Capitalizing on cultural diversity : the impact of cultural intelligence on performance through social integration and cultural distance

(1)

1

Capitalizing on cultural diversity: the impact of

cultural intelligence on performance through social

integration and cultural distance

Author: Lodewijk Christiaan Maurits Henneveld

Student ID: 10278257

Date of submission: 23-06-2017 – final version

MSc Business Administration – International Management track

University of Amsterdam

(2)

2

Statement of Originality

This document is written by University of Amsterdam student Lodewijk Henneveld who

declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources

other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion

of the work, not for the contents.

Acknowledgements

I would like to express my gratitude to Mrs Mihalache who has given honest and valuable

feedback. She remained understanding and enthusiastic throughout my internship at firm A

where other teachers would advise students to quit.

I also wish to thank Firm A and especially my mentor, Helen, who has helped me to grow

personally and professionally. All their help has proven invaluable to completing my thesis.

Furthermore, I am grateful to my good friend Koen de Brauw who took the time to proofread

my thesis and provided valuable insights to improve its legibility and quality.

Finally, I want to dedicate my thesis to my late father Peter Henneveld, who passed away

during my Masters. Thank you for showing me what‘s important, how to live for others and

teaching me to be how to be myself. I‘ll work hard every day to pass on what you‘ve given me.

Thank you for being my dad.

(3)

3

Table of figures

Figure 1: Cultural diversity effects ... 13

Figure 2: Research gap ... 18

Figure 3: Conceptual framework ... 25

Figure 4: Actual relationships ... 42

Table of figures

Table 1: Biculturalism ... 27

Table 2: Correlation ... 36

(4)

4

1 Abstract:

Studies on cultural diversity have found both positive and negative effects for companies. Cultural intelligence is presented as the mechanism to leverage the positive effect of differentiated perspectives, experience and information and mitigate the negative effect of lower social integration. The literature on the actual effect of cultural intelligence on performance was found to contradict itself, use different mechanisms to explain the effect, use different definitions of performance and be open to new

conceptualizations. In addition, no study to date has investigated the effect between cultural intelligence, social integration and cultural distance. This thesis hypothesizes that cultural intelligence directly improves both task- and contextual performance. In addition, cultural intelligence also proposed to indirectly improve both performance indicators using social integration as a mediator. Finally, the effects of cultural intelligence are theorized to be negatively moderated by cultural distance.

Using a sample of 90 foreign employees working in a Dutch professional service firm, this thesis found a positive, direct and significant relationship between cultural intelligence and task- and contextual performance. The proposed mediator and moderator relationships of social integration and cultural distance were found to be non-significant. However, there was a positive and significant effect of social integration on both performance measures. Also, cultural distance has a negative and significant effect task- and contextual performance. This thesis illustrates the importance of cultural intelligence as an enabler of the performance enhancing effects of cultural diversity. In addition, it links the negative effects of cultural diversity to cultural distance. Practitioners should embrace cultural intelligence training as a core business practice for dealing with cultural diversity and be aware of the importance of social integration for culturally distant employees.

(5)

5

2 Introduction:

Managing cultural diversity is a crucial issue for the increasingly globalized business world of the future. Renowned professor on globalization Pankaj Ghemawat writes in an article for McKinsey: ―According to one survey of senior executives, 76 percent believe their organizations need to develop

global-leadership capabilities, but only 7 percent think they are currently doing so very effective.‖

(Ghemawat, 2012: 1). Microsoft CEO, Satya Nadella is also convinced of its importance: ―Diversity and

inclusion is huge. It obviously starts by having a workforce that is diverse but inclusion is a cultural piece that we all have to work on everyday‖ (Press Trust of India, 2017). Dealing with diversity of any kind is a

problem of all ages but as emerging markets are becoming the most attractive (Ghemawat & Altman, 2014), globalization allows for the acquisition of foreign talent and where high innovation pressures demands new perspectives (Lewin, Massini, & Peeters, 2009) successfully managing culturally diverse teams is more crucial than ever (Finaccord, 2014). Diversity comes in many forms (age, gender, ethnicity, sexual preferences, religion or lifestyle choice) however none are so pervasive, comprehensive and intangible as culture. Culture is defined in this paper as ―the shared motives, values, beliefs, identities, and

interpretations or meanings of significant events that result from common experiences of members of collectives that are transmitted across generations‖ (House et al, 2002: 494-495). It is often referred to as

a lens through which we make sense of all the stimuli, both internal and external, that we receive (Chiao, Harada, Komeda, Li, Mano, Saito, Iidaka, 2010). Moreover, it often decides which stimuli we ignore thereby shaping our world before consciously becoming aware of it (Zhu, Zhang, Fan & Han, 2007; Goh & Park, 2009). As such, cultural differences can be very profound but remain unrecognized.

Research on cultural diversity in teams shows that there are both benefits and costs (Horwitz & Horwitz, 2007). Because team members have more diverse backgrounds they also have more diverse perspectives and information which increases creativity (Doz, Santos, & Williamson, 2004). In addition, more diversity makes the team interaction more stimulating and challenging and therefore more satisfying

(6)

6 (Stahl et al, 2010). However, the same study also finds that more diversity resulted in less social integration and more conflict. More cultural diversity can lead to conflict due to opposing views, different assumptions and communication barriers. In short, the challenge of diversity is to maximize differences for better outputs and minimize the issues that arise from such diversity. Recently, cultural intelligence was introduced to explain why some individuals are better at leveraging culturally adapting than others. Earley & Ang (2003: 9) originally coined Cultural Intelligence as ―a person’s capability for successful

adaptation to new cultural settings, that is, for unfamiliar settings attributable to cultural context‖.

Cultural Intelligence is abbreviated as CQ to emphasize it‘s conceptualization as a facet of intelligence. High CQ individuals quickly identify, classify and interpret unfamiliar behaviours and uncover their meaning. They then adapt their reasoning and behaviour to convey the right meaning, to the right person at the right time. Low CQ individuals do not (effectively) identify important behaviours or misinterpret their meaning. Because of that, they default to their own cultural lens and misinterpret behaviour or fail in applying the foreign cultural lens to assess the meaning of behaviour correctly. Depending on the extent of misalignment the interpretation of the other person can vary from mild confusion to aggressiveness which lowers trust and/or impairs communication. Theoretically, high CQ individuals should therefore be able to minimize the risks and maximize the benefits of cultural diversity resulting in higher performance.

But the literature on the subject is inconsistent, incongruent, incomparable and incomplete due to the fact CQ research is relatively new (Earley & Ang, 2003). Therefore, the following gaps have been identified. First of all, results are conflicting. Some studies find that CQ is related to task performance (Lee & Sukoco, 2010; Ang et al, 2007) but others do not (Ramalu, Wei & Rose, 2011; Malek & Budhwar, 2013). Also, the original conceptualization of Earley & Ang (2003) is being challenged by Thomas et al (2015). In addition, it is difficult to compare studies as multiple performance measures are used. The most common were task performance, contextual performance, job performance, cultural effectiveness and expatriate performance (Ramalu, Wei & Rose, 2011; Ang et al, 2007; Malek & Budhwar, 2013; Jyoti &

(7)

7 Kour, 2015; Chen et al, 2010; Thomas et al, 2015; Lee et al, 2014; Peng et al, 2015; Lee & Sukoco, 2010). Also, the main drawback of culturally diverse teams is the lower social integration (Stahl et al, 2010). Curiously, CQ has never been linked to social integration while the latter has been found to increased performance by multiple scholars (Guillaume, Brodback & Riketta, 2011; Price, Gavin & Florey, 2002). Finally, Ott & Michailova (2016) call for research on the effect of cultural distance on cultural intelligence. Literature has never investigated the influence of cultural distance on CQ.

So, cultural diversity is inescapable, important and double-edged factor for multinational performance. Cultural intelligence has the potential to leverage the positive side while mitigating its negative effects but there is no literary consensus on this. Therefore this thesis investigates the following research question:

Does cultural intelligence directly enable the positive effect and indirectly mitigate the negative effect of cultural diversity and does cultural distance influence these effects?

This thesis offers multiple contributions to the literature. First, it employs a questionnaire sample of 90 culturally foreign employees in a Dutch professional service firm. A hierarchical regression is used to investigate whether cultural intelligence has a direct effect on task- and contextual performance, social integration is a mediator between cultural intelligence and task- and contextual performance and if cultural distance moderates the effect of cultural intelligence. Second, this thesis is the first empirical study to employ the new conceptualization of cultural intelligence by Thomas et al (2015) after its introduction. This includes the usage of the SFCQ questionnaire. Third, this is the first time cultural intelligence is linked to higher social integration. Fourth, the call of Ott & Michailova (2016) to investigate the effect of cultural distance on cultural intelligence is answered for the first time. There are also important managerial contributions. First, the research design looks at professional services industry employees. These

employees are the core of their business and hold all their knowledge-based resources. Improving their performance directly improves their productivity and profits. Second, expat failure rates are still high

(8)

8 (Chalré, 2015) and can be improved based on the results of this study. Also, the results can inform

recruitment personnel, managers and expatriates on how to improve on-boarding procedures to enhance performance. Third, this thesis informs MNEs on how to maximize the potential gains of cultural diversity through cultural intelligence and social integration and also informs them on the downside of hiring culturally distance employees.

The remainder of this thesis is structured as followed. First, the literature on cultural diversity, cultural intelligence and performance, social integration and cultural diversity is reviewed to provide a theoretical base and identify the research gap. Second, a theoretical framework with associated hypotheses is presented. Third, the methodology section explains why and how certain data was collected. It continues with a discussion on the statistical models used to analyse the data. The fourth section reports the results of the analysis and either supports or rejects the hypotheses. Fifth, a discussion on the findings interprets the results. This section also explains the academic contributions, managerial implications, limitations of the thesis and suggests areas for future research. Finally, the last section acts as the conclusion of this thesis.

3 Literature Review:

This section reviews the scholarly debate on cultural diversity, the role of cultural intelligence and the issues that surround it. First, culture and cultural diversity are defined, followed by a review of the potential positive and negative effects on performance. Second, the literature on cultural intelligence is reviewed and its influence on cultural diversity effects is reported. Finally, major research gaps are identified and reported.

3.1

Culture

To understand the literature on cultural diversity it is necessary to quickly review the crucial literature on culture. The next section will explore the definition, content and function of culture. As

(9)

9 mentioned earlier, culture is defined as ―Shared motives, values, beliefs, identities, and interpretations or

meanings of significant events that result from common experiences of members of collectives that are transmitted across generations‖ (House et al, 2002: 494-495). Because the GLOBE definition is more

specific this thesis will choose it over the more dated one from Hofstede (1980). Hofstede (2011) identifies six key dimensions where cultures differ: Power Distance, Individualism-Collectivism, Masculine-Feminine, Uncertainty Avoidance, Long-Term Orientation and Indulgence. Hofstede (2011) provides the following definitions of the six dimensions. Power distance is the extent to which the less powerful members of organizations and institutions (like the family) accept and expect that power is distributed unequally. Uncertainty avoidance deals with a society's (dis)comfort with ambiguity. Individualism/collectivism is the degree to which people in a society are integrated into groups. Masculinity/feminimity refers to the distribution of values between the genders. Long-term/short-term orientation has two poles. The long-term pole reflects the degree to which perseverance, thrift, ordering relationships by status, and having a sense of shame is important. The short-term pole reflects reciprocating social obligations, respect for tradition, protecting one's 'face', and personal steadiness and stability. Finally, indulgence/restraint indicates whether a society allows relatively free gratification of basic and natural human desires related to enjoying life and having fun or restrains them with strict social norms. Important to note is that Hofstede and subsequent research received multiple critiques ranging from flawed assumptions (Shenkar, 2001), illusions about its mechanisms (Shenkar, 2001) and too broad and generalized concepts (Zaheer, Schomaker & Nachum, 2011). Its academic users have been criticized for a lack of theoretical vigour when applying the model (Kirkman, Lowe & Gibson, 2006). These issues arose because culture is hard to grasp and Hofstede made this intangible concept mathematically usable. This ease of use lulled researchers into using Hofstede without critically understanding and applying the framework (Zaheer, Schomaker & Nachum, 2011).

(10)

10 The cultural dimensions illustrate the breadth of differences between cultures. The depth is characterised by tangibility (Schein, 2004). At the top level reside visible artefacts which are observable with the naked eye e.g. office design, status marks and formality of interaction. At the second level are the espoused beliefs, values, rules and behavioural norms. These are commonly held and explicitly outspoken rules, behaviours and norms such as laws, slogans or proverbs. Finally, the deep-seated, tacit and taken-for-grated assumptions are the bottom. At this level, most individual members of a culture are not aware that they hold these assumptions. This makes them very hard to identify and change. Examples include the basic notions of reality, truth, right and wrong. Superficial symbols are easily copied however assumptions are even ambiguous to their holders. The function of culture was examined by Allaire & Firsirotu (1984). They conclude that culture functions as a system to bind human groups together and enhance their survival odds. This is achieved by creating a common ‗language‘ of meanings behind stimuli and behaviours which reduces the uncertainty of the others‘ intentions. The uncertainty reduction theory posits (Berger & Calabrese, 1975) that individuals experiencing uncertainty generate cognitive stress and therefore go into fight or flight mode. Therefore, cultural similarity makes cooperation easier while cultural differences are stressful, divisive and troublesome.

3.2

The effects of cultural diversity

The next section uses the theory on culture to explore what happens when cultural diversity occurs. Cultural diversity describes the existence of two or more culturally different individuals in a social unit (Cox & Blake, 1991). The first section starts with a discussion on the positive effects of cultural diversity.

Cultural diversity is often linked to increased problem-solving skills (Cox & Blake, 1991). Because foreign individuals were raised with a different culture and environment they acquire different perspectives, experiences and information than local individuals. This increases the total amount information available to teams and increases the processing techniques relative to homogenized teams and therefore more accurate problem-solving (Pieterse, van Knippenberg & Dierendonck, 2013). Along the

(11)

11 same lines, cultural diversity is also theorized to increases creativity. The broader spectrum of information at their disposal, more diverse network, unique perspectives, and increased adaptability allows for culturally diverse teams to come up with more creative solutions (Stahl et al, 2010). In addition, Stahl et al (2010) found that cultural diversity increases workgroup satisfaction. They ascribed this to a higher motivation to work together which made the process more satisfying. Another study by Watson, Kumar & Michaelsen (1993) found that heterogonous groups outperformed homogenous groups in the long term.

However, the literature also finds negative effects of cultural diversity. Overall, diversity leads to higher conflict, lower opinions and less social integration (Stahl et al, 2010a). These effects are largely due to a combination of the similarity-attraction effect (Williams & O‘Reilly, 1998), the social identity effect and social categorization effects (Tajfel, 1982). The similarity-attraction effect states that individuals have more positive attitudes towards people with similar values, beliefs and attitudes because they are more predictable and they are less likely to disagree with them. The social identity and categorization theory explains that humans categorize people based on common traits and then extends individual-level judgements to group-level and vice versa. In addition, because group membership strongly influences individual social identity which individuals always wants to keep positive (van Knippenberg, De Dreu & Homan, 2004). As a result individuals tend to regard their own group more highly. This is called the intergroup bias effect. This creates ‗in-groups‘ and ‗out-groups‘. Combining these two effects explains why individuals prefer more homogenous groups and why group cohesion suffers with cultural diversity. The reason that culture is particularly susceptible to this is explained by the Categorization-Elaboration Model (CEM) (Knippenberg, de Drue & Homan, 2004). Out-group categorization occurs when there is a high comparative fit (how much diversity is explained by the categories), a normative fit (how much does the categorization make sense in this context) and finally, high cognitive accessibility (how much is the categorization primed or learned through experience). Cultural heritage is a highly salient categorization marker. The negative outcomes associated with these

(12)

12 categories are then caused by the intergroup bias effect. Threats to group identity causes backlash and more intergroup bias. As cultural differences are often subconscious, the root of the issue is hard to identify and resolve (Stahl et al, 2010). Empirical data often supports the negative outcomes of cultural diversity. Both Stahl et al (2010) and Staples & Zhao (2006) found that cultural diversity increased conflict. In addition, both studies also found that cultural diversity lowered social integration. Because culture promotes social integration between members of that culture, it follows that members of different cultures are less socially integrated.

While research often theorizes the positive effects outweigh the negative (Stahl et al, 2010a; van Knippenberg, De Dreu & Homan, 2004; Gómez-Meija & Palich, 1997; Watson, Kumar & Michaelsen, 1993), the literature finds no direct relationship between cultural diversity and performance (Stahl et al, 2010a; Gómez-Meija & Palich, 1997). The likely explanation is that the positive and negative effects cancel each other out (Van Knippenberg, De Dreu & Homan, 2004). The previously cites authors studied this and found that the disadvantage of social discord and conflict created by diversity mitigates the advantages of having more diverse perspectives, information and norm-deviant behaviour. An experiment done by Watson, Kumar & Michaelsen (1993) seems to confirm their suspicions. Homogenous teams were found to outperform heterogeneous teams in the short term. In the long term however, the

relationship was reversed as groups seemed get over their differences. It seems likely that leveraging one‘s cultural differences requires some cultural adaptation in order to be either accepted by the dominant culture or create a common culture of understanding. This is in line with the categorization and intergroup bias effect explained by van Knippenberg, De Dreu & Homan (2004).

Finally, the literature shows cultural distance acts as a measure for cultural diversity (Drogendijk & Slangen, 2006). It is defined as the amount of difference between two cultures. Cultural distance is predominantly calculated using the cultural dimensions as defined by Hofstede (1980a). By calculating the absolute numerical differences between the cultural dimensions a relative distance level

(13)

13 can be determined. Cultural distance strengthens the positive and negative effects of cultural diversity as it increases the depth, breadth and complexity of adapting (Hutzschenreuter, Voll & Verbeke, 2011) while also increasing the opportunities for arbitrage, learning and creativity (Zaheer, Schomaker & Nachum, 2011). Cultural distance seems to act as a multiplier for cultural diversity. For example, Newman & Nollen (1996) found that RoA increased when sales managers lowered cultural distance when adapting their

values to host country values. Vice versa, in an experimental study higher cultural distance was linked to higher process losses and higher conflict (Thomas, 1999). Figure 1 illustrates the previous discussion.

3.3

Cultural Intelligence

The next section will summarize the literature on CQ and how in interacts with figure 1. First, CQ is defined and explained. Second, the literature on the link between CQ and performance is critically analysed.

Ang & Earley (2003) invented CQ with the intention of understanding how intercultural interactions worked and defined it as ―a person’s capability for successful adaptation to new cultural

settings, that is, for unfamiliar settings attributable to cultural context‖ (Ang & Earley , 2003:9). By

definition, the concept of CQ is related to the overcoming of cultural distance and the negative effects of cultural diversity. This makes CQ a promising construct in the debate on cultural diversity management. The original concept consists of four dimensions, meta-cognitive CQ, cognitive CQ, motivational CQ and behavioural CQ which became the norm (Ang & van Dyne, 2008). Meta-Cognitive CQ refers to an

(14)

14 individual‘s level of conscious cultural awareness and executive processing during cross-cultural interactions. It allows individuals to recognize their own and others‘ cultural programming and strategize to deal with the differences. Cognitive CQ refers to an individual‘s knowledge structures about cultural institutions, norms, practices and conventions in different cultural settings. This construct deals with the factual knowledge between cultures which is used to pick up important cues and know which behaviours are appropriate. Behavioural CQ refers to an individual‘s capability to enact a wide repertoire of verbal and nonverbal actions when interacting with people from different cultures. This concerns the exhibition of appropriate verbal and non-verbal behaviours and their proper execution. Finally, motivational CQ refers to an individual‘s capability to direct attention and energy toward learning about and functioning in situations characterized by cultural differences. It allows individuals to overcome their anxiety and uncertainty and sustain the drive to learn and interact in culturally diverse situations.

However, scholars are advancing new conceptualizations in response to the original. The first model by Earley & Ang (2003) was discussed earlier and uses the four dimensions of meta-cognitive, cognitive, motivational and behavioural CQ. In this model the aggregate of all four dimensions make up CQ and all four are equal. In contrast, the second model by Thomas et al (2015) posits that CQ is made up of three dimensions: cultural metacognition, cultural knowledge and cross-cultural skills. They argue that intelligence and intelligent behaviour are two different concepts. Cross-cultural skills and Cultural knowledge are used by the higher level concept of cultural metacognition to be effective in intercultural situations. Therefore, cultural metacognition takes a higher place than knowledge and cross-cultural skills. The main two differences between van Dyne et al (2012) and Thomas et al (2008) are the aggregate versus hierarchical natures. Thomas et al (2008) take the stance that a multi-dimensional construct must specify the relationships between the dimensions and the overall construct or it loses its utility. In addition, they abandon the idea that motivational CQ is related to cultural Intelligence as a willingness to behave in a

(15)

15 particular way does not equal an ability to interact effectively (Thomas et al, 2015). The debate on the better conceptualization is still going as no study to date has compared them directly in tandem.

3.4

The effects of cultural intelligence

A recent paper by Ott & Michailova (2016) summarises the field quite well and the next section utilizes their efforts considerably. CQ research only began in 2003 (Earley & Ang, 2003) and therefore results on the effects of CQ on performance are relatively limited. The review found the following issues. First, the CQ literature is conflicted on the question if cultural intelligence directly influences performance. Task performance was found to be positively related by Lee & Sukoco (2010) and Ang et al (2007) however, they were not at all related in the studies done by Ramalu, Wei & Rose (2011) and Malek & Budhwar (2013). There are two major differences between the significant and non-significant studies. First, the non-significant studies were both performed in Malaysia while the significant studies were not. Second, the non-significant studies measured the adaptation of many different foreigners adapting to one culture. Ang et al (2007) their sample consisted of different foreigners adapting to different countries. Lee & Sucoko (2010) investigated the intra-national adaptation of Indian bank managers adapting to regional cultures. More research is needed to resolve this. Either the Malaysian cultural specifically or the research design is responsible for the differences.

Second, different measures for performance used throughout the literature and they show conflicting results. A total of six measures were found. First, task performance (Ramalu, Wei & Rose, 2011; Ang et al, 2007; Malek & Budhwar, 2013) which is a function of the knowledge, skills, abilities and motivation directed at role prescribed behaviour such as formal job responsibilities and how much a person employs these to contribute to the core activities of a firm (Jyoti & Kour, 2015). Second, as contextual performance (Ramalu, Wei & Rose, 2011; Malek & Budhwar, 2013) which describes the activities that are directed at maintaining the interpersonal and psychological environment that needs to exist to allow the technical core to operate (Borman & Motowildo, 1997). Third, as job performance

(16)

16 (Chen et al, 2010; Thomas et al, 2015) which describes the extent to which individuals actually carry out their work assignments and duties in an effective manner (Thomas et al, 2015; Chen et al, 2015). Fourth, performance was equated to cultural effectiveness by Lee et al (2014) and Peng et al (2015) which they defined as the ability to interact with people from different backgrounds. Fifth, as expatriate performance by Lee & Sukoco (2010) which is a combination of task- and contextual performance. The performance indicators of job performance, cultural effectiveness and expatriate performance are aggregations or derivatives of task- and contextual performance. This can obscure effects because task performance is an individual-level variable and contextual performance is of a more social nature. In addition, because of the aggregation and specificity of job performance, cultural effectiveness and expatriate performance it is hard to compare the results and draw conclusions.

Third, there are multiple mechanisms for explaining the direct effect of CQ. Role expectation theory (Ang et al, 2007; Jyohti & Kour, 2015; Rose et al, 2010; Ramalu, Wei & Rose, 2011), self-efficacy theory (Chen et al, 2010; Rose et al, 2010; Ramalu, Wei & Rose, 2011), cultural judgement & decision making (Lee & Sukoco, 2010; Rose et al, 2010), anxiety/uncertainty management theory (Malek & Budhwar, 2013; Peng et al, 2015) behavioural flexibility (Rose et al, 2010) and once, no mechanism (Thomas et al, 2015) was employed. While role-expectation theory and cultural judgement & decision-making are meta-cognitive/cognitive in nature, anxiety/uncertainty management theory and self-efficacy theory are related to stress and self-confidence. Behavioural flexibility is part of behavioural CQ is seems. It is not clear if the mechanisms work in tandem or are mutually exclusive. This myriad shows the lack of academic consensus on the subject and invites scholarly attention.

Fourth, no study on CQ and performance has used cultural distance in its framework. Ott & Michailova (2016) conclude their literature with a call for more research into cultural distance. They reason that some of the found variance in the results might be explained by the ease/difficulty of certain nationalities adapting to other cultures. Empirical research on the effects of cultural distance on cultural

(17)

17 intelligence is non-existent but there is some related research. Ang & Earley (2003) theorized that cultural intelligence is used to cross the gap of cultural distance. Through adaptation an individual understands foreign culture and therefore the cultural gap becomes smaller. Research often suggested that cultural distance negatively moderated the positive effect of CQ on expatriate adjustment (Kim, Kirkman & Chan, 2008; Zhang, 2013; Johnson et al, 2006). The larger the cultural differences are, the harder it becomes to adapt to them. Therefore, expatriates adjustment takes longer or is less effective. An empirical study done by Chen et al (2010) puts cultural distance in a moderating role between cross-cultural motivation (similar to motivational CQ) and expatriate work adjustment. Their results indicate that higher cultural distance makes motivation less of a factor in successful expatriate work adjustment. They argue that as the distance becomes greater, motivation alone isn‘t enough to successfully adapt. While Chen et al (2010) comes close, no study to date (to the best knowledge of the author) has empirically studied whether cultural distance is a negative moderator on the effects of CQ. This is highly problematic as the context of the cultural adaptation is now ignored. The result is that adaptation situations treated as equal, something that Shenkar (2001) explicitly warns for. It is likely that adaptation from the Dutch culture to the Japanese cultures is more challenging than adapting to the Belgian culture.

Fifth, as mentioned earlier, the original conceptualization of CQ by Earley and Ang (2003) with its four sub-dimensions has been challenged by multiple authors. Thomas et al (2015) developed the three-dimension model and the short from measure of Cultural Intelligence (SFCQ). Thomas et al (2015) has theoretical merit and proved its validity, reliability and significance testing in the introductory paper but has not been tested by other scholars to date. The field would benefit from further testing (Ott & Michailova, 2016).

Sixth, the literature review has revealed that lower social integration is a negative effect of cultural diversity (Stahl et al, 2010) but no research has investigated if cultural intelligence affects social integration. By ignoring the negative side of cultural diversity it could be the source of conflicting results.

(18)

18 Research on the relationship tentatively points to a positive relationship between social integration and cultural intelligence. If communication is more difficult and more conflict arises then social integration should lower (Barrick, Stewart, Neubert & Mount, 1998). In addition, the social categorization and similar-attraction theory makes culturally different individuals likely to be considered an ‗out-group‘ lowering social integration (Tajfel, 1982; Knippenberg, de Drue & Homan, 2004). The absence of social interaction resulting from this is vital for teamwork and lowers performance (Kirkman et al, 2017). Further research should be conducted to test the relationship between the two variables.

3.5

The research gap

The literature suggests that cultural intelligence allows individuals to overcome the negative- and enable the positive aspects of cultural diversity but there is much that remains ambiguous. Does cultural

intelligence directly influence performance and how does this work? In addition, do social integration and cultural distance impact this relationship? Figure 2 visualizes the gap.

4 Theoretical framework

In the next section a theoretical framework is developed which is tested in order to fill the research gap. The first part focuses on the clarification of the link between CQ and performance which enables the positive aspect of cultural diversity. The second part focuses on the theoretical link between CQ, social integration and performance which mitigates the negative effect of cultural diversity. Third, the influence

Figure 2 Visualization of the interaction between the effects of cultural diversity and cultural intelligence. The question marks illustrate the research gap.

(19)

19 of cultural distance on the effect of CQ is discussed. These three parts culminate in five hypotheses and a theoretical framework.

4.1

Cultural intelligence and performance

This thesis argues that CQ is positively linked to performance by combining four mechanisms into one framework. Previous research employed role expectation theory cultural judgement & decision making theory, self-efficacy theory anxiety/uncertainty management theory and behavioural flexibility (Ang et al, 2007; Jyohti & Kour, 2015; Rose et al, 2010; Ramalu, Wei & Rose, 2011; Chen et al, 2010; Malek & Budhwar, 2013; Peng et al, 2015; Lee & Sukoco, 2010). Because the literature found that aggregate performance measures might conceal effects this thesis splits it into task performance and contextual performance. Role expectation theory explains that role expectations and assessment thereof are culture-specific (Ang et al, 2003). Foreigners who do not adapt to local expectations will therefore receive lower performance ratings. Higher CQ allows foreigners to assess local role expectations better than low CQ foreigners (Jyohti & Kour, 2015). Cultural judgement & decision-making theory refers to the human information processes for appraising cultural situations, applying appropriate cultural values and choosing the correct option to achieve a desired result (Ang et al, 2007). High CQ individuals combine the correctly assessed role expectations with the correct values to achieve optimal results (Lee & Sukoco, 2010). Self-efficacy theory is also employed. It explains that individuals with a high self-belief in their intercultural skills are more prone to engage in adaptive behaviour and less sensitive to stress (Chen et al, 2010). In a similar vein, anxiety/uncertainty management theory states that intercultural interactions increase anxiety and are inherently more uncertainty because communication failure risk is higher (Peng et al, 2015). High CQ individuals are better able to cope with this stress which increases interaction

frequency which leads to higher learning and more instances to show role congruent behaviours. Finally, cultural flexibility research by Shaffer et al (2006) found that individuals who can better substitute their home country verbal-, non-verbal and tone-of-voice behaviour for a locally appropriate equivalent showed

(20)

20 higher performance. Therefore, high CQ individuals improve their performance through adapting their behaviours. All the theories combined explain how high CQ leads to higher task- and contextual

performance which is coined as the ‗planning-resolve-execute‘ (PRE) mechanism. In the planning phase, high CQ individuals first correctly assess what the role expectations on both task- and contextual

performance. They then apply their cultural judgement and decision making skills to choose the best course of action. In the resolve phase, higher self-efficacy and better anxiety/uncertainty management lowers the barriers to frequently engage in intercultural behaviour instead of avoiding it. Finally, cultural flexibility allows higher CQ individuals to appropriately execute their appropriate behaviour by adapting their verbal, non-verbal and tone-of-voice behaviour. The PRE mechanism allows them to adapt

appropriately which increases performance. This will be tested using the following two hypotheses:

𝐻1: Higher cultural intelligence has a positive effect on task performance.

𝐻2: Higher cultural intelligence has a positive effect on contextual performance.

4.2

Social integration

In addition to the direct effects of CQ on performance, this thesis also proposes that CQ has a positive effect on performance through social integration as a mediator. First, it is explained how CQ increases social integration then how social integration increases task performance. Finally, the section finishes with a discussion on the positive effect of social integration on contextual performance. For the mediation effect to exist there must be a direct link between CQ and social integration and social

integration and task/contextual performance. The main hypotheses are listed here and their sub-hypotheses and theoretical background are explained further on below:

𝐻3: Cultural intelligence increases task performance through social integration as a mediator.

(21)

21 Social integration is defined by O‘Reilly, Caldwell & Barnett (1989:22) as the ―amount of attraction to the group, satisfaction with other members of the group, and social interaction among group members‖. This thesis reasons that CQ increases social integration using three main arguments. First, cultural diversity lowers social integration because individuals tend to like others who are the same due to the similarity-attraction effect (Williams & O‘Reilly, 1998). As individuals adapt their values, habits and behaviours to local equivalents using their CQ, they become more similar and therefore more attractive for social interaction. Also, in the categorization-elaboration model cultural diversity is a salient variable for out-group classification (Knippenberg, de Drue & Homan, 2004). Due to the interout-group-bias effect other groups are a threat social identity resulting in lower social integration. Again, through lowering in the differences through cultural adaptation other groups become less threatening and a foundation for cooperation is achieved. Second, research shows that expatriates living in Japan socially integrated more successfully when they have high emotional stability (Peltokorpi, 2008). As mentioned, intercultural interactions are inherently stressful leading to avoiding behaviour. Higher emotional stability allows individuals to cope with the added stress of intercultural interactions. CQ has been repeatedly linked to better stress/anxiety/uncertainty management (Malek & Budhwar, 2013; Peng et al, 2015; Ang et al, 2007) which is equivalent to higher emotional stability (Ang et al, 2007). Third, research found that higher cultural empathy/lower ethnocentrism is related to higher social integration (Pletokorpi, 2008). Higher cultural empathy means that individuals tend to image and understand the circumstances and reasons why other cultures do what they do. Ethnocentrism on the other hand is the belief in the superiority of ones‘ own culture. High ethnocentrism leads to the immediate rejection of different cultures while cultural empathy leads to deeper understanding and milder judgement of others‘ cultural habits. As cultures are deeply tied to social identity, it follows that negative judgement leads to strained social ties and lower social integration. Empirical research found strong negative correlations between CQ and ethnocentrism (Reichard, Dollwet & Louw-Potgieter, 2014) and a strong correlation with cultural empathy (Ward,

(22)

22 Fischer, Lam & Hall, 2009), supporting the idea that CQ positively influences social integration. This leads to the following hypotheses:

𝐻3𝑎: Higher cultural intelligence has a positive effect on social integration.

𝐻4𝑎: Higher cultural intelligence has a positive effect on higher social integration.

The next section explains that higher social integration has a positive effect on task performance. By definition social integration increases the amount of affection towards, satisfaction with and interaction with team members. This leads to four effects which increases task performance. First, it increases the frequency with which colleagues interact therefore allowing for more learning (Sidelinger, Bolen, McMullen & Nyste, 2015). More learning leads to more information, better skills and more experience which directly links with performance (Marton & Saljo, 1976). Second, higher social integration also increases trust which increases colleagues‘ willingness to share and access each other‘s‘ guarded knowledge (Cabrera & Cabrera, 2005). These ‗tricks of the trade‘ are often kept private in order to outperform colleagues. However, according to Mayer, Davis & Schoorman (1995) if perceived

benevolence goes up, trust goes up and knowledge is shared. Third, increased social integration increases the frequency of colleagues offering assistance with work when needed (Sheng, Tian & Chen, 2010). This assistance makes sense as increased ‗groupiness’ (van der Vegt, Bunderson & Kuipers, 2010) decreases the internal competition in order to increase external competition performance and therefore overall performance. In practice, this would be a local colleague taking some extra time to help out a foreigner employee when he gets stuck with an unfamiliar problem. Fourth, higher social integration lowers the trigger tendency of local colleagues‘ not-invented-here (NIH) syndrome. The NIH syndrome is the dismissal of ideas not on their merits but the external/out-group status of its proponent. Antons & Piller (2015) find that colleagues from geographically different parts of the firm or cultural backgrounds can trigger heuristic judgements that classify them as external and thus trigger the NIH syndrome. Higher

(23)

23 social integration transforms the external/out-group classification into an internal/in-group one. By not triggering the NIH syndrome, foreign employees can employ their differentiated knowledge, experience and skills which increases their task performance. This thesis therefore reasons that:

𝐻3𝑏: Higher social integration has a positive effect on task performance.

Finally, it is reasoned that higher cultural integration increases contextual performance. Contextual performance entails both group-level behaviour of helping out colleagues, volunteering for extra work and organising teambuilding exercises as well as organisation-level behaviours such as following the

regulations and procedures (even when personally inconvenient) and participating in marketing and recruitment events. More social integration leads to more interaction with colleagues and more instances to meet contextual performance expectations (Beal & Cohen, 2003). Also, by definition, more social integration leads to more affection within the team/division/company. Higher affection levels lead to more organizational citizenship which is linked to contextual performance (Motowildo, Borman & Schmit, 1997). These arguments are also in line with the findings of Guillaume, Brodback & Riketta (2011). Their meta-study revealed that deep-level dissimilarity has a negative impact on contextual performance through lower social integration. They reason, based on previous meta-studies and work (Harrison et al, 2006; Judge et al, 2001; Meyer et al 2002; Riketta, 2005; Seers, 1989), that lower social integration means unfavourable relationships with peers and lower attachment to their workgroup. This leads to fewer opportunities to meet performance expectations, a lower tendency to offer and accept input into their work, less access to information and resources and progressively leading to more withdrawal. Finally, Malek & Budhwar (2013) found a relationship between expatriate interaction adjustment to host country nationals and expatriate contextual performance. This indicates that adjusting to host country nationals increases the contextual performance of foreigners trying to integrate locally. This collectively leads to the following hypothesis:

(24)

24 𝐻4𝑏: Higher social integration has a positive effect on contextual performance.

4.3

Cultural distance

As a final hypothesis, it is proposed that the effects of CQ on task performance, contextual performance and social integration are negatively moderated by cultural distance. More cultural distance means a less positive effect of CQ. To illustrate the logic the mechanisms from 𝐻₁, 𝐻₂, 𝐻₃ and 𝐻₄ are used. The effects of the PRE mechanism, the similarity-attraction effect, the categorization-elaboration model, the amount of uncertainty/stress in intercultural interactions and the amount of cultural

empathy/ethnocentrism are all build around the assumption that adaptation closes the cultural gap between individuals. Cultural distance theory states that the gap between two cultures can be larger or smaller depending on their relative distance. Low cultural distance means that many aspects of two cultures are the same e.g. Spain and Portugal. High cultural distance means many aspect are different e.g. Egypt and Sweden. More differences mean more unfamiliar concepts and more learning/adapting to do and therefore a lower positive effect of cultural intelligence. This moderating effect has been found in multiple related studies on cultural distance with different dependant variables. Chen, Kirkman, Kim, Farh & Tangirala (2010) find that higher cultural distance negatively moderates the positive effect of CQ motivation on expatriate performance. In a similar vein, Stahl & Caligiuri (2005) found that expatriates trying to cope with cultural adaptation problems were having a more difficult time when cultural distance was high. Finally, Kossek, Huang, Piszczek, Fleenor & Ruderman (2015) find that higher cultural distance made colleagues rate expatriate performance lower. They imply that more adaptation is required in order to cross the distance gap. Therefore, the following hypotheses will be tested to investigate the proposed effects:

𝐻5𝑎−𝑐: Cultural distance negatively moderates the relationship between cultural intelligence and social

(25)

25

4.4

The model

Summarizing all previously mentioned hypotheses leads to the following conceptual framework as shown in figure 3:

5 Methodology

The next section starts with an explanation of the research design. It continues with a discussion on the population, sample and data collection. The section then discusses the variables, the measurement tools and their validity. Finally, the statistical techniques used to test the hypotheses are reported.

In order to test the hypotheses, this study will employ an empirical, quantitative approach. The CQ, social integration, task performance, contextual performance, cultural distance and demographics of individuals will be measured via self-administered questionnaires. Previous researchers have also adopted this approach to investigate the relationships between mediators, moderators, outcome variables and CQ (Ang & Earley, 2003; Ang et al, 2007; Thomas et al, 2015; Ramalu, Wei & Rose, 2011; Ramalu, Wei & Rose, 2011; Chen et al, 2010; Lee & Sukoco, 2010; Malek & Budhwar, 2013; Lee et al, 2014; Jyohti & Kour, 2015; Ott & Michialova, 2016). Also in line with these authors, a hierarchical regression is used to test the hypotheses. The hierarchical regression has the advantage that variables can be added one at a time to see how they interact. This allows for deeper analysis than a normal multiple linear regression.

(26)

26

5.1 Population, sample and data collection

This thesis has chosen to focus on knowledge-workers in the professional services industry who are working in a different cultural setting for multiple reasons. First, knowledge workers are often employed in teams evaluated on both task- and contextual performance. This benefits data collection efforts and increases the accuracy of self-reported data as respondent are focussed on both. Second, the professional service industry is quite internationalized which increases the depth and breadth of internationals working in a single firm. Otherwise, chances are low of getting enough respondents and nationalities to draw statistically valid conclusions. Third, many practitioners of IM theory are in the knowledge industry which increases the applicability of the results. The sample of this study is drawn from a real-life, multinational firm referred to as Firm A in compliance with their request for anonymity. Firm A is globally active on all continents, is part of the professional service industry, is a knowledge-intensive MNE and delivers tax, assurance, mergers & acquisitions and advisory services. As a result, its employee base very culturally diverse and often do assignments in other countries.

There are multiple advantages to using a single firm in a single country. First, all respondents are foreigners coming to the Netherlands which corrects for the illusion of symmetry defined by Shenkar (2001). In addition, the research design also negates the role of organisational culture as all respondents adapt to the same one: Firm A‘s Dutch offices. Third, all employees have gone through the same on-boarding experience and therefore adaptation assistance is the same for the entire sample. One

disadvantage is that employees adapt from cultures all over the world. Because cultures widely differ the adaptation process might vary between individuals and CQ might play a slightly different role per employee. This is unavoidable however, as there are often not enough employees of one culture going to one specific firm in the Netherlands to find statistically acceptable results. This potential problem is compensated for by adding the culture distance component into the framework.

(27)

27

Table 1 Possible combinations resulting in bicultural backgrounds. The horizontal axis deals with parental heritage and the vertical axis represents the childhood country of an individual.

Population is defined as Firm A employees who identify as a different culture than Dutch as their primary culture. In addition, the Firm A employees are required to have worked for more than one month at firm A. This means that employees have to adapt to the Dutch culture with their CQ and have at least a months‘ worth of performance to report. Important to note is the existence of bi-cultural individuals. Culture is taught to humans as a child by our parents and environment. Some individuals are raised by immigrant/expatriate parents, in a foreign country or both and as a result possess intricate knowledge of foreign parent or country culture (Tadmore, Nguyen, & Benet-Martínez, 2012).

Both Dutch One is foreign Both are foreign The Netherlands 1. Fully Dutch

4. Bi-cultural 2. Probably Dutch 5. Bi-cultural 3. Bi-cultural 6. Fully foreign Foreign country

As table 1 shows, six types of employees emerge: grown up in the Netherlands with Dutch parents (1), grown up in the Netherlands with one foreign parent (2), grown up in the Netherlands with two foreign parents (3), grown up in a foreign country with Dutch parents (4), grown up in the Netherlands with one Dutch parent (5) and grown up in a foreign country with foreign parents (6). Type 1 and 6 are fully Dutch or fully foreign. Type 2 has lived in the Netherlands and has one parent with a foreign heritage and probably does not have to adapt to Dutch culture. This results in exclusion from the sample. Types 3, 4 and 5 have had significant foreign influence and a little Dutch influence. Although adaptation might be easier for them it is unlikely that they have completely integrated Dutch culture. As such, they are included into the sample but controlled for by additional variables. Defining the total population size within Firm A is difficult as privacy and discrimination laws obligate that employee cultural backgrounds are not recorded. In order to bypass this, in-bound expatriates and local hires targeted. In-bound

expatriates are Firm A employees coming from foreign locations to work for an extended period of time in the Netherlands and numbered 26 in total. Local hires are foreigners who‘ve been hired on a local Dutch

(28)

28 contract and have either a working permit or a visa which was used to identify them. The local hires list Firm A provided contained 150 valid individuals so the total population size is 176.

In order to assess the hypothesized relationships between CQ and performance a questionnaire is used to collect data. According to Saunders, Lewis & Thornhill (2015), using a questionnaire has multiple benefits. First, a questionnaire allows for the gathering large samples in a short period of time in a cheap way. Second, questionnaires can objectively measure the same variable repeatedly to ensure consistent and unbiased data collection. Third, questionnaires are quick and relatively non-invasive for the practitioners, reducing sampling bias. Finally, questionnaires are easy to quantify which simplifies statistical analysis considerably and aids empirical testing. The population was asked to fill in the digital questionnaire. Multiple steps were taken in order to maximize the chances of a high response rate. First, the digital invitation was tailored to fit the corporate-brand style to instil a sense of reliability, familiarity and professionalism. Second, the invitation explained that the research was supported at the highest managerial levels of Firm A. Third, the invitation explained that the research was supported by the

HR/privacy department. Fourth, the questionnaire was stylized in the UvA corporate-brand using Qualtrics to instil academic confidence. Fifth, academic references to the validity of the measures were added to show academic validity of the questionnaire. Sixth, during the whole process, anonymity was assured and reassured as demographic and performance measures were asked of the respondents. The invitation was sent three times using one-week intervals to give respondents enough time and opportunity to partake in the study. In total, 118 responses were collected of which 28 were incomplete or had to be excluded on bicultural grounds resulting in 90 useful respondents for the testing of H1 to 4 and 87 usable respondents for H5 due unavailable cultural dimension data. This amounts to a response rate of 50,6%. According to Tabachnick and Fidell (2007), the minimum required sample size equals 50 + 8k where k is the number of independent variables. This means a minimum of 50 + 8*4 = 82 respondents is required per SPSS

(29)

29 analysis. There are only four variables as task- and contextual performance can‘t be in the SPSS model at the same time. While relatively close with 90 respondents the sample does passes the threshold.

Appendix 1 shows the sample frequencies of Gender, Age, Primary Culture, Bicultural, Secondary Culture, Dutch Parents, Firm A Rank and Tenure. The next section will highlight the important data here in lieu of readability. First, there was an exact split between male and female respondents. Second, the vast majority of respondents (58,9%) were between 32 and 39 years old. Third, the majority of respondents (53,3%) already worked for more than 5 years in the Netherlands. Fourth, 24,4% was at staff-level, 40% at senior-level and 22,2% at managerial level. Fifth, there were 34 different primary cultures of which the largest was India with 6 individuals so a good spread was achieved. However, there were 19 bicultural individuals with a Dutch primary culture on a total of 59 bicultural individuals. 21 different foreign secondary cultures were reported of which the U.S.A. was largest with 5 respondents. There were also 27 individuals who identified Dutch as a secondary culture. Finally, 10% of respondents had one Dutch parent and only 8,9% of respondents had both a Dutch father and mother. In conclusion, the sample is culturally diverse although some individuals identify as either primarily or secondarily Dutch. As these are controlled for in the analysis they should not pose an issue. In addition, half of the individuals have worked for more than five years in the Netherlands and the majority is in his/her thirties. Again, both issues are controlled for in the model and therefore mitigated.

5.2 Variables

Cultural Intelligence. As mentioned earlier, CQ is an individual-level trait and there are currently two

models which measure CQ. This paper has chosen to use the three-dimension model/Short Form CQ developed by Thomas et al (2015). The three dimensions are cultural knowledge, cross-cultural skills and cultural metacognition. When combined these reflect the high-level latent factor of CQ. This method was chosen for a number of reasons. First, the authors removed the motivational element of CQ. They reason that CQ is the ability to act cross-culturally and not the willingness to do so. Motivation is therefore not an

(30)

30 integral part of CQ. The second reason is that the three dimension model is a truly aggregated

conceptualization where the three dimensions together make up CQ. The popular conceptualization of Earley & Ang (2003) essentially splits CQ in four facets which operate on their own. Third, Thomas et al (2015) extensively validated the new conceptualization with regards to construct validity, measurement equivalence cross-culturally, discriminant, convergent and criterion-related validity and is therefore a reliable measure. Fourth, Ott & Michailova (2016) called for the exploration of new conceptualizations. In line with this call, it is to the benefit of the academic field to test the new scale. Fifth, the short-form CQ scale (SFCQ) is 10-item scale which benefits data collection. This is half the amount of other

conceptualizations which allows for a shorter questionnaire. While it is a more practical reason rather than academic, the employees of Firm A are likely to ignore long questionnaires leading to an insufficient sample size. Respondents answered rated their tendency to show certain behaviour on a one to five Likert-scale where 1= ‗not at all‘ and five = ‗extremely well‘. The 10 items were added up and averaged to compute the CQ score.

Social Integration. Social integration is considered an individual-level construct. It is measured by

adapting the nine-item scale from Smith et al (1994) which is based on the validated scale from Shaw (1971). The adaptation consisted of changing the questions to reflect Firm A instead of TMG used Smith et al (1994). The essence of the questions was not altered in any way. Responses were measured using a five point Likert scale where 1= ‗strongly disagree‘ and 5= ‗strongly agree‘. The overall social integration score per respondent was computed by reversing the negatively-keyed items, adding up all values and averaging it.

Cultural Distance. Cultural distance is an individual-level variable and is measured using Hofstede‘s six

cultural dimensions. Initially, cultural values were to be measured by individually filling out the Hofstede cultural questionnaire. However, Firm A strongly advised against this for two reasons. First, personally assessed Hofstede data requires the adding of 42 extra questions i.e. 15 extra minutes. The likelihood of

(31)

31 getting responses would become very low. In addition, the privacy department of Firm A was against the measuring personal values and permission wouldn‘t be granted. To overcome this, country-level data was coupled to the self-identified culture of individuals. This does trigger the assumption of spatial

homogeneity (culture is the same in all parts of a country) and illusion of stability (cultural variables remain stable over time) of Shenkar (2001). The problem of spatial homogeneity is partly mitigated by allowing respondents to choose their own culture. The illusion of stability is partially mitigated as we‘ve employed the latest Hofstede dataset but admittedly individual measurement would have been better. Hofstede was employed because it was the only database with enough breadth in data to support the amount of nationalities included in the sample. Unfortunately, there were still some missing data on cultural dimensions. The data was incomplete for Andorra, Azerbaijan, Belarus, Kazakhstan, Israel and Nepal. If only one or two values were missing cultural distance was calculated without them. If more were missing, a suitable culturally close nation was sought to supplement. If that was not possible, the data point was dropped from the C.D. hypothesis analysis. The following adaptations took place: Andorra supplemented with Spain, Azerbaijan omitted, Israel was only missing one variable and included, Kazakhstan supplemented with Russian as the respondent was biculturally Russian and its cultural has been heavily influenced due to their USSR past and finally, Nepal was only missing two values. Regrettably, the countries of Eritrea, Sudan, Tunisia and Turkmenistan had no values for any cultural dimension and were omitted from the C.D. analysis. In order to calculate the cultural distance, respondents submitted their primary- and if bicultural, their secondary self-identified culture. If the respondent was primarily Dutch but bicultural, the secondary culture was used. By cross-referencing with the Hofstede dataset (available upon request) on national cultural dimensions the absolute differences between the primary culture and the Netherlands were found and averaged.

Task- and contextual performance. Task- and contextual performance is measured using the task-

(32)

32 number of reasons. First, they employed it to related research which measured employees‘ cultural fit and performance. Second, other scholars have also employed the scale to their research such as Bakker, Demerouti & ten Brummelhuis (2012), Kahya, 2007 and Bakker & Xanthopoulou (2009). Third, its length is limited to sixteen questions in total. As questionnaire length is an important factor in this research it was deemed more appropriate than other measurement tools. The scale uses seven questions to measure contextual performance and nine questions to measure task performance. Respondents self-reported their agreement with the statements on a four-point Likert scale where 1= ‗Strongly disagree and 4= ‗Strongly agree‘. The scale has the benefit of employing both individual- (‗You volunteer to do things not formally required by the job‘), team- (‗You help others when their work load increases‘) and organisation level (‗You willingly attend functions not required by the organization, but helps in its overall image‘) questions to measure contextual performance. Both measures are summed and averaged to calculate the respondents‘ separate task- and contextual performance.

Control variables. In order to check for sample bias and alternative effects the following control variables

were added. First, we asked respondents what their self-identified gender was: male, female or other. Research has shown that gender can account for variance in contextual performance (e.g., Kidder & McLean Parks, 1993; Shore, Barksdale & Shore, 1995). Only male and female was chosen. Second, age might potentially explain performance due to previous life experience and was therefore included. Due to restrictions from Firm A eight age categories were introduced to limit the ability to use demographics to still identify respondents. Third, tenure was included because the length of stay in the Netherlands could explain performance. The longer an employee has worked in the Netherlands the more that person has adapted to local culture. The same identification restriction was applied to tenure and therefore ten categories were introduced. Fourth, higher rank is very likely to explain higher performance and should therefore be controlled for with regards to sample bias. Respondents‘ rank at Firm A was measured using seven categories where staff/senior staff is consultant/operational level, manager/senior manager middle

(33)

33 manager level and executive director/partner top level management. Finally, respondents‘ parental heritage was also measured as either ‗no Dutch parent‘, ‗one Dutch parent‘ or ‗both parents are Dutch‘. Adaptation to local Dutch culture is likely to be easier if one or both parents are Dutch and therefore taught a

respondent about his/her original culture. A respondent could think of him/herself being a foreign culture but subconsciously already be adapted to Dutch culture.

5.3 Statistical methods

In order to test the hypotheses, I will employ a hierarchical regression. Fields (2009) supports this writing that a multiple linear regression is a valid and robust test for measuring the strength and significance of the influence of an independent variable upon a dependent variable. By using hierarchical model, controlling and testing the influence of control, mediator and moderator variables is achieved. For the purpose of brevity, equation terms are only explained once. So assume that reoccurring terms are identical unless otherwise specified. The direct influence of CQ on task- and contextual performance is measured using the following equations:

Eq. 1: 𝑌𝑇𝑃= 𝛼 + 𝛽CQ→TP ∗ 𝑥𝐶𝑄+ ∑(𝛽𝑖 ∗ 𝑥𝑖 ) + 𝜀𝑇𝑃

Eq. 2: 𝑌𝐶𝑃= 𝛼 + 𝛽CQ→CP ∗ 𝑥𝐶𝑄+ ∑(𝛽𝑖 ∗ 𝑥𝑖 ) + 𝜀𝐶𝑃

Both equations follow the same principal, only the dependant variable is different. 𝑌_𝑇𝑃 and 𝑌_𝐶𝑃 are the levels of task or contextual performance, 𝛼 is the constant or intercept, 𝛽 is the slope of the regression line and represents the main effect of CQ on task- or contextual performance, 𝑥_𝐶𝑄 represents the level of CQ of an individual, ∑(𝛽𝑖 ∗ 𝑥𝑖 ) is the sum of all control variables‘ influence and 𝜀𝑇𝑃 is the

error term of the regression. The function illustrates that performance (𝑌_𝑇𝑃 or 𝑌_𝐶𝑃) increases by 𝛽 per 𝑥_𝐶𝑄 increase in cultural intelligence while controlling for the influence of all control variables. If 𝛽_CQ→TP or 𝛽CQ→CP is positive and significant, 𝐻₁or 𝐻2 respectively are supported.

Capitalizing on cultural diversity : the impact of cultural intelligence on performance through social integration and cultural distance