Article age, Article length, Special issue, and Quantitative articles are the variables that are statistically significant at a 95% confidence level, and the rest are not (Number of authors and Mixed methods). The variables Article Access, Article age, and Special issue maintain a positive relationship with the endogenous variable, which means, for example, that as Article access increases the number of article citations—as the variable that most influences Article age in the endogenous variable—the variables Article length and Quantitative articles have an inverse relationship, with the first most negatively influencing the number of citations. Finally, the set of exogenous variables explains the number of citations of articles published in the journal by 70.77% and does not contain an autocorrelation of order 1; the latter can be observed in the Durbin–Watson statistic.

For model 2, the variables that are statistically insignificant from the first model (Number of authors and Mixed methods) were excluded to avoid possible multicollinearity or to eliminate their influence on the final estimation result, and, as can be observed, the changes were marginal in all cases. In model 3, the variable Quantitative articles was excluded and its counterpart (Qualitative articles) was introduced, to determine how much this characteristic influences the number of citations and, as a result, conclusions similar to those of model 1 were obtained—that is, Article Access, Article age, Article length, and Special issue are statistically significant at a 95% confidence level. It was also found that the signs are the same as before, with the exception that the Qualitative articles variable is not significant in explaining the endogenous variable. The conclusions obtained in models 1 and 3 are that the variables Article Access, Article age, Article length, and Special issue best explain the number of citations for articles published in JIBS, and the method used in the analyses (Qualitative articles, Quantitative articles, or Mixed methods) has no influence on the number of citations, because in the first model they have a negative sign and in the third model they are statistically insignificant.

Conclusions

The objective of the present study was to determine if the use of qualitative methods was a predictor of article impact as reflected by the number of citations. The general conclusion is that qualitative methods had a positive influence on impact. In particular, articles in the field of international business were cited more if they employed qualitative methods. This is a significant finding because qualitative methods can consolidate theories and validate quantitative research results (Molina-Azorín, 2012). It suggests an avenue by which international business researchers might increase the impact of their work.

The results of this work are consistent with others in the literature on citation prediction—that is, a given variable has a positive effect on citation. Although they coincide in general terms, they differ because other study variables are used (Abrishami & Aliakbari, 2019; Confraria et al., 2017). This research supports the results according to which the methods used to carry out an investigation have a positive relationship with the level of relevance it has in the scientific community (Bergh et al., 2006; Molina-Azorín, 2012; Thelwall & Nevill, 2021).

The present study makes two main contributions to the study of international business. First, it demonstrates the importance of the need for qualitative approaches in an area that is dominated by quantitative research (Bansal et al., 2018). Second, it reveals the close relationship between citation rates and research methods.

Other factors can increase the number of times an article is cited. These include the number of hits the article receives, its age, its length, and whether it features in a special issue. It may be concluded that—although the qualitative method is an important part of the story—if authors want to increase the impact of their work, they should be aware of a range of influencing factors. It is also important to note that very long articles do not help to generate new citations and may therefore lessen impact.

In terms of future research, the present study’s methodology could be applied to other areas of management (e.g., human resources, strategy, and leadership). In addition, citations could be measured according to sub-method (e.g., case studies or storytelling). Variables other than those referred to herein could also be considered.

One of the main practical implications of the present study is that researchers in international business who want to publish higher impact articles should consider using qualitative methodologies. It can be difficult to obtain data on international business, so qualitative methods provide an alternative. Given the positive influence of qualitative methods on citation, researchers should improve their information-gathering techniques. They should also try to use titles that draw people´s attention and, in some cases, reduce the number of pages of their publications (in other words, be more concise).

The need for innovation and creativity is a constant in management science (Bogers et al., 2018). Qualitative methods allow researchers to blend different approaches, for example, case studies, ethnography, and/or interviews, and thus to be more innovative and creative (Ashworth et al., 2019). This can lead to greater knowledge not only of international business but management generally.