Business Owners’ Cultural Orientations: Conceptualization, Measurement, and Implications for Business Success in China and Germany Inaugural-Dissertation zur Erlangung des Doktorgrades der Philosophie des Fachbereiches 06 der Justus-Liebig-Universität Gießen vorgelegt von Christine König aus Gießen 2007 Dekan: Prof. Dr. Joachim C. Brunstein 1. Berichterstatter: Prof. Dr. Michael Frese 2. Berichterstatter: Prof. Dr. Peter Schmidt Acknowledgments First of all, I would like to express my gratitude to Michael Frese who gave me the opportu- nity to write my dissertation within the framework of a Chinese-German research project. Doing cross-cultural research was a fascinating and challenging experience for me. I am grateful to Michael Frese for guiding and inspiring me throughout my dissertation, and to Peter Schmidt for providing valuable comments. Moreover, I would like to thank Zhong-Ming Wang, Wei Wei Guo, Lin Xia, Xianfang Xue, Xiangyang Zhao, and Jiping Zhu, as well as Andreas Rauch and Jens Unger for their support and cooperation. My special thanks go to Holger Steinmetz and Ute Stephan for their help and encouragement. I am thankful to Sandy Rosenhauer, Eva Schraub, and the other student assistants who contributed to the research project, as well as to the business owners who shared their business experiences with me. I am also thankful to the Chinese National Science Foundation and the German Research Foundation for funding the research project. Finally, I would like to express my gratitude to my parents and my brother, as well as to Jens Bedke, Fabienne Gogol, and Henning Pape who were always there for me. Table of Contents 1. Introduction 1 1.1. Research Project 1 1.2. Dissertation 2 1.2.1. First Study 3 1.2.2. Second Study 4 1.2.3. Third Study 5 2. Scenario-Based Scales Measuring Business Owners’ Cultural Orientations 6 2.1. Individual-Level Measurement of Owners’ Cultural Orientations 6 2.2. Scenario-Based Measurement of Owners’ Cultural Orientations 9 2.3. Cross-Cultural Validity 11 2.4. Construct Validity 12 2.5. Method 14 2.5.1. Development 14 2.5.2. Participants and Procedure 15 2.5.3. Measures 16 2.5.4. Cross-Cultural Validation 17 2.5.5. Construct Validation 18 2.6. Results 19 2.6.1. Cross-Cultural Validity 19 2.6.2. Construct Validity 22 2.7. Discussion 31 2.7.1. Limitations and Implications for Future Research 32 2.8. Conclusion 33 2.9. Addendum 34 3. Business Owners’ Cultural Orientations as Moderators of the Relationships Between Vision Characteristics and Business Success 35 3.1. The Moderator Effects of Owners’ Cultural Orientations 35 3.2. Cross-Cultural Differences in the Moderator Effects of Owners’ Cultural Orientations 39 3.3. Method 41 3.3.1. Participants 41 3.3.2. Procedure and Measures 42 3.3.3. Method of Analysis 46 3.4. Results 48 3.4.1. Intercorrelations 48 3.4.2. Interaction Effects on Business Success 48 3.5. Discussion 70 3.5.1. Limitations 72 3.5.2. Implications for Future Research 72 3.6. Conclusion 73 4. The Relationships Between Business Owners’ Cultural Orientations and Business Success 74 4.1. Owners’ Cultural Orientations and Business Success 74 4.2. Method 77 4.2.1. Participants and Procedure 77 4.2.2. Measures 78 4.2.3. Method of Analysis 81 4.3. Results 84 4.3.1. Intercorrelations 84 4.3.2. Measurement Models 85 4.3.3. Structural Models 85 4.4. Discussion 92 4.4.1. Limitations and Implications for Future Research 94 4.5. Conclusion 94 5. Overall Discussion 96 6. References 99 7. Appendix 109 Chapter 1 Introduction 1 1. Introduction This dissertation comprises three studies of business owners’ cultural orientations. 1 They are presented in Chapters 2, 3, and 4. An introduction to the three studies is given in Chapter 1, and an overall discussion of them is provided in Chapter 5. 1.1. Research Project The three studies comprised in this dissertation were conducted within the research project ‘Psychological Factors of Entrepreneurial Success in China and Germany’. Supervised by Prof. Dr. Zhong-Ming Wang from the University of Zhejiang and Prof. Dr. Michael Frese from the University of Giessen, the research project was carried out by Chinese and German graduates and postgraduates of psychology and management. It was funded by the Chinese National Science Foundation and the German Research Foundation. Consisting of two parts (T1 and T2), the research project was longitudinal in design. It started in January 2004 and will end in March 2008. As a framework for the research project, we used a revised version of the ‘Giessen- Amsterdam Model of Entrepreneurial Success’ by Frese and Wang (2003, 2005), which builds on the original version by Rauch and Frese (2000). The model assumes that actions, which are mainly influenced by visions, goals, and strategies, mediate the relationships be- tween personality, human capital, and environment on the one hand and business success on the other hand (Frese & Wang, 2003, 2005). We studied the concepts included in the model in samples of Chinese and German owners. Both at T1 and T2, we interviewed the owners and asked them as well as up to three of their employees to complete questionnaires. Among other things, the interview schemes contained questions regarding the owners’ actions, their visions, goals, and strategies, and their objective business success, whereas the questionnaires con- tained measures of the owners’ personality and human capital, the environment their busi- nesses operated in, and their subjective business success. Further, the model assumes that actions, visions, goals, and strategies, as well as personality, human capital, environment, and business success are influenced by culture (Frese & Wang, 2003, 2005). Studying the con- 1 Business owners are defined as individuals who own and manage their businesses (Carland, Hoy, Boulton, & Carland, 1984). For simplification, they are referred to as ‘owners’ in the following. Chapter 1 Introduction 2 cepts included in the model in two cultures as different as China and Germany enabled us to explore this influence. According to the Global Leadership and Organizational Behavior Effectiveness (GLOBE) Study (House & Javidan, 2004), which introduced nine cultural dimensions and measured them in terms of practices and values, China and (Western) Germany differ in six of the nine cultural dimensions, measured in terms of practices: China is higher on institutional collectivism (Gelfand, Bhawuk, Nishii, & Bechthold, 2004), in-group collectivism (Gelfand et al., 2004), humane orientation (Kabasakal & Bodur, 2004), and performance orientation (Javidan, 2004), whereas Germany is higher on assertiveness (Den Hartog, 2004) and future orientation (Ashkanasy, Gupta, Mayfield, & Trevor-Roberts, 2004). This means that there is more “collective distribution of resources and collective action” (institutional collectivism), more “loyalty and cohesiveness” (in-group collectivism), more fairness, altruism, generosity, care, and kindness (humane orientation), and more striving for “performance improvement and excellence” (performance orientation) in China, whereas there is more confrontation and aggressiveness (assertiveness) and more “delaying gratification, planning, and investing in the future” (future orientation) in Germany (Javidan, House, & Dorfman, 2004, p. 30). 2 1.2. Dissertation Culture is manifested in practices and values of societies and organizations (Erez & Gati, 2004; House & Javidan, 2004). Cultural orientations are manifested in practices and values of individuals (Chirkov, Ryan, Kim, & Kaplan, 2003; Maznevski, DiStefano, Gomez, Noorderhaven, & Wu, 2002). The focus in entrepreneurship research has been on culture (Freytag & Thurik, 2007; George & Zahra, 2002). The focus of the three studies comprised in this dissertation is on cultural orientations. The first study focuses on the conceptualization and measurement of owners’ cultural orientations, whereas the second and the third study focus on the implications of owners’ cultural orientations for business success. The three studies were conducted to provide evidence on whether owners’ cultural orientations are useful concepts for entrepreneurship research. 2 China and Germany do not differ in uncertainty avoidance (Sully De Luque & Javidan, 2004), power distance (Carl, Gupta, & Javidan, 2004), and gender egalitarianism (Emrich, Denmark, & Den Hartog, 2004). This means that there is as much reliance on “social norms, rules, and procedures” to prevent incertitude (uncer- tainty avoidance), as much acceptance of power being distributed unequally (power distance), and as much promotion of gender equality (gender egalitarianism) in China as in Germany (Javidan et al., 2004, p. 30). Chapter 1 Introduction 3 1.2.1. First Study The first study focuses on the conceptualization and measurement of owners’ cultural orientations. It is cross-sectional in design and was conducted in China and Germany. We argued that many approaches to cross-cultural measurement in entrepreneurship research have been flawed and that there is a need for scales measuring owners’ cultural orientations at the individual level. We conceptualized owners’ cultural orientations as mani- fested in the practices and values owners use in their businesses. However, we considered practices to be more relevant for studying owners than values because practices are related to actions (Frese, 2006). Owners’ practices substantially influence the development of organiza- tional cultures (Schein, 2004). We developed scales measuring cultural orientations that refer to cultural dimensions introduced by the GLOBE Study (House & Javidan, 2004). We adapted the definitions given by Javidan et al. (2004, p. 30) to the practices owners use in their businesses: Uncertainty avoidance implies that owners support reliance on “social norms, rules, and procedures” to prevent incertitude. Power distance means that owners pro- mote acceptance of power being distributed unequally. Collectivism signifies that owners foster “collective distribution of resources and collective action” (institutional collectivism) as well as “loyalty and cohesiveness” (in-group collectivism). Assertiveness implies that owners support confrontation and aggressiveness. Future orientation signifies that owners foster “delaying gratification, planning, and investing in the future”. Humane orientation means that owners promote fairness, altruism, generosity, care, and kindness. Performance orientation implies that owners support striving for “performance improvement and excellence”. 3 We based the scales on scenarios rather than on Likert items because scenario-based scales tend to hold higher cross-cultural validity and higher construct validity than scales based on Likert items, which means that they tend to allow for more meaningful cross-cultural comparisons and more accurate descriptions and predictions of behaviors (Peng, Nisbett, & Wong, 1997). We validated the scales on Chinese and German owners. If we succeeded in demonstrating the scales’ suitability for owners from two cultures as different as China and Germany (Javidan et al., 2004), we could be optimistic that the scales would also be suitable for owners from other cultures. 3 Later on, we also developed a scale measuring gender egalitarianism, a cultural orientation that refers to another cultural dimension introduced by the GLOBE Study (House & Javidan, 2004). Gender egalitarianism means that owners promote gender equality (Javidan et al., 2004). Chapter 1 Introduction 4 1.2.2. Second Study The second study focuses on the implications of owners’ cultural orientations for busi- ness success. It is cross-sectional in design and was conducted in China and Germany. Visions represent images of desirable futures that provide meaning and direction (Bennis & Nanus, 1985; House & Shamir, 1993; Kouzes & Posner, 1987). They can be de- scribed by vision characteristics (Locke et al., 1991). One focus in entrepreneurship research has been on the effectiveness of vision characteristics, that is, on the relationships between vision characteristics and business success (Baum, Locke, & Kirkpatrick, 1998). Business success comprises financial dimensions, such as sales growth or growth in the number of employees, and operational dimensions, such as product and service quality or customer satisfaction (Combs, Crook, & Shook, 2005; Venkatraman & Ramanujam, 1986). Vision characteristics may be more effective, that is, more strongly related to business success, for some owners than for others. However, moderators of the relationships between vision char- acteristics and business success have not yet been identified in entrepreneurship research. We assumed that owners’ cultural orientations moderate the relationships between vision charac- teristics and business success. This assumption implied that the effectiveness of vision charac- teristics depends on whether they match owners’ cultural orientations ('match hypothesis', Tung, Walls, & Frese, 2006). We developed hypotheses regarding the moderator effects of six cultural orientations, namely, performance orientation, humane orientation, future orientation, assertiveness, power distance, and uncertainty avoidance. We hypothesized that each of the six cultural orientations moderates the relationship between one vision characteristic and business success. The six vision characteristics were ‘challenge’, ‘social responsibility’, ‘fu- ture orientation’, ‘growth orientation’, ‘clarity’, and ‘stability’ (Baum et al., 1998; Locke et al., 1991). Further, we assumed that cross-cultural differences in institutional collectivism (Gelfand et al., 2004), in-group collectivism (Gelfand et al., 2004), and the prevailing con- strual of the self (Markus & Kitayama, 1991) may lead to cross-cultural differences in the moderator effects of owners’ cultural orientations. We supposed that our hypotheses would receive more support in collectivist cultures than in individualistic cultures. Therefore, we tested our hypotheses in China, a culture that is high on institutional collectivism and in-group collectivism and in which the interdependent construal of the self prevails, and in Germany, a culture that is low on institutional collectivism and in-group collectivism and in which the independent construal of the self prevails (Gelfand et al., 2004; Markus & Kitayama, 1991). We supposed that our hypotheses would receive more support in China than in Germany. Chapter 1 Introduction 5 1.2.3. Third Study The third study also focuses on the implications of owners’ cultural orientations for business success. It is longitudinal in design and was conducted in Germany. The focus in entrepreneurship research has been on studying the relationships between culture and entrepreneurial concepts at the societal and the organizational level of analysis (cf. the review by Hayton, George, & Zahra, 2002) rather than the relationships between cultural orientations and entrepreneurial concepts at the individual level of analysis. Shifting the fo- cus, we studied the relationships between owners’ cultural orientations and business success. These relationships may be characterized as follows: First, owners’ cultural orientations may have effects on business success. Second, business success may have effects on owners’ cultural orientations. Third, there may be reciprocal effects. Moreover, the effects may be synchronous or lagged. Our aim was to determine the effects by which the relationships be- tween each of six cultural orientations and business success are characterized. The six cultural orientations were performance orientation, humane orientation, future orientation, assertive- ness, power distance, and uncertainty avoidance. _______ The chapters in which the three studies are presented can be read independently of each other. They contain separate introductions and discussions. Chapter 2 Cultural Orientation Scales 6 2. Scenario-Based Scales Measuring Business Owners’ Cultural Orien- tations Culture is manifested in practices and values of societies and organizations (Erez & Gati, 2004; House & Javidan, 2004). It constitutes an important concept in entrepreneurship research (Freytag & Thurik, 2007). Indeed, there has been a recent increase in cross-cultural studies on entrepreneurship (cf. the review by Hayton et al., 2002). Several good scales have been developed to measure societal culture, such as the ones by Hanges and Dickson (2004) and Schwartz (1994). However, these scales should only be used when research is oriented toward the societal level of analysis. For example, scholars should use these scales when relating societal culture to rates of business ownership. In contrast, when research is oriented toward the individual level of analysis, scales developed to measure individual cultural orien- tations should be used. For example, scholars should use these scales when relating individual cultural orientations to business success. Cultural orientations are manifested in practices and values of individuals (Chirkov et al., 2003; Maznevski et al., 2002). We developed and validated scenario-based scales measuring business owners’ cul- tural orientations. 1 Given their conceptual and methodological features, the scales differ from other scales commonly used in cross-cultural research: Instead of measuring culture at the societal level, they measure cultural orientations at the individual level, and instead of being based on Likert items, they are based on scenarios. 2.1. Individual-Level Measurement of Owners’ Cultural Orientations Many cross-cultural studies in entrepreneurship research have focused on individual owners (cf. the review by Hayton et al., 2002). However, instead of measuring individual owners’ cultural orientations, these studies imputed national culture scores found in other cross-cultural studies to each individual owner (“culture inferred from nationality” as Hayton et al., 2002, p.38, called it). The study by Steensma, Marino, Weaver, and Dickson (2000), which imputed Hofstede’s national culture scores as individual scores, may serve as an exam- ple. Hofstede (2001) has repeatedly warned against such imputations because they involve committing ecological fallacies. They are based on the wrong assumption that all people within a nation show the same level of cultural concepts. Moreover, imputing national culture 1 For simplification, ‘business owners’ are referred to as ‘owners’ in the following. Chapter 2 Cultural Orientation Scales 7 scores found in studies based on non-owners (such as the study by Hofstede, which was based on managers) to owners is problematic because owners are systematically different from non- owners across cultures (McGrath & MacMillan, 1992). Finally, the relationships between cultural concepts and outcome variables often differ depending on the level at which the cultural concepts are measured (Hofstede, 2002; Klein, Dansereau, & Hall, 1994). There are two approaches to dealing with these problems: The first approach is to measure owners’ cultural orientations using scales developed to measure societal culture. However, this leads to problems of analysis and interpretation (Chan, 1998; Klein et al., 1994). The use of societal-level scales at the individual level often involves loss of reliability and validity (Hofstede, 2002; Spector, Cooper, & Sparks, 2001). The second approach is to measure owners’ cultural orientations using scales developed to measure individual cultural orientations. However, among the cultural orientation scales, we are not aware of any that are suitable for owners. Therefore, we developed scales measuring the practices owners use in their businesses. We considered practices to be more relevant for studying owners than values because practices are related to actions (Frese, 2006). Owners are defined as individuals who own and manage their businesses (Carland et al., 1984). How owners go about managing their businesses becomes apparent in their practices (Schein, 2004). The practices owners use in their businesses provide starting points for the development of organizational cultures. Start- ing from owners’ practices, organizational cultures develop as a result of the interactions between owners and their employees (Schein, 2004). Thus, although owners cannot entirely determine organizational cultures, they can substantially influence them through their prac- tices. Owners support organizational cultures that they consider conducive to business success (Ogbonna & Harris, 2000). As starting points for their development, owners use practices that may or may not be in accordance with their personality traits (Schein, 2004). For example, owners who are personally low on humane orientation may nevertheless use humane-oriented practices when they expect humane-oriented organizational cultures to foster their employees’ motivation. Whereas personality traits are genetic and unalterable (Jang, Livesley, & Vernon, 1996), cultural orientations are acquired and can be altered. Thus, cultural orientations can be distinguished from personality traits. The scales measure cultural orientations that refer to cultural dimensions introduced by the Global Leadership and Organizational Behavior Effectiveness (GLOBE) Study (House & Javidan, 2004). The cultural dimensions represent a theoretical and empirical advancement over other cultural dimensions introduced earlier in cross-cultural research. We based the Chapter 2 Cultural Orientation Scales 8 scales on definitions given by Javidan et al. (2004, p. 30), but adapted the definitions to the practices owners use in their businesses: Uncertainty avoidance implies that owners support reliance on “social norms, rules, and procedures” to prevent incertitude. Power distance means that owners promote acceptance of power being distributed unequally. Collectivism signifies that owners foster “collective distribution of resources and collective action” (institu- tional collectivism) as well as “loyalty and cohesiveness” (in-group collectivism). Assertive- ness implies that owners support confrontation and aggressiveness. Future orientation signi- fies that owners foster “delaying gratification, planning, and investing in the future”. Humane orientation means that owners promote fairness, altruism, generosity, care, and kindness. Performance orientation implies that owners support striving for “performance improvement and excellence”. 2 Similar to the importance of organizational cultures for business success (O'Reilly & Chatman, 1966), cultural orientations have an important function for how businesses are managed. For example, when owners support uncertainty avoidance, there is little support for risk taking. Provided that risk taking is crucial for business success, owners’ support for uncertainty avoidance may reduce business success. Apart from one-to-one relationships between owners’ cultural orientations and business success, there may also be interactions (Tung et al., 2006). For example, owners’ support for uncertainty avoidance may particularly reduce business success when their businesses operate in high-tech environments with many competitors and owners, therefore, need to take risks. The development and the validation of the cultural orientation scales were embedded in a research project on Chinese and German owners. This enabled us to ascertain whether the scales are suitable for both Chinese and German owners and whether they allow for meaning- ful comparisons across China and Germany. If we succeeded in demonstrating the scales’ suitability for owners from two cultures as different as China and Germany (Javidan et al., 2004), we could be optimistic that the scales would also be suitable for owners from other cultures and that they would allow for meaningful comparisons across cultures other than China and Germany as well. 2 Later on, we also developed a scale measuring gender egalitarianism, a cultural orientation that refers to another cultural dimension introduced by the GLOBE Study (House & Javidan, 2004). Gender egalitarianism means that owners promote gender equality (Javidan et al., 2004). Please see the Addendum on page 34 for details. Chapter 2 Cultural Orientation Scales 9 2.2. Scenario-Based Measurement of Owners’ Cultural Orientations Cross-cultural scholars have recently suggested that culture and cultural orientations should be measured using scales based on scenarios rather than using scales based on Likert items (Heine et al., 2001; Heine, Lehman, Peng, & Greenholtz, 2002; Kitayama, 2002; Peng et al., 1997). Likert items and scenarios differ in the measurement of culture and cultural orientations (Peng et al., 1997). Likert items consist of general abstract statements, such as ‘I care for my family members’, and standardized scale responses, such as ‘strongly agree’ or ‘strongly disagree’. Hence, Likert items measure culture and cultural orientations via people’s self-evaluations on general abstract statements. In contrast, scenarios consist of concrete social situations, such as ‘Your poorly qualified nephew asks you to employ him in your business’, and behavioral options, such as ‘You employ your poorly qualified nephew’ or ‘You don’t employ your poorly qualified nephew’. Hence, scenarios measure culture and cultural orientations via people’s behavioral preferences in concrete social situations. Scales based on Likert items tend to hold lower cross-cultural validity than scales ba- sed on scenarios, which means that they tend to allow for less meaningful cross-cultural com- parisons (Peng et al., 1997). The cross-cultural validity of scales based on Likert items has been challenged for two main reasons: First, Likert items are more likely to be interpreted differently by people from different cultures than scenarios because general abstract state- ments and standardized scale responses offer a wider scope of interpretation than concrete social situations and behavioral options (Kitayama, 2002; Peng et al., 1997). For example, Chinese and German owners are more likely to differ in their interpretations of what it means to care for one’s family members than in their interpretations of what it signifies to be asked by one’s poorly qualified nephew to employ him in one’s business. Also, Chinese and Ger- man owners are more likely to differ in their interpretations of what it means to ‘strongly agree’ or to ‘strongly disagree’ than of what it signifies to employ or not employ one’s poorly qualified nephew. Different interpretations threaten the validity of cross-cultural comparisons (Kitayama, 2002; Peng et al., 1997). Second, Likert items are more affected by the reference group effect than scenarios (Heine et al., 2001; Peng et al., 1997). The reference group effect occurs when people have to refer to the standards of their reference groups (Heine et al., 2002). Likert items are affected by the reference group effect because people have to refer to the standards of their reference groups to give their self-evaluations on general abstract statements (Biernat, Manis, & Nel- son, 1991). For example, to tell how much they care for their family members, owners have to Chapter 2 Cultural Orientation Scales 10 consider how much other owners care for their family members. Scenarios are less affected by the reference group effect because people do not have to refer to the standards of their refer- ence groups to give their behavioral preferences in concrete social situations (Peng et al., 1997). For example, to tell whether or not they prefer to employ their poorly qualified neph- ews, owners do not have to consider whether or not other owners prefer to employ their poorly qualified nephews. The reference group effect occurs because people from different cultures have different reference groups that may differ in their standards (Heine et al., 2002). For example, Chinese owners refer to other Chinese owners, whereas German owners refer to other German owners. If Chinese owners generally care more for their family members than German owners, Chinese owners evaluate themselves with higher standards than German owners. Different standards threaten the validity of cross-cultural comparisons (Heine et al., 2002; Peng et al., 1997). In addition to these cross-cultural issues, scales based on Likert items tend to hold lo- wer construct validity than scales based on scenarios, which means that they tend to allow for less accurate descriptions and predictions of behaviors (Peng et al., 1997). The construct validity of scales based on Likert items has been challenged because people’s self-evaluations on general abstract statements less accurately describe and predict their behaviors than peo- ple’s behavioral preferences in concrete social situations (Chan & Schmitt, 1997; Motowidlo, Dunnette, & Carter, 1990). For example, owners’ self-evaluations on how much they care for their family members less accurately describe and predict the active support they provide to them than owners’ preferences on whether or not to employ their poorly qualified nephews. As we wanted the cultural orientation scales to hold cross-cultural validity and con- struct validity, we based them on scenarios rather than on Likert items. However, this in- volved potential loss of reliability in terms of coefficient alpha (Cronbach, 1951) and compos- ite reliability, a reliability estimate used in structural equation modeling (Fornell & Larcker, 1981). Coefficient alpha and composite reliability estimate internal consistency. Scales based on scenarios tend to show lower internal consistencies than scales based on Likert items (Chan & Schmitt, 1997; Motowidlo et al., 1990). Consisting of concrete social situations and behavioral options, scenarios capture more situational and behavioral aspects than Likert items that consist of general abstract statements and standardized scale responses. Therefore, scenarios have higher specific variances that result in lower intercorrelations. We accepted potential loss of reliability in terms of coefficient alpha and composite reliability because we considered it outweighed by the superior cross-cultural validity and construct validity held by Chapter 2 Cultural Orientation Scales 11 scenario-based scales. Moreover, there is an alternative to coefficient alpha and composite reliability, and that is test-retest reliability. Test-retest reliability is assumed to be a more appropriate reliability estimate for scenario-based scales because it does not estimate internal consistency (Chan & Schmitt, 1997; Motowidlo et al., 1990). 2.3. Cross-Cultural Validity The cultural orientation scales are useful for cross-cultural scholars if they hold cross- cultural validity and, thus, allow for meaningful comparisons across cultures. In particular, the scales must enable scholars to meaningfully compare the means of the cultural orientations as well as the relationships between them. Five forms of invariance should be supported for the scales (Steenkamp & Baumgartner, 1998; Vandenberg & Lance, 2000): Configural invariance implies that a measure holds an equal configuration of factors and indicators across cultures. The same indicators load on the same factors. Given configural invariance, scholars can compare constructs across cultures because the constructs have the same meaning. Configural invariance provides the basis for all other forms of invariance (Horn & McArdle, 1992). Metric invariance means that the indicators have equal factor load- ings across cultures. Scalar invariance signifies that the indicators also have equal intercepts across cultures. Given metric and scalar invariance, scholars can conduct meaningful cross- cultural comparisons of observed and latent construct means (Horn & McArdle, 1992; Mere- dith, 1993). Factor variance invariance implies that the factors have equal variances across cultures. Given metric and factor variance invariance, scholars can conduct meaningful cross- cultural comparisons of relationships between constructs (Schmitt, 1982; Steenkamp & Baumgartner, 1998). Error variance invariance means that the indicators hold equal error variances across cultures. Given metric, factor variance, and error variance invariance, a measure is equally reliable across cultures (Steenkamp & Baumgartner, 1998). Configural, metric, scalar, and error variance invariance are forms of measurement invariance, which concerns the relationships between the factors and the indicators (Byrne, Shavelson, & Muthén, 1989). Factor variance invariance is a form of structural invariance, which concerns the factors themselves (Byrne et al., 1989). In contrast to full invariance, partial invariance signifies that some, but not all, pa- rameters are equal across cultures (Reise, Widaman, & Pugh, 1993). Partial invariance is more likely to be supported in cross-cultural research than full invariance (Steenkamp & Baumgartner, 1998). Fortunately, partial invariance hardly affects the meaningfulness of Chapter 2 Cultural Orientation Scales 12 cross-cultural comparisons. Provided that metric and scalar invariance are partially given, construct means can still be meaningfully compared across cultures (Byrne et al., 1989). Provided that metric and factor variance invariance are partially given, relationships between constructs with equal variances can still be meaningfully compared across cultures (Byrne et al., 1989). 2.4. Construct Validity We established a nomological net (Cronbach & Meehl, 1955) in which we related the cultural orientations to other constructs that are theoretically associated with them. Among the constructs covered in the research project, we considered nine constructs to be appropriate validation constructs for the cultural orientations: (1) Achievement striving implies that own- ers work hard to achieve their goals (McClelland, 1961). (2) Deliberation means that owners carefully consider their decisions (Costa & McCrae, 1992). (3) Error communication signifies that owners turn to their employees when they have made errors (Rybowiak, Garst, Frese, & Batinic, 1999). (4) Meta-cognitive activity implies that owners plan, monitor, and revise their performance (Schmidt & Ford, 2003). (5) Task-oriented personal initiative means that owners take proactive and self-starting approaches to seizing opportunities and preparing for chal- lenges (Frese, Kring, Soose, & Zempel, 1996). (6) Relationship-oriented personal initiative signifies that owners take proactive and self-starting approaches to improving and expanding their business relationships (Zhao, Giardini, & Frese, 2005). (7) Social satisfaction implies that owners are satisfied with the social relationships they have with their employees. (8) Number of co-owners who are actively involved in the management of the business. (9) Num- ber of family members who work in the business. Uncertainty avoidance. We made the hypotheses that uncertainty avoidance is nega- tively related to task-oriented and to relationship-oriented personal initiative. Task-oriented and relationship-oriented personal initiative require willingness to take risks because taking proactive and self-starting approaches to seizing opportunities and preparing for challenges or to improving and extending business relationships brings about changes and, therefore, in- creases uncertainty (Fay & Frese, 2001; Frese, Fay, Hilburger, Leng, & Tag, 1997). Owners who consider it beneficial to rely on “social norms, rules, and procedures” to prevent incerti- tude (Javidan et al., 2004, p. 30) are unwilling to take risks (Sully De Luque & Javidan, 2004). This renders them unlikely to show task-oriented and relationship-oriented personal initiative. Chapter 2 Cultural Orientation Scales 13 Power distance. We hypothesized a negative relationship between power distance and error communication. To turn to their employees when they have made errors, owners must be ready to acknowledge fallibility (Hofstede, 1984; Rybowiak et al., 1999). Owners who deem it advantageous that their employees accept power being distributed unequally (Javidan et al., 2004) are interested in demonstrating infallibility to assert their superior positions (Carl, Gupta, & Javidan, 2004). Therefore, they are unlikely to communicate their errors. Collectivism. We made the hypotheses that institutional collectivism is positively re- lated to the number of co-owners who are actively involved in the management of the busi- ness, and that in-group collectivism is positively related to the number of family members who work in the business. Owners who consider it beneficial to act collectively (Javidan et al., 2004) are likely to join with others in managing their businesses (Gelfand et al., 2004; Hofstede, 1984), whereas owners who deem it advantageous to be loyal and cohesive (Javidan et al., 2004) are likely to employ their family members (Gelfand et al., 2004; Hofstede, 1984). Assertiveness. We hypothesized negative relationships between assertiveness and de- liberation as well as between assertiveness and meta-cognitive activity. Carefully considering decisions and planning, monitoring, and revising performance imply cautiousness and fore- thought (Costa & McCrae, 1992; Schmidt & Ford, 2003). Owners who deem it advantageous to be confrontative and aggressive (Javidan et al., 2004) are rather impulsive (Den Hartog, 2004). Therefore, they are unlikely to show deliberation and meta-cognitive activity. Future orientation. We made the hypothesis that future orientation is positively related to achievement striving. Owners who consider it beneficial to delay gratification and to invest in the future (Javidan et al., 2004) are likely to work hard to achieve their goals (Ashkanasy et al., 2004; McClelland, 1961). Moreover, we made the hypotheses that future orientation is positively related to deliberation and to meta-cognitive activity. Carefully considering deci- sions and planning, monitoring, and revising performance reflect planful action (Costa & McCrae, 1992; Schmidt & Ford, 2003). Owners who deem it advantageous to plan (Javidan et al., 2004) act planfully (Ashkanasy et al., 2004). This renders them likely to show deliberation and meta-cognitive activity. Humane orientation. We hypothesized positive relationships between humane orienta- tion and error communication as well as between humane orientation and social satisfaction. To turn to their employees when they have made errors and to be satisfied with the social Chapter 2 Cultural Orientation Scales 14 relationships they have with their employees, owners must perceive these relationships as trusting and compassionate (Hofstede, 1984; Rybowiak et al., 1999). Given that fairness, altruism, generosity, care, and kindness (Javidan et al., 2004) enhance mutual trust and com- passion, owners who promote these qualities probably perceive their relationships with their employees as trusting and compassionate (Kabasakal & Bodur, 2004). Therefore, they are likely to communicate their errors and to be socially satisfied. Performance orientation. We made the hypothesis that performance orientation is positively related to achievement striving. Owners who consider it beneficial to strive for excellence (Javidan et al., 2004) are likely to work hard to achieve their goals (Javidan, 2004; McClelland, 1961). Moreover, we hypothesized a positive relationship between performance orientation and meta-cognitive activity. Owners who deem it advantageous to strive for per- formance improvement (Javidan et al., 2004) are likely to plan, monitor, and revise their performance (Javidan, 2004; Schmidt & Ford, 2003). Finally, we made the hypotheses that performance orientation is positively related to task-oriented and to relationship-oriented personal initiative. Taking proactive and self-starting approaches to seizing opportunities and preparing for challenges or to improving and extending business relationships enhances busi- ness success (Frese, Krauss, & Friedrich, 2000; Koop, De Reu, & Frese, 2000). Owners who support striving for “performance improvement and excellence” (Javidan et al., 2004, p. 30) are interested in enhancing business success (Javidan, 2004). This renders them likely to show task-oriented and relationship-oriented personal initiative. 2.5. Method 2.5.1. Development Following a parallel approach (Harkness, Van de Vijver, & Johnson, 2003), we devel- oped the cultural orientation scales in a team of Chinese and German scholars. We combined our expertise to make sure that the scales were suitable for both Chinese and German owners. We developed the scales in English. The translations into Chinese and German were produced and checked by competent bilinguals. We created scenarios that consist of social situations and behavioral options. The so- cial situations describe problems owners may encounter in their businesses. They all begin with ‘Imagine that …’ and end with ‘What do you do?’. Each social situation represents one of the cultural orientations. For example, the problem whether or not to employ one’s poorly qualified nephew represents in-group collectivism. The behavioral options describe behaviors Chapter 2 Cultural Orientation Scales 15 owners may show to solve the problems. They all begin with ‘You …’. Two behavioral op- tions follow each social situation. The first option represents a low score on the cultural orien- tation, whereas the second option represents a high score. For example, not employing one’s poorly qualified nephew represents low in-group collectivism, whereas employing him repre- sents high in-group collectivism. Between the two behavioral options, there are two mirror- inverted three-point scales that are directed towards the first and the second option, respec- tively. The two scales range from ‘somewhat true of me’ (3/4) over ‘very true of me’ (2/5) to ‘extremely true of me’ (1/6). To complete the scales, owners have to make themselves aware of how they generally behave in their businesses. Going through the scenarios, they have to make mental simulations of their behaviors in the social situations. For each social situation, they have to decide which of the two behavioral options applies more to them. They can indicate their decision by ticking a point on the respective three-point scale. In a pilot study, we tested the scenarios on 100 Chinese and German business students. Based on the data obtained from the business students, we conducted exploratory factor ana- lyses to judge whether the scenarios appropriately measured the cultural orientations. Judging them as appropriate, we included all those scenarios in the scales that had high factor loadings on the cultural orientations. At the end of their development, the scales comprised 40 scenar- ios. Institutional collectivism and in-group collectivism were each assessed by three scenar- ios. Uncertainty avoidance and power distance were each measured by five scenarios, whereas assertiveness, future orientation, humane orientation, and performance orientation were each assessed by six scenarios. 2.5.2. Participants and Procedure The sample comprised Chinese and German owners. Their businesses belonged to four industries, namely, information technology, hotel and catering, automobile, and construction. To participate in the study, the owners had to meet two criteria: First, they had to own (with shares of at least 10%) and manage their businesses. Second, they had to have at least one employee. There is a qualitative difference between owners who work alone and owners who have employees. The step from working alone to having employees implies a change in self- perception, responsibility, and managerial demands (Frese & de Kruif, 2000). We searched for participants in four provinces (Hubei, Hunan, Jiangsu, and Zhejiang) and two municipali- ties (Chongqing and Shanghai) in Eastern China and in one province (Hesse) in Western Germany. As a first strategy, we used the yellow pages as well as lists provided by the Chi- Chapter 2 Cultural Orientation Scales 16 nese local government and the German chamber of commerce. As a second strategy, we relied on personal contacts with and recommendations of owners. The first strategy was more effec- tive in Germany, whereas the second strategy was more effective in China. Of the 458 owners who met the criteria for participation in China, 298 (65%) partici- pated in the study. Of the 697 owners who met the criteria for participation in Germany, 290 (42%) participated in the study. We interviewed the owners and asked them to complete a questionnaire that included the cultural orientation scales. Of the 588 owners who participated in the study, 461 (78%) completed the questionnaire. They served as participants for the validation of the scales. Among them were 260 Chinese (56%) and 201 Germans (44%). Most of the Chinese and the German owners did not only own and manage their businesses but had also founded them (82%, n = 213, and 68%, n = 137, respectively). The Chinese owners had 198 employees on average. Their businesses belonged particularly to the automobile industry (33%, n = 85), followed by the hotel and catering industry (26%, n = 68), the information technology industry (21%, n = 56), and the construction industry (20%, n = 51). The German owners had 12 employees on average. Their businesses belonged particularly to the construc- tion industry (41%, n = 82), followed by the information technology industry (23%, n = 47), the hotel and catering industry (21%, n = 43), and the automobile industry (15%, n = 30). Six months after they had completed the scales for the first time, we asked 25 German owners to complete them a second time. The 22 German owners (88%) who agreed to do so served as participants for the assessment of the scales’ test-retest reliabilities. The sub-sample was representative of the German sample. 2.5.3. Measures Apart from the cultural orientation scales, the questionnaire included scales and single items that measured the validation constructs: 3 Achievement striving and deliberation were each assessed by two items developed by Costa and McCrae (1992). Sample items were ‘I work hard to accomplish my goals’ and ‘I think things through before coming to a decision’, respectively. The items were rated on five-point scales ranging from ‘strongly disagree’ (1) to ‘strongly agree’ (5). Error communication was measured by four items adapted from Ry- bowiak et al. (1999). A sample item was ‘If I cannot rectify an error by myself, I turn to my 3 The composite reliabilities of the scales measuring the validation constructs could only be assessed through specifying and estimating models. Therefore, they are presented in the result section. Chapter 2 Cultural Orientation Scales 17 employees’. The items were rated on five-point scales ranging from ‘does not apply at all’ (1) to ‘applies completely’ (5). Meta-cognitive activity was assessed by ten items adapted from Schmidt and Ford (2003). A sample item was ‘I think about what skills need the most prac- tice’. The items were rated on five-point scales ranging from ‘strongly disagree’ (1) to ‘strongly agree’ (5). Task-oriented personal initiative was measured by seven items developed by Frese et al. (1997), whereas relationship-oriented personal initiative was assessed by seven items developed by Frese, König, and Rauch (2005). Sample items were ‘I actively attack problems’ and ‘I actively seek to improve my business relationships’, respectively. The items were rated on five-point scales ranging from ‘does not apply at all’ (1) to ‘applies completely’ (5). Social satisfaction was measured by a single item: ‘How satisfied are you with your social relationships with your employees?’ The item was rated on a seven-point scale ranging from ‘very unsatisfied’ (1) to ‘very satisfied’ (7). The number of co-owners who are actively in- volved in the management of the business and the number of family members who work in the business were each assessed by a single item. 2.5.4. Cross-Cultural Validation To ascertain whether the cultural orientation scales hold cross-cultural validity, we tested configural, metric, scalar, factor variance, and error variance invariance. We specified a model of configural invariance in which we restricted the configuration of the cultural orienta- tions and their scenarios to be equal across the Chinese and the German samples. The model of configural invariance comprised the scenarios that appropriately measured the cultural orientations. We included all those scenarios in the model that had high factor loadings and low modification indices. We assigned scales and origins to the cultural orientations by set- ting the factor loading of one scenario per cultural orientation to one and fixing its intercept to zero. Starting from the model of configural invariance, we specified nested models of metric, scalar, factor variance, and error variance invariance. In the nested models, we successively constrained the factor loadings and intercepts of the scenarios, the variances of the cultural orientations, as well as the error variances of the scenarios to be equal across the Chinese and the German samples. We estimated the models by conducting multi-group confirmatory factor analyses. We used LISREL 8 (Jöreskog & Sörbom, 1993) and the maximum likelihood esti- mation method on the basis of variance-covariance matrices and mean vectors. To evaluate model fit, we relied on the chi-square test (Jöreskog, 1971) along with the root mean square error of approximation (RMSEA, Browne & Cudeck, 1993) and the comparative fit index (CFI, Bentler, 1990). We interpreted RMSEA values close to .060 and CFI values close to .95 Chapter 2 Cultural Orientation Scales 18 as indicators of good model fit (Hu & Bentler, 1999). To compare two nested models, we relied on the chi-square difference test (Bollen, 1989). Given a non-significant increase in chi- square between the less and the more constrained model, full invariance was supported. Given a significant increase in chi-square between the less and the more constrained model, we investigated whether partial invariance was supported. We examined modification indices and relaxed the equality constraint for those parameters that were unequal across the Chinese and the German samples. We assessed the scales’ composite reliabilities in the Chinese and the German samples. Moreover, we assessed the scales’ test-retest reliabilities in the German sub- sample. 2.5.5. Construct Validation To ascertain whether the cultural orientation scales hold construct validity, we as- sessed the relationships between the cultural orientations and their validation constructs. We specified a model of configural invariance in which we restricted the configuration of the cultural orientations and their scenarios as well as the configuration of the nine validation constructs and their items to be equal across the Chinese and the German samples. Apart from the scenarios that appropriately measured the cultural orientations, the model of configural invariance comprised the items that appropriately measured the nine validation constructs. We included all those items in the model that had high factor loadings and low modification indices. We assigned scales and origins to the nine validation constructs by setting the factor loading of one item per validation construct to one and fixing its intercept to zero. The model of configural invariance provided the relationships between the cultural orientations and their validation constructs. To ascertain whether the relationships could be meaningfully compared across the Chinese and the German samples, we tested not only configural invariance but also metric and factor variance invariance. Starting from the model of configural invariance, we specified nested models of metric and factor variance invariance. In the nested models, we successively constrained the factor loadings of the scenarios and the items, as well as the variances of the cultural orientations and the nine validation constructs to be equal across the Chinese and the German samples. We estimated the models by conducting multi-group con- firmatory factor analyses. To determine the significance of the difference between two corre- lation coefficients, we used the Fisher r-to-z transformation. We assessed the composite reli- abilities of the scales measuring the validation constructs in the Chinese and the German samples. Chapter 2 Cultural Orientation Scales 19 2.6. Results 2.6.1. Cross-Cultural Validity The results obtained in the tests of configural, metric, scalar, factor variance, and error variance invariance are presented in Table 2-1 on page 26. The model of configural invariance (Model A) comprised 23 scenarios that appropriately measured seven cultural orientations. In- group collectivism was assessed by two scenarios. 4 Uncertainty avoidance, assertiveness, future orientation, and performance orientation were each measured by three scenarios. Hu- mane orientation and power distance were assessed by four and five scenarios, respectively. The model of configural invariance provided adequate fit (χ 2 (418) = 603.45; RMSEA = .044; CFI = .94). Hence, configural invariance was supported. Given configural invariance, the seven cultural orientations can be meaningfully compared across the Chinese and the German samples. In the model of full metric invariance (Model B), the factor loadings of the scenarios were constrained to be equal across the Chinese and the German samples. They are presented in Table 2-2 on page 27. The increase in chi-square between the model of configural invari- ance (Model A) and the model of full metric invariance (Model B) was not significant (∆χ 2 (16) = 24.82, n.s.), and the latter model achieved adequate fit (χ 2 (434) = 628.26; RMSEA = .044; CFI = .93). Hence, full metric invariance was supported for each of the seven cultural orientations. In the model of full scalar invariance (Model C), the intercepts of the scenarios were restricted to be equal across the Chinese and the German samples. They are presented in Table 2-2 on page 27. The increase in chi-square between the model of full metric invariance (Model B) and the model of full scalar invariance (Model C) was highly significant (∆χ 2 (16) = 74.61, p < .01). Full scalar invariance was thus not supported for each of the seven cultural orientations. Examination of the modification indices revealed that the significant increase in chi-square was due to unequal intercepts of two scenarios measuring power distance, one scenario measuring assertiveness, and one scenario measuring humane orientation. The inter- cepts of the three scenarios measuring power distance and humane orientation were higher in 4 Due to low factor loadings and high modification indices, the scenarios created to assess institutional collectivism turned out to be inappropriate. Therefore, they were not comprised in the model of configural invariance. Chapter 2 Cultural Orientation Scales 20 the Chinese sample, which means that, regarding these three scenarios, the Chinese owners ticked higher scale points. The intercept of the scenario measuring assertiveness was higher in the German sample, which means that, regarding this scenario, the German owners ticked higher scale points. In the model of partial scalar invariance (Model D), we relaxed the equal- ity restriction for the unequal intercepts. The increase in chi-square between the model of full metric invariance (Model B) and the model of partial scalar invariance (Model D) was not significant (∆χ 2 (12) = 13.66, n.s.), and the latter model achieved adequate fit (χ 2 (446) = 641.92; RMSEA = .044; CFI = .93). Hence, full scalar invariance was supported for uncer- tainty avoidance, in-group collectivism, future orientation, performance orientation, whereas partial scalar invariance was supported for power distance, assertiveness, and humane orienta- tion. Given full metric and partial scalar invariance, the observed and latent means of the seven cultural orientations can be meaningfully compared across the Chinese and the German samples. In the model of full factor variance invariance (Model E), the variances of the seven cultural orientations were constrained to be equal across the Chinese and the German samples. They are presented in Table 2-2 on page 27. The increase in chi-square between the model of partial scalar invariance (Model D) and the model of full factor variance invariance (Model E) was highly significant (∆χ 2 (7) = 26.30, p < .01). Full factor variance invariance was thus not supported. Examination of the modification indices revealed that the significant increase in chi-square was due to unequal variances of uncertainty avoidance and assertiveness. They were higher in the Chinese than in the German sample, which means that, regarding these two cultural orientations, the Chinese owners were more heterogeneous than the German owners. In the model of partial factor variance invariance (Model F), we relaxed the equality con- straint for the unequal variances. The increase in chi-square between the model of partial scalar invariance (Model D) and the model of partial factor variance invariance (Model F) was not significant (∆χ 2 (6) = 4.79, n.s.), and the latter model provided adequate fit (χ 2 (451) = 646.71; RMSEA = .044; CFI = .93). Because most, but not all, variances were equal across the Chinese and the German samples, partial factor variance invariance was supported. Given full metric and partial factor variance invariance, the relationships between power distance, in-group collectivism, future orientation, humane orientation, and performance orientation can be meaningfully compared across the Chinese and the German samples. The relationships involving uncertainty avoidance and assertiveness should be compared with caution. Chapter 2 Cultural Orientation Scales 21 In the model of full error variance invariance (Model G), the error variances of the scenarios were restricted to be equal across the Chinese and the German samples. They are presented in Table 2-2 on page 27. The increase in chi-square between the model of partial factor variance invariance (Model F) and the model of full error variance invariance (Model G) was highly significant (∆χ 2 (23) = 263.16, p < .01). Full error variance invariance was thus not supported for each of the seven cultural orientations. Examination of the modification indices revealed that the significant increase in chi-square was due to unequal error variances of eleven scenarios measuring uncertainty avoidance, power distance, assertiveness, future orientation, humane orientation, and performance orientation. They were higher in the Chi- nese than in the German sample, which means that, regarding these eleven scenarios, the Chinese owners produced larger variances due to measurement error than the German owners. In the model of partial error variance invariance (Model H), we relaxed the equality restriction for the unequal error variances. The increase in chi-square between the model of partial factor variance invariance (Model F) and the model of partial error variance invariance (Model H) was not significant (∆χ 2 (12) = 18.14, n.s.), and the latter model achieved adequate fit (χ 2 (463) = 664.85; RMSEA = .044; CFI = .93). Hence, full error variance invariance was supported for in-group collectivism, whereas partial error variance was supported for uncertainty avoidance, power distance, assertiveness, future orientation, humane orientation, and performance orien- tation. The cultural orientation scales are presented in the Appendix. The scales’ composite reliabilities in the Chinese and the German samples are presented in Table 2-3 on page 28. The scale measuring uncertainty avoidance showed higher composite reliability in the Chi- nese sample, whereas the scales measuring power distance, assertiveness, future orientation, humane orientation, and performance orientation displayed higher composite reliabilities in the German sample. Only the scale measuring in-group collectivism showed equal composite reliability across the Chinese and the German samples. This is due to the fact that in-group collectivism is the only cultural orientation for which both full metric and full error variance invariance are given and that, at the same time, has equal factor variances across the Chinese and the German samples. The scales’ test-retest reliabilities in the German sub-sample are also presented in Table 2-3 on page 28. Each of the scales displayed higher test-retest reliabil- ity than composite reliability. This supports the assumption that test-retest reliability is a more appropriate reliability estimate for scenario-based scales than composite reliability (Chan & Schmitt, 1997; Motowidlo et al., 1990). Chapter 2 Cultural Orientation Scales 22 2.6.2. Construct Validity The relationships between the seven cultural orientations and their validation con- structs are presented in Table 2-4 on page 29. The relationships were provided by the model of configural invariance. The results obtained in the tests of configural, metric, and factor variance invariance are presented in Table 2-5 on page 30. Apart from the 23 scenarios that appropriately measured the seven cultural orientations, the model of configural invariance comprised 19 items that appropriately measured the nine validation constructs. Social satis- faction, the number of co-owners who are actively involved in the management of the busi- ness, and the number of family members who work in the business were each assessed by one item. Achievement striving, deliberation, and error communication were each measured by two items. Meta-cognitive activity and task-oriented personal initiative were each assessed by three items. Relationship-oriented personal initiative was measured by four items. The models of configural, full metric, and partial factor variance invariance achieved adequate fits. Hence, configural, full metric, and partial factor variance invariance were supported. Given con- figural invariance, the seven cultural orientations and the nine validation constructs can be meaningfully compared across the Chinese and the German samples. Given full metric and partial factor variance invariance, the relationships between five of the seven cultural orienta- tions and seven of the nine validation constructs can be meaningfully compared across the Chinese and the German samples. The relationships involving uncertainty avoidance and assertiveness as well as the number of co-owners who are actively involved in the manage- ment of the business and the number of family members who work in the business should be compared with caution. The composite reliabilities of the scales measuring the validation constructs are presented in Table 2-3 on page 28. Uncertainty avoidance. The correlations between uncertainty avoidance and its two validation constructs were consistent with our hypotheses. Uncertainty avoidance was nega- tively correlated to task-oriented personal initiative (r = -.22, p < .05 / r = -.22, p < .01) 5 and to relationship-oriented personal initiative (r = -.45, p < .01 / r = -.14, p > .05). The more Chinese and German owners supported uncertainty avoidance, the less they showed task- oriented and relationship-oriented personal initiative. The negative correlation between uncer- tainty avoidance and relationship-oriented personal initiative was significantly higher in the 5 Whenever two correlation coefficients are given, the first correlation coefficient refers to the Chinese sample, whereas the second correlation coefficient refers to the German sample. Chapter 2 Cultural Orientation Scales 23 Chinese sample (z = -3.64, p < .01). In China, business relationships may be more delicate to handle and, therefore, may require more willingness to take risks than in Germany. This could explain why Chinese owners who supported uncertainty avoidance showed even less relation- ship-oriented personal initiative than their German counterparts. There was one non- hypothesized correlation that was as high as the hypothesized correlations. Uncertainty avoid- ance was negatively correlated to achievement striving (r = -.28, p < .05 / r = -.17, p > .05). The more Chinese and German owners supported reliance on social norms, rules, and proce- dures to prevent incertitude, the less they worked hard to achieve their goals. Post hoc, the negative correlation could be explained as follows: Achievement striving may require will- ingness to take risks. Owners who support uncertainty avoidance are unwilling to take risks. This renders them unlikely to show achievement striving. Power distance. In accordance with our hypothesis, power distance was negatively correlated to error communication (r = -.19, p < .01; r = -.50, p < .01). The more Chinese and German owners promoted power distance, the less they communicated their errors. The nega- tive correlation was significantly higher in the German sample (z = 3.78, p < .01). In Ger- many, acknowledgement of fallibility may be regarded as less compatible with leadership than in China. This could explain why German owners who promoted power distance com- municated their errors even less than their Chinese counterparts. Collectivism. The correlation between in-group collectivism and its validation con- struct was only partially in line with our hypothesis. 6 In the Chinese sample, in-group collec- tivism was positively correlated to the number of family members who work in the business (r = .19, p < .05). The more Chinese owners fostered in-group collectivism, the more they employed their family members. In the German sample, however, in-group collectivism was not correlated to the number of family members who work in the business (r = .00, p > .05). Employing one’s family members may be regarded as collectivism in China but disregarded as nepotism in Germany. This could explain why the hypothesized correlation existed in the Chinese but not in the German sample (z = 2.03, p < .05). There were several non- hypothesized correlations that were higher than the hypothesized correlation. No post hoc explanations could be provided for them. 6 As the model of configural invariance comprised no scenarios created to assess institutional collectiv- ism, we could not test our hypothesis on the positive correlation between institutional collectivism and the number of co-owners who are actively involved in the management of the business. Chapter 2 Cultural Orientation Scales 24 Assertiveness. Consistent with our hypotheses, assertiveness was negatively correlated to deliberation (r = -.30, p < .01 / r = -.18, p < .05) and to meta-cognitive activity (r = -.26, p < .01 / r = -.16, p < .05). The more Chinese and German owners supported assertiveness, the less they showed deliberation and meta-cognitive activity. Future orientation. The correlations between future orientation and its three validation constructs were in line with our hypotheses. Future orientation was positively correlated to achievement striving (r = .31, p < .01 / r = .13, p > .05), to deliberation (r = .40, p < .01 / r = .12, p > .05), and to meta-cognitive activity (r = .25, p < .01 / r = .16, p < .05). The more Chinese and German owners fostered future orientation, the more they showed achievement striving, deliberation, and meta-cognitive activity. The correlation between future orientation and achievement striving (z = 2.01, p < .05) and the correlation between future orientation and deliberation (z = 3.21, p < .01) were significantly higher in the Chinese sample. In China, hard work and careful consideration may be regarded as more essential to implementing long-term projects than in Germany. This could explain why Chinese owners who fostered future orien- tation showed even more achievement striving and deliberation than their German counter- parts. Humane orientation. In accordance with our hypotheses, humane orientation was posi- tively correlated to error communication (r = .22, p < .01 / r = .25, p < .01) and to social satisfaction (r = .19, p < .01 / r = .14, p > .05). The more Chinese and German owners pro- moted humane orientation, the more they communicated their errors, and the more they were socially satisfied. There were two non-hypothesized correlations that were as high as the hypothesized correlations. First, humane orientation was positively correlated to achievement striving (r = .36, p < .01 / r = .14, p > .05). The more Chinese and German owners promoted fairness, altruism, generosity, care, and kindness, the more they worked hard to achieve their goals. Post hoc, the correlation could be explained as follows: Achievement striving may require motivating others to help achieve one’s goals. Owners may promote humane orienta- tion because their employees may be more motivated to help achieve their goals when they are treated in a humane-oriented way. The correlation was significantly higher in the Chinese sample (z = 2.50, p < .05). In China, employees’ motivation to help achieve owners’ goals may be lower than in Germany. This could explain why Chinese owners who showed achievement striving promoted even more humane orientation than their German counter- parts. Second, humane orientation was positively correlated to deliberation (r = .40, p < .01 / r = .16, p > .05). The more Chinese and German owners promoted fairness, altruism, generos- Chapter 2 Cultural Orientation Scales 25 ity, care, and kindness, the more they carefully considered their decisions. Post hoc, the corre- lation could be explained as follows: Deliberation may reflect responsibility towards others who are affected by one’s decisions. Owners who promote humane orientation act responsibly towards their employees. This renders them likely to show deliberation. The correlation was significantly higher in the Chinese sample (z = 2.77, p < .01). In China, responsibility towards employees may be more pronounced than in Germany. This could explain why Chinese own- ers who promoted humane orientation showed even more deliberation than their German counterparts. Performance orientation. The correlations between performance orientation and its four validation constructs were consistent with our hypotheses. Performance orientation was positively correlated to achievement striving (r = .42, p < .01 / r = .52, p < .01), to meta- cognitive activity (r = .35, p < .01 / r = .19, p < .05), to task-oriented personal initiative (r = .39, p < .01 / r = .21, p < .05), and to relationship-oriented personal initiative (r = .28, p < .01 / r = .34, p < .01). The more Chinese and German owners supported performance orientation, the more they showed achievement striving, meta-cognitive activity, and task-oriented and relationship-oriented personal initiative. The correlation between performance orientation and task-oriented personal initiative was significantly higher in the Chinese sample (z = 2.10, p < .05). In China, seizing opportunities and preparing for challenges may be regarded as more essential to reaching excellence than in Germany. This could explain why Chinese owners who supported performance orientation showed even more task-oriented personal initiative than their German counterparts. (To be continued on page 31.) Chapter 2 Cultural Orientation Scales 26 Table 2-1 Tests of Configural, Metric, Scalar, Factor Variance, and Error Variance Invariance Models Comparisons χ 2 (df) ∆χ 2 (∆df) RMSEA CFI A Configural invariance - 603.45 (418) ** - .044 .94 B Full metric invariance A versus B 628.26 (434) ** 124.82 (16) n.s. .044 .93 C Full scalar invariance B versus C 702.87 (450) ** 174.61 (16) ** .050 .91 D Partial scalar invariance B versus D 641.92 (446) ** 113.66 (12) n.s. .044 .93 E Full factor variance invariance D versus E 668.22 (453) ** 126.30 1(7) ** .046 .93 F Partial factor variance invariance D versus F 646.71 (451) ** 114.79 1(6) n.s. .044 .93 G Full error variance invariance F versus G 889.87 (474) ** 263.16 (23) ** .062 .86 H Partial error variance invariance F versus H 664.85 (463) ** 118.14 (12) n.s. .044 .93 Note. * = p < .05, ** = p < .01, n.s. = not significant. RMSEA = root mean square error of approximation, CFI = comparative fit index. Chapter 2 Cultural Orientation Scales 27 Table 2-2 Variances of the Cultural Orientations, and Factor Loadings, Intercepts, and Error Variances of the Scenarios Cultural orientations (variances) Scenarios Unstandardized factor loadings Standardized factor loadings Intercepts Error variances UA1 1.00 0.50 -0.00 1.29 / 0.44 UA2 1.22 0.54 -0.23 1.10 Uncertainty avoidance (0.40 / 0.19) UA6 1.20 0.60 -0.38 0.80 PD1 0.87 0.50 1.05 / 0.64 1.71 / 1.09 PD2 0.91 0.58 0.75 / 0.22 1.28 / 0.70 PD3 1.00 0.77 -0.00 0.42 PD4 0.95 0.65 -0.37 0.80 Power distance (0.62) PD6 0.85 0.52 -0.69 1.23 C5 1.00 0.52 -0.00 1.16 In-group collectivism (0.42) C7 0.87 0.53 -0.22 0.82 A3 1.09 0.75 -0.71 / -0.39 0.28 A5 1.00 0.54 -0.00 0.93 / 0.46 Assertiveness (0.38 / 0.21) A6 1.35 0.70 -0.91 0.86 / 0.16 FO2 1.00 0.47 -0.00 1.02 FO4 1.15 0.52 -0.81 1.29 / 0.69 Future orientation (0.28) FO6 1.30 0.69 -1.26 0.56 HO1 1.00 0.62 -0.00 0.68 HO3 0.74 0.56 1.40 0.59 / 0.39 HO4 0.88 0.59 0.66 / 0.34 0.62 Humane orientation (0.43) HO6 0.96 0.59 -0.24 0.90 / 0.54 PO2 1.00 0.45 -0.00 2.09 / 0.72 PO3 0.82 0.54 -1.52 0.75 / 0.45 Performance orientation (0.37) PO5 0.77 0.32 -0.09 2.54 / 1.33 Note. In cases of unequal variances, factor loadings, intercepts, and error variances across the Chinese and the German samples, two values are given. The first value refers to the Chinese sample, whereas the second value refers to the German sample. Chapter 2 Cultural Orientation Scales 28 Table 2-3 Reliabilities of the Scales Measuring the Cultural Orientations and the Validation Constructs Composite reliabilities Test-retest reliabilities Cultural orientations / validation constructs China Germany Germany Uncertainty avoidance .60 .46 .74 Power distance .73 .77 .78 In-group collectivism .43 .43 .78 Assertiveness .70 .74 .76 Future orientation .56 .61 .74 Humane orientation .66 .71 .73 Performance orientation .35 .53 .75 Achievement striving .70 .66 - Deliberation .55 .56 - Error communication .69 .75 - Meta-cognitive activity .75 .75 - Task-oriented personal initiative .75 .69 - Relationship-oriented personal initiative .81 .77 - Note. Composite reliability is defined as the quotient between the added squared standardized factor loadings and the sum of the added squared standardized factor loadings and the added error variances (Fornell & Larcker, 1981). Chapter 2 Cultural Orientation Scales 29 Table 2-4 Relationships Between the Latent Cultural Orientations and the Latent Validation Constructs 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 01 Uncertainty avoidance .34** -.04 .32* -.20 -.33** -.01 -.28* -.02 .06 -.09 -.22* -.45** -.09 .11 -.13 02 Power distance .54** .64** .62** -.50** -.53** -.39** -.23** -.36** -.19** -.10 -.07 -.11 -.16* .01 .02 03 In-group collectivism .24* .13 .70** -.50** -.63** -.60** -.32** -.30* -.29** -.31** -.17 -.07 -.14 -.02 .19* 04 Assertiveness .18* .11 .21* -.28* -.27** -.31* -.24* -.30** -.01 -26** -.17 -.14 .02 .02 .02 05 Future orientation -.48** -.41** -.40** -.23** .72** .58** .31** .40** .14 .25** .15 .24* .18* .02 .11 06 Humane orientation -.38** -.53** .06 -.40** .39** .62** .36** .40** .22** .39** .28** .15 .19** -.10 -.02 07 Performance orientation -.26* .09 -.57** -.21* .33** .07 .42** .42** .14 .35** .39** .28* .26* .23 .08 08 Achievement striving -.17 -.06 -.21 -.07 .13 .13 .52** .74** .27** .72** .66** .51** .23** .21* .00 09 Deliberation -.08 -.09 -.01 -.18* .12 .16 .04 .36** .14 .60** .67** .45** .26** .14 -.02 10 Error communication -.17* -.50** -.05 -.11 .08 .25** -.25** .07 .06 .24** .18* .31** .23** .01 .02 11 Meta-cognitive activity -.05 .04 -.20 -.16* .16* .11 .19* .33** .31** -.02 .77** .51** .15* .14 -.08 12 Task-oriented PI -.22** -.13 -.21 -.10 .10 .09 .21* .59** .20 .13 .32** .66** .17* .21* .03 13 Relationship-oriented PI -.14 -.06 -.31* -.10 .14 .07 .35** .49** .16 .08 .33** .50** .20** .21* .06 14 Social satisfaction -.10 -.12 -.03 -.05 -.07 .14 -.11 .07 .04 .18** .07 .04 -.02 .05 .05 15 Number of co-owners .01 .01 .00 .09 -.12 -.10 -.02 -.02 -.00 -.01 -.02 .05 .04 .05 -.10 16 Number of family members -.01 -.00 .00 -.00 .01 .00 -.00 .01 -.00 .00 .01* .01 .01 .00 -.00 Note. * = p < .05, ** = p < .01. The values above the diagonal refer to the Chinese sample, whereas the values below the diagonal refer to the German sample. Chapter 2 Cultural Orientation Scales 30 Table 2-5 Tests of Configural, Metric, and Factor Variance Invariance Models Comparisons χ 2 (df) ∆χ 2 (∆df) RMSEA CFI A Configural invariance - 1922.02 (1404) ** - .040 .92 B Full metric invariance A versus B 1950.42 (1430) ** 128.41 (26) n.s. .040 .92 C Full factor variance invariance B versus C 2121.26 (1446) ** 170.84 (16) ** .045 .82 D Partial factor variance invariance B versus D 1965.43 (1442) ** 115.01 (12) n.s. .040 .92 Note. * = p < .05, ** = p < .01, n.s. = not significant. RMSEA = root mean square error of approximation, CFI = comparative fit index. Chapter 2 Cultural Orientation Scales 31 2.7. Discussion We developed and validated scenario-based scales measuring seven cultural orienta- tions of owners, namely, uncertainty avoidance, power distance, in-group collectivism, asser- tiveness, future orientation, humane orientation, and performance orientation. The seven cultural orientations are manifested in the practices owners use in their businesses. Owners’ practices provide starting points for the development of organizational cultures. The assessment of their invariance across China and Germany suggests that the scales hold cross-cultural validity. Full configural, full metric, and partial scalar invariance were supported, as were partial factor variance and partial error variance invariance. Hence, the scales enable scholars to meaningfully compare the means of the seven cultural orientations across China and Germany. Moreover, the scales enable scholars to conduct meaningful Chinese-German comparisons of the relationships involving power distance, in-group collec- tivism, future orientation, humane orientation, and performance orientation. The relationships involving uncertainty avoidance and assertiveness should be compared with caution. Cultural response bias occurs when people from different cultures differ in their re- sponse sets (Triandis, 1994). The scales measuring uncertainty avoidance, in-group collectiv- ism, future orientation, and performance orientation were not affected by cultural response bias because the Chinese and the German owners did not differ in their response sets on the scenarios assessing these cultural orientations. The scales measuring power distance, asser- tiveness, and humane orientation were marginally affected by cultural response bias because the Chinese and the German owners differed in their response sets on one or two scenarios assessing these cultural orientations. The assessment of the relationships between the seven cultural orientations and their validation constructs suggests that most of the scales hold construct validity. According to tests of invariance, the relationships between five of the seven cultural orientations and seven of the nine validation constructs can be meaningfully compared across China and Germany. The relationships involving uncertainty avoidance and assertiveness, as well as the number of co-owners who are actively involved in the management of the business and the number of family members who work in the business should be compared with caution. Both in China and Germany, uncertainty avoidance, power distance, assertiveness, future orientation, hu- mane orientation, and performance orientation demonstrated the hypothesized relationships with their validation constructs. Hence, it can be assumed that the scales measuring these Chapter 2 Cultural Orientation Scales 32 cultural orientations hold construct validity. As for uncertainty avoidance and humane orien- tation, there were three non-hypothesized relationships that were as high as the hypothesized relationships. However, as post hoc explanations could be provided for these relationships, they do not challenge the construct validity of the scales measuring uncertainty avoidance and humane orientation. We intended to develop two scales measuring the two forms of collectivism, namely, institutional collectivism and in-group collectivism. However, we could only partially imple- ment our intention. We did not succeed in developing a scale measuring institutional collec- tivism because the three scenarios created to assess this form of collectivism turned out to be inappropriate. We succeeded in developing a scale measuring in-group collectivism because two of the three scenarios created to assess this form of collectivism turned out to be appro- priate. In-group collectivism demonstrated the hypothesized relationship with its validation construct in China but not in Germany. Hence, it can be assumed that the scale measuring in- group collectivism holds construct validity in China. The validation construct may not have been appropriate for in-group collectivism in Germany. There were several non-hypothesized relationships that were higher than the hypothesized relationship. As no post hoc explanations could be provided for these relationships, they challenge the construct validity of the scale measuring in-group collectivism both in China and Germany. 7 2.7.1. Limitations and Implications for Future Research Some of the scales do not cover all facets of the cultural orientations specified in the definitions. The 40 scenarios created to assess the cultural dimensions captured all of their facets. However, in the cross-cultural validation of the scales, only 23 scenarios turned out to be appropriate and were included in the scales. Some of the scales show low internal consistencies. The scales measuring in-group collectivism and performance orientation display low composite reliabilities both in China and Germany, whereas the scale measuring future orientation and the scale measuring uncer- tainty avoidance display low composite reliabilities in China and Germany, respectively. These scales are based on few scenarios, and short scales usually suffer from low internal consistencies. However, we assume with Chan and Schmitt (1997) and Motowidlo et al. 7 Later on, we redeveloped the scales measuring institutional collectivism and in-group collectivism. Please see the Addendum on page 34 for details. Chapter 2 Cultural Orientation Scales 33 (1990) that test-retest reliability is a more appropriate reliability estimate for scenario-based scales than composite reliability. All of the scales show high test-retest reliabilities. The Chinese and the German samples comprised both founders and non-founders. We compared the intercorrelation matrices obtained in samples that included only founders to the intercorrelation matrices obtained in samples that included only non-founders. The correla- tions between the compared intercorrelation matrices were r = .90 (p < .01) in China and r = .92 (p < .01) Germany. Thus, we can rule out that our results were distorted by the fact that the Chinese and the German samples comprised both owners who had and owners who had not founded their businesses. Given that we developed and validated the scales for Chinese and German owners, their use may be limited in two respects: First, the scales are suitable for owners from China and Germany but may not be suitable for owners from other cultures. As long as it has not been ascertained whether the scales can be used to meaningfully compare owners from other cultures, comparisons should be conducted with caution. China and Germany are two quite different cultures. Therefore, we are optimistic that future studies will demonstrate the scales’ suitability for owners from other cultures. Second, the scales are suitable for owners but may not be suitable for managers. As long as it has not been ascertained whether the scales allow for meaningful comparisons of managers, they should only be used to compare owners. Own- ers and managers have quite a lot in common. Therefore, we are optimistic that future studies will demonstrate the scales’ suitability for managers. 2.8. Conclusion The cultural orientation scales are useful for cross-cultural scholars and entrepreneur- ship scholars. Cross-cultural scholars can use the scales to investigate how owners from dif- ferent cultures differ in their cultural orientations. Moreover, they can use the scales to inves- tigate cross-cultural differences in the effects of owners’ cultural orientations. Entrepreneur- ship scholars can use the scales to assess the practices owners use in their businesses. Thereby, they can assess how owners go about managing their businesses and how they sup- port the development of organizational cultures. Moreover, entrepreneurship scholars can use the scales to study the relationships between owners’ cultural orientations and entrepreneurial concepts at the individual level of analysis. The scales are also useful for owners. They can be used in training to make owners aware of the practices they use in their businesses. The Chapter 2 Cultural Orientation Scales 34 awareness of how they go about managing their businesses and how they support the devel- opment of organizational cultures may lead owners to challenge and improve their practices. 2.9. Addendum Later on, we developed a scale measuring gender egalitarianism, and we redeveloped the scales measuring institutional collectivism and in-group collectivism. Again, we based the scales on scenarios. We could not ascertain whether the scales hold cross-cultural validity because they were completed by German owners only. Also, we could not ascertain whether the scales hold construct validity because we did not consider the constructs covered in the research project to be appropriate validation constructs for the three cultural orientations. Thus, we need to validate the scales measuring gender egalitarianism, institutional collectiv- ism, and in-group collectivism in a future study. Through specifying and estimating a model, we could, at least, assess the scales’ com- posite reliabilities. The 149 German owners who completed the scales served as participants for the assessment. The model comprised the scenarios that appropriately measured the three cultural orientations. Gender egalitarianism, institutional collectivism, and in-group collectiv- ism were each assessed by three scenarios. The model provided good fit (χ 2 (24) = 26.18; RMSEA = .022; CFI = 1.00). The scales’ composite reliabilities ranged from .79 for institu- tional collectivism over .84 for in-group collectivism to .95 for gender egalitarianism. The scales measuring the three cultural orientations are presented in the Appendix. Chapter 3 Cultural Orientations as Moderators 35 3. Business Owners’ Cultural Orientations as Moderators of the Rela- tionships Between Vision Characteristics and Business Success Visions represent images of desirable futures that provide meaning and direction (House & Shamir, 1993). They can be described by vision characteristics (Locke et al., 1991). One focus in entrepreneurship research has been on the effectiveness of vision characteristics, that is, on the relationships between vision characteristics and business success (Baum et al., 1998). Business success comprises financial dimensions, such as sales growth or growth in the number of employees, and operational dimensions, such as product and service quality or customer satisfaction (Combs, Crook, & Shook, 2005; Venkatraman & Ramanujam, 1986). Vision characteristics may be more effective, that is, more strongly related to business suc- cess, for some business owners than for others. 1 However, moderators of the relationships between vision characteristics and business success have not yet been identified in entrepre- neurship research. We assume that owners’ cultural orientations moderate these relationships, and we further assume that there are cross-cultural differences in the moderator effects of owners’ cultural orientations. 3.1. The Moderator Effects of Owners’ Cultural Orientations We assume that owners’ cultural orientations moderate the relationships between vi- sion characteristics and business success. Owners’ cultural orientations are manifested in the practices and values owners use in their businesses (König, Steinmetz, Frese, Rauch, & Wang, 2007). We consider practices to be more important moderators than values because practices are related to actions (Frese, 2006). Owners’ practices substantially influence the development of organizational cultures (Schein, 2004), which, in turn, shape employees’ practices (Aycan, Kanungo, & Sinha, 1999). Organizational cultures develop as a result of the interactions between owners and employees (Schein, 2004). Due to the substantial influence of owners’ practices on these interactions, organizational cultures shape employees’ practices such that employees’ practices conform to owners’ practices (Schein, 2004). Thus, owners’ cultural orientations, which are manifested in owners’ practices, are reflected in employees’ practices. 1 Business owners are defined as individuals who own and manage their businesses (Carland et al., 1984). For simplification, they are referred to as ‘owners’ in the following. Chapter 3 Cultural Orientations as Moderators 36 The assumption that owners’ cultural orientations moderate the relationships between vision characteristics and business success implies that the effectiveness of vision characteris- tics depends on whether they match owners’ cultural orientations ('match hypothesis', Tung et al., 2006): When there is a match between vision characteristics and owners’ cultural orienta- tions, vision characteristics are effective because both owners’ practices (in which owners’ cultural orientations are manifested) and employees’ practices (in which owners’ cultural orientations are reflected) are conducive to their effectiveness. In contrast, when there is no match between vision characteristics and owners’ cultural orientations, vision characteristics are not effective because neither owners’ practices nor employees’ practices are conducive to their effectiveness. Based on the assumption that owners’ cultural orientations moderate the relationships between vision characteristics and business success, we developed hypotheses regarding the moderator effects of six cultural orientations. The six cultural orientations refer to cultural dimensions introduced by the Global Leadership and Organizational Behavior Effectiveness (GLOBE) Study (House & Javidan, 2004). We adapted the definitions given by Javidan et al. (2004, p. 30) to the practices owners use in their businesses: Performance orientation implies that owners support striving for “performance improvement and excellence.” Humane orien- tation means that owners promote fairness, altruism, generosity, care, and kindness. Future orientation signifies that owners foster “delaying gratification, planning, and investing in the future.” Assertiveness implies that owners support confrontation and aggressiveness. Power distance means that owners promote acceptance of power being distributed unequally. Uncer- tainty avoidance signifies that owners foster reliance on “social norms, rules, and procedures” to prevent incertitude. To each of the six cultural orientations, we assigned one vision charac- teristic, namely, the one that matches the cultural orientation best. The six vision characteris- tics have been assumed to be related to business success (Baum et al., 1998; Locke et al., 1991). We hypothesize that each of the six cultural orientations moderates the relationship between the vision characteristic that matches it best and business success. Performance orientation. The vision characteristic ‘challenge’ implies that visions make great demands on owners and employees (Locke et al., 1991; Nanus, 1992). Therefore, its effectiveness depends on whether owners and employees work hard. We suppose that the vision characteristic ‘challenge’ matches high performance orientation. The more owners support performance orientation in their businesses, the more effective the vision characteris- tic ‘challenge’ is, because the more organizational cultures emphasize striving for “perform- Chapter 3 Cultural Orientations as Moderators 37 ance improvement and excellence” (Javidan et al., 2004, p. 30), the more likely owners and employees are to work hard (Javidan, 2004). In contrast, we suppose that the vision character- istic ‘challenge’ does not match low performance orientation. The less owners support per- formance orientation in their businesses, the less effective the vision characteristic ‘challenge’ is, because the less organizational cultures emphasize striving for “performance improvement and excellence” (Javidan et al., 2004, p. 30), the less likely owners and employees are to work hard (Javidan, 2004). Hypothesis 1: Owners’ performance orientation moderates the relationship between the vision characteristic ‘challenge’ and business success. The more owners support performance orientation in their businesses, the stronger the relationship is. Humane orientation. The vision characteristic ‘social responsibility’ means that vi- sions are concerned with the well-being of others (Conger, 1989; Strange & Mumford, 2002). Thus, its effectiveness depends on whether owners and employees stand up for others. We hypothesize that there is a match between the vision characteristic ‘social responsibility’ and high humane orientation. The more owners promote humane orientation in their businesses, the more effective the vision characteristic ‘social responsibility’ is, because the more empha- sis organizational cultures put on fairness, altruism, generosity, care, and kindness (Javidan et al., 2004), the more likely owners and employees are to stand up for others (Kabasakal & Bodur, 2004). In contrast, we hypothesize that there is no match between the vision character- istic ‘social responsibility’ and low humane orientation. The less owners promote humane orientation in their businesses, the less effective the vision characteristic ‘social responsibil- ity’ is, because the less emphasis organizational cultures put on fairness, altruism, generosity, care, and kindness (Javidan et al., 2004), the less likely owners and employees are to stand up for others (Kabasakal & Bodur, 2004). Hypothesis 2: Owners’ humane orientation moderates the relationship between the vi- sion characteristic ‘social responsibility’ and business success. The more owners pro- mote humane orientation in their businesses, the stronger the relationship is. Future orientation. The vision characteristic ‘future orientation’ signifies that visions refer to long-term perspectives (Locke et al., 1991). Therefore, its effectiveness depends on whether owners and employees plan ahead. We suppose that the vision characteristic ‘future orientation’ matches high future orientation. The more owners foster future orientation in their businesses, the more effective the vision characteristic ‘future orientation’ is, because the Chapter 3 Cultural Orientations as Moderators 38 more organizational cultures emphasize “delaying gratification, planning, and investing in the future” (Javidan et al., 2004, p. 30), the more likely owners and employees are to plan ahead (Ashkanasy et al., 2004). In contrast, we suppose that the vision characteristic ‘future orienta- tion’ does not match low future orientation. The less owners foster future orientation in their businesses, the less effective the vision characteristic ‘future orientation’ is, because the less organizational cultures emphasize “delaying gratification, planning, and investing in the future” (Javidan et al.,