Cross-cultural research in management control & accounting A meta-analysis of research to date

(1)

Cross-cultural research in management control & accounting

A meta-analysis of research to date

Jasper de Ruiter (S1690892) Master Thesis University of Groningen Faculty of Economics and Business

MSc Business Administration

Specialization: Organisation &Management control

February 2013

Word count (excluding appendixes): 14.303 Total 19.657

Supervisor: dr. C.P.A. Heijes

(2)

Abstract

This study investigates the developments and problems in cross-cultural research in management control and accounting. This is done via a meta-analysis of 34 different studies. There are four main findings of this meta-analysis. First it was found that there is an intensification of publications as time progresses of cross-cultural research in management control and accounting . Second it was found that the field of research heavily relies on the work of Hofstede. The third finding is that many of the studies have unsupported or partially supported hypotheses. Finally the predominant method of data gathering of cross-cultural research in management control and accounting is still the usage of survey questionnaires.

Suggestions are made based upon these findings to improve future research.

Keywords:

(3)

1. Introduction

The continuing globalisation of markets and the increasing importance of multinationals have led to an ever growing body of research on the effects of national culture on all aspects of management. This globalisation has led to the opportunity but in other cases necessity for companies to become internationally active. A specific field of research in this topic is the influence of national culture on management control systems (MCS) and accounting practices. This field of research has tried to explain differences between countries and their effect on the management control and accounting practices. For the companies that are expanding into unknown cultures this raised the question whether to adopt their current MCS and accounting practices or adapt their MCS and accounting practices according to the cultural influences of overseas nations.

The MCS is an important tool for organizations to increase the probability that employees act and make decisions that are in the best interest of the organization (Chow, Shields & Wu, 1999). Accounting information can also be used for making business decisions or for internal and external reporting. The influence of culture is also visible on accounting practices that can differ between different countries. These differences can be influenced by national culture. By understanding the influence of culture on MCS and accounting practices these can be optimized for the national culture of the country. A poor fit between national culture and the MCS can lead to management control failures. These failures can lead to large financial losses, reputation damage and possible even to organization failure (Merchant & van der Stede, 2007).

National culture has been a difficult to understand construct in relation to MCS and accounting practices. Many authors have different opinions of the influence of national culture on MCS and accounting. This field of research has been troubled by problems with the operationalization of national culture. A vast variety of different operationalization’s in the form of value dimensions have been created over the past decades (E.g. Kluckhohn & Strodtbeck, 1961; Hall, 1976; Hofstede, 1980; Ronen & Shenkar, 1985; Gray, 1988; Trompenaars & Hampden-Turner, 1997; House et al., 1997). These difficulties with operationalizing national culture in cross-cultural research have led to important limitations and questions regarding the value of these studies.

In this thesis the relation between national culture, MCS and accounting practices will be the central topic of research.

1.1 Subject of the research

Two of the main constructs used throughout this thesis are national culture and management control systems. Therefore it has to be clear what exactly is meant by this.

National culture has had many different definitions by various authors over time but has also been defined in diverse ways in other research fields. National culture is defined as the values, beliefs and assumptions learned in early childhood that distinguish one group of people from another (Beck & Moore, 1985). One of the most cited definitions of national culture in management and accounting is that of Hofstede (1989): “the collective programming of the mind that distinguishes the members of one organisation from others” (Bond, 2002). Both of these definitions of national culture are in line with each other.

(5)

has shifted to one that embraces a much broader scope of information (Chenhall, 2003).

1.2 Goal of the research

The ultimate goal of this research is to identify the problems and trends in cross-cultural research related to management control and accounting. By identifying these trends and problems they can be overcome by future researchers to improve their work and research design. This in turn could lead to a better understanding of the phenomenon of the influence of national culture on accounting and MCS.

The goal is however not only to provide guidance for future research but also to provide a critical note about research that has been done to date. With these criticism practitioners could get a better understanding of the problems in the field of cross-cultural research related to management control and accounting. These practitioners can be managers who are implementing their management control system in a different culture and want to know the limitations of the empirical research on this topic. Understanding these limitations could help practitioners to value the research done to date.

1.3 Research questions

The main research question of this thesis will be:

What are the trends and problems in cross-cultural research in management control and accounting literature to date and what suggestions for future research can be made? This research question will be answered by using several sub questions.

-Which cultural concepts are used in cross-cultural research in management control and accounting literature and what is the legitimization that the researchers provide for this choice?

-Are there many unsupported or partially supported hypotheses in cross-cultural research in management control and accounting literature?

-What is the predominant data collection method in in cross-cultural research in management control and accounting literature?

1.4 Research method and conceptual model

The research method will be a meta-analysis of the research done from 1980 onwards to date on the topic of cross-cultural research related to management control and accounting. A meta-analysis is a research that combines the outcomes and results of several different studies in the hope of

identifying patterns among study results, sources of disagreement among those results, or other interesting relationships that may come to light in the context of multiple studies (Thompson,& Pocock, 1991). As mentioned by Hunter and Schmidt (2004) the purpose of an meta-analysis is to estimate what the results would have been had all the studies been conducted without

methodological limitations or flaws. The goal of this thesis is however slightly different, the focus will not exclusively be on the outcomes of the studies but rather be to investigate the trends and

problems in cross-cultural research in management control and accounting literature.

(6)

that results can be summarized into a table that can be quickly overseen, instead of having to read through numerous articles. By summarizing the results and other data from multiple studies by performing a meta-analysis is would be possible to create the information needed to answer the research questions. A meta-analysis offers the framework to critically evaluate and compare results (Hunter & Schmidt, 2004). A meta-analysis offers the best research method to answer the research questions of this study and is therefore used.

In order to create this meta-analysis relevant literature will be used from mainstream English scientific journals from 1980 and onwards. The selection criteria for these articles will be presented in the meta-analysis section.

Creating a conceptual model for a study which is designed as a meta-analysis is difficult as there are many factors and trends that can be found. However a simple model can be created as depicted in figure 1. National Culture Management Control Systems Accounting Practices

Figure 1: Conceptual model

The relation between national culture and management control systems and accounting practices will be an important part of this research. However the thesis itself is not about this relationship but focusses on the problems and trends of this field of research that can be identified in the meta- analysis .

The thesis will consist of several sections. In the next section the different used taxonomies of culture will be discussed to give an overview of main taxonomies that can be used in cross-cultural research. In section three the meta-analysis will be presented, this section also deals with the design of the

meta-analysis. After conducting the meta-analysis the research findings will be presented in section four. Section five will consist of a discussion of the research findings. In section five it will also

be discussed if there are any lessons learned from similar previous studies.

(7)

2. Taxonomy of culture

In this part the main different taxonomies of national culture will be discussed. Though not all of these are used in the cross-cultural studies that are included in this study. They are however included to give a complete overview of the main taxonomies that can be used in cross-cultural research.

One of the oldest definitions used in cross-cultural research is the taxonomy of Strodtbeck and Kluckhohn that dates back to 1961. Antropologist Hall (1976) created the high low context continuum based on communication styles. One of the most used taxonomies of culture are the four dimensions of Hofstede (1980) and his later additions. Ronen and Shenkar (1985) created a cultural classification of countries. Trompenaars and Hampden-Turner have made a cultural taxonomy based upon seven dimensions. In an extensive research by House et al. (1997) project GLOBE was created to conceptualize and develop measures of culture. Finally the work of Gray (1988) that focusses on accounting cultures will be discussed. Additions to the field of national culture taxonomies since the work of Hofstede (1980) have been less validated than the work of Hofstede (Chow et al., 1999).

These different taxonomies of culture will be discussed in chronological order.

2.1 Kluckhohn and Strodtbeck

Kluckhohn and Strodtbeck (1961) have proposed six different cultural dimensions. These dimensions are (1) the nature of people (2) the person’s relationship to nature (3) the person’s relationship to others (4) the modality of human activity (5) temporal focus of human activity and (6) conception of space (Singh, 2004; Kluckhohn & Strodtbeck, 1961). The first dimension is to measure whether people are orientated in a selfish or more societal orientated way. Another part of this dimension are the assumptions people make about the basic beliefs and nature of others. The second dimension is to measure how people see nature, either as a source for resources or if they live in harmony with nature. By looking at the relationship that people have to others it is possible to see whether people focus on their own lives or support others. The fourth dimension, the modality of human activity, is strongly related to the status of people. For example do people derive their status from who they are or because of what they do? The fifth dimension answers the question whether the society is focused on the past, present or future.

Some cultures are more focused on the past for example by focussing on tradition and ancestors. The sixth dimension, the conception of space, is to look into who owns the space. Space can be owned privately by individuals or by governments. Another aspect of this dimension is privacy, some cultures would be less focused on privacy than others.

By answering questions related to these dimensions it would be possible to create different cultural groups based on the different preferences people have (Hills, 2002). The questions used for this classification of different cultural groups are (Birnberg & Snodgrass, 1988; Kluckhohn & Strodtbeck, 1961):

(8)

The theory of Kluckhohn and Strodtbeck remains widely used to this day, however mostly in the field of psychological related cross-cultural studies (Hills, 2002). In the cross-cultural studies related to management control or accounting only Birnberg and Snodgrass (1988) used the value dimensions of Kluckhohn and Strodtbeck.

2.2 High-low-context continuum of Hall

In 1976 anthropologist Hall created the High-low context continuum. This is a taxonomy of culture that divides countries in either high or low context cultures. This is measured by looking at the verbal and non-verbal communication used in different cultures. The low context cultures are typically identified by a focus on straight communication for example one is supposed to use facts, figures and abundant details to communicate (Granlund & Lukka, 1998; Hall, 1976). Whereas in the high context cultures communication is more implicit and means that the listener has to focus more than in low context cultures to interpret the message effectively (Granlund & Lukka, 1998; Hall, 1976). Examples of low context countries include Germany and the United States of America while the Far East countries are generally identified as high context cultures (Hall, 1976)

2.3 Hofstede

In 1980 Hofstede concluded that the concept of culture suffered a lot from vagueness. Hofstede (1980) identified four cultural dimensions that could be used to classify national culture. Hofstede found the dimensions by theoretical reasoning and by using one of the largest surveys ever conducted over two individual time frames. The four dimensions used by Hofstede (1980) are power distance, uncertainty avoidance, individualism versus collectivism and masculinity. Power distance indicates the extent to which power is distributed unequally among people in institutions or societies (Hofstede, 1980). A high score on power distance for example means that there are extensive hierarchical structures in place. A low score on power distance on the other hand could for example mean that everybody has equal rights. The second dimension is uncertainty avoidance, this indicates the extent that society feels threatened by uncertain or confusing situations (Hofstede, 1980). Some ways to avoid uncertainty are by providing greater career stability, more formal rules and not tolerating deviant ideas and behaviour, believing in absolute truths and the attainment of expertise (Hofstede, 1980). An example of uncertainty avoidance is that Dutch managers prefer a fixed salary unlike US managers who appreciate bonuses more (Jansen, Merchant & van der Stede, 2009). In the study of Jansen et al. (2009) the Dutch managers showed a higher level of uncertainty avoidance than the US managers. The third dimension is individualism versus collectivism, this dimension indicates whether people feel individualistic or collectivistic. Individualist would only take care of themselves and of their direct family. A collectivist is the direct opposite of an individualist and is characterised by a strong social framework in which other people from their group or organisation are supposed to take care of each other (Hofstede, 1980). The fourth dimension, masculinity is to indicate to which extend the “typical” male values are in use in the society. Hofstede (1980) identified assertiveness, the

acquisition of money and things, and not caring for others or the quality of life as typical masculine behavioural traits.

(9)

orientation are mostly Asian and are orientated on family and traditions (Hofstede & Minkov, 2010). Countries that score low on long term orientation are the US and UK, these countries are more able to change and are more creative.

More recently in 2010 Hofstede and Minkov added a sixth value dimension. This dimension is indulgence versus restraint. This dimension is used to indicate the difference between free and less free countries. Indulgence means that the society is free and natural human drives related to enjoying life and having fun are stimulated (Hofstede & Minkov, 2011). Restraint on the other hand means that this is more restricted and the society is more regulated by means of strict social norms (Hofstede & Minkov, 2010). The sixth value norm has not yet been used in cross-cultural

management and accounting research.

Each of the discussed dimensions would get a score from zero to 120 points, Hofstede (1980) indicated that only a 20 point differences between nations can be considered a substantial cultural difference.

2.4 Ronen and Shenkar

In 1985 Ronen and Shenkar created a cultural classification of countries. This clustering of countries was done on the basis on patterns of similarity in in employees attitudes towards work and how well it met their needs (Ronen & Shenkar, 1985). In total eight country clusters where identified and four countries remained independent of the clusters. These clusters are Near Eastern, Arabic, Far Eastern, Latin American, Latin European, Anglo, Germanic, and Nordic. The independent countries are Brazil, Japan India and Israel. The cluster classifications were made after a widespread review of various earlier researches where the respondents answered questions about the importance of various work goals, the extent to which work satisfies certain needs, organizational and managerial issues, and the nature of work roles and interpersonal relationships. However the work of Ronen and Shenkar does have limitations in total only about 50 countries are included and large parts of the world such as the majority and Africa and the Middle East are not included in the research (Ronen & Shenkar, 1985). The work of Ronen en Shenkar is derived directly from the seminal work of Hofstede (1980) (Johnson, Cullen, Sakano & Bronson, 2001).

2.5 Trompenaars and Hampden-Turner seven dimensions of culture

Trompenaars and Hampden-Turner (1997) have created a model with seven dimensions to measure culture. The first five dimensions deal with the interactions and interrelations between different people. The dimensions of Trompenaars and Hampden-Turner (1997) are:

1. Universalism vs. particularism. The goal of this dimension is to describe how different people judge each other’s behaviour. An Universalist is an individual who feels that the same rules apply to all people in all situations. And therefore everybody needs to follow these rules. The opposite of an universalist is a particularist. The particularist does look into specific situations and personal backgrounds. The particularist treats his family and friends as best as possible. Other people are on their own. The particularist clearly makes a distinction between different individuals. In reality there can be a trade-off between these two values and they can reinforce each other (Trompenaars & Hampden-Turner, 1997).

(10)

of modern culture for example in mass media. The two dimensions are more complementary in the work of Trompenaars and Hampden-Turner (1997) than in the work of Hofstede (1980). For example in an individualist culture individuals are punished for their failures, a result of this can be that the individualist could become a better team player.

3. Neutral vs. emotional. The third dimension of Trompenaars and Hampden-Turner (1997) is used to describe the feelings of people. Neutral people are controlled and do not act irrational. They don’t display their feelings to a great extent, but this does not mean that neutral people are cold or don’t have feelings. Emotional people show their emotions clearly by laughing, smiling, grimacing, scowling and gesturing. Emotional people can also act directly upon their feelings.

4. Specific vs. diffuse. People that live in a specific culture show the following characteristics: directness, being to the point, purposeful in relations and high moral standards. Specific-orientated people are often more open in public but closed in private environments (Trompenaars &

Hampden-Turner, 1997).In diffuse-orientated cultures people are more closed in public but open in private situations. The characteristics of the diffuse-orientated people are indirect, circuitous and seemingly aimless in forms of relations (Trompenaars & Hampden-Turner, 1997).

5. Achievement vs. ascription.The fifth dimension is to see whether people are given their status or have earned it. In cultures that are focused on achievement status is directly related to individual performance. In ascription- orientated cultures status is based upon seniority and hierarchy. Status can also refer to religion or origin and can be based upon traditions in ascription based cultures. 6. Sequential vs. synchronic. The sixth dimension is how people deal with time. Trompenaars and Hampden-Turner (1997) identify two ways that people can plan their activities in a sequential or synchronic way. A sequential individual has an order of their activities worked out and starts a new one after completing the past stage. They schedule their time into great detail. In synchronic time management the past, present and future are interrelated. People who plan their time synchronic are planning multiple activities at the same time and are usually working on several projects at once. 7. Internal vs. external control. The seventh dimension is how different cultures see their environment. The environment can be controlled internally. In this culture the organisation is controlled by internal mechanisms. The external control is the opposite and sees the actions of the organisation as a reaction to the environment and is thus controlled by the external environment.

2.6 GLOBE nine dimensions of culture

Over 170 researchers worked on project Global Leadership and Organizational Behaviour

Effectiveness or in short GLOBE. The main author and starter of the GLOBE project is House et al. (1997). The goal amongst others was to conceptualized and develop measures of culture. The result is nine different dimensions that conceptualize culture.

Project GLOBE was partly inspired but not based on the previous work of Hofstede. And one of the goals of the project was to be a correction of Hofstede’s value dimensions. The GLOBE researchers constructed their dimensions of national culture just like Hofstede from variables that correlate across nations. The researchers of project Globe felt that some of Hofstede’s dimensions lacked face validity, the dimensions of Hofstede measure something different than their name suggest. (Minkov and Hofstede, 2011) Some of the dimensions of project GLOBE have the same name as in Hofstede’s taxonomy but measure different underlying constructs.

(11)

concentrated at the higher levels of an organization or government.

2. Uncertainty Avoidance- the degree to which people attempt to avoid uncertainty by depending on for example rules and regulations.

3. In-Group Collectivism- the degree to which people express pride, loyalty and cohesiveness in their families.

4. Institutional Collectivism- the level in which institutions collectively distribute their knowledge and resources and the collective distribution of resources.

5. Gender Equality- the extent to which gender equality (or inequality) is embedded in a culture. 6. Assertiveness- the extent to which people are assertive, provocative, or aggressive in social Relationships with other people.

7. Performance Orientation- the level that people try to improve their own performance and that of others.

8. Future Orientation- the degree that people focus on their future. This can be done via activities such as planning or investing for the future.

9. Humane Orientation- the degree to which people encourage each other to be fair, selfless, sociable, generous, compassionate and kind.

The dimension gender equality and assertiveness replaces Hofstede’s masculinity dimension. Also collectivism is replaced with two instead of one dimension.

2.7 Gray accounting values

Although Gray (1988) is not really a definition of national culture it more a definition of accounting culture. Gray suggests the accounting values of professionalism, uniformity, conservatism and secrecy can be used to explain and predict differences between national accounting models. The work of Gray is based on the work of the value dimensions of Hofstede (1980) and review of various accounting literature. The work of Gray is mainly used in the (management) accounting literature.

The first dimension professionalism versus statutory control indicates the level of control that is exercised either by the accountant´s own professional values or by rules, laws and regulations. The second dimension, uniformity versus flexibility indicates whether the country prefers a uniform application of accounting practices or a more tailor made set of accounting practices that fit

(12)

3. Meta-analysis

In this section it will be discussed how the meta-analysis was performed, in order to provide the underlying reasons for some of the specific choices made in this study.

The used articles are all from main English scientific journals from 1980 onwards studying cross-cultural relationships in management control and management accounting. It was chosen to only use these articles from main English scientific journals to ensure a level of quality of the studies. Both management control and accounting were used to show if there would differences or

similarities between the two closely related fields of research. Another reason for this choice is that there is a larger set of studies presented in the overview than when only focussing on management control. The focus of the articles has to be on national culture and it has to be a cross-cultural study. Articles that show a focus on company culture are therefore not used. Because of this the list of articles is not an exhaustive list of articles on cultural aspects related to management but only includes the articles that shows a strong focus on the before mentioned points. This is also why the list of articles differs for example used by Harrison and McKinnon (1999) in a similar study which used different selection criteria.

The reason only articles from 1980 and onwards are selected has to do with the fact that the field of cross-cultural research is virtually none existing before this date and that the pre-1980 studies have a large chance of being outdated with the rapidly changing environment.

Keywords used during the search for appropriate literature were: national culture, cross-cultural, management control (system) and management accounting. It is important to note that although closely linked on some points, studies focussing on corporate culture are not included.

The overview as presented in table 1 consist of 34 articles of which 32 are an empirical study and two are a literature study. Although a focus of the search definitely was on empirical studies that

compared multiple countries the critical review of McArthur (2006) is included as it summarised multiple field studies. The study of Heidhues and Patel (2011) is included because although it is a literature research it also is a case study using data from previous studies.

The choice was made for the overview to show some general information, such as the authors, title and the year the article was published. The year the article was published can be an important factor in discovering if there are trends over time. This is also the reason that the articles are sorted by date of publishing and not in alphabetical order. The research method of the study is also included. This is done to see whether there are specific research methods such as for example survey

questionnaires that form the backbone of most of the research done to date. Furthermore the used conceptualization / taxonomy of culture or more specific the cultural dimensions used are included. Hofstede (4 dimensions) means that only the four dimensions are used as first introduced by

Hofstede in 1980 (power distance, individualism, uncertainty avoidance and masculinity). Hofstede added a fifth dimension in the nineties; Confucian dynamism or as later referred to as long term orientation. In 2010 Hofstede and Minkov added a sixth value dimension but this has not yet been used. This means that in table 1, Hofstede (5 dimensions) indicates that power distance,

individualism, uncertainty avoidance, masculinity and long term orientation are used.

(13)

unsupported hypothesis. The hypotheses are also included to be able to see if the way the

hypotheses are formulated could be of importance. It is also done to see if most studies work with hypotheses. Table 1 can be found in appendix 1.

(14)

4. Presentation of the research findings.

In the following section the results of the meta-analysis will be presented. The results will be presented as factual as possible by creating tables and by providing percentages. This is done in order to be able to present the results with as little bias as possible and without providing a personal opinion. The presented findings in this section will be discussed in section five. In section five possible explanations for these findings will also be given.

4.1 Intensification of research as time progresses

When looking at table 1 one of the first observations is that in the 33 year timeframe from 1980 till now the majority of the articles have been published in last half of this period. Out of the 34 only 8 articles are from before 1996. This means that over 75% has been published in the last 16 years.

4.2 Use of cultural concepts

Out of the 34 articles 22 use one or all of the dimensions of Hofstede. It has to be noted that not all articles treat culture as whole but focus on some dimensions of culture. This of course has a negative influence on the reliability of those studies as their reliability can be compromised by focussing on parts of culture. Out of the 34 articles only 18 use all of the dimensions of culture of their selected taxonomy. In case of Hofstede this means the use of four or more dimensions, as more dimensions have been added to the work of Hofstede over time. Additionally it has been argued by Hofstede that the dimension of long versus short term orientation is only relevant when comparing Asian countries

(Hofstede & Bond, 1988).

Number of times used Chosen taxonomy of national culture:

23 Taxonomy of Hofstede (1 or more/ all dimensions used)

5 Gray accounting values

2 Ronen and Shenkar, cultural classification

1 Hall, High-low context continuum

1 Taxonomy of Kluckhohn and Strodtbeck

1 Six cultural dimensions: Materialism, Time, Collectivism/Individualism, Uncertainty Avoidance Power Distance, and Efficiency

1 Beliefs about the purpose of corporations, masculinity (Hofstede) and long term orientation (Hofstede)

1 Hampden-Turner and Trompenaars (used in combination with Hofstede)

4 No specific definition of culture

39 Total

Table 2, taxonomies used.

Because the research of Sudarwan and Fogarty (1996),Tsakumis (2007) and Chand, Cummings and Patel (2012) use both the work of Hofstede and Gray the total of number of used taxonomies is higher than the 34 articles presented in table 1. Richards (2000) also uses multiple different taxonomies of culture namely the work of Hofstede (1980); Ronen and Shenkar (1985); Hampden-Turner and Trompenaars (1997).

(15)

of culture from Hofstede without using complementary literature or without making modifications. However when looking at the other used operationalization’s of culture the influence of Hofstede becomes even larger. As discussed earlier, the work of Gray (1988), Ronen and Shenkar (1985) can be directly linked to the work of Hofstede. Two other studies have adopted parts of the work of

Hofstede. Brice (2012) uses six dimensions to operationalize culture from different authors:

materialism came from Kelley, Whatley and Worthley (1986), time as researched by Levine and Wolf 1985, Power distance, individualism/ collectivism and uncertainty avoidance were used from

Hofstede (1980) and efficiency is from McClelland (1961). Jansen, Merchant and van der Stede (2009) used masculinity and long term orientation based on the work of Hofstede but also added beliefs about the purpose of corporations based on the work of Boselie, Paauwe and Jansen (2001). When also counting these studies an impressive 82% of all studies are influenced by Hofstede.

That leaves only three other authors of cultural taxonomies being cited in the selected literature. Namely the work of Hall, Kluckhohn, Strodtbeck, Hampden-Turner and Trompenaars. However the work of Hampden-Turner and Trompenaars was used in combination with amongst others the work of Hofstede. Four other studies treated cultures as whole and have not analysed national cultures in separate dimensions.

4.3 Argumentations for using the cultural concepts

It is obvious that the work of Hofstede is clearly the most used in cross-cultural management studies, something that Harrison and McKinnon also concluded in 1999. However amongst others Harrison and McKinnon (1999) also point out that the work of Hofstede has its limitations. A next step therefore would be to see why so many of the authors would have chosen to do this and why others opted to choose (also) a different taxonomy or a modified version of the work of Hofstede. It is striking to see that out of all the taxonomies explained in the previous section, that the Project GLOBE, the work of House et al. (1997s) isn´t even be used once.

Table 3 presents the reasons that the researchers have given for choosing the taxonomy of Hofstede. Again some authors mention multiple reasons so the total number of reasons is more than the amount of authors that use Hofstede.

Number of times mentioned Reason:

5 Take the work of Hofstede for granted, testing

which taxonomy of culture is the best is beyond the scope of the research.

8 Using the work of Hofstede because it shows

divergence between two or more nations.

7 The work of Hofstede is empirically tested and

also proven by other authors.

6 Because Hofstede is the “most” widely used and

cited taxonomy/ Hofstede is seen as the predominant approach to culture

2 Hofstede is chosen because this means the

(16)

1 Hofstede has not yet been used for the specific type of research

2 No reason given.

31 Total

Table 3 Reasons why the work of Hofstede was used.

From table 3 it can be seen that the most used reason for the choice of the work of Hofstede is because it shows divergence on one or more dimensions between nations. This immediately raises the question whether this could not have done using a different taxonomy of national culture. The other three predominant arguments used are that the work of Hofstede is been empirically proven by multiple others; testing the work of Hofstede and other taxonomies to find out which one is the best is beyond the scope of the research; because Hofstede is the most widely used and cited taxonomy. Especially the reason that the most widely used and cited taxonomy is a bit of a vague argument when not combined with other arguments for example with the empirical support that has been found for the work of Hofstede or the need to compare of generalize outcomes.

Some of the authors provide critique on the work of Hofstede, the most noteworthy of these are HassabElnaby and Mosebach (2005). They argue that Hofstede is not actually measuring culture because he separated subjects by nationality rather than culture and cultures do not necessary equate to nations. It has also been suggested that, given the proxies used, Hofstede was measuring socio-economic factors rather than culture HassabElnaby and Mosebach (2005, p22). However they still use Hofstede and give as argument for this: “Because the isolation of the impact of the social environment, in a business setting, is the goal of this study, the use of Hofstede’s dimensions of culture is appropriate” HassabElnaby and Mosebach (2005, p22) .

The reason Salter and Sharp use Hofstede is that in the work of Hofstede only a small difference on individualism between USA and Canada is identified. This is the direct opposite of the other seven researches that chose to use Hofstede because he identified a large divergence between nations. Salter and Sharp (2001) is also the only study that focusses specific on the role of small cultural differences.

The reasons that the work of Hofstede is used can be nicely presented in a table, however the reasons some of the authors chose to use a different taxonomy than Hofstede are more divers and will be discussed below.

The arguments that other taxonomies of culture are used than Hofstede have a more uniform reason in case of the work of Gray (1988). All of the authors that used Gray did this because of the strong link between accounting and the work of Gray. Almost all of the authors that use Gray also emphasize the work of Hofstede and the relation between the work of Hofstede and Gray. The two studies that used the cultural classification of Ronen and Shenkar (1985 ) did this because in the case of Richards (2000) as part in combination with two other taxonomies to identify differences between different cultural groups. The other authors that used Ronen and Shenkar are Johnson, Cullen, Sakano

and Bronson (2001) however they unfortunately don´t give a reason for this. The authors that treated culture as a whole didn’t present any arguments for this but is has to be assumed that this would have to do with the limited scope and length that these studies had.

It is unfortunate that Birnberg and Snodgrass (1988) who were the only ones to the taxonomy of Kluckhohn and Strodtbeck (1961) don’t mention their motive for this.

(17)

styles.

Jansen, Merchant and van der Stede (2009) also use beliefs about the purpose of corporations from Boselie et al. (2001), because this dimension also shows a difference between the U.S.A. and the Netherlands. For the same reason Jansen, Merchant and van der Stede (2009) use masculinity and long term orientation from Hofstede.

Brice (2012) gives no specific reason for his choice to use more dimensions than Hofstede (1980) suggests. Although it is assumed Brice (2012) uses these six dimensions because one of the goals is to separate ethnic culture from national culture.

Richards (2000) uses a combination of Hofstede (1980), Ronen and Shenkar (1985) and Hampden-Turner and Trompenaars (1997) because he argues all of these taxonomies have been rigorously tested.

4.4 The amount of unsupported and partially supported hypotheses

It is also interesting to see how many of the studies have confirmed their hypotheses. In total the 34 studies had 112 hypotheses. Out of these only 60.7% has been fully supported. The other 39.3% was either partially supported, could not be tested or was not supported at all. There are two hypotheses that could not be tested due to statistical problems or that the wanted effects could not be

measured.

The studies that used research questions instead of hypotheses are not included in this overview. As in table 1 can be seen some studies indicate hypotheses as mainly supported or overall supported, these are counted in table 4 as supported as the authors argue that they have more arguments in favour for their hypotheses than against it. When there is on support found for a single country or the support is only found under specific conditions this is counted as a partial support.

Number of hypotheses Results:

68 Supported

11 Partially supported

31 Not supported

2 Could not be tested

112 Total

Table 4 results of hypotheses

4.5 Reasons for the unsupported and partially supported hypothesis

The reasons for why so many of the hypotheses could not be supported or only be partially

supported would require some additional investigation as presented below in table 5 to 8. However in most cases this has to do with the results not being significant. Even though as can be seen in table 1 many of the included studies have used a comparison of Asian versus Western countries to try and make use of the largest differences on dimensions of culture. On the other hand the majority of the supported hypotheses is also supported because the results are statistical significant.

Reasons supported hypotheses Number of times mentioned

Statistical significant results found 59

(18)

The hypotheses is written in such a way that finding no result or a non-significant result means the hypotheses is supported

3

Total 68

Table 5 reasons for supported hypotheses

With the category supported by research findings is meant that the results of the research support the hypotheses. This could be all kinds of results from other literature results to data that could not be tested in a way to develop a significant result.

Reasons not supported hypotheses Number of times mentioned

Results are not statistical significant 16

No evidence found 6

The results are (not) significant and point in the contrary direction as predicted

7 The hypothesis is written in such a way that

finding a significant result means that the hypothesis is not supported

1

No differences are found between the groups 1

Total 31

Table 6 reasons for not supported hypotheses

With the category no evidence found is meant that there are no results from the research that provide evidence for the hypotheses. These are result that cannot be translated in a statistical significant or insignificant result. A remarkable result of table 6 is that 7 hypotheses actually found results in the direction that was contrary to the others expectations.

Reasons for partially supported hypotheses Number of times mentioned Not all results of the different parts of

hypothesis are significant

5 Different results for different cultural groups or nations

5 Not all parts of the hypothesis point in the right direction

1

Total 11

Table 7 reasons for partially supported hypotheses

The category of “different results for different cultural groups or nations” means that there are only significant results found for a part of the total. For example in the research of Awasthi, Chow and Wu (1998) they only found support for hypothesis 1 for the USA. Although the authors predicted that the results would be the same for the different cultural groups these result might actually be support for the existence of differences between cultures.

Reasons hypotheses could not be tested Number of times mentioned

Unacceptably low reliability 1

Due to research design 1

Total 2

Table 8 reasons hypotheses could not be tested 4.6 Type of data gathering

(19)

used in the selected articles have been worked out into table 9.

Number of times used Type of data gathering:

4 Interviews

21 Survey questionnaires

6 Experiment

4 Financial and performance data

2 Data from other literature

1 Field study

38 Total

Table 9 type of data gathering

(20)

5. Discussion of research findings

In this part the previous research findings of the literature analysis will be discussed. The findings can be divided into four main groups. These are 1. Research intensifies in the past 16 years;

2. Heavy reliance on the work of Hofstede; 3. Large number of not supported or partially supported hypotheses 4; One-sided method of data gathering. These findings will all in turn be discussed.

5.1 Intensification of research in the past 16 years.

The first finding that research intensifies over time is perhaps not the most striking finding of this study however it is worth looking into some deeper. As presented in the previous section 75% of the articles used for the analysis where published in the last 16 years. A reason for this could possibly be the opening of international markets in this time period, making the research more relevant. This was initiated by the transition from a planned economy to a mixed economy with an increasingly open market environment in China in the 1980’s. The fall of the Berlin wall in 1989 which

consequently lead to the dissolution of the Soviet Union was another step in the globalisation process. These important historical events led to an increase in international trade. Many

multinationals entered markets of these former communistic countries. This raised the importance of the question whether management control practices can be uniformly used in different national cultures. Scientific research usually takes a few years to present results of these globalization processes, so it is completely logical that the increased importance of cross-cultural management research has grown rapidly since the mid 1990’s.

In the case of the articles that are more focused on the role of accounting practices an explanation can also be the increased importance of globalisation. This had as consequence that there is ever increasing role of international rules and regulations such as International Financial Reporting Standards (IFRS) and the increasing acceptance of International Standards on Auditing (ISA) (Heidhues & Patel, 2011). The growth of multinational organisations (MNO’s) also means that for internal management accounting and control decisions the data needs to be comparable between the different business units. An important question in this process is the influence of national culture, this in its turn led to an intensifications in the field of research.

Another important factor in the explanation is not only the opening of the international markets but also the importance of emerging markets. Many companies see the need to introduce their products in these economies or to move their production process into low wage countries and so become MNO’s.

Another possible explanation can be found by the introduction of the work of Hofstede. Until the work of Hofstede was published in 1980 there was no real way to operationalize culture and measure culture besides the work of Strodtbeck and Kluckhohn (1961) and Hall (1976). Which both are harder to use because they don’t provide a classification of countries like the work of Hofstede. The work of Hofstede started to really get accepted in the nineties by authors as the work became more popular and validated. By applying the work of Hofstede the researchers could create new studies in a far less time-consuming way. The work of Hofstede was the research tool that so many authors had been anticipating. Combine this with an urge to publish by many researchers and this could explain the intensification of research in the past 16 years.

5.2 Limitations to the work of Hofstede

(21)

As showed in table 2 many of the cross-cultural studies still rely heavily on the work of Hofstede. When looking at the different reasons given for this there seems to be a limited amount of researchers that actually provide a solid reasoning for their choice to rely their work on Hofstede. As argued by Baskersville (2003) the acceptance of the work of Hofstede in management and accounting literature could be because of an absence of sufficient consideration for the reasons behind the rejection of such a universalist approach in anthropology and sociology.

As pointed out earlier the authors have given multiple different arguments for their choice of the work of Hofstede. The most commonly used arguments for choosing the work of Hofstede where: because it shows divergence on one or more dimensions between nations; the work of Hofstede has been empirically proven by multiple others; testing the work of Hofstede and other taxonomies to find out which one is the best is beyond the scope of the research; because Hofstede is the most widely used and cited taxonomy. The last four of these arguments all share the same component and that is that is the ease that using Hofstede gives. Because there is empirical proof provided by other authors the researchers don’t have to provide it. Secondly reviewing other options is too time consuming so it is easier to use Hofstede. And the fact that Hofstede is so widely used and cited doesn’t automatically mean that all these citations are positive. However researchers still give this as an argument to use Hofstede. These are arguments are each pretty weak on their own, hence that most researchers provide multiple arguments why they use Hofstede.

However there are also arguments that suggest that the use of Hofstede might be less suitable for cross-cultural studies. Criticism on this topic is provided by multiple authors (Jones, 2007; Baskerville, 2003; Baskerville-Morley, 2005; McSweeney, 2009 A&B). The work of Jones (2007) gives a good overview of the pros and cons of the work of Hofstede. The pros of the research of Hofstede provided by Jones (2007) can be classified into three groups. The first is the relevance of the work of Hofstede. Hofstede is considered to be a pioneer in the field of cultural taxonomy and provided a framework that fitted right into what the research field was looking for (Jones, 2007). As result the work of Hofstede is widely used and a predominant approach to culture, this argument is also found in table 3.

The second argument that Jones (2007) mentioned in favour of the work of Hofstede is its

rigourness. The work of was based on rigorous design with systematic data collection and coherent theory (Jones, 2007). The third argument is the relative accuracy of the work of Hofstede. In a similar research as that of Hofstede, Søndergaard (1994) manages to replicate the main findings of the work of Hofstede. These last two arguments can also be found in table 3.

But as mentioned before there is also a lot of criticism on the work of Hofstede. The arguments against Hofstede can also be divided into different groups.

(22)

2009A).

A second argument against the use of the work of Hofstede is that he used nations as unit of analysis. However a culture is not necessarily bounded by national borders (Jones, 2007). The work of

Hofstede conflicts nation and states (McSweeney, 2009A). First it is important to define a state, as state is a territorial juridical unit (McSweeney, 2009A). If all states would consist of one nation then there would not have been a problem. However many states such as the United Kingdom exist of multiple different nations. However in the work of Hofstede many of the states with multiple nations are depicted as having a single national culture (McSweeney, 2009A). In essence it comes down to the fact that countries can and will have more than one culture.

The third criticism on the work of Hofstede is that there are many political influences in his work (Jones, 2007). The time of the data gathering of the IBM survey of Hofstede has had an influence on the outcomes. The research of Hofstede was conducted at a time period that Europe was at the height of the cold war and still influenced by the aftermath of the Second World War. At the time there also was a rise of communism in Europe, Asia and Africa. As a result of political instabilities some communist and undeveloped countries are left out of the research (Jones, 2007). Especially the results of the masculinity dimension (Søndergaard, 1994) and uncertainty avoidance

dimension(Newman & Nollen, 1996) would be influenced by this.

The fourth criticism is that the data Hofstede used was derived from a single company (IBM). As a result this could not provide information over an entire nation’s culture (Jones, 2007).

The fifth argument criticising the work of Hofstede is that it was first published in 1980 over three decades ago, this could indicate that the data on which he based the study could be outdated. Recent events such as rapidly changing global environments, internationalization and convergence could change cultures (Jones, 2007). Hofstede himself even admitted that the data of IBM is old and now obsolete as concluded by Baskerville-Morley in 2005. Another assumption of the continued use of the data and results of Hofstede is that culture would be stable and not change significantly over time. This assumption is wrong according to McSweeney, 2009A and only a result of the inability of researchers to adopt changing culture into the work of Hofstede.

It has been argued by Baskerville (2005) that four or five dimensions are not enough to explain the complete culture of a country. Hofstede responded to this by admitting that new dimensions should continue to be added to his original work (Hofstede, 1998). Hofstede in fact added a sixth dimension to his original work in 2010.

A final point that has provided a lot of criticism on the work of Hofstede is its statistical integrity. For example the sample size can be extremely small. In most cases (with the exemption of only six countries) the sample of respondents used in both surveys was under a 1000 persons (McSweeney, 2002). Another point is that Hofstede has had some problems with his survey. An analysis of the survey used by Hofstede (Jones, 2007) pointed out that there were only 32 questions used for only 40 countries. An analysis built on so few countries takes great advantage of chance and increases the likelihood of sample error Dorfman and Howell (1988).

Of course there are more points of criticism on the work of Hofstede however discussing all of these is beyond the goal of this research. The arguments that could be of major influence to the relevance of choosing the work of Hofstede are only discussed.

(23)

scale of Hofstede are mainly found between Western and Asian countries.

Now that it is clear that there are also limitations to the use of the dimensions of culture from Hofstede, this raises the question why other work isn’t used more in cross-cultural studies. For example why isn’t the work of GLOBE used or one of other taxonomies as discussed in section two. A reason for this can be found in the arguments why Hofstede was used. Some authors mentioned the reason for choosing the work of Hofstede was the comparability with other studies. Using a different taxonomy would eliminate this comparability. Also other taxonomies never received the attention and popularity as the work of Hofstede. A consequence of this is that other taxonomies have not been as validated as much as Hofstede. A critical analysis of all the other taxonomies of culture is beyond the goal of this research. The other taxonomies also received criticism and none of them can be seen as perfect. This creates a kind of vicious cycle that makes authors choose for Hofstede. Because other authors use it, they will use it too. Although Hofstede has limitations, so do the other taxonomies, so they will use Hofstede and so on. This makes that for the next years we can probably still expect to see new cross cultural research in the field of management and accounting being published that still relies heavily on Hofstede’s dimensions.

5.3 Non supported and partially supported hypotheses

The third finding is the large number of not supported of partially supported hypotheses. As provided in section 4.3 there are a number of reasons given as to why the hypotheses could not be supported namely: results are not statistical significant; no evidence found; the results are (not) significant and point in the contrary direction as predicted; the hypothesis is written in such a way that finding a significant result means that the hypothesis is not supported; no differences are found between the groups.

The first reason is that the hypotheses are not statistical significant. This could have to do with the research methods. For example the almost exclusive use of surveys. The use of surveys will be discussed more in depth in part 5.4. Another reason for not finding any support for the hypotheses could be the research design. In cross-cultural research related to management and accounting the use students as sample is quite common (McSweeney, 2009A). These students are used to fill in surveys or in some cases for experiments. This conflicts the level of analysis as the taxonomies of culture are created and designed for the use at the level of national culture. Deriving data at a lower level of analysis can have a huge impact on the outcomes of the research. The reason for this is that the relationships identified at one level of analysis may be more significant or insignificant at different levels of analysis, or may even act in an opposed direction (Klein &Kozlowski, 2000). Many of the studies who use this type of research data are therefore inaccurately described as cross-national studies (McSweeney, 2009A). Another consequence of the use of students is that the sample size is often very small which would mean that the sample is not representative for the national culture (McSweeney, 2009A).

(24)

lead to finding no significant results or support for the hypotheses. This because the multiple cultures within the nation could distort the sample (Jones, 2007). Or the failure to recognize all different cultures could mean that the wrong hypothesis and assumptions are created by the researchers.

5.4 Type of data gathering

The fourth finding is as discussed previously the type of data gathering, the predominant mode of research is the use of survey questionnaires. Although the advantages (quick, inexpensive, easy to administer) of this method make is seem like a perfect research tool, there are however critical limitations to the use of survey questionnaires. Poor survey construction and administration can undermine otherwise well- designed studies. The answer choices provided on a survey may not be an accurate reflection of how the participants truly feel. While random sampling is generally used to select participants, response rates can bias the results of a survey. The sample size could also lead to biased results. However these are all problems found in all survey based researches, other problems are more of importance to cross-cultural research. Such as problem that arises when translating surveys into other languages or by the use of expressions only know to a single culture (Schaffer & Riordan, 2003).

Failures in the design of the survey could have huge implications on the outcomes of the research and could also be a possible explanation for the non-supported hypotheses. An instrument design failure in the survey could be the scale of the answers.

An example of a study where the data collection was different and done meticulously is by Efferin and Hopper (2007). They didn’t use a survey but instead relied on interviews, observations and documentary analysis. This was done to overcome the limitations with survey research. These methods allowed Efferin and Hopper (2007) a more in depth analysis during interviews and observations than would be possible with a survey. The downside of this method are the costs and time involved to travel to different countries. This ethnographic data collection method was combined by Efferin and Hopper (2007) with the grounded theory method in order to study how cultures, ethnic differences, history, politics, and commercial considerations shaped management controls. This combination of research methods creates more reliable and less biased results than traditional surveys (Efferin and Hopper, 2007).

5.5 What was learned from previous research?

Some of the above findings have already been addressed by previous research. The two most noteworthy of these researches are Harrison and McKinnon in 1999 and Chow, Shields and Wu, also in 1999.

Harrison and McKinnon identified four shortcomings in cross-cultural management and accounting research. Firstly a failure to consider the totality of the cultural domain in the theoretical exposition. With this Harrison and McKinnon (1999) mean that in many cross-cultural research only study a specific dimension of culture are used in the conceptualization and theorization phase of cross-cultural research. A second finding of Harrison and McKinnon is the tendency to threat cross-cultural dimensions as equally important across different nations. It is argued that cultural dimensions can be divided into core or peripheral values indicating the importance of these values to the national culture. This complex issue is often left out (Harrison & McKinnon, 1999). The third finding of

(25)

unidimensionally of those dimensions. A risk of this is that relations between different cultural dimensions can be overlooked and moderating effects can be overseen (Harrison & McKinnon, 1999). The fourth finding of Harrison and McKinnon (1999) is the excessive reliance on the value dimension conceptualization of culture with all its limits. On the basis of these four shortcomings Harrison and McKinnon (1999) argue that future research should take these points into account perhaps by using a finer partitioning of culture and cultural dimensions than the work of Hofstede has provided. Harrison and McKinnon (1999) also have mentioned that the reliance on mail survey methods have its

limitations and it would be better for more personal contact of the researcher with the studied subjects to understand the deeper underlying aspects of culture. Such as done by Efferin & Hopper (2007).

It is interesting to see that almost 14 years after the article of Harrison and McKinnon was published, which has been widely cited since, if this has had any effects on the field of research. Since

1999 the cultural classification used is almost exclusively the work of Hofstede as can be seen in table 1. Also none of the studies included in table 1 have adopted the core vs peripheral method of

Harrison and McKinnon (1999). Also on the point of the reliance of survey methods there seems to be no improvement since the work of Harrison and McKinnon (1999) was published.

The article of Chow also identified several pitfalls. These are Failure to control for contingency variables such as uncertainty, technology, organizational size and regulations. Also the heavy reliance on survey instruments is mentioned by Chow et al. (1999). Limitations of the work of Hofstede such as that the dimensions are insufficiently precise to ensure correct and consistent applications (Chow et al., 1999) Chow et al. (1999) make the suggestion that instead of relying on Hofstede for future research needs to look at alternative ways of operationalizing or structuring national culture. Finally notions by Chow et al. (1999) are made that not only national culture is of influence on the design of management control and accounting, it is suggested that the competitive environment should also be included. When looking at table 1 it is clear that the suggestions of Chow, Shields and Wu have not been adopted by the field of research. The predominant approach to culture is still the work of Hofstede.

(26)

6. Conclusion and suggestions for future research

This research started out with the question: What are the trends and problems in cross-cultural research in management control and accounting literature to date and what suggestions for future research can be made? Through the process of a meta-analysis of cross-cultural research four points have been identified.

The first was the trend of increasing amount of studies conducted in the past 16 years as opposed to the period before this. An explanation has been given that this might have to do with the intensified globalisation in this period which has led to an increase in relevance and importance of the research topic. As alternative explanation it was discussed that the work of Hofstede might have actually been related to the intensification of the research field in the past 16 years.

The second finding was the heavy reliance on the taxonomy of culture created by Hofstede. Part of the analysis was to look at which reason where given for basing the research on the work of Hofstede. These reasons where then discussed from which it can be concluded that not all the arguments provided by the authors are strong and relevant. In many cases the reasoning came down to a point that a full discussion of an alternative operationalization of culture was beyond the scope of the research or that it was used because it is one of the most cited academic works ever (Bond, ). In some cases no reason at all was given for using the work of Hofstede. The work of Hofstede was then discussed to see what the limitations of the work of Hofstede are.

The third finding was that there is a large amount of unsupported or partially supported hypotheses. This could actually be a consequence of using the work of Hofstede because as discussed this has numerous invalid assumptions and generalizations. Using these assumptions or by failure to test for other variables that can influence the relation between national culture and management control and accountancy could lead to insignificant results.

The last finding was the heavy reliance on survey methods. Although survey methods have clear advantages such as low costs, they also have major drawbacks. The limitations posed by the use of survey methods could lead to the unsupported or partially supported hypotheses. However slowly the methods used in cross-cultural research related to management control and accounting seem to shift away from the sole use of survey questionnaires.

Suggestions for future research

On the basis of this research two distinct recommendations can be made for future cross-cultural management control and accounting research. Firstly researcher should look beyond the work of Hofstede. A lesson on this point can be learned from the research field of anthropology that has never adopted and has never relied the work of Hofstede like research in management &

accounting literature (McSweeney, 2009 A). Although the suggestion to move away from Hofstede does not mean that any other of the current taxonomies are significantly better, this advice should more be seen as suggestion to try and reinvent the current limited amount models of culture used in management control and accounting. When the scope of the research does not allow to try and create a new view on national culture, the limitations of the work of Hofstede and the assumptions underlying his theory should be critically noted as these will have a profound effect on the results of the study.

(27)

view of the relation between culture and management control and accounting. Suitable research methods for this are observation and personal interviews which allow to researcher to study more hidden aspects of culture that would never be discovered using survey questionnaires with their limitations. A downside of using survey and questionnaires is that they are more time consuming and expensive. If budget and time constraints mean the only available tools are surveys it could be an option to look at an intermediate solution called field surveys. These are as discussed by Harrison and McKinnon (1999) a survey where the researcher is present at the location to make observations during the survey and give the ability to directly observe any anomalies. During this observations some of the deeper lying feelings of culture could be uncovered.

(28)

7. Limitations

This study like any other also has its limitations. Firstly only a suggestion can be made to improve cross-cultural research in the field of management control and accounting. To see whether these suggestions would have a positive outcome on future research is impossible without firstly applying the suggestions made.

Secondly there are also limitations regarded to the research design. Although as discussed earlier a meta-analysis has many advantages there are also downsides. Researchers are more likely to publish positive studies than negative ones (Thompson & Pocock, 1991). This could also be the case for cross- cultural research in management control and accounting literature. If this is the case than the amount of non-significant result in cross-cultural management and accounting literature could possibly be even higher.

The limitations of this meta-analysis could be the selected literature, although a careful search in order to find all relevant literature was conducted there is always a possibility that some studies are left out. On a relatively small sample of 34 articles the effect of an omitted article could result in different outcomes. In order to clarify why the articles are chosen, and as a consequence other articles not, the search process is included in the meta-analysis section.

Another limitation of a meta-analysis is the interpretation of the author. A wrongfully interpreted result can lead to a biased finding. To overcome this problem wherever there categorizations are made the legitimization of the choice behind this is described.

Another limitations is that only studies from scientific English journals are included. This was done because translating different studies would be too time consuming for the available time for this research. However this does mean that studies from several parts of the world where English is less common amongst scholars are left out. These studies might have had a significant impact on the results. For example the amount of studies from the Middle Eastern, Eastern Europe and Africa are vastly outnumbered by studies originating from Western developed countries.

Cross-cultural research in management control & accounting A meta-analysis of research to date