Educational and Psychological Measurement http://epm.sagepub.com Examining Work Ethic Across Populations: A Comparison of the Multidimensional Work Ethic Profile Across Three Diverse Cultures David J. Woehr, Luis M. Arciniega and Doo H. Lim Educational and Psychological Measurement 2007; 67; 154 DOI: 10.1177/0013164406292036 The online version of this article can be found at: http://epm.sagepub.com/cgi/content/abstract/67/1/154 Published by: http://www.sagepublications.com Additional services and information for Educational and Psychological Measurement can be found at: Email Alerts: http://epm.sagepub.com/cgi/alerts Subscriptions: http://epm.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations (this article cites 17 articles hosted on the SAGE Journals Online and HighWire Press platforms): http://epm.sagepub.com/cgi/content/abstract/67/1/154#BIBL Downloaded from http://epm.sagepub.com at SAGE Publications on May 17, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Examining Work Ethic Across Populations Educational and Psychological Measurement Volume 67 Number 1 February 2007 154-168 © 2007 Sage Publications 10.11770013164406292036 http://epm.sagepub.com hosted at http://online.sagepub.com A Comparison of the Multidimensional Work Ethic Profile Across Three Diverse Cultures David J. Woehr The University of Tennessee Luis M. Arciniega Instituto Tecnológico Autónomo de México Doo H. Lim The University of Tennessee The authors examined the measurement equivalence of the Multidimensional Work Ethic Profile (MWEP) across the diverse cultures of Korea, Mexico, and the United States. Korean- and Spanish-language versions of the MWEP were developed and evaluated relative to the original English version of the measure. Confirmatory factor analytic results indicated measurement invariance across samples drawn from each country. Further analyses indicated potential substantive differences for some of the seven subscales of the MWEP across samples. The implications of these findings and directions for future research are presented. Keywords: work ethic; measurement equivalence; Multidimensional Work Ethic Profile; cross-cultural measurement M odern formulations of the work-ethic construct stem from the work of Max Weber (1904-1905/1958). In the now classic two-part essay “The Protestant Ethic and the Spirit of Capitalism,” Weber advanced the thesis that the introduction and rapid expansion of capitalism and the resulting industrialization in Western Europe and North America were in part the result of the Puritan value of asceticism and the belief in a calling from God (Byrne, 1990; Charlton, Mallinson, & Oakeshott, 1986; Fine, 1983; Furnham, 1990b; Green, 1968; Lehmann, 1993; Maccoby, 1983; Nord, Brief, Atieh, & Doherty, 1988; Poggi, 1983). It was the application of these values that Weber believed led to the “work ethic”: the complete and relentless devotion to one’s economic role on earth (Lessnoff, 1994). Stemming from Weber’s thesis, a great deal of research has focused on the nomological network surrounding work ethic as well as Authors’ Note: Correspondence concerning this article should be addressed to David J. Woehr, The University of Tennessee, Department of Management, Knoxville, TN 37922; djw@utk.edu. 154 Downloaded from http://epm.sagepub.com at SAGE Publications on May 17, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Woehr et al. / Work Ethic Across Populations 155 the comparison of work ethic across cultures (see Furnham, 1990b, for a summary). Recent literature has suggested that one of the primary factors limiting this research is the lack of common conceptualizations and measurement systems for the work-ethic construct. In an attempt to address this limitation, Miller, Woehr, and Hudspeth (2002) recently developed the Multidimensional Work Ethic Profile (MWEP). The MWEP is a multidimensional inventory assessing conceptually distinct components of work ethic. Our goal in the present study was to extend the work of Miller et al. by further examining the construct validity of MWEP scale scores. Specifically, we (a) report on the development of Spanish- and Korean-language versions of the MWEP; (b) provide a detailed assessment of the degree of measurement equivalence between the English-, Korean-, and Spanish-language versions with a multinational sample; and (c) explore potential substantive differences with respect to work ethic as measured with the MWEP between Korean, Mexican, and U.S. samples. The MWEP Miller et al. (2002) recently presented a historical and conceptual review of the work-ethic construct. Drawing on the large body of literature stemming from Weber’s (1904-1905/1958) original work, they posited that work ethic is not a single unitary construct but a constellation of attitudes and beliefs pertaining to work behavior. They suggested that the work-ethic construct (a) is multidimensional; (b) pertains to work and work-related activity in general, not specific to any particular job (yet may generalize to domains other than work, such as school, hobbies, etc.); (c) is learned; (d) refers to attitudes and beliefs (not necessarily behavior); (e) is a motivational construct reflected in behavior; and (f) is secular, not necessarily tied to any one set of religious beliefs. On the basis of previous literature as well as original empirical research, Miller et al. identified seven components or dimensions that they argued constitute the work-ethic construct. The dimensions posited are centrality of work, self-reliance, hard work, leisure, morality/ethics, delay of gratification, and wasted time (see Table 1 for a full definition of each dimension). Miller et al. (2002) also argued that previous measures of work ethic had been deficient to the extent that they did not sufficiently assess and/or differentiate among the various facets of work ethic. Consequently, they developed and provided initial support for a multidimensional work-ethic inventory: the MWEP. The MWEP purports to measure seven conceptually distinct (i.e., divergent) facets of work ethic. (Each of the seven facets, or dimensions, tapped by the MWEP, along with definitions and sample scale items, is presented in Table 1.) Although Miller et al. provided a great deal of evidence pertaining to the psychometric evaluation of the MWEP, they provided little or no evidence with respect to the appropriateness of the MWEP as a measurement tool across cultures (i.e., cross-cultural measurement invariance). This question of measurement invariance is a particularly critical one. Several authors (Furnham, 1990a; Jones, 1997; Miller et al., 2002; Niles, 1999) have Downloaded from http://epm.sagepub.com at SAGE Publications on May 17, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 156 Educational and Psychological Measurement Table 1 MWEP Dimensions, Dimension Definitions, and Sample Items Dimension Definition Sample Items Centrality of work Belief in work for work’s sake and the importance of work Self-reliance Striving for independence in one’s daily work Belief in the virtues of hard work • “Even if I inherited a great deal of money, I would continue to work somewhere.” • “It is very important for me to always be able to work.” • “I strive to be self-reliant.” • “Self-reliance is the key to being successful.” • “If you work hard you will succeed.” • “By simply working hard enough, one can achieve their goals.” • “People should have more leisure time to spend in relaxation.” • “The job that provides the most leisure time is the job for me.” • “People should be fair in their dealings with others.” • “It is never appropriate to take something that does not belong to you.” • “The best things in life are those you have to wait for.” • “If I want to buy something, I always wait until I can afford it.” • “I try to plan out my workday so as not to waste time.” • “Time should not be wasted, it should be used efficiently.” Hard work Leisure Proleisure attitudes and beliefs in the importance of nonwork activities Morality/ethics Believing in a just and moral existence Delay of gratification Orientation toward the future; the postponement of rewards Attitudes and beliefs reflecting active and productive use of time Wasted time Note: MWEP = Multidimensional Work Ethic Profile. argued that previous measures of work ethic are inadequate both conceptually and psychometrically. In addition, although a great deal of research has examined cultural differences with respect to work ethic, little if any of this research has examined the extent to which differences in work ethic are the result of actual differences or simply the result of a lack of measurement equivalence of the measures used. If this is the case, research that has attempted to make substantive interpretations on the basis of these measures is at best seriously flawed. In a recent review and integration of the literature on measurement equivalence in organizational research, Vandenberg and Lance (2000) stated, “Violations of measurement equivalence assumptions are as threatening to substantive interpretations as is an inability to demonstrate reliability and validity” (p. 6). Specifically, a lack of equivalence between groups indicates that a measure is not functioning the same across the groups, and any substantive interpretation of similarities or differences is suspect at best. Thus, although measurement equivalence is critical for any measurement instrument, it is even more so Downloaded from http://epm.sagepub.com at SAGE Publications on May 17, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Woehr et al. / Work Ethic Across Populations 157 for the MWEP, because it is presented as a measurement of work ethic that overcomes the problems associated with previous measures of the construct. Although a number of approaches have been used to evaluate measurement equivalence (cf. Hui & Triandis, 1985; Vandenberg & Lance, 2000), there has been general agreement that the multigroup confirmatory factor analytic (CFA) model (Jøreskog, 1971) provides the most powerful and versatile technique for testing cross-group measurement invariance. On the basis of their review, Vandenberg and Lance (2000) called for an increased application of measurement equivalence techniques before substantive comparisons were considered. Given the limitations of classical test theory approaches for assessing measurement equivalence, they also recommended the use of multigroup CFA techniques. Thus, we took this approach to the evaluation of multiple versions of the MWEP. Spanish- and Korean-Language Versions of the MWEP On the basis of the 2000 census data, the U.S. Census Bureau recently revealed that Hispanics constitute the largest minority group in the United States (Grieco & Cassidy, 2001). Hispanics also represent the largest linguistic minority group in the United States. In addition, recent migration continues to produce sizable percentages of Latin American immigrants and refugees with limited proficiency in English. This continuing trend increases the need, by researchers in the United States as well as in Latin America and Spain, for Spanish-language versions of measures of various constructs typically explored in psychological and organizational research. One such construct is that of work ethic. To facilitate cross-cultural research pertaining to the attitudes and beliefs constituting work ethic, we sought to develop and evaluate a Spanish-language version of the recently developed MWEP (Miller et al., 2002). Parallel to the English and Spanish versions, the Korean-language version of the MWEP was developed to examine the cross-cultural viability of the MWEP and to facilitate further cross-cultural research with respect to the work-ethic construct. As one of the newly industrializing countries in East Asia and as one of the most Westernized countries in Asia (Scarborough, 1998), South Korea has experienced an incredible record of economic growth and integration into the high-tech, modern world economy since the early 1960s. The economic success of South Korea, however, was affected by the Asian financial crisis between 1997 and 1999, and the closed economic system between government and business was compelled to move toward a system more open to foreign countries (Central Intelligence Agency, 2004). During the rapid modernization of the past several decades and the more recent economic crisis, South Korea has experienced major social changes, moving from its traditional society, in which Confucianism and Buddhism were dominant, to a modern and Westernized society. We believed that the development of a Korean-language version of the MWEP would provide a valuable tool to conduct a cross-cultural Downloaded from http://epm.sagepub.com at SAGE Publications on May 17, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 158 Educational and Psychological Measurement analysis of the role of work ethic on various workplace issues, including work performance, employee morale, motivation, and work stress. These three cultures differ not only structurally (e.g., language, predominant religion, economic and educational systems) but also with respect to characteristic cultural values (e.g., these countries differ considerably with respect to Hofstede’s [1991] dimensions of culture, especially individualism and power distance). Thus, to the extent that measurement equivalence was supported, we also sought to explore potential substantive differences with respect to work ethic as measured with the MWEP between Korean, Mexican, and U.S. samples. Method Participants Participants in the present study were drawn from three distinct populations. U.S. sample. The U.S. sample consisted of 238 employees recruited from four different private, nonmilitary organizations: (a) a bank (n = 58), (b) a car dealership (n = 89), (c) a newspaper (n = 27), and (d) a manufacturing company (n = 64). Fifty-one percent of the participants were male, and the mean age was 39.05 years (SD = 11.90 years, range = 18 to 76 years). Participants voluntarily and anonymously completed the original English-language version of the 65-item MWEP measure developed by Miller et al. (2002). (A copy of the complete scale is available from the first author and is included as an appendix in Miller et al.) Previous research has demonstrated measurement equivalence with respect to the MWEP across each of the four organizational samples, so we treated them as a single sample in the present study (Woehr & Miller, 1998). Mexican sample. The Mexican sample consisted of 208 full-time working adults living in the metropolitan area of Mexico City, all having at least 2 years of undergraduate studies. Forty-six percent of the participants were male, and the mean age was 28.86 years (SD = 8.21 years, range = 20 to 61 years). Participants voluntarily and anonymously completed the Spanish-language version MWEP developed specifically for this study. Korean sample. The Korean sample consisted of 412 adults working full-time and living in the metropolitan area of Seoul, South Korea. The participants were drawn from three large multinational corporations, but the vast majority of participants (74%) were from one of the three companies. Seventy-five percent of the participants were male, and 80% were between the ages of 27 and 35 years. Participants voluntarily and anonymously completed the Korean-language version MWEP developed specifically for this study. Downloaded from http://epm.sagepub.com at SAGE Publications on May 17, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Woehr et al. / Work Ethic Across Populations 159 Measures The MWEP. The MWEP is a 65-item self-report measure tapping seven conceptually distinct dimensions (see Table 1). Each of the seven dimensions is assessed with 10 items, with the exception of delay of gratification (7 items) and wasted time (8 items). Responses to all items are made on 5-point, Likert-type scales (1 = strongly disagree, 5 = strongly agree). Scores for each of the dimension scales are computed as the mean item response times 10 (i.e., all dimension scores are expressed on the same scale). Translation procedure. To develop the Spanish and Korean versions of the MWEP, a back-translation process was followed on the basis of the method proposed by Brislin (1986). First, two teams composed of two undergraduate students and two university professors (one of the authors as well as an additional colleague), all proficient in English but whose mother tongues were Spanish or Korean, independently translated each of the items of the questionnaire into Spanish and Korean. The four translators on each team were asked to use wording and grammar that could be understood by any adolescent. Once the four versions in each language were gathered and compared by each team, consolidated versions were developed through team consensus. Each of the two consolidated versions was then back-translated into English by four fully bilingual undergraduate students (two students per each language version) independently. Any discrepancies between the original version in English and the back-translated Korean and Spanish versions were analyzed and corrected by the researchers. During the translation processes, some items in English were indirectly translated into Korean because of the differences in sentence structure, phrase expressions, and grammar between the two languages. For example, the Korean translators used the appropriate meaning “If we work too hard, it may not leave a time to rest” in translating the original sentence “Work consumes our time too much and doesn’t leave a room for a break” into Korean. Regarding the translation for the Spanish version, we found no back-translation discrepancies. Finally, a group of eight undergraduate students were asked to classify each of the items according to the theoretical content of each of the seven variables measured by MWEP. This task was successfully completed. (Both the Spanish and Korean versions of the instrument are available from the authors.) Analysis To test the measurement equivalence of the MWEP across the three samples, we used a multigroup CFA application of LISREL 8.51 (Jøreskog & Sørbom, 2001) to test two models representing configural (Model 1) and metric (Model 2) invariance. Both models were operationalized as seven-factor models (corresponding to the seven MWEP dimensions) with the factors allowed to correlate and uncorrelated Downloaded from http://epm.sagepub.com at SAGE Publications on May 17, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 160 Educational and Psychological Measurement errors. Model 1 posited an equivalent factor structure (i.e., items related to the same factors) across groups and thus represented a test of configural invariance. If Model 1 was not supported, the interpretation was that the groups differed in terms of the number or composition of factors represented in the measurement instrument, and thus no further tests were warranted (i.e., the measure was not equivalent across groups). Model 2 was based on the same measurement model as Model 1 but with more constraints placed on the model parameters. That is, factor pattern coefficients for like items were constrained to be equal across groups. Here, the question was whether the relations between specific items and the underlying constructs or factors tapped by the items were the same across groups. Thus, Model 2 provided a test of metric invariance. If Model 2 was not supported, the interpretation was that the groups differed in terms of the extent to which items were viewed as indicative of the various dimensions, and thus no further tests were warranted (i.e., the measure was not equivalent across groups). Given that the models represented a parameter nested sequence, we initially used the χ2 test statistic and a difference-of-χ2 test to evaluate the fit of each model in the series. We also focused on four additional overall fit indices: Steiger’s (1990) root mean square error of approximation (RMSEA), McDonald’s (1989) noncentrality index (NCI), Steiger’s (1989) gamma hat, and the comparative fit index (CFI; Bentler, 1990). The RMSEA provides an overall test of model fit that compensates for the effect of model complexity. Browne and Cudeck (1993) suggested using the RMSEA as a measure of lack of fit per degree of freedom. Browne and Cudeck suggested that an RMSEA value of .05 or less indicates a close fit and that values up to .08 represent reasonable errors of approximation in a population. In addition, both the NCI and gamma hat represent absolute fit indices providing an indication of overall model fit, and the CFI is an incremental (comparative) measure of fit providing an indication of fit relative to a null model. All three of these indices range from 0 to 1, with higher values indicating better fit and values of .90 or greater typically interpreted as indicating acceptable levels of fit. More important, Cheung and Rensvold (2002) demonstrated that across a wide range of fit measures, the changes in fit from one group to another reflected in the CFI, gamma hat, and NCI provide the most robust statistics for testing the between-group invariance of CFA models. In a large-scale simulation, Cheung and Rensvold demonstrated that when testing across two groups, a change in the value of CFI smaller than or equal to .01 indicates that the null hypothesis of invariance should not be rejected (i.e., measurement equivalence). Similarly, critical values for the gamma hat and NCI are .001 and .02, respectively. Finally, given the relatively small sample sizes for each of the three groups, the large number of scale items, and the difficulties inherent in factor analyzing categorical item-level data (for detailed discussions of these problems, see Bernstein & Teng, 1989; Nunnally & Bernstein, 1994), we did not analyze item-level responses. Rather, we constructed three “item parcels” (composites based on subsets of items) Downloaded from http://epm.sagepub.com at SAGE Publications on May 17, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Woehr et al. / Work Ethic Across Populations 161 to serve as manifest indicators for each of the seven work-ethic factors. The rationale for this approach was to avoid the difficulties associated with categorical itemlevel data and to achieve a higher level of reliability for each of the scores on which the confirmatory factor analyses were based than would be realized from responses on each of the 65 individual items. The literature provides a good deal of support for this approach and suggests that the use of composite-level indicators leads to far more interpretable and meaningful results than an analytic approach based on large numbers of individual items (e.g., Bagozzi & Heatherton, 1994; Gibbons & Hocevar, 1998; Hall, Snell, & Foust, 1999; Landis, Beal, & Tesluk, 2000; Paik & Michael, 1999). In the present study, we formed the item parcels by summing 2 to 4 randomly selected items from each of the dimension scales. Specifically, for each scale containing 10 items, three parcels were formed as one set of 4 items and two sets of 3 items. If the scale contained 8 items, three parcels were formed as two sets of 3 items and one set of 2 items. If the scale contained 7 items, three parcels were formed as one set of 3 items and two sets of 2 items. Although the set of items comprising each of the indicators was randomly selected within dimensions, the same sets of items were used to form indicators across samples. Previous research has demonstrated that this random approach to parceling items is appropriate to the extent that all items are equivalent measures of the focal construct (Hall et al., 1999; Landis et al., 2000). Miller et al. (2002) provided evidence supporting the unidimensionality of each of the MWEP dimensions as well as the equivalence of items corresponding to each dimension. Thus, this random-item-parceling approach is appropriate with respect to the MWEP items. Results Fit indices for each of the two models (described above) are presented in Table 2. As expected, Model 1 provided the best overall level of fit relative to the other models. In addition, examination of the fit indices for Model 1 indicated an acceptable overall level of fit (i.e., although the χ2 value was statistically significant, the RMSEA was not statistically significantly different from a value of .05, indicating a close fit in the population). Thus, it appears that the seven-factor model with three item composite indicators per factor provided an adequate representation of the data in each of the three samples. Model 2 further restricted Model 1 by constraining like factor pattern coefficients to be equal across all three samples. Examination of the fit indices for Model 2 indicated relatively little decrement in fit relative to Model 1. Specifically, although the change in χ2 from Model 1 to Model 2 was statistically significant, the additional fit indices were nearly identical to those for Model 1. In addition, the change in fit from Model 1 to Model 2 was within the critical values suggested by Cheung and Rensvold (2002) for both the CFI and the NCI. Only the Downloaded from http://epm.sagepub.com at SAGE Publications on May 17, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 162 Downloaded from http://epm.sagepub.com at SAGE Publications on May 17, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 1,076.69* 1,039.86* 1,041.41* 1,024.49* 504 532 518 518 518 Model 1: configural invariance (same factor structure) Model 2: metric invariance (equal factor pattern coefficients) Model 2a: metric invariance (equal factor pattern coefficients) for U.S. and Mexican samples Model 2b: metric invariance (equal factor pattern coefficients) for U.S. and Korean samples Model 2: metric invariance (equal factor pattern coefficients) for Mexican and Korean samples 27.02* 43.94* 42.39* 79.22* — χ2 Differencea .062 (ns) .063 (ns) .063 (ns) .063 (ns) .062 (ns) RMSEAb .93 .93 .93 .93 .93 CFI .00 .00 .00 .00 — CFI Differencea .74 .74 .74 .73 .75 NCI .00 –.01 –.01 –.02 — NCI Differencea .993 .993 .993 .992 .994 Gamma Hat .001 .001 .001 .002 — Gamma Hat Differencea Note: MWEP = Multidimensional Work Ethic Profile; RMSEA = root mean square error of approximation; CFI = comparative fit index; NCI = noncentrality index. a. Difference for each model is relative to Model 1. b. p < .01 (provides a test of close fit of the model in the population, with a statistically significant result indicating a lack of fit). *p < .05. 997.47* df Model χ2 Table 2 Results for the Sequence of Measurement Invariance Tests for the MWEP Measure Woehr et al. / Work Ethic Across Populations 163 Table 3 Means and Standard Deviations for the Seven MWEP Scales by Sample U.S. Sample (n = 238) Centrality of work Delay of gratification Hard work Leisure Morality/ethics Self-reliance Wasted time Totalb Mexican Sample (n = 208) Korean Sample (n = 412) M SD M SD M SD η2a 37.44A 34.79A 38.47A 29.96A 44.51A 35.81A 38.04A 259.02A 6.17 7.20 6.98 7.03 5.57 6.86 5.82 30.52 39.26B 33.85A 39.19A 29.39A 44.41A 35.20A 38.23A 259.54A 4.91 6.50 5.49 5.91 3.83 6.54 5.07 23.11 40.98C 34.24A 38.51A 24.97B 40.82B 38.84B 39.56B 257.91A 4.49 5.52 4.87 5.39 4.40 4.27 4.23 20.33 .08 .01 .01 .14 .13 .08 .02 .01 Note: Means within the same row with different superscripts are statistically significantly different (p < .05). MWEP = Multidimensional Work Ethic Profile. a. Effect size estimate for group differences across all three groups. b. Total score is the sum of the seven work-ethic subscales. gamma hat measure exceeded the critical value suggested by Cheung and Rensvold. However, it is important to note that the critical values recommended by Cheung and Rensvold were based on model tests across two groups, and authors suggested that “suitability of the recommended GFIs for testing across three or more groups is an interesting topic for future study” (p. 251). In the present study, the models were evaluated across three groups. However, Models 2a to 2c in Table 2 represented tests of metric invariance across all possible two-group comparisons in the present study (i.e., United States vs. Mexico, United States vs. Korea, and Mexico vs. Korea). For each of these models, all of the fit measures indicated that the null hypothesis of invariance should not be rejected. Thus, it appears that the MWEP measure was conceptually equivalent across all three groups. To this point, the results suggest that the English, Korean, and Spanish versions of the MWEP are conceptually equivalent to the extent that item composites relate to the same factors to the same extent. Given the demonstrated level of measurement equivalence between the English, Korean, and Spanish versions of the MWEP, we next examined potential differences across groups with respect to the actual scores on the dimension subscales (as opposed to the latent factors). Mean scores by sample for each of the MWEP dimensions are presented in Table 3. Examination of mean differences across samples indicated no statistically significant differences across groups for two of the subscales (i.e., delay of gratification and hard work). For four of the seven subscales, no statistically significant differences were found between the U.S. and Mexican participants, but both groups were statistically significantly different when compared with the Korean participants (the Korean participants were statistically significantly higher Downloaded from http://epm.sagepub.com at SAGE Publications on May 17, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 164 Educational and Psychological Measurement Table 4 Work-Ethic Dimension Intercorrelations (coefficient α reliability estimates) by Sample Dimension U.S. sample (n = 238) 1. Centrality of work 2. Delay of gratification 3. Hard work 4. Leisure 5. Morality/ethics 6. Self-reliance 7. Wasted time Mexican sample (n = 208) 1. Centrality of work 2. Delay of gratification 3. Hard work 4. Leisure 5. Morality/ethics 6. Self-reliance 7. Wasted time Korean sample (n = 412) 1. Centrality of work 2. Delay of gratification 3. Hard work 4. Leisure 5. Morality/ethics 6. Self-reliance 7. Wasted time 1 2 3 4 5 6 7 (.81b) .45 .55 .31 .36 .33 .61 (.81c) .61 .10 .30 .43 .55 (.89a) .06 .42 .57 .59 (.87b) .06 –.04 .22 (.83b) .24 .47 (.86b) .37 (.77c) (.78c) .41 .44 .27 .47 .25 .60 (.75c) .34 .02 .24 .30 .37 (.83b) –.03 .33 .29 .45 (.82c) .28 –.11 .09 (.64f) .09 .45 (.85b) .21 (.72e) (.80b) .27 .56 .09 .43 .47 .64 (.73e) .40 –.25 .25 .34 .35 (.82b) –.01 .36 .46 .52 (.84b) .05 –.10 –.04 (.70e) .41 .43 (.79b) .53 (.68e) Note: Ninety-five percent confidence intervals for the reliability estimates are represented as follows: a = ±.02; b = ±.03; c = ±.04; d = ±.05; e = ±.06; f = ±.07. on the self-reliance and hard work dimensions and statistically significantly lower on the leisure and morality/ethics dimensions). Finally, all three groups were statistically significantly different on the centrality of work dimension, with the U.S. participants providing a statistically significantly lower value relative to the Mexican participants, who in turn were statistically significantly lower than the Korean participants. Finally, correlations among the seven MWEP dimension scores as well as internal consistency reliability estimates for each of the dimension scale scores are presented in Table 4. Examination of the estimates indicated that reliabilities were generally acceptable for all scale scores across samples (i.e., mean coefficient α = .79, range = .64 to .89). More specifically, examination of the 95% confidence intervals (Fan & Thompson, 2001; Henson & Thompson, 2002) around the reliability estimates (also presented in Table 4) indicated that all but five of the reliabilities were statistically significantly higher than the value of .70 often cited as indicative of a reasonable level Downloaded from http://epm.sagepub.com at SAGE Publications on May 17, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Woehr et al. / Work Ethic Across Populations 165 of reliability (Nunnally & Bernstein, 1994). The remaining five reliability estimates were not statistically significantly different from .70. In addition, estimates were relatively consistent across samples. That is, none of the estimates for each scale was statistically significantly different across samples. The only exception was for the morality/ethics dimension, for which the estimate in the U.S. sample was statistically higher than in both the Mexican and Korean samples, and the estimate in the Korean sample was statistically higher than the estimate in the Mexican sample. In summary, our results suggest that the newly developed Spanish- and Koreanlanguage versions of the MWEP demonstrate high degrees of measurement equivalence with the original English version. Specifically, a seven-factor measurement model corresponding to the seven MWEP dimensions with three manifest indicators (based on three summed item composites) and factor pattern coefficients constrained to be equal across groups provided a good representation of the MWEP data across all three versions administered in distinct samples. In addition, scale score reliability estimates were generally high and consistent across versions. Finally, we found statistically significant, albeit relatively small, substantive differences on the workethic dimensions across samples. Discussion Jones (1997) suggested that two indices of a theory’s importance are the length of time it continues to attract attention and the number of scholars who investigate it. One might add to these the extent to which a theory captures attention in popular as well as scientific culture. Using these indices, the ideas of Max Weber (1904-1905/1958) with regard to the construct of work ethic have certainly been among the most influential. Work ethic has been the focus of scholars in history, sociology, anthropology, political science, psychology, and organizational behavior. In addition, a great deal of literature has sought to address cross-cultural comparisons with respect to work ethic (Furnham, 1990b). Miller et al. (2002) facilitated the research on work ethic by providing a comprehensive and empirically supported multidimensional measure of the work-ethic construct. Our goal in the present study was to extend this work by developing and evaluating two new versions of Miller et al.’s measure of work ethic. In addition, we sought to apply a detailed CFA approach to the assessment of the measurement equivalence of the measure of work ethic across Korean, Mexican, and U.S. samples. Our results indicated that the original English and newly developed Korean and Spanish versions of the MWEP were equivalent measures across the three diverse populations we sampled. In essence, we found that the English-, Spanish-, and Koreanlanguage versions of the MWEP measured the same set of seven work-ethic dimensions originally proposed by Miller et al. (2002). We also found some evidence of substantive differences between U.S., Mexican, and Korean participants with respect to work ethic. Downloaded from http://epm.sagepub.com at SAGE Publications on May 17, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 166 Educational and Psychological Measurement What are the practical implications of these findings? First, our results support the construct validity of scores based on the two new versions of the MWEP. The two alternate-language versions assess Miller et al.’s (2002) conceptualization of work ethic to the same extent as the original English version. In addition, our results indicate that the Spanish and Korean versions of the MWEP may legitimately be used to make cross-group comparisons between English-, Spanish-, and Korean-speaking participants. Finally, we found evidence of differences both within and across groups. This suggests a variety of avenues for future research. Identifying both predictors as well as outcomes of these differences is certainly one such possibility. To what extent do differences in work ethic as measured by the MWEP translate into difference in work-related behavior? Are these relationships the same across cultures? These are only a few questions future research might address. Finally, our primary focus in the present study was on the development and psychometric evaluation of Korean- and Spanish-language versions of the MWEP. We emphasize a detailed approach to the assessment of measurement equivalence. This approach represents a valuable tool not only for assessing the conceptual equivalence of measures but also for addressing substantive cross-group comparisons. The measurement equivalence of measures used across cultures is a critical and often overlooked assumption in much cross-cultural research. We add our voice to those of previous authors in stressing the importance of demonstrating measurement equivalence before evaluating any substantive group differences. Given our focus in the present study, however, we did not pose any specific hypotheses with respect to work-ethic differences between Korean, Mexican, and U.S. samples. On the contrary, we sought out participants with as much within-group diversity as possible. All three groups were composed of participants from different organizations, with different educational levels, and with different socioeconomic backgrounds. Nevertheless, our results highlight some cross-group differences. These results should be considered preliminary. Our hope is that this study serves as a springboard for future research focusing on cross-cultural comparisons pertaining to work ethic. References Bagozzi, R. P., & Heatherton, T. T. (1994). A general approach to representing multi-faceted personality constructs: Application to self-esteem. Structural Equation Modeling, 1, 35-67. Bentler, P. M. (1990). Comparative fit indexes in structural models. Psychological Bulletin, 107, 238-246. Bernstein, I. H., & Teng, G. (1989). Factoring items and factoring scales are different: Spurious evidence for multidimensionality due to item categorization. Psychological Bulletin, 105, 467-477. Brislin, R. W. (1986). The wording and translation of research instruments. In W. J. Lonner & J. W. Berry (Eds.), Field methods in cross-cultural research (pp. 137-164). Newbury Park, CA: Sage. Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. A. Bollen & J. S. Long (Eds.), Testing of structural equation models (pp. 136-162). Newbury Park, CA: Sage. Byrne, E. F. (1990). Work, inc.: A philosophical inquiry. Philadelphia: Temple University Press. Downloaded from http://epm.sagepub.com at SAGE Publications on May 17, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Woehr et al. / Work Ethic Across Populations 167 Charlton, W., Mallinson, T., & Oakeshott, R. (1986). The Christian response to industrial capitalism. London: Sheed & Ward. Cheung, G. W., & Rensvold, R. B. (2002). Evaluating goodness-of-fit indexes for testing measurement invariance. Structural Equation Modeling, 9, 233-255. Central Intelligence Agency. (2004). The world fact book. Retrieved April 20, 2003, from http://www.cia .gov/cia/publications/factbook/geos/ks.html Fan, X., & Thompson, B. (2001). Confidence intervals about score reliability coefficients, please: An EPM guidelines editorial. Educational and Psychological Measurement, 61, 517-531. Fine, R. (1983). The Protestant ethic and the analytic ideal. Political Psychology, 4, 245-264. Furnham, A. (1990a). A content, correlational, and factor analytic study of seven questionnaire measures of the Protestant work ethic. Human Relations, 43, 383-399. Furnham, A. (1990b). The Protestant work ethic: The psychology of work-related beliefs and behaviours. London: Routledge. Gibbons, R., & Hocevar, D. (1998). Levels of aggregation in higher level confirmatory factor analysis: Application for academic self-concept. Structural Equation Modeling, 5, 377-390. Green, A. W. (1968). Sociology: An analysis of life in modern society. New York: McGraw-Hill. Grieco, E. M., & Cassidy, R. C. (2001). Census 2000 brief: Overview of race and Hispanic origin (C2KBR/01-1). Washington, DC: U.S. Census Bureau. Hall, R. J., Snell, A. F., & Foust, M. S. (1999). Item parceling strategies in SEM: Investigating the subtle effects of unmodeled secondary constructs. Organizational Research Methods, 2, 233-256. Henson, R. K., & Thompson, B. (2002). Characterizing measurement error in scores across studies: Some recommendations for conducting “reliability generalization” studies. Measurement and Evaluation in Counseling and Development, 35, 113-127. Hofstede, G. (1991). Cultures and organizations: Software of the mind. Berkshire, UK: McGraw-Hill. Hui, C. H., & Triandis, H. C. (1985). Measurement in cross-cultural psychology: A review and comparison of strategies. Journal of Cross-Cultural Psychology, 16, 131-152. Jones, H. B. (1997). The protestant ethic: Weber’s model and the empirical literature. Human Relations, 50, 757-778. Jøreskog, K. G. (1971). Simultaneous factor analysis in several populations. Psychometrika, 36, 409-426. Jøreskog, K. G., & Sørbom, D. (2001). LISREL 8.50 [Computer software]. Chicago: Scientific Software International. Landis, R. S., Beal, D. J., & Tesluk, P. E. (2000). A comparison of approaches to forming composite measures in structural equation models. Organizational Research Methods, 3, 186-207. Lehmann, H. (1993). The rise of capitalism: Weber versus Sombart. In H. Lehmann & G. Roth (Eds.), Weber’s Protestant ethic: Origins, evidence, contexts (pp. 195-208). Washington, DC: Cambridge University Press. Lessnoff, M. H. (1994). The spirit of capitalism and the Protestant ethic: An enquiry into the Weber thesis. Brookfield, VT: Edward Elgar. Maccoby, M. (1983). The managerial work ethic in America. In J. Barbash, R. J. Lapman, S. A. Levitan, & G. Tyler (Eds.), The work ethic—A critical analysis (pp. 183-196). Madison, WI: Industrial Relations Research Association. McDonald, R. P. (1989). An index of goodness-of-fit based on non-centrality. Journal of Classification, 6, 97-103. Miller, M. J., Woehr, D. J., & Hudspeth, N. (2002). The meaning and measurement of work ethic: Construction and initial validation of a multidimensional inventory. Journal of Vocational Behavior, 60, 451-489. Niles, F. S. (1999). Toward a cross-cultural understanding of work-related beliefs. Human Relations, 52, 855-867. Nord, W. R., Brief, A. P., Atieh, J. M., & Doherty, E. M. (1988). Work values and the conduct of organizational behavior. Research in Organizational Behavior, 10, 1-42. Downloaded from http://epm.sagepub.com at SAGE Publications on May 17, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 168 Educational and Psychological Measurement Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric theory (3rd ed). New York: McGraw-Hill. Paik, C., & Michael, W. B. (1999). A construct validity investigation of scores on a Japanese version of an academic self-concept scale for secondary school students. Educational and Psychological Measurement, 59, 98-110. Poggi, G. (1983). Calvinism and the capitalist spirit: Max Weber’s Protestant ethic. Amherst: University of Massachusetts Press. Scarborough, J. (1998). The origins of cultural differences and their impact on management. Westport, CT: Quorum. Steiger, J. H. (1989). EzPath: Causal modeling. Evanston, IL: SYSTAT. Steiger, J. H. (1990). Structural model evaluation and modification: An interval estimation approach. Multivariate Behavioral Research, 25, 193-180. Vandenburg, R. J., & Lance, C. E. (2000). A review and synthesis of the measurement invariance literature: Suggestions, practices, and recommendations for organizational research. Organizational Research Methods, 3, 4-69. Weber, M. (1958). The Protestant ethic and the spirit of capitalism (T. Parsons, Trans.). New York: Scribner. (Original work published 1904-1905) Woehr, D. J., & Miller, M. J. (1998). Understanding work ethic: A conceptual and psychometric examination across multiple samples. In Proceedings of the Sixth Bi-Annual Conference of the International Society for the Study of Work and Organizational Values: Cross-cultural perspectives on work values and organizational behavior: East meets West (pp. 397-401). Shreveport, LA: International Society for the Study of Work and Organizational Values. Downloaded from http://epm.sagepub.com at SAGE Publications on May 17, 2007 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. International Journal of Cross Cultural Management http://ccm.sagepub.com Brazilian National Culture, Organizational Culture and Cultural Agreement: Findings from a Multinational Company Adriana V. Garibaldi de Hilal International Journal of Cross Cultural Management 2006; 6; 139 DOI: 10.1177/1470595806066325 The online version of this article can be found at: http://ccm.sagepub.com/cgi/content/abstract/6/2/139 Published by: http://www.sagepublications.com Additional services and information for International Journal of Cross Cultural Management can be found at: Email Alerts: http://ccm.sagepub.com/cgi/alerts Subscriptions: http://ccm.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations (this article cites 13 articles hosted on the SAGE Journals Online and HighWire Press platforms): http://ccm.sagepub.com/cgi/content/abstract/6/2/139#BIBL Downloaded from http://ccm.sagepub.com at SAGE Publications on May 17, 2007 © 2006 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Cultural Perspectives CCM International Journal of 2006 Vol 6(2): 139–167 Cross Cultural Management Brazilian National Culture, Organizational Culture and Cultural Agreement Findings from a Multinational Company Adriana V. Garibaldi de Hilal Federal University of Rio de Janeiro – COPPEAD (Graduate School of Business), Brazil The present study looks into the organizational culture of a Brazilian company, concentrating on its main Brazilian branches as well as on its European, Latin American, Central American, North American and Asian branches, making a total sample of 36 cities and 1742 respondents. Results indicate the influence of national culture on organizational culture, as the dimensions found clearly reflect the ambiguity and double-edged ethic characteristic of Brazilian culture. This study also shows the importance of hierarchy, and of relational networks, which stresses the relevance of the cultural element in organizational structure and functioning. In brief, understanding the double-edged ethic that governs Brazilian culture helps us understand apparently different, ambiguous or even contradictory behaviors reflected in the organizational culture practices of a Brazilian company with international operations. Moreover, there is little empirical research that directly deals with what combination of factors makes individuals agree or disagree over their cultural viewpoints. Consequently, we consider that this study attempts to deal with that issue as the cultural clusters were obtained using a multivariate approach, using demographic variables and the identified organizational dimensions. Thus results suggest the organizational context may increase or reduce the probability of nationality affecting the cultural agreement of group members. ABSTRACT KEY WORDS • cross cultural research • cultural agreement • Latin American multinational • national culture • organizational culture During the 1980s and 1990s, culture became a widely discussed subject in organizations, when western organizational scientists became interested in the culture of their countries and on the links between culture and organizational forms of life (Morgan, Copyright © 2006 SAGE Publications www.sagepublications.com DOI: 10.1177/1470595806066325 Downloaded from http://ccm.sagepub.com at SAGE Publications on May 17, 2007 © 2006 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 140 International Journal of Cross Cultural Management 6(2) 1997) in order to explain the superior performance of Japanese companies as compared to North American ones. Some authors argued that the key to competitiveness lay in the possibility of organizational culture (OC) change (Deal and Kennedy, 1982; Ouchi, 1981; Pascale, 1981; Sathe, 1985), with the first specific studies dating back to the beginning of the 1970s (Clark, 1972; Pettigrew, 1973) and Schein (1985) formally articulating the conceptual framework to analyze and intervene in the culture of organizations. Culture is treated as a variable in the perspective of popular authors such as Peters and Waterman (1982), Deal and Kennedy (1982), and Schein (1985). On the other hand, culture is treated as a metaphor in the symbolic perspective expressed in the work of anthropologists like Geertz (1993), whose focus is on meaning. Other debate focuses on the possibilities or not of measuring OC. Organizational symbolists advocate the dense description of organizations, involving qualitative in-depth case studies, as opposed to qualitative but thin descriptions based on interviews with managers. However, it can be argued that thick descriptions are difficult to replicate and that the result is highly subjective, as ethnographic researchers consider that objectivity in organizational research is a myth (Ott, 1989). Within the quantitative approach, on the other hand, the biggest advantage of the use of survey techniques to study OC is the fact that the same methodology can be applied to different organizations in the same way, thus providing a basis for comparisons or generalizations. One of the drawbacks is the fact that there is no protection against overgeneralization (Denison, 1984). Those who prefer qualitative research argue that cultural processes are the result of unique social constructions and that they are, therefore, impossible to measure with quantitative standardized measures (Cooke and Rousseau, 1988). Among the intercultural scientists (e.g. Laurent, 1983; Maznevski, 1994; Tromp- enaars, 1993), Hofstede (1997, 2001) is one of the best-known authors. He has an intermediary position and argues that both methodological approaches have limitations and, as a consequence, should be seen as complementary. According to Hofstede (1997), the core of OC is in the practices shared by its members. Consequently, national cultures would differ mainly on their basic values, while OCs would differ more superficially in terms of their practices, which would be the visible parts of culture and could be manageable within certain limits. However, most studies on OC have considered the organization as a whole (Martin, 1992). This kind of research emphasizes the existence of a unique general OC and focuses on harmony and organizational consensus, instead of on conflict and subcultural consensus (Martin, 1992). Although most researchers accept the existence of organizational subcultures (Trice and Beyer, 1993), they have emphasized the homogeneity of culture and its cohesive function instead of its differentiation potential (Gregory, 1983). Given the perceived need for further research within organizations that takes into account clusters of ideologies, cultural forms and behaviors that identify groups of people within organizations (Trice and Beyer, 1993), the present study adopted a sub-cultural perspective, perceiving the organization as composed by a multiplicity of different subcultures. This perspective emphasizes the existence of differences, although it does not deny the possibility of the existence of consensus in relation to certain values (Martin, 1992). A critical factor that defines the relationship and the existence of sub-cultures is how much they reflect their own particular values. Here, possibilities offer a spectrum that can include great differences (which can imply deeply rooted conflicts and culture clashes) as well as groups that share similar values and have similar ways of perceiving Downloaded from http://ccm.sagepub.com at SAGE Publications on May 17, 2007 © 2006 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Garibaldi de Hilal: Brazilian Culture and Cultural Agreement 141 and interpreting the organizational context events (which would foster inter-group cooperation). According to Kilduff (1993), organizational members from different nationalities will probably tend to develop specific cultural perspectives while they modify and transform the cultural routines of the organization so as to adapt them to their own cultural bias, possibly using different practical approaches to deal with organizational problems. Thus the complexity of the internal environment of organizations with international operations increases the probability that their culture tends to differentiation; that is, multiple systems of meanings, and therefore of practices, tend to coexist simultaneously. The different systems of meanings, or the different sub-cultures, usually greatly affect the operations of those organizations. Groups with different cultural viewpoints tend to interpret and to respond to the same organizational events in different ways – accepting, modifying, questioning or even ignoring the rules and procedures that emanate from the dominant culture (Jermier et al., 1991). Additionally, ethnocentrism, which is the tendency to evaluate others according to our own cultural point of view (Rocha, 1991), increases the probability of misunderstandings that can result in undesirable conflict levels, thus affecting the performance of the organization (Gregory, 1983). Moreover, in a relational society such as the Brazilian one, we have to consider the influence of relational networks on the organizational culture and sub-cultures. Within the relational perspective interaction among actors can lead to a certain homogeneity of attitudes and practices, thus significantly influencing the existence of cultural agreement (Burkhardt, 1994), and therefore helping to define OC clusters. Consequently, based on Hofstede’s framework for OC, the first purpose of this study was to identify the OC dimensions (i.e. values and practices; where practices involve sym- bols, rituals and heroes) of a Brazilian company, concentrating on its main Brazilian branches as well as on its European, Latin American, Central American, North American and Asian branches. The second purpose of this study was to determine if values and practices are uniform in the sampled organization or if there are different organizational culture clusters. Organizational Culture Authors such as Schein (1992) and Pettigrew (1985) present OCs as implying shared values, and confusion derives from the fact that such literature does not make a clear distinction between the values of the leaders and those of the other members of the organization. Hofstede (1997) defines OC as the collective programming of the mind that distinguishes the members of an organization from those of another, and he argues that the values of the founders and of the main leaders undoubtedly shape OCs, but the ways in which these cultures affect the ordinary members of the organization would be limited to shared practices. Thus the values of founders and leaders would become the practices of the other members of the organization. In brief, what an individual has to learn when he or she joins an organization is mainly a question of practices, as most values are developed and learned in the family and at school. Hofstede et al. (1990) empirically derived six independent dimensions that describe the numerous organizational practices: (1) process oriented versus results oriented; (2) employee oriented versus job oriented; (3) parochial versus professional; (4) open system versus closed system; (5) loose control versus tight control; and (6) normative versus pragmatic. These six dimensions are descriptive but not prescriptive: no position in each of the six dimensions is intrinsically good or bad. What is good or bad depends on each case, on what is desired for the organization and on the Downloaded from http://ccm.sagepub.com at SAGE Publications on May 17, 2007 © 2006 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 142 International Journal of Cross Cultural Management 6(2) strategic options. Moreover, Hofstede pointed out that the OC dimensions he identified would not necessarily apply to any kind of organization in any country. OCs are gestalts and their knowledge can only be totally appreciated by insiders. However, a conceptual framework allows us to make significant comparisons between cultures of different organizations, or between the sub-cultures of different parts of the same organization. Brazilian Culture Overview Brazil is the only other country in the western hemisphere that has the continental proportions, the regional contrasts and the demographic diversity that can be compared to the US and Canada. According to Hess (1995), Brazil, in spite of its western-like institutions, is a country where western culture has mixed and mingled with non-western cultures for centuries. This mixture of western and non-western, as well as modern and traditional, is what Da Matta (1997a) has called the Brazilian dilemma, or what Brazilians call the Brazilian reality. Brazil is a country where institutions operate through personal relationships as much as general rules. Diversity is not the best word for describing Brazil and Brazilians; mixture is better. Brazil is a nation of mixed races (miscegenation), religions (syncretism), and cultures (diasporas, borderlands). In cultural anthropology and studies of Brazilian national culture, Da Matta (1997a, b) has influenced a number of scholars (such as David Hess, 1995; Roberto Kant de Lima, 1995; Livia Neves Barbosa, 1995; Rosane Prado, 1995; and Martha de Ulhoa Carvalho, 1995) with his framework for interpreting Brazilian culture. Hess (1995) describes Brazil as the product of a particular colonial legacy that includes a class of wealthy landowners who supported a highly centralized Portuguese state. In turn the state imposed a latifundia, or plantation agricultural system in Brazil, where the plantations were controlled by patriarchs who exercised nearly absolute authority over their dominions. According to Buarque de Holanda (1995), the colonial legacy also includes the origins of the traditional Latin American personalism,1 the lack of social cohesion and the looseness of the institutions. Additionally, the Tocquevillian legacy of comparative analysis influenced a number of 20th-century thinkers such as Louis Dumont (1980). Dumont’s studies focused on two key dimensions for comparing values and patterns of social relations across societies: hierarchy and equality, and holism and individualism. In the ascribed form of hierarchy used by Dumont, one’s social position is assigned at birth or is limited by one’s family position. In a traditional hierarchical society, laws apply differently to different groups of people. Of course, there are remnants of the ascribed kind of hierarchy even in the most modern of societies, but the legal recognition of such hierarchy is considered an affront to the fundamental value of equality. The concepts of holism and individualism are closely related to those of hierarchy and equality. In a hierarchical society everyone occupies a definite position, and people’s identity is rooted in their association with a particular position in society. Da Matta’s approach to Brazilian culture departs from these key concepts as developed by Dumont. Da Matta uses the term persons to describe the category of identity, in which one is defined by one’s position in the family or in a hierarchically ordered social group. In contrast, in an individualistic society identity is rooted in one’s own life history and choices and people are individuals linked by the rules of the game, which are assumed to apply equally to all (or universally). Although in an individualistic society people certainly have personalistic loyalties, one’s identity as an individual rather than as a person tends to prevail. Likewise, in a personalistic or relational society, there are domains of society that operate according to individualistic and egal- Downloaded from http://ccm.sagepub.com at SAGE Publications on May 17, 2007 © 2006 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Garibaldi de Hilal: Brazilian Culture and Cultural Agreement 143 itarian principles, but, in general, personal loyalties tend to prevail. Da Matta argues that Brazil is somewhere between the two ideal polar extremes (hierarchical and holistic, and egalitarian and individualistic). He rejects the model of two Brazils, in which a traditional culture located in the lower classes of the cities and in the rural areas is opposed to a modern Brazil in the upper classes and in the big cities, showing how in societies like Brazil, Dumont’s distinctions can be applied simultaneously throughout the society. Instead of working with an either or model, he opted for a both and model, as both tendencies are present in any number of social groups, institutions and practices. Thus Brazilians are constantly negotiating between a modern, egalitarian code and a traditional one. In some situations, modern practices predominate. However, frequently, hierarchical and personalistic/relational practices encompass modern ones. Thus Brazil is neither modern nor traditional but both. Da Matta also developed an analysis of intermediary terms or symbols, as in Brazil there is a tendency to move toward a middle ground of mediation and ambiguity. Those mediating terms become sites for the conflict of values and the encompassment of the modern by the traditional. This seems to apply to Latin American countries as a whole, as Latin American history and social structure seem to imply. For example, the injustices of the Latin American authoritarian and hierarchical system are blunted by the existence of a number of mediating institutions: extended kin networks, nepotism, the famous Brazilian jeitinho or the Argentine gauchada (the art of bending rules), and all sorts of social practices that would appear corrupt in North America and Western Europe. In short, personal relationships form the flip side of official hierarchies. Personalism is more than a cultural system that gives people a social address in the hierarchical society; it is also a resource that people can use to get around the official rules of the hierarchical society. Of course, personalism does not work the same way for everyone. The networks of the weak are usually smaller and less influential. As a result, although personalism can be used as a resource to subvert hierarchy, as an overall system it ends up reproducing the general hierarchical order (Hess, 1995). Perhaps the most well known of Da Matta’s studies of mediation is his discussion of the street and the home (1997a). The space of the home is identified with the hierarchical and relational/personalistic moral world, whereas that of the street is egalitarian and individualistic. Of course, in Latin America, and especially in Brazil, the two worlds of home and street interact considerably. As a social space, the home, and institutions modeled on the home, such as the workplace, are places where relations among family members and servants or among superiors and subordinates institute hierarchies of race, class, age and gender. The street is a different sort of place where those hierarchies are suspended. The street is the place where the egalitarian and individualistic principles of the marketplace or legal system are in operation. The home is the place where people find their identity, while the street is the place of individual anonymity. In certain situations the home encompasses the street and all matters are treated in a personal, familiar domestic way; in others the street encompasses the home: the domain of personal relations is totally submersed and the axis of impersonal laws and rules prevails. There is, therefore, a double-edged ethic that operates simultaneously and that determines different behaviors that apply to the street (where behavior is free of the sense of loyalty, free of the meaning of us, ruled by the criteria of individualism, by laws and by the rules of the market) and to the home (where behavior is ruled by personal relations, the sense of loyalty and emotions, by reciprocity and friendship). In brief, in a dynamic sense, behaviors continually oscillate in Brazil in particular Downloaded from http://ccm.sagepub.com at SAGE Publications on May 17, 2007 © 2006 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 144 International Journal of Cross Cultural Management 6(2) and in Latin America in general: people can express apparently different or even contradictory opinions and behaviors depending on whether they position themselves in the street or in the home. Methodology The methodology used in this study was based on a research design that combined quantitative research with a qualitative exploratory procedure. The research took place in a Brazilian company that, owing to issues of confidentiality, will be called company XYZ. The main criterion used in its selection was the fact that it has approximately 81,500 employees, thus allowing us to replicate the research design used by Hofstede et al. (1990) in their OC study. The field research was done in 36 cities, including 17 in Brazil2 as well as the international branches located in Europe (namely Milan, London, Securities UK (also in London), Lisbon, Madrid, Paris, Frankfurt, Vienna and Amsterdam); in Latin America (namely Buenos Aires, the capital city of Argentina, Asuncion and Ciudad del Este – both cities located in Paraguay – Santiago, the capital city of Chile and La Paz, the capital city of Bolivia); in Central and North America (namely Panama, Grand Cayman, Miami and New York); and in Asia (namely Tokyo). The qualitative exploratory research aimed at collecting information and trying to gain some insights about the specific features of XYZ in order to adjust the contents of the questionnaire to the specificities of this organization, and as input for the interpretation of the quantitative data. It consisted of six one and a half hour in-depth semi-structured interviews conducted in Portuguese by the researcher. For the interviews, six managerial level employees were selected using as selection criteria the fact that they were reputed to be communicative and had the necessary experience and knowledge. The survey sample was defined in two stages. For the first stage, involving the selection of the cities or units where the data were collected, we used intentional sampling to cover the five Brazilian geographical areas. The initial intention of using stratified probabilistic sampling in the Brazilian units had to be discarded because the data collection had to be tailored to the needs of each regional division to cause the minimum interference in the work flow. Consequently, in each of the Brazilian units we selected an intentional sample of, on average, 74 employees (including, on average, 37 managerial employees and 37 non-managerial employees per unit). In the European, Latin American, Central American, North American and Asian units, as there were fewer employees per functional category, the research took the form of a census and hence included all the managerial and non-managerial employees present at the time of the survey. The total surveyed sample was of 1968 respondents. After eliminating the questionnaires that were annulled or not returned, we were left with a final sample of 1742 respondents (including, on average, 33 managerial employees and 33 non-managerial employees per Brazilian unit). The survey consisted of 131 pre-coded questions3 plus an open question. Questions were extracted from various publications on Hofstede’s questionnaire on OC. However, certain questions were developed based on the results of the qualitative phase of the study. The questionnaire also included five questions on demographics: sex, age, number of years working for the company, educational level and nationality. To assure that the questionnaire mostly reproduced an instrument already used by Hofstede et al. (1990), Geert Hofstede was personally consulted and directly involved. The anonymous self-administered questionnaire was prepared in several versions: Portuguese, English, Spanish, Dutch, German, Italian, French and Japanese. Following Downloaded from http://ccm.sagepub.com at SAGE Publications on May 17, 2007 © 2006 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Garibaldi de Hilal: Brazilian Culture and Cultural Agreement 145 Adler (1982) we used back-translation. After being translated, the questionnaire was pretested with a small group of retired XYZ employees to check that the understanding was the desired one. Following Robinson (1950) and Shweder (1973), in order to obtain etic or comparable OC dimensions we prepared an ecological matrix, calculating the mean of each item or variable for each of the 36 units.4 Subsequently, we applied ecological factor analysis (principal component) with orthogonal varimax rotation to provide a factor solution explaining the maximum share of variance to the fewest number of factors. Moreover, ecological factor analyses are characterized by flat matrixes (matrixes with fewer cases than the number of variables).5 However, the original database had in fact 1742 respondents, and not just 36 cases, thus being considered adequate using Hofstede et al.’s (1990) criterion. Considering the fact that the questionnaire mostly reproduced an instrument already used by Hofstede et al. (1990) and that the small modifications introduced did not affect its design or any of the variables held to be key by Hofstede, we considered that the constructs were already validated. Moreover, internal consistency was also verified based on the existing literature (such as Blake and Mouton, 1964; Burns and Stalker, 1961; and Peters and Waterman, 1982). It should be noticed that the six practice dimensions identified by Hofstede et al. (1990) together explained 73% of the variance. The first step was to calculate the 131 × 131 product moment correlation matrix of the 20 mean scores for each possible pair of questions, verifying that the matrix was appropriate for multivariate analysis as, on average, it presented mean correlations between the variables. For analytic purposes we followed Hofstede et al.’s (1990) recommendations. First, we divided the questions into three categories (57 questions on values; 61 on perceived practices and typical-member scores; and 13 on reasons for promotion and dismissal) and conducted separate factorial analyses for each category. As the ecological correlations tend to be stronger than individual correlations we expected to find high percentages of explained variance. In order to avoid giving undue attention to trivial things in ecological factorial analysis, the number of factors should be fewer than the number of cases and fewer than what is technically possible based on eigenvalues larger than 1, only taking into account variables with loadings higher than 0.5 or 0.6. Then the scores of each of the 13 identified dimensions (4 on values, 6 on practices and 3 on heroes) were put in a 0–100 scale using the formula: Final score (0,100) = (original score × 50 / 3.090245) + 50 (in order to have approximately 99.9% of the observations within the interval of the normal curve). Following this scores were ranked to better visualize the relative position of each unit in relation to the others (with 1 indicating the highest score and 36 the lowest). It should be noticed that three scores presented values outside the 0–100 range, and were considered as outliers: namely Amsterdam with a score for factor V3 (work centrality) of –12; Madrid with a score for factor P3 (individualistic × relational) of –19; and, Frankfurt with a score for factor V4 (need for survival) of 106, indicating a strong relationship directly or inversely (depending on the value being positive or negative) linked to the meaning of the corresponding factor. Next we calculated the product-moment correlation matrix of the 13 dimensions plus the 5 demographic variables for the 36 units, in order to identify significant correlations at the .05 level. In order to identify clusters of cultural agreement we submitted the 13 OC dimensions plus the 5 demographic variables for the 36 units to a hierarchical cluster analysis, using Ward Method and square Euclidean distance. From the resulting Downloaded from http://ccm.sagepub.com at SAGE Publications on May 17, 2007 © 2006 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 146 International Journal of Cross Cultural Management 6(2) dendrogram we selected four clusters. To explain the main features of each cluster we used the option reports – case summaries (SPSS 11.0 program), computing, for each cluster, the minimum, maximum and mean value of each of the 13 OC dimensions and of the 5 demographic variables. Finally, in order to verify if the identified clusters were in fact significant, we submitted the original 131 variables (except the demographic variables as they had already been used in the cluster analysis), to Multiple Discriminant Analysis – MDA – stepwise procedure and Wilks Lambda Method, in spite of this involving a certain degree of circularity. Results The first purpose of the analysis was to identify the OC dimensions (i.e. values and practices, where practices involve symbols, rituals and heroes). The second purpose of this study was to identify clusters of cultural agreement. Based on the 131 survey pre-coded questions, the 131×131 product-moment correlation matrix showed that values correlated with other values and also with practices; perceived practices and typical-member scores correlated among each other; and the reasons for promotion and dismissal correlated among each other, but also with other items. Before analyzing the value, practice and hero dimensions a word of caution is necessary. In the Brazilian branches (PC4 to PC20), almost 100% of the employees are Brazilian and the vast majority are local, from the region where the unit is located. Exceptions are PCs 18 and 19 (namely Campo Grande and Cuiabá in the CenterWest region of Brazil) that only have 40% and 50%, respectively, of local employees, the rest belonging to the other Brazilian regions. However, the same situation is not necessarily true of the non-Brazilian branches, where the local employees reflect the percentage of employees originally from the country where the unit is located. Thus the composition of local as opposed to Brazilian employees in the non-Brazilian branches has to be taken into account when analyzing the value, practice and hero dimensions. Table 1 indicates the approximate percentages of local and Brazilian employees in each of the non-Brazilian branches. Moreover, very few of the Brazilian employees belong to headquarters – usually only the top managers – while all the others are usually hired locally according to local regulations, and are hardly ever transferred between branches. Value Dimensions We obtained the following four independent factors that together explained 52.41% of the variance: V1 – Need for security V2 – Need for authority V3 – Work centrality (the importance of work) V4 – Need for survival. Tables 2 to 5 show the variables with loadings approximately higher than 0.50 or 0.60 that were considered to explain each factor. It should be noticed that items with negative loadings were reworded negatively. Factor V1 – need for security – which is a combination of collectivism and elements of uncertainty avoidance (see items in Table 2) shows that in relational societies, people are part of personal networks from which they derive their identity, thus justifying the importance given to good physical working conditions, to having training opportunities, to cooperation between co-workers and to having a good relationship with the hierarchical superior. On the other hand, relational and hierarchical societies usually justify the fact that it is not considered important for employees to be consulted by their direct superior in decisions, or to have the freedom to adopt their own approach to work. Downloaded from http://ccm.sagepub.com at SAGE Publications on May 17, 2007 © 2006 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Garibaldi de Hilal: Brazilian Culture and Cultural Agreement 147 Table 1 Percentage of local and of Brazilian employees per branch PC Unit 1 2 3 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 Santiago Asuncion Buenos Aires Ciudad del Este La Paz Panama Grand Cayman Miami New York Tokyo London Securities UK Lisboa Madrid Paris Milan Frankfurt Vienna Amsterdam Table 2 % of local employees % of Brazilian employees 90 93 88 93 90 76 20 19 28 29 58 36 88 45 64 86 75 50 85 10 7 6 7 10 18 40 70 42 62 35 55 12 55 18 14 25 40 15 % of other origin 6 6 40 11 30 9 7 9 18 10 Total % 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 V1 – Need for security (high) Code Loading OT9 OT15 OT19 CG17 OT8 OT7 OT20 OT11 CG10 0.840 0.829 0.809 0.740 0.734 0.707 0.693 0.677 0.670 CG23 CG28 OT5 OT4 CG21 FV6 CG22 CG16 0.665 0.648 0.633 0.626 0.624 0.597 0.593 0.568 Description Being consulted by direct superior in his or her decisions not important Having an element of adventure and variety in the job not important Having training opportunities important For young people to be critical of their teachers is all right Working with people who cooperate well with one another important Having freedom to adopt your own approach to job not important Having good fringe benefits important Having opportunities for high earnings important The employee that quietly does his or her duty is not one of the greatest assets of an organization Both parties compromising a bit best resolves conflicts with opponents Large corporation is a more desirable place to work than small company Having good working relationship with superior important Having good physical working conditions important People like work Would continue working if did not need the money Parents satisfied when children become independent Good personal relationships not more important than high income Downloaded from http://ccm.sagepub.com at SAGE Publications on May 17, 2007 © 2006 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 148 International Journal of Cross Cultural Management 6(2) Table 3 V2 – Need for authority (high) Code Loading FV1 CG1’9 CG6 OT12 FV2 CG9 CG15 OT10 OT1 OT13 FV4 OT2 CG5 0.723 0.710 0.689 0.664 0.660 0.651 0.633 0.633 0.628 0.610 0.599 0.577 0.564 OT18 0.552 OT6 0.520 Table 4 Description Does not prefer consultative/participative manager Rules should not be broken Decisions by individuals are usually of higher quality than group decisions Serving your country important Own manager autocratic or paternalistic Management authority should not be questioned It is often necessary to bypass hierarchical lines Making contributions to success of organization not important Having sufficient time for personal and family life not important Living in an area desirable to you and family not important Subordinates frequently afraid to express disagreement with superiors Having challenging things to do not important Main reason for having hierarchical structure is that everyone knows who has authority over whom Working in a well defined work situation where requirements are clear is important Having security of employment important V3 – Work centrality (high) Code Loading CG8 CG26 CG18 0.818 0.679 0.638 Description CG20 0.627 OT16 0.620 CG12 0.599 FV7 0.580 OT14 0.525 OT11 0.520 2nd loading Table 5 Parents should stimulate their children to be best in class Having a job you like not more important than having a successful career The individual who pursues his or her own interest makes the best possible contribution to society as a whole When people fail it is often their own fault Working in a prestigious and successful company important Competition between employees does not do more harm than good Feel proud working for this organization Having opportunities for advancement to higher level jobs important Having opportunities for higher earnings important V4 – Need for survival (high) Code Loading CG25 FV5 CG1 CG27 CG3 CG7 0.618 0.607 0.557 0.528 0.516 0.509 Description Even a lousy job is better than no job at all Intend to continue working for this organization until retirement Most people can be trusted One’s job is more important than one’s leisure time Most organizations would be better off if conflict could be eliminated forever A good manager should have precise answers to most of the questions that subordinates may raise about their work Downloaded from http://ccm.sagepub.com at SAGE Publications on May 17, 2007 © 2006 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Garibaldi de Hilal: Brazilian Culture and Cultural Agreement 149 The need for security also justifies the fact that having an element of adventure and variety in the job is not important, that a large corporation is a more desirable place to work than a small company, that having good fringe benefits and opportunities for high earnings is important and that compromise is the best way to resolve conflict with opponents. However, the items that state that the employees would continue working even if they did not need the money, that parents should be satisfied when children become independent, that people like work and that good personal relationships are not more important than a high income, all typical of individualistic and capitalistic modern societies, show an apparent contradiction with the high need for security. In fact those statements are representative of the traits of duality and ambiguity usually present in Latin American societies. In connection with V2 – need for authority (Table 3) – which clearly relates to power distance, the following comments are appropriate. The fact that it is often necessary to bypass the hierarchical lines fits the famous ‘jeitinho brasileiro’ or ‘gauchada argentina’; that is, the Latin way of bypassing rules in order to get things done, typical of relational societies where excessive formalism is bypassed, in practice, alleviating pressures and emphasizing the importance of personal relationships. That is in apparent conflict with the item that states that rules should not be broken, typical of hierarchical societies. However, those statements are representative of the traits of duality and ambiguity usually present in Latin American societies. The item that states that making a real contribution to the organization’s success is not important shows a vision of the organization as the ‘street’ in opposition to the ‘home’. According to Da Matta (1997a), we live in a society where there is a permanent state of confrontation between the public world of universal laws and the market and the private universe of the family, relatives and friends. Additionally, in connection with the preferred and perceived type of manager, which includes the typology of autocratic, paternalistic, consultative and participative manager (key element of the classical power distance dimension) our study indicates a clear preference for the autocratic and paternalistic types. However, careful analysis of the scores of the 36 units makes us realize that, while those statements would be mainly valid for the other Latin American branches (which have the highest scores and therefore a higher need for authority), the same does not apply to any of the Brazilian units, where the scores are consistently below the midpoint of the scale (50), indicating that both the preferred and perceived managers tend to be consultative or participative. Moreover, it is worth mentioning that sometimes the limited experience of the respondents can influence their perception of the type of manager they in fact have. One also has to consider that management training courses normally glorify consultative or participative management as being superior and more modern rather than more autocratic or paternalistic management styles, not considering the cultural profile of the audience. This could suggest the existence of consultative or participative rituals, without necessarily implying the implementation of managerial models that are actually consultative or participative. From this perspective, consultative or participative management styles might be reduced to a ritualistic representation of participation just because they are perceived as the politically correct discourse by the managerial establishment. The items that state that decisions by individuals are usually of higher quality than group decisions, that management authority should not be questioned, that subordinates are frequently afraid to express disagreement with their superiors and that the main reason for having a hierarchical structure is that Downloaded from http://ccm.sagepub.com at SAGE Publications on May 17, 2007 © 2006 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 150 International Journal of Cross Cultural Management 6(2) Table 6 Value scores and rankings PC V1 V2 V3 V4 RV1 RV2 RV3 RV4 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 69 70 75 64 66 61 61 65 61 60 62 60 56 60 61 51 63 66 63 61 41 37 36 31 43 45 15 38 28 32 23 17 32 43 48 36 88 83 91 45 40 38 30 42 39 45 41 39 28 37 37 40 49 36 41 31 72 53 63 45 41 70 57 53 56 58 44 35 49 51 76 58 38 53 45 50 55 52 40 32 43 57 54 45 48 64 60 78 52 53 59 54 67 49 37 44 33 72 70 63 55 51 49 53 52 16 68 –12 33 60 42 47 59 48 55 32 41 40 62 47 43 59 46 47 46 57 36 43 56 53 42 75 106 44 51 48 40 35 45 13 64 74 77 35 3 2 1 7 5 14 12 6 13 18 10 16 19 17 15 20 8 4 9 11 25 27 29 32 24 22 36 26 33 30 34 35 31 23 21 28 2 3 1 17 26 29 35 21 28 19 23 27 36 30 31 25 16 32 22 34 5 13 7 18 24 6 10 12 11 8 20 33 15 14 4 9 31 16 26 22 12 20 30 34 29 10 14 27 25 6 8 1 18 17 9 13 5 24 32 28 33 2 3 7 11 21 23 15 19 35 4 36 34 7 27 19 9 16 12 35 28 29 6 18 24 8 20 17 21 10 31 25 11 13 26 3 1 23 14 15 30 33 22 36 5 4 2 32 Note: R indicates the ranking of the factor. everyone knows who has authority over whom are typical of societies with high power distance, justifying clear authority lines and traditional hierarchy. Factor V3 – work centrality – which clearly relates to masculinity (see items in Table 4), links the importance given to work- ing in a prestigious company with icons of modern capitalistic societies, such as the value attributed to competitiveness (parents should stimulate their children to try to be the best in class, competition between employees is not harmful, the importance attributed to one’s career, and the fact that people’s failure Downloaded from http://ccm.sagepub.com at SAGE Publications on May 17, 2007 © 2006 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Garibaldi de Hilal: Brazilian Culture and Cultural Agreement 151 is considered their own fault). This symbiosis suggests that the desired values of work and competitiveness are inscribed in the relational universe, in permanence and tradition, where the prestige of the organization grants prestige to its members, being the basis of their identification system. Factor V4 – need for survival (Table 5) – which contains other elements of uncertainty avoidance, shows how in relational and traditional societies (where employees intend to continue working for the organization until retirement and where most people can be trusted), modern capitalistic values are encompassed (namely the item that states that one’s job is more important than one’s leisure time). Moreover, the fact that even a lousy job is better than no job at all clearly indicates the specific difficulties of the labor market, frequently perceived as permanent. The scores and rankings of Factors V1 to V4 are in Table 6. In connection with V1 – need for security – it should be noted that the apparent contradiction between the scores of Asuncion – PC2 – and Ciudad del Este – PC21 – both Paraguayan cities, could be explained by the fact that Ciudad del Este is a small town on the borders of Paraguay, Brazil and Argentina, with immigrants of different origins, which give the city a peculiar profile. Practice Dimensions Of the 12 independent factors obtained we decided to keep the following 6 that together explain 71.25% of the variance: P1 – Employee oriented versus job oriented P2 – Results oriented versus process oriented P3 – Isolated versus relational P4 – Egalitarian versus hierarchical P5 – Parochial versus professional P6 – Informal versus formal. Tables 7 to 12 list the variables with loadings approximately higher than 0.50 or 0.60 that were considered relevant to explain each factor. In order to name the six dimensions we indicated in bold type the four items that we considered key to define each dimension. The 24 key items (4×6) were submitted to an ecological factorial analysis of principal component using varimax rotation and together explained 82.35% of the accumulated variance of the mean scores between units. The scores and rankings of the different company units are shown in Table 13. Factor P1 – employee oriented v job oriented (Table 7) – shows that the organization is perceived as interested only in the work of the employees and not in their well-being and, that in general, important decisions are taken by individuals. It also shows the fascination that Brazilian organizations have for hierarchy and tradition, as indicated by the items that state that decisions are centralized at the top and that changes are implemented by management decree. In such an environment of individuals, impersonal rules substitute for relationships. Thus the fact that the organization does not have relevant links with the local community and contributes little to society could be explained by the historical indifference of the forms of association that imply solidarity, as stated by Buarque de Holanda (1995). To exemplify, in individualistic societies, such as North American society, the concept of community is founded on the equality and homogeneity of all its members. In Latin America in contrast, the community is heterogeneous, hierarchical and complementary. Its basic unit is not the individual, but relationships and persons, and groups of friends. Moreover, when employees become embedded in the relational networks, the perception of formal hierarchies would decrease, with personal relationships forming the flip-side of official hierarchies. Factor P2 – results oriented v process oriented (Table 8) – shows that the major emphasis is on following organizational procedures correctly, and that following the correct procedures is more important than Downloaded from http://ccm.sagepub.com at SAGE Publications on May 17, 2007 © 2006 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 152 International Journal of Cross Cultural Management 6(2) Table 7 P1: Employee oriented v job oriented Code Loading* Description PP30 PP49 PP26 PP43 PP48 PP18 PP32 PP23 PP28 CT6 CT3 PP25 PP6 PP42 PP53 0.848 0.841 0.797 0.771 0.725 0.725 0.718 0.697 0.648 0.580 0.565 0.562 0.550 0.526 0.571 2nd loading No special ties with local community Pragmatic not dogmatic in matters of ethics Organization contributes little to society Decisions centralized at top Changes implemented by management decree Job competence is what counts regardless of how it was acquired Managers keep good people for own department People only told when they have made a mistake Organization only interested in work people do Typical member sloppy Typical member direct People’s private life is their own business Important decisions made by individuals Little attention to physical work environment Each day is pretty much the same * Items with negative loadings are reworded negatively. Table 8 P2: Results oriented v process oriented Code Loading* Description PP1 PP21 PP53 PP3 PP14 PP11 PP45 PP41 PP37 PP6 0.808 0.770 0.690 0.686 0.639 0.635 0.599 0.565 0.516 0.505 2nd loading Major emphasis on correctly following organizational procedures People identify primarily with own branch or location Each day is pretty much the same Uncomfortable in unfamiliar situations – people avoid taking risks Many people wonder about purpose and importance of their work Organization and people closed and secretive Correct procedures are more important than results Not aware of competition of other organizations Our branch worst of organization All important decisions taken by individuals * Items with negative loadings are reworded negatively. Table 9 P3: Isolated v relational Code Loading* CT2 CT7 CT4 CT1 PP29 CT5 0.854 0.853 0.788 0.746 0.708 0.604 Description Typical member warm Typical member relational Typical member flexible Typical member initiating Newcomers are helped to adapt quickly to job and group Typical member fast * Items with negative loadings are reworded negatively. Downloaded from http://ccm.sagepub.com at SAGE Publications on May 17, 2007 © 2006 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Garibaldi de Hilal: Brazilian Culture and Cultural Agreement 153 Table 10 P4: Egalitarian v hierarchical Code Loading* Description PP50 PP36 PP47 PP2 PP54 PP20 PP7 0.730 0.717 0.708 0.683 0.649 0.597 0.514 2nd loading Ordinary members never meet their top managers Meeting times only kept approximately People from the right background better chance of being hired Little concern for personal problems of employees Administrative discontinuity Top managers resent being contradicted Subordinates have to work according to detailed instructions from superiors * Items with negative loadings are reworded negatively. results. This factor also shows that the organization is perceived as a closed system as indicated in the items that state that the organization and its people are closed and secretive and that they are not aware of competition from other organizations. Yet, just like in a set of mirrors, the dimensions present their own opposites as shown in Factor P3 – isolated v relational (Table 9) – where the view of the ‘home’, of the relational axis is clearly represented. XYZ is an organization that exists in a complex system of social relationships, of links among its members. In the Brazilian case, in certain situations the ‘street’ is encompassed inside the ‘home’, treating the organization as if it were a large family. The result is a discourse where personal relationships constitute the framework of the whole system. In the ‘street’, society is encompassed by the axis of impersonal laws, hiding the domain of personal relationships. Brazil can be read or understood from both perspectives, and both possibilities are institutionalized in the organization. Consequently, the ethic that applies depends on how the organization is perceived (as the ‘street’ or as the ‘home’, i.e. isolated or relational), thus implying the concept of a double-edged ethic. There are interpretation codes and ways of behavior that are opposite and that are valid only for certain people, actions and situations. Factor P4 – egalitarian v hierarchical (Table 10) – complements factor P1 (employee oriented v job oriented), also showing the fascination that Brazilian organizations have for hierarchy and tradition, describing an organization that could be interpreted using the code of ‘the street’, the code of laws and of individualism, as indicated by the items that state that ordinary members never meet their top managers, that top managers resent being contradicted and that subordinates have to work according to detailed instructions from their superiors. Factor P5 – parochial v professional (Table 11) – shows how long-term planning, rational thinking and the fact that quality prevails over quantity can be embedded in a closed system, represented by the item that states that only very special people fit into the organization. Factor P6 – informal v formal (Table 12) – shows formality in discourse, in dealing with each other as well as regarding the dress code. Such formality fits the hierarchically process-oriented structure, as shown in the item that states that subordinates have to work according to detailed instructions from their superiors. All in all (Table 13), the analysis of the value and practice dimensions appears to denote that the relational universe provides the appropriate environment that would ‘facilitate’ existence in societies with high Downloaded from http://ccm.sagepub.com at SAGE Publications on May 17, 2007 © 2006 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 154 International Journal of Cross Cultural Management 6(2) Table 11 P5: Parochial v professional (closed system) Code Loading* PP8 PP46 PP40 PP16 PP34 0.830 0.821 0.704 0.657 0.642 Description We think three years ahead or more Our top managers only decide on the basis of facts We let quality prevail over quantity Everybody is conscious of costs of time and materials Only very special people fit into our organization * Items with negative loadings are reworded negatively. Table 12 P6: Informal v formal Code Loading* PP39 PP10 PP33 PP51 P...
Purchase answer to see full attachment