Article body

Introduction

The aim of strategic human resource (HR) management is to encourage employees to adopt attitudes and behaviours that are likely to enhance organizational performance (Delaney and Huselid, 1996). In theory, HR management practices used in synergy have a positive influence on the employees’ attitudes and behaviours that will enhance organizations’ performance and competitive advantage (Lawler, 1986; Boxall and Macky, 2009). However, social capital (including employees’ work values and the organization’s values) also plays an important role in explaining organizational high performance (Gittell, Seidner and Wimbush, 2010; Den Hartog and Verburg, 2004).

Given that social capital can decrease as a result of conflicts and other social dysfunctions, Kupperschmidt (2000) and Crumpacker and Crumpacker (2007) suggest that HR managers should manage different generations of employees in different ways, in order to reduce the “clash” between generations. Proponents of this approach postulate that organizational performance is negatively affected by intergenerational differences in work values, because of the conflicts between employees that arise from these differences. However, the majority of the articles that claim there is a “clash of generations” are speculative in nature. Recently, other authors have questioned the assumption that marked intergenerational differences exist (Saba, 2009). According to Audet (2007) and Giancola (2006), the “clash of generations” appears to be a “myth.” Several studies have also found intercultural differences in work values (Hofstede, 2001; Trompenaars, 1993), but others have also questioned whether marked intercultural differences exist (Gerhart and Fang, 2005). If it is true that there are no major intergenerational or intercultural differences, it would certainly be counterproductive to adjust HR management practices for minor differences.

The main goal of this study was to examine the work values of individuals from different generations, within two very different and understudied cultures: French Canadians from Quebec Province (or Quebecers) and Arabs living in the United Arab Emirates. Another goal of this study was to validate the structuring of work values using an abridged version of the Work Values Inventory (WVI) of Wils, Luncasu and Waxin (2007).

Conceptual Framework

Work Values

Empirical results from past research on work values are difficult to synthesize due to the lack of consensus on how work values are defined and measured. These problems mainly arise due to incomplete and inconsistent theorization regarding work values. Work values have alternately been defined as (1) desirable workplace behaviours, (2) reflections of the significance of work, (3) the meaning of work, or (4) the basis for business ethics (Lyons, Higgins and Duxbury, 2009). This variation in definitions helps to explain the diversity of instruments used to measure work values. However, the major problem in work value theorization remains our lack of knowledge about the theoretical structuring of work values.

Recently, Wils, Luncasu and Waxin (2007), Waxin and Wils (2007) and Waxin, Wils and Labelle (2009) proposed and empirically validated a work value model based upon Schwartz’s (1992) studies of fundamental values. According to these authors, work values are conceptualized as desirable and trans-situational goals that act as guiding principles for an individual’s work life. Like Schwartz’s fundamental values (1992), these work values are organized around a circular structure made up of two orthogonal axes encompassing 10 work value dimensions (Wils, Luncasu and Waxin, 2007; Waxin, Wils and Labelle, 2009). Each of these axes has two poles that oppose each other: self-enhancement and self-transcendence are opposite poles on one axis, and conservation and openness-to-change are opposite poles on a second axis. Self-enhancement includes value dimensions related to power and achievement, whereas self-transcendence includes value dimensions related to benevolence and universalism. Conservation includes value dimensions relating to tradition, conservatism and security, whereas openness-to-change includes value dimensions related to self-direction, stimulation and hedonism. Before studying work value differences between different generations and cultures, we first have to validate the structuring of values (Schwartz, 1992). As a result, two hypotheses are formulated at the axial level since De Clercq, Fontaine and Anseel (2008) found that the higher order factor structure (axial level) is more robust.

H1a: The work value structure is validated in the Quebec sample at the axial level, i.e. the work values items are displayed around the four work values poles, forming the two bipolar axes described by Wils, Luncasu and Waxin.

H1b: The work value structure is validated in the U.A.E. sample at the axial level, i.e. the work values items are displayed around the four work values poles, forming the two bipolar axes described by Wils, Luncasu and Waxin.

The Different Generations

The word “generation” has multiple meanings. It can refer to groups of individuals, psychological characteristics, or ways of thinking (Attias-Donfut, 1988) and can be applied to different concepts such as demographic generations, social generations, or historical generations (Chauvel, 2003). In the field of management, the word “generation” mainly refers to a cohort defined in the demographic sense of the term. A cohort is considered “the aggregate of individuals who experienced the same event within the same time interval” (Ryder, 1965: 845). However, management scholars focus more on the boundaries between the generations than on the events marking them. Birth year is the main criterion used to define the boundaries between the generations since it makes it possible to isolate individuals who shared the same life events over the course of a given period. In rich Western countries, such as the United States and Canada, the boundary between the Baby Boom Generation and Generation X is the beginning of the 1960s (when the birth rate began to decrease in these countries), whereas the boundary between Generations X and Y corresponds to the beginning of the 1980s (when the birth rate increased in these countries). Lyons (2003) defines the following boundaries between generations, based on demographic analyses: Veterans (born before World War II), Baby Boomers (1946–1961), Generation X (1962–1979), and Generation Y (1979–2004).

Very few studies have attempted to define the boundaries between different generations in the United Arab Emirates or in the Arab world in general. Whiteoak, Crawford and Mapstone (2006) distinguish two generations without naming them or justifying where the boundary is between them: older individuals (30 years old or over) and younger individuals (under 30 years old).

Defining boundaries between generations assumes that notable differences exist between the generations, which Audet (2007: 19), however, states is a “myth.” Determining whether actual differences exist between defined generations has interested researchers for years.

The Relationship between Work Values and the Different Generations

Inglehart (1971) advanced the “theory of cohorts,” which is based upon two hypotheses: the scarcity hypothesis and the socialization hypothesis. The scarcity hypothesis proposes that individuals in a given country tend to attach utmost importance to their most pressing needs. During periods of scarcity, when survival takes precedence over everything else, individuals attach great importance to materialistic values (e.g., money, economic security, respect for authority). Conversely, during periods of prosperity, individuals attach increasing importance to post-materialistic values such as autonomy, tolerance, creativity and protection of the environment. The socialization hypothesis stipulates that changes in the values of a society occur gradually, as generations succeed one another, because the values of each generation reflect the conditions that prevailed during their years preceding adulthood. Thus, Inglehart (2008) observed a cultural shift from materialistic values to post-materialistic values in Western European countries after World War II since basic survival was no longer threatened (higher living standards and social programs).

Although the theory of cohorts proposed by Inglehart (1971) has often been referred to (Egri and Ralston, 2004), few authors in the field of management have found empirical results that clearly distinguish between different generations. Empirical results from past research examining the influence of generation on work values are difficult to synthesize for four reasons: (1) the weak theorization of work values, (2) the different instruments used to measure work values, (3) the different methodologies used to examine generational value differences, and (4) the different cultures that the studies sample.

The first problem is that previous studies conducted on work value differences between generations have not explained the theoretical structuring of work values a priori as Schwartz (1992) did for fundamental values, or Wils, Luncasu and Waxin (2007) did for work values. Instead, empiricism (i.e., exploratory factor analysis) is often used in these studies, rather than theorization. For example, the personal-growth factor in Chen and Choi’s (2008) empirical study includes values pertaining to two different work value poles in Wils, Luncasu and Waxin’s theoretically structured work value model: the work value items “altruism,” “stimulation” and “self-direction” are all included in the same “personal growth” factor in Chen and Choi’s (2008) study, whereas “altruism” is associated with the self-transcendence pole and “stimulation” and “self-direction” are associated with the openness-to-change pole in Wils, Luncasu and Waxin’s model. In another example, in Lyons, Duxbury and Higgins’s (2006) empirical study, their extrinsic work values factor includes items relating to high wages (which is associated with the self-enhancement work value pole in Wils, Luncasu and Waxin’s model) and job security (which is associated with the conservation work value pole in Wils, Luncasu and Waxin’s model).

The second problem is that different studies use very different instruments to measure work values. Some studies do not list the precise work value items included in their measurement instrument (Hui-Chun and Miller, 2003). Some studies use a combination of measurement instruments (Cennamo and Gardner, 2008). Some studies use instruments which measure concepts that are closely related to work values, such as motivation factors, but are not actual work values themselves (Wong et al., 2008; Jurkiewicz, 2000). Other studies measure concepts that are unrelated to work values, such as distrust (Taylor and Thompson, 1976). The construct validity of the majority of the instruments used in most work values studies is unpublished and therefore unknown.

The third problem is that different authors use different methodologies to compare generational work value differences. For example, some authors (Lyons, Duxbury and Higgins, 2007) follow Schwartz’s recommendations for analyzing the differences in values between different groups, such as different generations. These recommendations include controlling for mean score and eliminating questionnaires whose responses may be skewed by social desirability. Most authors do not follow these recommendations, however. Some authors control for the effect of gender (Lyons, Duxbury and Higgins, 2007), while others do not. Some authors eliminate questionnaires with more than 15% missing values (Cennamo and Gardner, 2008), while others do not. Some authors compare single work values items (e.g., status in Cennamo and Gardner, 2008) while others compare groupings of values (e.g., the professional development factor in Chen and Choi (2008), made up of the following four work values items: management, independence, prestige and variety). Some authors ask respondents of different generations to state the degree of importance of the different work values whereas other authors ask respondents to rank the values in order of importance (Jurkiewicz, 2000).

The fourth problem is that different studies on work values have been conducted in many different cultures. Culture of origin affects work values (Hofstede, 2001; Trompenaars, 1993), so that generalizing about generational differences in work values based upon the results of studies conducted in different cultures may be difficult. Moreover, boundaries between the generations can vary according to the different events that occurred in each culture. So, altogether, based on past research, it is certainly very difficult to determine whether or not different generations do have different work values.

Intergenerational Differences in Work Values in Western Countries

Only a few studies have investigated differences in work values between generations in Western countries, and their results are often contradictory. Our review of the literature on work values was thus broadened to include the literature on fundamental values as well. We excluded the speculative literature that is quite prolific (Wong et al., 2008) and which has recently been synthesized by Saba (2009).

The empirical literature shows that the different generations often do not differ substantially from each other in terms of values. Several studies have found no significant differences in values between the generations (Jurkiewicz, 2000 in the United States; Wong et al., 2008 in Australia).

Other studies found some differences between generations, but the differences were smaller than the authors’ expectations, and were only detected because of the high statistical power in the study due to a large sample size. Lyons, Duxbury and Higgins (2007) sampled 1,194 Canadians and found that Canadian Baby Boomers have a few fundamental values that differ from those of Generation X (for example, Baby Boomers are less open to change than generation X), but the authors acknowledged that “the value differences observed in this study were small in magnitude: generational differences explained approximately 7% of the variance in people’s values” (p. 350). Cennamo and Gardner (2008) found that Generation Y attached greater importance to freedom than Baby Boomers in New Zealand, but overall, these authors found fewer differences between the generations than they expected.

Differences in values that are observed in studies conducted in Western countries, are not consistent between countries. For example, the results of Lyons, Duxbury and Higgins (2007) validate their hypothesis that the work values of Veterans and Baby Boomers differ from those of Generations X and Y in Canada. However, Cennamo and Gardner (2008) found that in New Zealand, there are some differences between Generations X and Y as well. As another example, Lyons, Duxbury and Higgins (2007) found that Generation X and Y were just as attached to conservation as Veterans in Canada, but Egri and Ralston (2004) found that US Baby Boomers were more attached to conservation than Generation X. Contrary to their expectations, Lyons, Duxbury and Higgins (2007) found no significant difference in openness to change between Baby Boomers and Generation Y in Canada, but Cennamo and Gardner (2008) found that Generation Y attached greater importance to freedom than Baby Boomers in New Zealand.

Our literature review found that many studies do not find value differences between generations in Western countries. Studies that did find differences found only small differences, and the results were inconsistent between different countries. It is possible that within a given generation, values preferences are quite heterogeneous, and therefore it is difficult to detect clear differences between generations. Lyons, Duxbury and Higgins (2007) conclude that the values within each generation are heterogeneous, in that individuals within a given generation vary greatly in their work value preferences. According to Inglehart (2008) there are no “major” intergenerational differences in values between the generations born after World War II. He states that “the post–war cohorts show relatively similar values” (Inglehart, 2008: 145). Based upon our literature review, our second hypothesis is:

H2: Among Quebecers, there are no differences in the scores on the four work value poles between the different generations.

Intergenerational Differences in Work Values in Arab Countries

The Western world has been characterized by continuity in values for recent generations (Boudon, 2002), but a break or discontinuity in values occurred between recent generations and those of Veterans (Inglehart, 2008). A comparable break seems to have occurred in the Arab world. The Arab world has undergone a major discontinuity, or break, in values as a result of modernization, industrialization and “Westernization.” This major change in the United Arab Emirates (U.A.E.) has been magnified by a recent increase in trade with the West and the arrival of many Western expatriates. Moreover, part of this country’s oil wealth has been invested in education, as demonstrated by the creation of “university cities” in the different Emirates. One would expect that more education and greater modernization would contribute to the Westernization of youth, and an increase in self-transcendence (in post-materialistic values such as tolerance or protection of the environment) as proposed by Inglehart (1971). This thesis, however, has not been empirically validated. Whiteoak, Crawford and Mapstone (2006) surveyed 241 U.A.E. respondents and found that younger Arabs (under 30 years old) have more individualistic values than older Arabs (30 years old or over). These findings must be interpreted with caution, however, due to the weaknesses of the measurement tools used in their study. Whiteoak, Crawford and Mapstone’s (2006) instrument for measuring “individualism” includes items that are associated with different work value poles in our model. For example, one item in their study was “One’s loyalty should be first and foremost to oneself and one’s family.” “Loyalty to oneself” would be associated with the self-enhancement pole in our study, but “loyalty to family” would be associated with the conservation pole in our study. Furthermore according to Whiteoak, Crawford and Mapstone (2006: 82), “little is known about the influence of modernization and Westernization on attitudes and values in the U.A.E. and, in general, the Arab world.”

Given the methodological weaknesses of Whiteoak, Crawford and Mapstone’s (2006) study and the robustness of Inglehart’s theory (1971), it is reasonable to think that the older Arab generation will attach more importance to materialistic values and individualism (i.e., self-enhancement) than the new Arab generation. Therefore we pose the following hypotheses:

H3a: The new Arab generation scores lower on the self-enhancement pole of Wils, Luncasu and Waxin than the older Arab generation.

H3b: The new Arab generation scores higher on the self-transcendence pole of Wils, Luncasu and Waxin than the older Arab generation.

Authors that have studied generational differences in work values have found no or small differences in the Western world. Since the Arab world has not been studied very much in this regard, it is likely that there are no major generational differences in most other work values between the two Arab generations. Therefore we pose the following hypothesis:

H4: There are no differences in the scores on the other work value poles (openness to change and conservation) between the two Arab generations.

The Relationship between Work Values and Cultural Origin

Several studies have shown that individual work values differ according to cultural origin (Hofstede, 2001; Trompenaars, 1993; Akhtar, 2000). Very few studies have specifically addressed the work values of Quebecers or Arabs and none has compared the values of these two cultural groups.

Differences in Work Values in Quebec and in Arab Countries

Quebec culture, as a North American culture, is characterized by low uncertainty avoidance (Hofstede, 2001), and autonomy (Schwartz, 2006), which represent values associated with our openness-to-change pole. According to Inglehart (2008), individuals from developing countries attach more importance to materialist values (such as security) and individuals living in rich countries attach more importance to post-materialist values (such as self-expression), a finding that was validated by Egri and Ralston (2004). Quebecers are attached to personal growth, social solidarity and equality between individuals (Lachapelle and Gagné, 2003; Langlois, 1990), which are values associated with Wils, Luncasu and Waxin openness-to-change and self-transcendence poles.

Abdalla (1997: 397) states that “in the Arab culture, the dimensions of work ethic are probably different from those of the Western culture.” According to Schwartz (1999), Arab cultures cluster around the hierarchy and conservatism dimensions in his study, because they value things such as authority, social power, obedience, self discipline and devotion. According to Hofstede (2001), Arab cultures score higher on power distance and uncertainty avoidance than North American cultures. According to Whiteoak, Crawford and Mapstone (2006), Arabs value conformity, family tradition and loyalty to friends and several empirical studies have highlighted the importance of tradition and tribal identity in Arab countries (Ali and Al-Kazemi, 2005 in Kuwait; Ali and Al-Shakis, 1985 in Saudi Arabia). Sharabi’s study (2010) also found that in Israel, Arabs demonstrate more traditional values than Jews (i.e., Arabs attach greater importance to community and religion and less importance to leisure activities). So, based on the above mentioned studies, Arabs attach great importance to the conservation and self-enhancement poles of Wils, Luncasu and Waxin.

Effect Size of between-Country Differences in Work Values

However Gerhart and Fang (2005) showed that past research findings on cultural differences in work values between countries should be interpreted with caution, as “country differences explain only a small portion of variance in individual-level cultural values,” especially on small samples (Gerhart and Fang, 2005: 975). In other words, the size effect of the variable “country” on values is at best moderate. Gerhard and Fang (2005) focused on the work of Hofstede (2001) and raised important questions about the statistical analyses used. Based upon meta-analysis of the studies that assess the impact of culture, they found that the variable “culture” explained 2–4% of variance rather than up to 50% as originally claimed. These authors also showed that the variable “culture” only explains a small percentage of the variance in “individualism.” Given that Quebec and the U.A.E. are demographically diverse, and given our relatively small samples, work value differences between these two countries may be modest. Therefore, we pose the following hypothesis:

H5: There are no differences in the scores on the four work value poles between the Quebecers and the U.A.E. Arabs.

Methodology

We used a quantitative methodology (questionnaire survey) to validate the structuring of work values and test the hypotheses. In Quebec, participants answered an online questionnaire in French posted on the l’Observateur web site, and in the U.A.E., participants filled out a distributed questionnaire in English. Convenience samples were used in both cases.

The Quebec Sample

In Quebec, the invitation to complete the survey has been sent by Internet to 1864 individuals identified by the firm l’Observateur specialized in conducting opinion polls. The respondents, who participated on a voluntary basis, are heterogeneous in terms of age and gender and cover a wide range of occupations, industries and sectors. Among the 1864 individuals, 439 filled the questionnaire. Of these 439 questionnaires, only the 278 respondents who stated that they were born in Quebec were selected for this study, since we were interested in the values of native (French) Quebecers. In accordance with Schwartz’s recommendations (1992), we eliminated 75 questionnaires in which respondents chose the maximum rating (8: “essential to me”) for more than 37.5% of their answers on the questionnaire, since this is said to be indicative of a social desirability bias. We also eliminated 17 questionnaires in which a particular rating was chosen in more than 62.5% of answers by a respondent, since, according to Schwartz, this indicates that the respondent did not make a serious effort to rate the values. Thus, the Quebec sample that we used was made up of 186 questionnaires. Respondents were mostly women (74%). Twenty-eight percent (28%) had a high school diploma, 31% had a college diploma and 59% had a university degree.

The U.A.E. Sample

In the U.A.E., the questionnaire has been distributed to University students in management (third and fourth year) in a class room setting by the professor, and to employees and managers by personal contacts in 20 organizations located in the U.A.E., from different sectors. Participation was voluntary. This data collection strategy was justified by the fact that it is difficult to collect data through questionnaires in Arab countries (Robertson, Al-Khatib and Al-Habib, 2002; Kemp, 2010).We distributed 480 questionnaires, and received 445 completed ones. Of these, 65 questionnaires were eliminated due to a social desirability bias and 10 were eliminated due to a lack of effort in rating the values. The respondents were of different nationalities, and only Arab respondents were selected in this study. The final sample was made up of a total of 252 questionnaires (105 employees, 147 students), 47% of which were filled by women. The U.A.E. Arab respondents came from the following countries: 34.5% from the U.A.E., 17.5% from the other Gulf Arab states (Saudi Arabia, Kuwait, Qatar, Oman, Bahrain) and 48% from other Arab countries (Egypt, Syria, etc.). At the university as well as in organizations, English was the respondents’ working language. All the selected respondents were living in the U.A.E. and 98% were Muslims. Among employees, 37% had a postgraduate degree, 45% had a university degree, 18% a college degree, and 10% a high school degree. Regarding their professional status, 39% were employees, 18% up front management, 20% in middle management, and 23% senior management.

Measures

The work values of respondents were measured using an abridged version of the original Work Value Inventory (WVI) (Wils, Luncasu and Waxin, 2007). This short version of the WVI includes 28 work values distributed among four poles. The self-enhancement (SE) pole includes the following work values: social power (POWE), hierarchical authority (AUTH), money (MONE), success (SUCC), ambition (AMBI) and competition (COMP). The self-transcendence (ST) pole includes the following work values: friendship (FRIE), honesty (HONE), team spirit (TEAM), responsibility (RESP), devotion (DEVO), justice (JUST), tolerance (TOLE), protection of the environment (ENVI), and work/life balance (BALA). The conservation (CO) pole includes the following work values: respect for rules at work (RESR), respect for code of conduct (RESC), humility (HUMI), respect for authority (RESA), security (SECU), and good working conditions (COND). The openness-to-change (OC) pole includes the following work values: initiative (INIT), autonomy (AUTO), independence (INDE), stimulation (STIM), variety (VARI), risk (RISK), and pleasure at work (PLEA). Each respondent had to indicate how important each of the 28 work value items is to them. In the U.A.E., the importance of work values is measured using the WVI scale from −1 to 7: −1 = opposed to my values, 0 = not important at all, 7 = essential to me (Wils, Luncasu and Waxin, 2007). Due to constraints related to the computerized formatting of the online questionnaire, it was impossible to keep the original WVI rating scale for the Quebec study. Nevertheless, a very similar scale was used from 1 to 8: 1 = opposite to my values, 2 = not important, 8 = essential to me.

Respondents were asked to report their birth year. Generations in the Quebec study were recorded according to Lyons, Duxbury and Higgins (2007), Baby Boomers: 1946–1961, Generation X: 1962–1979, Generation Y: 1979–2004. This classification scheme has been adopted by several authors (Cennamo and Gardner, 2008). The Quebec sample was made up of 27% Generation Y, 47% Generation X and 26% Baby Boomers. Veterans were excluded due to few respondents (9). Generations in the U.A.E. study were classified according to Whiteoak, Crawford and Mapstone (2006), New Generation (under 30 years old): 76% of respondents and Older Generation (30 years old or over): 24% of respondents.

Statistical Analysis

To validate the structuring of work values, we used the multidimensional scaling analysis (with SYSTAT), based on “Guttman–Lingoes smallest space analysis or SSA,” as explained in Wils, Luncasu and Waxin (2007). Then we computed the average score of each respondent for each of the four work value poles (self-enhancement, self-transcendence, conservation and openness-to-change), by averaging the scores of the work values items included within each pole.

To investigate generational differences in work values, on both samples, we computed multivariate analyses of covariance (MANCOVA), with the four work value pole scores as dependent variables, and the category-based generation variable as independent variable (or factor). We also entered gender as a factor since it might have had an impact on the importance attached to values (Cennamo and Gardner, 2007; Lyons, Duxbury and Higgins, 2007). In accordance with Schwartz’s recommendations (1992), we controlled for the mean score of respondents (as a covariate) to take account of the fact that some respondents might have used only part of the rating scale (scores skewed towards the bottom or the top of the scale).

To investigate cultural differences in work values, we merged the Quebec and U.A.E. data bases. Given that the value-rating scales used in the two studies were slightly different, we standardized the respondents’ scores on the four work value poles. Then we performed a multivariate analysis (MANCOVA) with two factors (country and gender), controlling for age and mean score (also standardized).

Results

Testing the Structuring of Work Values (H1)

For the Quebec data, the validation of the work value structure required four SSAs. Four work value items (devotion, team spirit, humility and work/life balance) were eliminated, respectively on the first, second, third and fourth multidimensional analysis, because they were not found on the expected poles, but in adjacent ones. The 24 remaining values are distributed on the final map as illustrated in Figure 1, with 0.23 as the coefficient of alienation. The alpha coefficients were, respectively, 0.78 for the five work values of the conservation pole, 0.78 for the six work values of the self-enhancement pole, 0.70 for the seven work values of the openness to change pole, and 0.55 for the six work values of the self-transcendence pole. The use of ipsative measures (Wils, Luncasu and Waxin, 2007) showed that the correlation dynamics between the opposite poles was confirmed: the conservation pole opposed the openness-to-change pole (r = −0.64; p < 0.000), whereas the self-enhancement pole opposed the self-transcendence pole (r = −0.59; p < 0.000). However, the axes were not entirely independent since weak correlations existed between some adjacent poles. We conclude that the abridged version of our WVI is validated for the Quebec sample: H1a is supported.

Figure 1

Final Work Value Structure (SSa # 4) for the Quebec Sample (N = 186)

Final Work Value Structure (SSa # 4) for the Quebec Sample (N = 186)

-> See the list of figures

For the U.A.E. data, the analysis only required two SSAs. In the first multidimensional analysis, the responsibility work value was found on the conservation pole instead of the self-transcendence pole, so it was eliminated. The remaining 27 values were distributed on the final map as illustrated in Figure 2, with 0.25 as the coefficient of alienation. The alpha coefficients were, respectively, 0.68 for the seven work values of the conservation pole, 0.65 for the six work values of the self-enhancement pole, 0.68 for the seven work values of the openness to change pole, and 0.65 for the seven work values of the self-transcendence pole. The correlation dynamics between the opposite poles was also confirmed: the conservation pole opposed the openness to change pole (r = −0.65; p < 0.000) whereas the self-enhancement pole opposed the self-transcendence pole (r = −0.70; p < 0.000). However, the axes were not entirely independent. We conclude that the abridged version of our WVI is validated for the U.A.E. sample: H1b is supported.

Figure 2

Final Work Value Structure (SSa # 2) for the U.a.e. Sample (N = 252)

Final Work Value Structure (SSa # 2) for the U.a.e. Sample (N = 252)

-> See the list of figures

Testing the Generational Differences in Work Values (H2 to H4)

For the Quebec study, the multivariate analysis of covariance indicated that there were no significant main effects for generation (Wilks’s Lambda = 0.95; non-significant), which confirmed Hypothesis 2 (table 1A). As was the case in Lyons, Duxbury and Higgins’s study (2007), a significant main effect was observed for gender (Wilks’s Lambda = 0.95; p < 0.048). The univariate analysis of variance indicated that men scored lower on the conservation pole than women (6.25 vs. 6.53; F = 5.3; p < 0.02). A significant covariate effect was observed for mean score (Wilks’s Lambda = 0.07; p < 0.000).

Although there is no significant difference between generations in their scores on the four work value poles, there are significant generational differences for some specific work values, as other authors have shown. We performed an ANCOVA with each work value as dependent variable, generation and gender as factors, and mean score as a covariate. Our results show that only five work values vary significantly between generations: stimulating work, fun at work, autonomy, respect for the rules at work, and respect for code of conduct (see table 2). Baby Boomers attach less importance to stimulation and fun at work, but more importance to respecting rules and code of conduct than generations X and Y. Generation Y attaches less importance to autonomy than the two previous generations. However, the size effect is modest (partial eta squared between 0.04 and 0.07).

For the U.A.E. study, the multivariate analysis of covariance revealed that there was a significant main effect for generation (Wilks’s Lambda = 0.90; p < 0.000; partial eta squared = 0.095). As table 1B shows, the New Generation scores lower on the self-enhancement pole, which supports H3a. The New Generation also scored higher on the self-transcendence pole, which supports our hypothesis 3b. However, there are no significant generational differences in work values on the other axis: openness to change versus conservation, which supports H4.

A significant main effect was observed for mean score (Wilks’s Lambda = 0.015; p < 0.000). The effect of gender was almost significant (Wilks’s Lambda = 0.96; p < 0.056). It is interesting to note that Arab women attach more importance to self-transcendence than Arab men (4.11 vs. 3.76; F = 6.9; p < 0.009), but less importance to self-enhancement (4.37 vs. 4.75; F = 6.9; p < 0.009). In both cases, the effect is modest (partial eta squared = 0.003).

Table 1A and B

Mean Work Value Pole Scores (M*) and Standard Deviations (SD) for Each Generation in Quebec and in the U.A.E.

Mean Work Value Pole Scores (M*) and Standard Deviations (SD) for Each Generation in Quebec and in the U.A.E.
* Marginal means estimated after having controlled for mean score (MANCOVA )

-> See the list of tables

Table 2

Mean Work Value (M*) and Standard Deviation (SD) for Each Generation in Quebec

Mean Work Value (M*) and Standard Deviation (SD) for Each Generation in Quebec
* Marginal means estimated after having controlled for mean score (ANCOVA )

-> See the list of tables

As the Arab generations differ significantly on two poles, it is not surprising that the number of differences in specific work values is higher in the U.A.E. than in Quebec (10 versus 5). The new Arab generation attaches more importance to five values belonging to the self-transcendence pole: work life balance, justice, tolerance, friendship and honesty. Conversely, the new Arab generation attaches less importance than the older one to three work values belonging to the self-enhancement pole: social power, hierarchical authority and competition. Moreover, the new Arab generation attaches more importance to security and less to initiative than the older generation (see table 3). The size effect is modest for the U.A.E. (partial eta squared between 0.02 and 0.07).

Table 3

Mean Work Value (M*) and Standard Deviation (SD) for Each Generation in the U.A.E.

Mean Work Value (M*) and Standard Deviation (SD) for Each Generation in the U.A.E.
* Marginal means estimated after having controlled for mean score (ANCOVA )

-> See the list of tables

Testing the Cultural Differences in Work Values (H5)

The multivariate analysis of covariance indicated that there were no significant main effects for cultural origin, which supports H5. As table 4 shows, the mean scores by cultural origin were more or less the same.

Table 4

Mean Work Value Pole Scores and Standard Deviations for Each Cultural Sample

Mean Work Value Pole Scores and Standard Deviations for Each Cultural Sample

* Marginal means estimated after having controlled mean score and age (MANCOVA )

-> See the list of tables

A significant effect existed for gender (Wilks’s Lambda = 0.96; p < 0.001). Univariate analysis shows that women score higher on the conservation (F = 5.9; p < 0.015) and self-transcendence (F = 4.09; p < 0.044) poles, whereas men score higher on the self-enhancement pole (F = 14.27; p < 0.000). A significant effect was also found for mean score (Wilks’s Lambda = 0.01; significant at p < 0.000) and age (Wilks’s Lambda = 0.96; significant at p < 0.006). The scores on the self-transcendence pole decreased slightly with age (F = 5.8; p < 0.017) while the scores on the conservation pole increased slightly with age (F = 8.9; p < 0.003).

Finally, the interaction between gender and cultural origin was also significant (Wilks’s Lambda = 0.96; p < 0.001; partial eta square = 0.042). The univariate analysis of variance showed that this interaction was significant only for the self-transcendence pole (F = 9.65; p < 0.002). Men from Quebec scored higher than Arab men (the standardized scores are 0.13 vs. −0.26), whereas women from Quebec scored lower than Arab women on this pole (0.04 vs. 0.19).

Although there are no significant differences between countries in their scores on the four work value poles, there are significant cultural differences for two specific work values. We performed an ANCOVA with each specific work value as dependent variable, country and gender as factors, and age and mean score as covariates which demonstrated that Quebecers attach more importance to friendship than Arabs (0.15 vs. −0.08; F = 4.1; p = 0.04; partial eta squared = 0.01), and Arabs attach more importance than Quebecers to respect for rules at work (0.75 vs. −.016; F = 4.9; p = 0.027; partial eta squared = 0.012).

Discussion and Conclusion

Contributions

Contrary to Schwartz’s recommendation (1992), most of the authors studying fundamental values or work values have not validated the structuring of their values items. A first contribution of this research is that it validated the structuring of work values included in the short version of the WVI, on both Quebec and Arab samples, through multidimensional analyses. The short version of the WVI worked in a similar way to the original WVI, i.e. we have found four poles of work values arranged around two almost orthogonal axes. This instrument for measuring work values satisfies a pressing need, given that the construct validity of most current instruments that measure work value is unknown (the construct validity of the WVI was shown in Waxin, Wils and Labelle, 2009). Consequently, the availability of a short version of the WVI is likely to encourage research on the role of work values in HR management, and to allow for a synthesis of results. In particular, the WVI is a first step for assessing person-organization fit (De Clerck, Fontaine and Anseel, 2008).

Ali and Al-Owaihan (2008) claim that the Islamic work ethic excludes the possibility of negative correlation between the opposite poles of values, but we did find this in our study. For example, in Ali and Al-Owaihan’s (2008) study, the value “success” (associated with the self-enhancement pole in our model) is not at odds with the values of “justice” and “cooperation” (associated with our self-transcendence pole). The problem with the measuring instrument developed by Ali (1988: 579) is that it includes work value items that are associated with different poles of values, such as: “Good work benefits both one’s self and others,’’ or “Work is not an end in itself, but a means to foster personal growth and social relations” (authors’ italics). Given that Ali and Al-Owaihan (2008) did not validate the empirical structuring of their model, future studies should strive to better understand the correlation dynamics between work values in the Arab world.

Another major contribution of this research is that we demonstrated that the different generations in our Quebec sample have quite similar work values. This implies that the “clash of generations” predicted in the speculative literature may be unfounded. It may be a waste of time and effort to implement different HR management practices for different generations in Quebec according to our sample’s results. Unlike other empirical studies that have examined generational differences in work values, we validated the structuring of work values before comparing the different generations, controlled for the confounding factors of mean score and gender and eliminated biases due to social desirability and lack of effort in answering the questionnaire, all of which make these results more robust.

According to Hofstede (2001), Arab countries are characterized by high power distance. Although our Arab sample is not representative of the entire Arab world, it is interesting to note that our results seem to point to a change in values, with the new generation scoring higher on the self-transcendence pole than the older generation. A likely explanation for this interesting result is that the re-investment of oil revenues in education and social programs has increased the transition from a poor society to a rich society. According to Inglehart (2008), this transition would entail a shift from materialist values to post-materialist values, which may explain why the older generation scores higher on materialistic values, and the newer generation scores higher on post-materialistic values in our Arab sample. This result is important because it may reveal a change in work values in segments of the Arab world that should be investigated further.

Our results indicate that the new Arab generation scores lower on the “self-enhancement” pole than the older generation which contradicts the results found by Whiteoak, Crawford and Mapstone (2006). The fact that values associated with “individualism” in their study and the values associated with our “self-enhancement” pole do not exactly correspond may account for our divergent results. Further studies should be performed using other measures of individualism, such as that of Ramamoorthy and Flood (2002), in order to retest the relationships between work values, the Islamic work ethic and individualism.

Our results support the findings of Gerhart and Fang (2005), and demonstrate that the size effect of intercultural differences is small. In our study, cultural origin has no significant, direct effect on the average score on each work value pole. This result occurred because in each country there were respondents with a wide variety of work value preferences, so that respondents did not score consistently high on any single work value pole in either country. This result supports Gerhart and Fang’s (2005) findings that the impact of culture of origin is often exaggerated in HR management research. Gerhard and Fang (2005) examined the assumptions behind the use of culture as an explanatory variable of country differences in HR management. They focused on the work of Hofstede and raised important statistical questions about the treatment of data. Based upon meta-analyses of the studies that assess the impact of culture, they produced a reduced estimate of explained variance down from 50% to 2–4%. Our results show that scores for a few single work value items were significantly different between the cultures, albeit with a small size effect. Another interesting result was that there was a significant interaction between cultural origin and gender for the self-transcendence pole in our samples. This result indicates that the relationship between work values and cultural origin is complex. Further studies are thus needed to explore the relationship between cultural origin, gender and work values.

Limitations

A few limitations of our study need to be considered when interpreting our results. Caution should be used when generalizing our results, because we used convenience sampling. Firstly, the U.A.E. respondents were mainly students at an American University and included several Arab nationalities rather than one single nationality. These students are wealthy and open enough to choose an American style education. Our Arab sample should not be considered as representative of the Arab world. Secondly, education and job categories were not controlled for. Further studies might assess the impact of these variables on work values. Thirdly, due to the scarcity of studies in the Arabic world, defining boundaries between generations is a challenge that future research needs to address. Fourth, the Quebec sample should not be considered as a representative sample of Western culture. Another limitation stems from the use of a cross-sectional research design which does not allow to disentangle generational effects from those caused by age (maturation) effects.

Practical Implications

Despite these limitations, our study has several practical implications, since it suggests that HR management practices should not be different for different generations in Quebec. As pointed out by Pfeffer and Sutton (2006), HR practitioners must make decisions based on facts rather than beliefs. Our study also confirms the importance of selective hiring within high performance work systems (HPWS), because within each country and each generation, there were respondents with a wide variety of work value preferences, which would make it possible to recruit selectively in order to achieve the best fit between the work values of employees and the philosophical principles underlying HPWS. Although appointment in Gulf countries often depends on “wasta” or personal connections, our study supports the idea that these HR strategies may be just as viable in emerging as in developed countries.