The interplay between cost accounting knowledge and presentation formats in cost-based decision-making:

(1)

THE INTERPLAY BETWEEN COST ACCOUNTING KNOWLEDGE AND PRESENTATION FORMATS IN COST-BASED DECISION-MAKING:

A CROSS-CULTURAL EXTENSION

by

LOTTE SIMONE JANSSEN

E-mail: L.S.Janssen@student.rug.nl Student number: 1534319

May, 2010

MSc in International Business & Management Specialization International Financial Management

University of Groningen

&

MSc in Business and Economics Uppsala University

Supervisors:

Dr. M. P. (Martijn) van der Steen (University of Groningen) Dr. S. (Sandra) Tillema (University of Groningen)

Dr. H. J. (Rian) Drogendijk (Uppsala University)

(2)

2

List of tables and figures

List of tables

Table 1: Summary of variables’ operational definitions and descriptive statistics………….. 27

Table 2: Hypothesis 1 t-tests: scores on Hofstede et al.’s (2008c) seven cultural dimensions of Hong Kong & The Netherlands and t-statistics………... 29 Table 3: Hypotheses 2 t-tests with (LN)PROFIT as dependent measure.……… 30

Table 4: Hypotheses 3 t-tests with (LN)PROFIT as dependent measure

Panel A: Difference in profit performance between tabular versus graphical ABC presentation formats within low and high KNOWSPLIT………... 31 Panel B: Difference in profit performance between tabular versus graphical ABC

presentation formats within low and high KNOWGROUP………. 31

Table 5: Supplementary analyses of the level of perceived complexity and information search Panel A: Self-perceived complexity (five-point Likert scale) and the effect of format

(tables versus graphs) per knowledge category (KNOWSPLIT)…………. 44 Panel B: ANCOVA analyses on information search………... 44

List of figures

Figure 1a: Actual screenshot of the initial ABC report issued prior to the task for the tabular format (between subjects factor)……… 24 Figure 1b: Actual screenshot of the initial ABC report issued prior to the task for the graphical format (between subjects factor)……….... 25

Figure 2: Hypotheses 4 tests: low-knowledge versus high-knowledge groups (KNOWSPLIT) Panel A: ANCOVA-models with LN(PROFIT) as dependent measure and

KNOWSPLIT and PRESENTATION as independent measures………….. 33 Panel B: Mean-by-mean comparison of the P x K interaction………... 34

Figure 3: Hypotheses 4 tests: low-knowledge (mean less one SD) versus high-knowledge (mean plus one SD) groups (KNOWGROUP)

Panel A: ANCOVA-models with LN(PROFIT) as dependent measureand low versus high KNOWGROUP and PRESENTATION as independent measures…. . 36 Panel B: Mean-by-mean comparison of the P x KG interaction.………... 37

Figure 4: Sensitive checks on low, medium, and high knowledge groups

Panel A: Effect of presentation format on low, medium, and high knowledge group 40 Panel B: Plots of the effect of presentation format on low, medium, and high

knowledge group……….. 40

Appendix

Table A1: Underlying functions of the experimental task and ABC cost allocation assumptions Panel A: Underlying functions of the experimental task (by substituting all are

f(x_i, p_i))... 49 Panel B: ABC-customer cost pools (CCP) and cost driver assumptions (i = Customer A, B, C)………. . 49

(4)

4

The interplay between cost accounting knowledge and presentation formats in cost-based decision-making:

A cross-cultural extension

Abstract

The influence of national culture on cost-based decision-making has repeatedly been underestimated or neglected. This study provides a cross-cultural extension of a study based on The Netherlands by Cardinaels [Cardinaels, E. (2008). The interplay between cost accounting knowledge and presentation formats in cost-based decision-making. Accounting, Organizations and Society, 33(6), 582-602.]. For the cross-cultural comparison, we conduct an experiment in both The Netherlands and Hong Kong. Presented with either tabular or graphical Activity Based Costing reports, respondents perform a complex pricing and resource allocation task that affects firm profitability. The results provide strong evidence that the effect of tables versus graphs on profit performance is dependent on both knowledge level and national culture. The effect of presentation format on the performance of Dutchmen is dependent on the decision maker’s level of accounting knowledge, whereas Hong Kong Chinese benefit from graphs in general. Companies can benefit considerably from tailoring a cost report’s presentation format to the decision maker’s national culture and knowledge level; low-knowledgeable Hong Kong Chinese can generate an additional 385,811 Euros profit (19% of the average profit within Hong Kong) just by altering the tabular format to a graphical format; and high-knowledgeable Dutchmen can generate an additional 354,420 Euros profit merely by changing the presentation format from a graphical to a tabular format.

1. Introduction

A significant literature has developed aimed at understanding individuals’ decision- making behaviour in accounting settings. Its central objective is to assess decision performance, to establish the factors that determine decision performance, and to test theories of cognitive processes which produce the decisions (Libby & Luft, 1993). Many of these theories are developed in Western economies. However, management accounting theories developed in Western economies may not be generalizable to the Chinese environment (Tsui, 2001). Unfortunately, however, the influence of national culture on cost-based decision-making has repeatedly been underestimated or neglected.

Accordingly, research on this relationship has been scarce. Only some studies call for future research on the impact of differences among the users of information or, more specifically, on the impact of national culture (Amer, 1991; Chandra & Krovi, 1999;

Ganzach, 1993; So & Smith, 2004).

The current study aims to fill part of the gap, by conducting a cross-cultural extension of Cardinaels’ (2008) study. Cardinaels (2008) examines the joint effects of one user characteristic, namely; a decision maker’s level of cost accounting knowledge, as well as presentation format in cost-based decision-making on profit performance.

Cardinaels (2008) finds evidence of an effect reversal across knowledge levels: decision makers with a low level of cost accounting knowledge perform better when presented

(5)

5

with graphical Activity Based Costing (ABC) presentation formats, whereas decision makers with a high level of cost accounting knowledge obtain superior profits under tabular ABC presentation formats. Cardinaels’ (2008) research is based on The Netherlands. The current study tests whether Cardinaels’ (2008) findings are independent of culture. For the cross-cultural comparison, we compare The Netherlands to Hong Kong.

Studying the combined effects of culture, level of accounting knowledge, and presentation format on profit performance is important for multiple reasons. Most important, Cardinaels (2008) mentions that a few studies suggest looking at differences among the users of information in decision-making (Amer, 1991; Chandra & Krovi, 1999; Ganzach, 1993). Cardinaels (2008) examines whether a decision maker’s level of accounting knowledge can explain when certain report formats are related to stronger performance than others in cost-based decision tasks. Although he does find a significant interaction between level of accounting knowledge and presentation format, there may be additional characteristics among the users of information that affect decision-making and may alter the relationship found by Cardinaels (2008).

The question then remains, what other characteristic can have an influence on decision- making. Characteristics are features that help to distinguish between people. An example of a user characteristic is the level of accounting knowledge. Another characteristic that clearly differentiates people from each another is culture. Hofstede (2010) defines culture as the mental programming of the mind distinguishing the members of one group or category of people from another. The ‘category’ can refer to nations, regions within or across nations, ethnicities, religions, organizations, occupations, or the genders (Hofstede, 2010). From an international business perspective, national culture is probably the most logical level of analysis from which to begin to explore the cultural domain (Thomas, 2008). Individuals from different cultural groups may react differently to the same stimuli or require different stimuli to reach the same behaviour (Birnberg &

Snodgrass, 1988). Therefore, the relationship found by Cardinaels (2008) may not hold for other cultures. Hence, we test whether Cardinaels’ (2008) findings are independent of culture.

Furthermore, this research can provide insights to the current debate on accounting standards as to whether it is advisable to adopt a single set of accounting standards across cultures (Meeks & Swann, 2009). Harmonization of accounting continues to take place throughout the world, though differences remain in how accounting information is reported between companies in different countries (Yuan, Entwistle, & Stolowy, 2007).

The study of Yuan et al. (2007) indicates that differences at the level of presentation can become a psychological obstacle when students or managers are not well prepared to confront them. Therefore, the influence of culture on the relationship between different presentation formats and decision-making in accounting settings is important, especially considering the growing globalization.

Cardinaels (2008) argues to be the first to examine whether specific knowledge (i.e. cost accounting knowledge) is a pivotal characteristic which can explain when tabular versus

(6)

6

graphical formats increase or decrease performance (i.e. profits from cost-based decisions). Cardinaels (2008) does, however, acknowledge there is another study including presentation format, knowledge, and performance (Vera-Muñoz, Kinney, &

Bonner, 2001). However, these variables carry a different meaning as within Cardinaels’

(2008) study. In Vera-Muñoz et al.’s (2001) study, presentation format does not refer to the visual presentation of the data, but to a cash-flow versus an earnings-based format and is based on tabular formats only. Moreover, they measure domain experience and assume that this leads to knowledge. Finally, task performance in Vera-Muñoz et al.

(2001) does not refer to profit performance, but to the quality of the problem representation.

Numerous prior studies only include two of our four variables (culture, presentation format, knowledge, and performance). Especially research about the influence of culture is scant. The studies of Tsakumis (2007) and Tsui (2001), however, test the effect of culture on decision-making and performance respectively. Various studies examine the influence of knowledge on performance (e.g. Cloyd, 1995; 1997; Dearman & Shields, 2001) sometimes in relation to task performance (e.g. Tan & Kao, 1999; Tan, Ng, & Mak, 2002). Cloyd (1995) does call for future research on the influence of both information presentation and knowledge on performance. Some studies on the interaction between presentation format and performance call for future research that would also get closer to the current study. Wright (1995), for instance, made recommendations for further research on the influence of differences in spatial perceptual ability among decision makers when using graphs. Moreover, So & Smith (2004) made suggestions for further research on the influence of user characteristics, such as culture and personality.

Surprisingly, many of the studies which include two of our four variables, as well as Vera-Muñoz et al. (2001) and Cardinaels (2008), do not even mention the possibility of national culture having an impact. As a result, the effect of culture on cost-based decision-making remains an open question.

Cardinaels (2008) found that decision makers with a high level of cost accounting knowledge perform better when presented a graphical format as opposed to a tabular format, given that knowledge in a domain leads to an increased analytical focus and according to Lucas (1981), analytical decision makers, or those who especially look for details, benefit from tables. On the other hand, heuristic decision makers, or those who look at the entire problem, benefit from graphs that stress an overview of the same data (Lucas, 1981; Vessey, 1991a). Moreover, theory on representational congruence predicts a more favourable effect on decision performance when the external presentation format is congruent with the user’s cognitive style or internal representation (Chandra & Krovi, 1999). Interestingly, international differences in cognitive style have been seen in terms of an East-West dichotomy (e.g. Nisbett, 2003; Nisbett et al., 2001; Redding, 1980;

Torrance & Sato, 1979). The Western approach tends to be analytical with a focus on the sequential links between cause and effect while Eastern thinking is leaning towards an intuitive assessment of the entire universe of events as a system of interdependent parts (Redding, 1980). Since The Netherlands is part of the ‘analytic West’ and Hong Kong of the ‘intuitive East’ and given that Cardinaels (2008) found evidence that analytical (high-

(7)

7

knowledge) decision makers benefit from tables and heuristic (low-knowledge) decision makers benefit from graphs, the research question is: What is the influence of culture on cost-based decision-making?

2. Prior research

2.1. Presentation format, knowledge, and performance

Cardinaels (2008) studied the joint effects of presentation format (graph versus table) and a decision maker’s level of accounting knowledge on profit performance. After controlling for differences in work experience and general ability, he found evidence of an effect reversal across knowledge levels. That is, decision makers with a low level of cost accounting knowledge perform better with a graphical ABC format, whereas decision makers with a high level of cost accounting knowledge benefit from a tabular ABC format. Cardinaels (2008) provides a possible explanation for these findings by suggesting that graphical formats have a tendency to reduce task complexity for a low- knowledgeable user, while tables support the information search of a high-knowledgeable decision maker.

In his article, Cardinaels (2008) reviews the literature that examines two of his three variables (presentation format, knowledge, and performance). Furthermore, he acknowledges the importance of taking user characteristics into consideration. Cardinaels (2008) examines the effect of accounting knowledge as a user characteristic that may explain when certain report formats are related to stronger performance than others in cost-based decision tasks. The current study, however, argues that culture may be a critical characteristic, since culture affects decision-making (Birnberg & Snodgrass 1988). Therefore, in what follows we review the accounting literature on culture in general and on culture in relation to presentation format, knowledge, and performance respectively.

2.2. Culture

According to Birnberg and Snodgrass (1988), culture can be seen as one of the forces steering human decision-making. It does so by setting the preconditions for human behaviour. Culture can influence the selection of stimuli to which a person pays attention, or it can influence any value judgments regarding stimuli. Certain sets of stimuli may be sought out while others are ignored due to a cultural filter. A second effect indicates that when a stimulus has been perceived, its relevance and credibility to the cultural group’s member will also be affected by that cultural filter (Birnberg & Snodgrass, 1988). This indicates the importance of investigating the influence of culture on cost-based decision- making, as in the current study.

(8)

8

In the accounting literature, research on the influence of national culture is scarce. Many of the studies focusing on national culture examine the relationship between culture and business ethics (e.g. Cohen, Pant, & Sharp, 1995; 1996; Dunn, 2006; Gernon, 1993) or the use or performance of management accounting practices (e.g. Chow et al., 2002;

Jansen, Merchant, & Van der Stede, 2009; Tsui, 2001; Williams & Seaman, 2001).

Important lessons can be learned from these studies; especially the influence of culture on management accounting practices gives useful insights to the current study. These studies show that culture has an influence on individuals’ perceptions and decision-making and that companies can benefit considerably from tailoring their practices and control systems to the local culture.

Berry et al. (2008) wrote a review of recent literature (1995 until 2008) on management control and highlighted seven emerging themes, of which three are related to the current study: Decision-making for Strategic Control; Performance Management for Strategic Control; and Culture and Control. Especially the theme ‘Culture and Control’ is linked to the current study; however, there has been relatively little research on this particular theme (Berry et al., 2008). Yet, already in 1995, Otley, Broadbent, and Berry (1995) made suggestions for further research on the impact of globalization in introducing greater complexity, uncertainty and risk; and most important, on control in multinational organizations in relation to the influence of national and organizational cultures (Berry et al., 2008). Even further back in time, Birkett (1967) considered accounting in combination with a variety of issues such as time, space, symbols, and culture.

Surprisingly, research progress on the influence of culture in accounting has been slow.

Moreover, most of the studies on culture and control focus on organizational culture (Ahrens & Mollona, 2007; Chapman, Hopwood, & Shields, 2007; Dent, 1991; Quattrone

& Hopper, 2005), not national culture (Berry et al., 2008). It seems that the suggestions of Otley et al. (1995) have retained much validity.

Though little is known concerning the influence of culture on decision-making in accounting, a recent study does address this issue, namely the study of Tsakumis (2007).

Tsakumis’ (2007) study about the influence of culture on accountants’ application of financial reporting rules makes a comparison between U.S. accountants and Greek accountants. Based on the studies of Hofstede (1980) and Gray (1988), the U.S. and Greece were chosen to represent low and high conservatism and secrecy countries respectively. Gray (1988) postulates that the accounting values of conservatism and secrecy are related to and derived from Hofstede’s (1980) cultural values. The result of the study is that Greek and U.S. accountants arrive at significantly different decisions when presented with identical figures and financial reporting guidelines. This finding suggests that uniform accounting standards may not result in similar disclosure decisions being made in different countries. It is possible that dissimilar disclosure judgments obstruct the comparability of financial statements across countries (Tsakumis, 2007).

Nevertheless, Tsakumis (2007) still does not investigate the effect of culture on the influence of financial information formats. Perhaps it is not desirable to standardize financial information formats, since individuals from different cultures appear to respond

(9)

9

differently to information and this alters their decisions. Correspondingly, Tsakumis (2007) calls for further research to verify whether his result – that culture has an influence on decision-making – is generalizable to other decision contexts and cultural areas.

2.3. Culture and presentation format

Extensive research has been performed concerning the design of management control systems (e.g. Chenhall, 2003; Dent, 1987; Ferreira, 2002; Ferreira & Otley, 2005), even in relation to culture (e.g. Ouchi, 1980; Simons, 1995). However, in these studies

‘design’ refers to the structure of a control method and not to the visual presentation of the system. Besides, little is known about which design features (in terms of structure) are more effective and appropriate in different situations (Merchant, Van der Stede, &

Zheng, 2003) which leaves many questions unanswered. Culture is an important contextual variable which permeates the whole control system by influencing individuals’

choices and behaviours. Thus, the study and understanding of the operation of a control system benefits from the consideration of the influence of culture (Ferreira & Otley, 2005). It emerges that a vital problem with the design of management control systems is the difficulty of anticipating the behavioural reactions to the controls (Berry et al., 2008).

Therefore, investigating the relationship between culture and varying presentation formats adds significant value to the accounting literature.

An example of a study that examines the influence of accounting information presentation on decision-making is the study of Courtis (2004). Courtis (2004) shows that some colours are associated with more (less) favourable perception formation and with more (less) investment allocations. These findings confirm that information presentation can influence decision-making. Furthermore, Courtis (2004) emphasizes the need for further research in different cultural settings, since culturally diverse individuals perceive colours differently. Courtis (2004), however, does not address the topic of tabular versus graphical presentation formats.

Several studies in the accounting literature do address the topic of tabular versus graphical presentation formats. A number of these studies, addressing the presentation of multivariate financial information, compare tabular presentation with graphics, including line graphs, bar charts, and schematic faces. These studies have shown both favourable effects of graphs (e.g. Stock & Watson, 1984; Wright, 1995) and beneficial effects of tables for questions and tasks that are sufficiently complex (Davis, 1989; So & Smith, 2004). The relative influence of graphical versus tabular formats remains inconclusive.

For instance, findings from DeSanctis and Jarvenpaa (1989) support the incremental benefit of graphs, but only when subjects are provided with practice, and Wright (1995), only when the task is complex. The general finding is that multidimensional graphics take less time to complete, and are more accurate than tabular formats, although much of the evidence for the latter is inconsistent (Bricker & Nehmer, 1995; Smith & Taffler, 1996;

Smith, Taffler, & White, 2001; Stock & Watson, 1984). Also, the benefits of graphical and pictorial formats reported in previous studies are not supported in So and Smith’s

(10)

10

(2004) study. So and Smith (2004) study the interactive effect of presentation format and information complexity on multivariate decision accuracy, to determine the most effective presentation format for the performance of multivariate decision tasks of varying complexity. They find a significant interaction: when information complexity is low, presentation format has no impact on accuracy, however, when information complexity is high, the tabular-alone format shows the highest accuracy. An explanation for the contrasting results concerning tabular versus graphical formats is that user characteristics are ignored (Chandra & Krovi, 1999). Also, Amer (1991: 30-31) argues that subject variation could explain the failure to achieve certain effects of presentation formats.

These conflicting findings confirm that the effects of graphs versus tables remain inconclusive. Furthermore, to our knowledge, there are no studies within the accounting literature addressing the effect of tables versus graphs in relation to culture. So & Smith (2004), however, state that there may be other relevant user characteristics, such as culture and personality, which may influence the effect of presentation format on decision-making. Therefore, they call for further research on these characteristics, which supports the need to investigate the affect of culture. We respond to these concerns by studying the combined effect of culture, presentation format, and knowledge level on performance.

2.4. Culture and knowledge

Accounting research on the effects of both national culture and knowledge is scarce.

Although these studies on national culture and knowledge address different accounting topics and different cultures than in the current study, these studies do show that national culture and knowledge levels can influence decision-making simultaneously. This confirms the importance of assessing the influence of culture in combination with knowledge.

One of these studies on national culture and (experience-related) knowledge is the study of Rong-Ruey and Thomas (2001). Rong-Ruey and Thomas (2001) examine the influence of these variables on cue weighting in internal control evaluation over payroll.

Accounting research has gathered much evidence in the U.S. that students and auditors with varying levels of experience considered the cue (information) related to the segregation of duties as the most important cue. Rong-Ruey and Thomas (2001) examined whether this pattern would still be valid in other cultures. Therefore, they performed an experiment in Taiwan, which was similar to the experiments conducted in the U.S. They chose Taiwan, since it culturally differs from the U.S. on several factors, such as power distance and individualism (Hofstede, 1983; Bond & Hwang, 1986).

Moreover, in contrast to the U.S., the family business dominates in Taiwan due to culture (see Hwang, 1990; Negandhi, 1973) and Taiwanese businesses are on average smaller than U.S. businesses. On the other hand, basic knowledge necessary for internal control evaluation is gained through college education and the teaching materials used in Taiwan are U.S. auditing textbooks (Duh, 1993). Therefore, Rong-Ruey and Thomas (2001)

(11)

11

predicted that Taiwanese students or novices with no experience with internal control evaluation are more likely to consider separation of duties as the most important cue as suggested by U.S. textbooks. On the contrary, experienced auditors were predicted to consider the cue related to the approval control by the superior(s) as most important, since they have obtained knowledge in addition to that basic knowledge. This additional knowledge reflects the effects of culture on internal control design for Taiwanese businesses. Rong-Ruey and Thomas (2001) essentially confirmed these predictions. Also, they showed that cue weighting on approval procedure was not significantly correlated with the length of experience, though was significantly correlated with the position layer of the auditor. That is, once an auditor acquires the necessary knowledge, any extra length of experience will have no impact on judgment. Rong-Ruey and Thomas’ (2001) study provides evidence of a cultural consequence on audit judgments.

Another study on the influence of both national culture and (professional) knowledge is the study of Ho and Chang (1994). Ho and Chang (1994) examined whether national culture or professional knowledge affects auditors and students more in making diagnostic conjunctive probability judgments. Comparisons were made between Taiwanese auditors’, American auditors’, Taiwanese students’, and American students’

diagnostic probability judgments. Each respondent received an auditing task, followed by a medical task. Experienced auditors were presumed to have experience and domain- specific knowledge in the audit task, but not in the medical task. Besides, Frederic and Libby (1986) showed that experienced auditors possess more knowledge than novices and have a different memory structure. Students, on the other hand, were lacking experience-related knowledge in both the audit and medical domains (Ho & Keller, 1994).

Ho and Chang’s (1994) findings show that in the audit task professional knowledge, not national culture, plays the dominant role in auditors’ diagnostic probability judgments:

American and Taiwanese auditors’ judgments were highly similar and judgments of auditors compared to students were quite different within Taiwan and were even more different within the U.S. However, judgments made by American students compared to Taiwanese students were significantly different, which demonstrates the influence of culture. Moreover, although American and Taiwanese auditors arrived at similar final probability assessments, they used quite different judgment processes. Furthermore, in the generic medical context, national culture strongly affected both auditor’s and students’ diagnostic probability judgments. American auditors and students made similar judgments, which differed from judgments made by Taiwanese respondents. Also, Taiwanese auditors and students made similar judgments (Ho & Chang, 1994).

These findings show that individuals with low levels of knowledge (students, as well as auditors in the medical task) are influenced by culture. On the other hand, for auditors in the audit task (with a high level of knowledge), the national culture effect found in the medical task is diminished by their professional standards and experience-related knowledge. Reasons why culture did not have a significant influence here are that Taiwanese auditors were from Taiwanese accounting firms affiliated with the U.S. Big Six and after joining these firms, they had similar professional training and continuing

(12)

12

professional education as American auditors.¹ Additionally, the Taiwanese accounting and auditing profession stringently comply with the Generally Accepted Accounting Principles (GAAP), Statement of Auditing Standards (SAS), and other regulations and standards promulgated in the U.S. The influence of national culture is considerably mitigated by this strong ‘profession’ or ‘firm’ culture (Ho & Chang, 1994).

These studies show that the relative influence of national culture and knowledge remains inconclusive. This indicates the need for further research on the relative influence of culture and knowledge.

2.5. Culture and performance

To our knowledge, studies on the relationship between culture and performance within the accounting literature can only be found in relation to Management Accounting Systems (MAS). Lessons can be learned from these studies, which examine the influence of culture on other types of decision-making than in the current study.

The research of Tsui (2001) is an example of such a study. Tsui (2001) examines the interaction effects of MAS and budgetary participation on managerial performance of Western and Chinese managers. The relationship between MAS information and managerial performance of Chinese managers was negative for high levels of participation, but positive for Western managers. That is, the interaction effects were different, depending on the cultural background of the managers. These findings suggest that the management accounting theories developed in Western economies may not be generalizable to the Chinese environment. However, performance in Tsui’s (2001) study is measured in terms of managers’ perceptions of their own performance and not in terms of profit performance, as in the current study. Nevertheless, Tsui (2001) does recommend the use of different measures of performance for future studies.

3. Theory and hypotheses development

We conduct a cross-cultural extension of Cardinaels’ (2008) study. Based on the theory on representational congruence (Chandra & Krovi, 1999), graphical formats are likely to improve the profit performance of Dutch as well as more knowledgeable (analytical) decision makers, whereas tabular formats are likely to improve the profit performance of Hong Kong Chinese and less knowledgeable (intuitive) decision makers, due to a fit of that particular presentation format with the decision maker’s internal representation.

1 In Ho and Chang’s (1994) study, respondents were presented with seven plausible event items that might have caused the fluctuation within the audit task and medical task respectively. Respondents had to make diagnoses in the one task (e.g. audit task) and predictions in the other task (e.g. medical task), in terms of probabilities. In the audit task, only one of the seven event items showed significant differences between American and Taiwanese auditors (see Ho & Chang’s, 1994).

(13)

13 3.1. Dutch versus Hong Kong Chinese culture

In order to test whether Cardinaels’ (2008) findings are independent of culture, we make a comparison between two culturally diverse nations. To increase the chance to find differences in decision-making behaviour, we compare The Netherlands (a Western culture) to a significantly different Eastern culture, namely the Hong Kong Chinese culture (e.g. Hofstede, 1991; House et al., 2004; Schwartz, 1992; Trompenaars &

Hampden-Turner, 1998). The aim of this study is not to contribute to the long-standing debate regarding how best to indentify and measure dimensions of culture that can be used to distinguish one culture from another. The aim of this study is merely to verify that there are cultural differences between the two national groups in the current study and to examine if this has an influence on the relationship found by Cardinaels (2008).

Whilst appreciating the content of other cultural studies, we will verify whether the two cultural groups indeed differ significantly by making use of the dimensions developed by Hofstede (1980; 1991) and Hofstede et al. (2008a; 2008c). We make use of Hofstede (1980; 1991) and Hofstede et al.’s (2008a; 2008c) framework, since it is probably the most cited cultural framework. This does not mean it is the best framework. In fact, there has been extensive critique on this framework, which has been beneficial, since it has led to several reconsiderations and improvements in the methodology and formulation being used. Accumulated experience has enabled Hofstede et al. (2008a, 2008c) to add two additional dimensions to the framework. Besides, Hofstede (1980; 1991) and Hofstede et al.’s (2008a; 2008c) framework has been widely used in various disciplines to study the influence of culture on individuals’ performance and decision-making, including the accounting literature (e.g. Chow, Shields, & Wu, 1999; Tsakumis, 2007; Tsakumis, Curatola, & Porcano, 2007; Tsui, 2001).

Hofstede (1980) suggests that culture consists of shared mental programs that control individuals’ response to their environment. Hofstede et al.’s (2008c) taxonomy consists of seven dimensions (of which the final two are added in the year 2008), namely: (1) power distance; (2) individualism versus collectivism; (3) masculinity versus femininity;

(4) uncertainty avoidance; (5) long-term versus short-term orientation; (6) indulgence versus restraint; and (7) monumentalism versus self-effacement. We will now shortly describe each of these dimensions.

(1) Power distance is the degree in which the less powerful members of a society expect and accept that power is distributed unequally. (2) Individualism denotes a society in which the ties between individuals are loose: a person is expected to look after himself or herself and his or her immediate family only. Collectivism represents a society in which people from birth onwards are integrated into strong, cohesive in-groups, which continue to protect them throughout their lifetime in exchange for unquestioning loyalty. (3) Masculinity stands for a society in which social gender roles are clearly different: men are supposed to be tough, assertive, and focused on material success; women are supposed to be more modest, tender, and concerned with quality of life. Its opposite pole, femininity, represents a society in which social gender roles overlap: both men and women are supposed to be tender, modest, and concerned with quality of life. (4)

(14)

14

Uncertainty avoidance is the extent to which the members of a society feel threatened by uncertain, ambiguous, unknown, or unstructured situations. (5) Long-term orientation stands for a society which fosters virtues oriented towards future rewards, in particular adaptation, thrift, and perseverance. Short-term orientation denotes a society which fosters virtues related to the past and present, in particular respect for tradition, preservation of ‘face’, and fulfilling social obligations. (6) Indulgence represents a society which allows relatively free gratification of some desires and feelings, especially those related with leisure, spending, consumption, merrymaking with friends, and sex. Its opposite pole, restraint, stands for a society which controls such gratification, and where people feel less able to enjoy their lives. (7) Finally, monumentalism represents a society which rewards people who are, metaphorically speaking, like monuments: proud and unchangeable. Its opposite pole, self-effacement, stands for a society which rewards humility and flexibility (Hofstede et al., 2008c).

Before we can test whether culture has an influence on the relationship found by Cardinaels (2008), we should verify whether significant cultural differences indeed exist between the two nations. Scores on the cultural dimensions used to range from 0 to 100, based on Hofstede’s original sample of 40 nations (Hofstede, 1980). However, the original research was later expanded to more nations so that values of over 100 have been rated (Hofstede, 1991). Since previous research has shown that the Hong Kong Chinese and Dutch culture differ significantly (Hofstede, 1991), we hypothesize the following:

H1: The Hong Kong Chinese culture differs significantly from the Dutch culture.

3.2. Congruence with presentation format: An East-West dichotomy

Culture can be seen as one of the forces steering human decision-making. Culture can influence the selection of stimuli to which a person pays attention, or it can influence any value judgments regarding stimuli (Birnberg & Snodgrass, 1988). Furthermore, P. C.

Chu, Spires, and Sueyoshi (1999) question the generalizability of descriptive decision theories across cultures, which largely stem from the West, and suggest the need for descriptive theories that integrate cultural factors. This indicates the need for empirical research to investigate the influence of culture on decision-making.

Although it is recognized that cultural differences exist and that culture affects behaviour, few studies have addressed cultural differences in individual business decision-making behaviour. While both national and organizational culture are present and affect individual behaviour in business situations, national culture has shown to have a stronger influence (Yamamura et al., 1996). Moreover, Laurent (1986) indicated that national differences were not reduced by the existence of a dominant corporate culture and, in fact, emerged to become even more pronounced.

An example of a study which lacks the influence of culture though examines the influence of diverse presentation formats on performance, is the study of Arunachalam, Pei, and Steinbart (2002). Arunachalam et al. (2002) show that, what they call,

(15)

15

‘improperly designed graphs’ can alter decision accuracy in making financial investment decisions, even though these graphs also show precise, accurate numeric data. These

‘improperly designed graphs’ can cause viewers to make choices that differ from the decision they would arrive at when presented ‘properly designed graphs’. Their finding is compatible with Hopwood’s (1996) view that the goal of corporate annual reports is not only to notify existing and potential investors, but also to influence or persuade viewers to create a more favourable impression about the corporation. Previous research has shown that ‘properly designed graphs’ can facilitate task performance (e.g. Davis, 1989;

DeSanctis & Jarvenpaa, 1989; Vessey, 1991b; Wright, 1995). Coherently, the findings of Arunachalam et al.’s (2002) study indicate that ‘improperly designed graphs’ can diminish performance. Arunachalam et al. (2002) emphasize the need for further research on the effects of using graphs to communicate financial information. Moreover, they recommend testing whether their findings generalize to more experienced decision makers, which would get closer to the study of Cardinaels (2008). Surprisingly, both Arunachalam et al. (2002) and Cardinaels (2008) do not even mention the possibility of culture having an influence.

Nevertheless, culture has been shown to dramatically affect a wide range of psychological processes. At the most basic perceptual level, members from different cultures have shown different susceptibilities to optical illusions and colour perception (Bontempo, Bottom, & Weber, 1997). Although research about the influence of culture on cost-based decision-making is scarce, few studies suggest taking into account differences among the users of information (Amer, 1991; Chandra & Krovi, 1999;

Ganzach, 1993). Furthermore, the theory on representational congruence (Arnold et al., 2004; Chandra & Krovi, 1999) predicts a more favourable impact on decision performance when the external presentation format fits the user’s cognitive style or internal representation, that is, when the external and internal representation are congruent (Chandra & Krovi, 1999).

If there is congruence between an already existing cognitive style (internal representation) and the external representation, the decision maker is better able to match the external to the internal representation, thereby leading to a better or more consistent mental representation, which facilitates the decision-making process thereby resulting in an efficient decision. On the other hand, if there is a mismatch between cognitive style and external representation (i.e. information is presented in a way that is not congruent with the cognitive style), then the decision maker is forced to combine incompatible structures resulting in memory loss and inefficiency. Therefore, representational congruence should lead to efficient retention and recall and consequently, to better performance (Chandra & Krovi, 1999).²

Cognitive, or thinking, style refers to “an individual’s characteristic and consistent approach to organizing and processing information” (Tennant, 1988). The most

2 Theory on cognitive fit suggests that a match between the external representation and the task results in a better decision-making performance. Chandra and Krovi (1999) extend the model of cognitive fit to also account for the congruence between the external information and the internal representation of the user and argue that this extended model is robust across tasks.

(16)

16

commonly recognized distinction is between analytical and intuitive styles. In a work context, an analytic person has the tendency to be compliant, favours structure and is most effective when handling problems that demand a step-by-step solution; whereas an intuitive individual tends to be nonconformist, favours open-ended tasks and performs best on problems favouring a holistic approach (Lynch, 1986). International differences in cognitive style have been seen in terms of an East-West dichotomy (e.g. Nisbett, 2003;

Nisbett et al., 2001; Redding, 1980; Torrance & Sato, 1979). The Western approach tends to be analytical with a focus on the sequential links between cause and effect, while Eastern thinking is leaning toward an intuitive assessment of the entire universe of events as a system of interdependent parts (Redding, 1980).

Interestingly, according to Lucas (1981), analytical decision makers, or those who especially look for details, benefit from tables. Contrarily, heuristic decision makers, or those who look at the entire problem, benefit from graphs that stress an overview of the same data (Lucas, 1981; Vessey, 1991a).

Since The Netherlands is part of the ‘analytic West’ and since Hong Kong is part of the

‘intuitive East’ and given that analytical decision makers benefit from tables and heuristic decision makers benefit from graphs, which increases their profit performance, we argue that Dutchmen benefit from tables and Hong Kong Chinese benefit from graphs. That is, based on the theory on representational congruence (Chandra & Krovi, 1999), a tabular format is likely to improve the profit performance of Dutchmen, whereas a graphical format is likely to improve the profit performance Hong Kong Chinese, due to the congruence between that particular presentation format and the user’s cognitive style. We define profit performance as profits from respondents’ cost-based decisions (see Cardinaels, 2008). Considering the above, we hypothesize the following:

H2a: Dutch decision makers achieve a higher profit performance given a tabular ABC presentation format than given a graphical ABC format.

H2b: Hong Kong Chinese decision makers achieve a higher profit performance given a graphical ABC presentation format than given a tabular ABC format.

3.3. Congruence with presentation format: A knowledge level dichotomy

Research that investigates the relative influence of graphical versus tabular formats remains inconclusive (Vessey, 1991a). A few studies, however, suggest taking into account decision makers’ characteristics, in an attempt to solve this controversy (Amer, 1991; Chandra & Krovi, 1999; Ganzach, 1993). Furthermore, the theory on representational congruence (Arnold et al., 2004; Chandra & Krovi, 1999) predicts a more favourable impact on decision performance when the external presentation format fits the subject’s cognitive style or internal representation (Chandra & Krovi, 1999).

According to Lucas (1981), analytical decision makers, or those who especially look for details, benefit from tables. Indeed, tables offer an analytical view of the data that enables

(17)

17

item-by-item evaluation (Vessey, 1991a). Contrarily, heuristic decision makers, or those who look at the entire problem, benefit from graphs that stress an overview of the same data (Lucas, 1981; Vessey, 1991a). As an analytical approach takes effort, a high level of knowledge in a field is needed to engage in analytical processing (Alba & Hutchinson, 1987: 418). In accordance with an analytical focus, high-knowledgeable users enact a more focused information search prior to solving a problem (Cloyd, 1997; Wood &

Lynch, 2002) and approach rules to discriminate between more and less relevant information (Bonner, 1990; O’Donnell, Koch, & Boone, 2005). On the contrary, Chi, Feltovich, and Glaser (1981), as cited in Vera-Muñoz et al. (2001: 409), suggest that novices as contrasted to experienced physicists, tend to focus more on surface features (heuristic style). Moreover, Benbasat and Schroeder (1977) suggest that low- knowledgeable decision makers focus on an overview and screen all available reports, while high-knowledgeable users look for particular details by requesting a limited amount of particular reports. An unanticipated finding by DeSanctis and Jarvenpaa (1989) suggests the possibility that graphical formats do not fit the way in which certain decision makers process information, since some users in their research started to convert bar charts into numerical tables.

Given that knowledge in a field leads to an analytical focus, we predict lower performance with a graphical format for high-knowledge decision makers. Since graphics highlight an overview of the data, changes in particular variables may go unnoticed; and given that knowledgeable decision makers look for particular details (item-by-item evaluation) we might expect them to perform better with tables (Mackay & Villarreal, 1987). On the other hand, we expect low-knowledge decision makers to benefit from graphs, since graphs emphasize an overview of the data. Therefore, we hypothesize the following:

H3a: Decision makers with low cost accounting knowledge achieve a higher profit performance given a graphical ABC presentation format than given a tabular ABC format.

H3b: Decision makers with high cost accounting knowledge achieve a higher profit performance given a tabular ABC presentation format than given a graphical ABC format.

3.4. Taking both culture and knowledge level into consideration

Although low-knowledge decision makers are hypothesized to benefit from graphs and high-knowledge decision makers from tables (see H3a and H3b), in hypotheses 2 we argue that Dutchmen benefit from tables and Hong Kong Chinese benefit from graphs.

That is, high-knowledge decision makers as well as Dutch decision makers are hypothesized to benefit from tables and both low-knowledge decision makers and Hong Kong Chinese decision makers are hypothesized to benefit from graphs. Therefore, a difference might exist between Dutchmen and Hong Kong Chinese in the relationship stated in H3a and H3b.

(18)

18

Since theory suggests that both high-knowledgeable users and Dutchmen have a match with tables and a mismatch with graphs, we expect a strong effect of presentation format on the high-knowledgeable Dutchmen. The low-knowledgeable Dutchmen, however, have a match with presentation format on the one hand (related to their knowledge level or culture) and a mismatch with presentation format on the other hand (with the remaining factor; culture or knowledge level). Since Cardinaels’ (2008) study found evidence for the relationship stated in H3a and H3b and given that his research was conducted in The Netherlands, we expect that these relationships will also hold within our Dutch sample. Therefore, we employ the exact same hypothesis as Cardinaels (2008), namely:

H4a: Dutch decision makers with high (low) cost accounting knowledge achieve a higher (lower) profit performance given a tabular ABC presentation format than given a graphical ABC format.

Since the current study examines whether the findings of Cardinaels (2008) are independent of culture, we will test whether the relationship stated in H4a also holds for Hong Kong Chinese. Theory suggests that both low-knowledgeable users and Hong Kong Chinese have a match with graphs and a mismatch with tables. Therefore, we expect a strong effect of presentation format on low-knowledgeable Hong Kong Chinese.

The high-knowledgeable Hong Kong Chinese, however, have a fit with presentation format on the one hand (related to their knowledge level or culture) and a misfit with presentation format on the other hand (with the remaining factor; culture or knowledge level).

Research on the relative influence of culture and knowledge levels is scarce (see section 2.4. Culture and knowledge), even outside the accounting literature. An example of such a study outside the accounting literature is the research of Bontempo et al. (1997).

University students in The Netherlands, the U.S., Hong Kong, and Taiwan, as well as an amount of Taiwanese security analysts rated the riskiness of 30 monetary lotteries. Cross- cultural differences in risk perception were greater than differences due to occupation.

Their findings suggest that cultural upbringing and environment can play a bigger role in shaping perceptions of financial risk than professional training or expertise. That is, risk perception was formed less by present situational influences, incentives, or loss functions and more by the larger, steadier, and more longstanding cultural environment (Bontempo et al., 1997).

However, as we discussed in section ‘2.2. Prior research’, Rong-Ruey and Thomas’

(2001) study found that experience-related knowledge (students and novices versus experienced auditors) has an influence on individuals’ cue weighting within Taiwan.

However, they document that within the U.S. students and auditors with varying levels of experience considered the same cue as most important. That is, knowledge levels do not have an influence within the U.S., but do have an influence within Taiwan. This shows that the influence of knowledge levels differs among cultures.

(19)

19

Contrarily, in the study of Ho and Chang (1994) experience-related knowledge does have an influence on individuals’ probability judgments in both the U.S. and Taiwan.

Furthermore, culture influences probability judgments of individuals that do not possess experience-related knowledge (auditors in the medical task as well as students in both the audit and medical task). However, culture does not significantly influence auditors with experience-related knowledge (that is, in the audit task); U.S. auditors and Taiwanese auditors arrive at highly similar judgments. This shows that the influence of culture differs among knowledge levels.

This shows that the relative influence of national culture and knowledge levels remains inconclusive. Besides, knowledge levels in these studies carry a different meaning as within the current study, namely; experience-related knowledge instead of cost- accounting knowledge. Moreover, the current study controls for experience. In Ho and Chang’s (1994) study, students’ knowledge levels are not being controlled for experience. Nor are different levels of knowledge within the group of auditors or within the group of students being accounted for, while students from the U.S. were MBA students and those from Taiwan were undergraduate students. The study of Rong-Ruey and Thomas (2001) also has its limitations. For instance, they only examined Taiwanese respondents and compared their findings to similar studies conducted in the U.S.

Moreover, these studies concern other types of decision-making and do not take presentation format into consideration. Finally, these studies predicted differences between cultures, based on specific characteristics of the Taiwanese culture.

Consequently, we cannot make any assumptions about the effect of presentation format within the Hong Kong Chinese high-knowledge subsample. However, since we can make assumptions about the effect of tables versus graphs within the Hong Kong Chinese low- knowledge subsample and since Cardinaels (2008) found evidence for the relationship stated in H4a within the Netherlands, we will test whether the same relationship holds within Hong Kong. Therefore, based on the same arguments as given for H4a, we hypothesize the following:

H4b: Hong Kong Chinese decision makers with high (low) cost accounting knowledge achieve a higher (lower) profit performance given a tabular ABC presentation format than given a graphical ABC format.

4. Experimental method

In order to test the hypotheses, we built an experiment similar to that of Cardinaels (2008). The Simulation Game (experimental task) comprises a complex management accounting problem. Respondents have to conduct eight trials of price and resource allocation (i.e. sales visits) decisions for three customers of a fictitious company. Their objective is to improve profits. Presentation format is the between subjects factor: every respondent receives ABC-driven customer profitability data, presented in either a tabular or a graphical format. Besides the Simulation Game, the respondents’ level of cost accounting knowledge is measured as well as some common control variables such as

The interplay between cost accounting knowledge and presentation formats in cost-based decision-making:

Table of contents

List of tables and figures

The interplay between cost accounting knowledge and presentation formats in cost-based decision-making:

A cross-cultural extension