Culture as perceived context: An exploration of the distinction between dignity, face and honor cultures

Abstract

Researchers are making increasing use of the distinction between cultural logics emphasizing dignity, face, and honor. Students from eight nations including two from Latin America rated items tapping the extent to which they believed that most persons in their nation endorsed these types of mindset. Their ratings did not accord with prior beliefs as to which cultures exemplify dignity, face, and honor. However, the predictions that analytic cognition would be more prevalent in dignity cultures and contrasting types of holistic cognition would be more prevalent in face and honor cultures were supported. The belief that the logic of dignity was prevalent within one's nation was significantly associated with higher life satisfaction.

Resumen

Las investigaciones se están centrando cada vez con mayor énfasis en el uso de la distinción entre lógicas culturales, haciendo más hincapié en la dignidad, el mantenimiento del status quo estructural de una sociedad (face) y el honor. Estudiantes de 8 países, incluyendo 2 de América Latina, clasificaron elementos sobre cuán importante para la aceptación social es la dignidad o el honor para la mayoría de las personas en su nación. Las calificaciones obtenidas no concuerdan con creencias previas en cuanto a cuáles culturas promueven más la dignidad y/o el honor como guías de su comportamiento. Sin embargo, las predicciones de que los patrones de cognición analítica serían más frecuentes en culturas que se centran en la dignidad, y que los patrones de cognición holística serían más frecuentes en culturas que se centran en la honra fueron apoyadas. Adicionalmente, la creencia de que la lógica de la dignidad era frecuente dentro de cada nación se asoció significativamente con una mayor satisfacción de vida.

The identification of cultural dimensions that was initiated by Hofstede (1980) has provided a substantial basis for interpreting nation-level differences in a broad range of social behaviors. However, the mechanism whereby nation-level context could influence individual-level behaviors has come under increasing scrutiny. The measures of values, beliefs and norms that are typically used to define and construct nation-level dimensions are found to show much greater variability within nations than between nations (Fischer & Schwartz, 2011). This renders less plausible the assumption that individual behaviors within a nation are guided by an implicit or explicit awareness of the values, beliefs and norms that are most typically endorsed within their nation.

While it remains very likely that conformity to prevailing norms is a frequent occurrence within more specific subcultural groups within a nation (Gelfand & Harrington, 2015), we lack an adequate explanation of the mechanisms through which nation-level differences arise. Nation-level effects could for instance be considered as an aggregation of relevant subcultural effects. Alternatively, they might be accomplished through the impact over time of the social norms and institutions that arise as adaptations to salient environmental constraints. Relevant adaptations can include both behaviors and the ways of thinking about one's social context that facilitate particular behaviors.

One way to address these ambiguities is to obtain direct measures of intersubjective perceptions of the values, beliefs, and norms that prevail within a nation. Fischer (2006) asked students in ten nations to rate the values that they personally endorsed most strongly and also to rate the values that were most strongly endorsed by persons in their nation. The two sets of values correlated on average at only .28. In a further study conducted only in New Zealand, the students’ subsequently reported behaviors were more strongly predicted by the perceived national norm than by their own values. In a later study, Fischer et al. (2009) found that across samples from eleven nations, personal values and norms for ‘what most people believe’ concerning aspects of individualism and collectivism were cumulatively predictive of subsequent reported behaviors. Shteynberg, Gelfand and Kim (2009) found perceived national norms to be more strongly predictive of blame attributions than personal attitudes in Korea and the United States. The present study extends this increasing focus on the predictive power of subjective norms by applying it to a characterization of cultural differences based on contrasts between honor, face and dignity that is becoming increasingly influential, but for which no overt measures have yet been established.

A second way to explore the impact of social norms is to examine variations in social cognition. We have substantial evidence that respondents in the interdependent cultures of East Asia engage more frequently in holistic modes of cognition, while those in more independent cultures more often employ analytic modes of cognition (Nisbett, Peng, Choi, & Norenzayan, 2000; Norenzayan, Choi, & Peng, 2007). While the initial studies in this area by Nisbett and his colleagues focused on contrasts between East Asia and North America, more recent studies indicate that variations in analytic versus holistic cognition can be found within samples from many parts of the world (Varnum, Grossmann, Kitayama & Nisbett, 2011). Furthermore, experimental priming can induce shifts in cognitive style (Oyserman & Lee, 2008). Cultural differences in such styles are therefore probably best considered as habits induced by both proximal and distal circumstances, whether these circumstances be predominantly normative or personal (Muramoto, 2013). In contexts characterized by strong norms, it can be of value to scan all aspects of one's environment in order to be aware of every aspect affecting one's relations with others. Where norms are weaker, there can be greater possibilities for distinguishing different aspects of one's context and making choices that are more contingent on personal preference.

Dignity, face, and honor

Leung and Cohen (2011) have proposed a distinction between three ‘cultural logics’ of dignity, face, and honor. In dignity cultures, individuals are construed as relatively equal, with each having a stable and internal sense of worth. Face cultures are relatively more hierarchical, with greater emphasis on in-group harmony and modesty. The distinction between dignity and face cultures parallels Hofstede's (1980) much studied contrast between the individualism of Western European and North American nations and the collectivism of East Asian nations. Honor cultures give greater emphasis to the need to establish and defend the virtue and honor of oneself and one's group. The cultural logic of honor is seen as exemplified within Mediterranean, Latin American and South Asian cultures. Understandings of honor have been compared between Mediterranean honor and North American and North European dignity cultures (Rodriguez Mosquera, Manstead, & Fischer, 2002; Gelfand & Harrington, 2015; Günsoy, Cross, Uskul, Adams, & Gercek-Swing, 2015; Mosquera, Fischer, Manstead, & Zaalberg, 2008; Uskul, Cross, Sunbay, Gercek-Swing, & Ataca, 2012), and between Mediterranean honor cultures and East Asian face cultures (Güngör, Karasawa, Boiger, Dincer, & Mesquita, 2014). Empirical comparisons between dignity cultures and Latin American honor cultures have been much less frequent, and have included Brazil (Vandello & Cohen, 2003) and Chile (Vandello, Cohen, Grandon & Franiuk, 2009), but not Mexico. Some studies have now simultaneously contrasted all three types of culture (Leung & Cohen, 2011; Severance et al., 2013; Uskul, Oyserman, Schwarz, Lee, & Xu, 2013). However in none of these studies has any direct measure been employed to test the validity of the researchers’ postulation that their samples do exemplify the hypothesized cultural contrasts involved.

The perspective of the present study is that members of a given nation may have diverse perceptions of the culture of that nation, given the existence of sub-cultural groupings within any nation, as well as the variability of individual personal life experiences. Consequently, respondents from all sampled nations may identify values, beliefs, and behaviors that they believe to be typical of dignity or face or honor cultures within their own nation. Comparisons of mean scores for nations may therefore be at risk of invalidity not just on account of lack of measurement equivalence, but also from the extent to which samples fully represent the range of variation in perspectives. However, we may expect that whatever values, beliefs and behaviors are perceived to be present within each sample will be construed in ways that are characteristic of the cognitive style that is prevalent within that type of culture. Thus, the cognitive style can provide a more direct indication of mindset than the content of what is perceived. We next consider in turn the basis for predicting such effects in each type of culture.

Hypotheses

Within dignity cultures, analytic modes of cognition are most prevalent (Nisbett et al., 2000). Members of individualistic cultures have been shown frequently to think analytically and to distinguish between behavior patterns seen as inconsistent with one another. Consequently, they can be expected to differentiate between all three alternative ways of characterizing a culture.H1

In dignity samples, respondents will distinguish between dignity, honor, and face values. Those who perceive their nation to be based on dignity values will perceive it not to be characterized by reliance on either face values or honor values.

Conversely, members of the face cultures of East Asia more frequently think holistically (Nisbett et al., 2000). Given their major preference for preservation of harmony and face, they are likely to see all values and behaviors that could generate harmony rather than conflict as contributing toward the creation and maintenance of a face culture. As dignity values emphasize equality, they can provide a basis for interpersonal harmony. In contrast, honor values imply assertion and defense against threat, which are unlikely to be seen as contributing toward harmony.H2

In face samples, respondents will contrast dignity and face values with honor values. Those who perceive their nation to be based on face values will also perceive reliance on dignity values as contributing to face.

The preferred cognitive styles of members of honor cultures have been less fully explored. Uskul, Kitayama, and Nisbett (2008) found contrasting patterns of holistic and analytic cognition within distinctive Turkish samples. Knight and Nisbett (2007) found more holistic cognition among Southern Italians (who are said to more strongly endorse honor values) than among northern Italians. The nation-level dimension of Monumentalism-Flexumility has been identified by Minkov (2011) from analyses of items within the World Values Survey. Members of nations scoring high on Monumentalism score high on national pride and distinguish sharply between evaluations of what is approved and what is disapproved (Smith, 2011). The highest scores on this dimension are found predominantly in Arab and African samples. Nations scoring high on Flexumility score moderately, and are found in East Asia. It appears that this dimension may differentiate honor cultures from face cultures. Given a major preference for the preservation of honor, reliance on face and dignity values would be rejected as unable to uphold honor.H3

In honor samples, respondents will contrast dignity and face values with honor values. Those who perceive their nation to be based on honor values will perceive it not to be based on dignity or face values.

While there may be considerable variability in the way that individuals perceive their own national culture, it could be argued that those who perceive it in a way that is concordant with the majority perspective will achieve a more positive adaptation to their circumstances. However, this would only be the case where the majority culture is in fact positively evaluated, which may not always be the case. Bearing this in mind, we can cautiously test whether outcomes are more positive for those whose reading of their context concurs with the regions of the world that have been characterized by prior theorizing and research respectively as favoring dignity, face, or honor values. Indices are available for depression and for life satisfaction. There are certainly numerous other more proximal factors involved in whether one is depressed or satisfied with life, so there is no reason to expect that the predicted effects will be strong.H4

Depression will be lower and life satisfaction will be higher among respondents who characterize their nation in the way that is consistent with prior characterizations of the nations whose cultures favor dignity, face, and honor, respectively.

Method

Sample

Respondents were students enrolled in universities in the nations shown in Table 1. Responses from non-nationals and those born outside the country were discarded. The Brazilian sample comprised 42 percent whites, 41 percent Pardo (i.e., mixed-race) and 12 percent Africans. The Malay sample comprised 55 percent Malays, 36 percent Chinese and 6 percent Indians. No other sample contained ethnic minorities in excess of 10 percent. Data were collected online in Beijing, Finland, Turkey, the UK and in one of the two universities sampled in Brazil. Data were collected on paper in Chengdu, Lebanon, Malaysia and the other sampled university in Brazil. The questionnaire was constructed in English and then translated into the language of instruction in each of the universities sampled. Lebanese respondents therefore used the English language version. Back-translation was employed to ensure translation accuracy.

Table 1.

Details of samples.

Sample	N	Mean age	% Male
Brazil	290	26.7	51.4
China – Beijing	236	19.5	27.1
China – Chengdu	101	20.1	47.5
Finland	182	26.7	10.4
Lebanon	143	19.0	50.7
Malaysia	186	22.2	38.2
Mexico	210	21.8	47.8
Turkey	161	21.3	32.3
UK	137	20.0	15.3

Measures

Cultural clusters

Data were collected from a minimum of two nations within each of the postulated culture types. UK and Finland were deemed to be exemplars of dignity cultures, since they score in the top third of Hofstede's (1980) sample on his measure of individualism. China and Malaysia were selected as exemplars of face cultures, as they score in the top third of Hofstede's measure for collectivism. Four nations were selected as exemplars of honor cultures, separating Mediterranean honor cultures from Latin American honor cultures. As noted above, almost all studies of honor cultures have sampled Mediterranean nations, and we have little information as to their similarity or difference from Latin American honor cultures.

Measurement equivalence

In order to test the hypotheses validly, metric equivalence is desirable for the measures used to test within-sample relationships, namely depression, life satisfaction and perceived culture. For each scale, model fit was assessed using the Comparative Fit Index (CFI), Root Mean Square Error of Approximation (RMSEA), and Standard Root Mean Squared Residual (SRMR). Values of SRMR <.08 (or <.10), RMSEA < .06 (or <.08), and CFI > .95 (or >.90) have been proposed as criteria for “good” (or “acceptable”) fit (Hu & Bentler, 1999). Metric or scalar invariance were considered to be supported if a model that assumes that level of invariance showed adequate fit. Analyses were conducted in Mplus Version 6 (Muthén & Muthén, 2010).

Perceived culture

Perceived culture was measured through the use of 18 survey items tapping the values perceived to be endorsed in one's nation. This measure was originally constructed on the basis of theoretical analyses and research into aggression and negotiation behaviors (Severance et al., 2013). For each culture type, six items each asked ‘to what extent do most [nationality] people believe that ….’. Sample items are ‘… how much a person respects himself is far more important than how much others respect him’ (dignity); ‘…. people should minimize conflict in social relationships at all costs’ (face); and ‘… you must punish people who insult you’ (honor). The full survey is provided as an Appendix A.

To investigate whether the proposed three factor solution was a better fit to the data than collapsing any two of the scales into a two factor solution, we first conducted pan-cultural analyses investigating the fit of a three factor model against several two factor models, one for each of the possible subscale pairings. The three factor solution was indeed a better fit to the data than any of the two factor solutions (ΔCFI all >0.14).

Next, we conducted multiple-group confirmatory factor analyses to investigate measurement invariance of the three factor perceived culture scales. After removing two dignity items, two honor items, adding one covariance across all national groups, and five nation-specific covariances, partial metric invariance was supported; χ² (738) = 1216.037, p < .001, CFI = .90, RMSEA = .06 and SRMR = .087. Thus we may validly examine correlations between factor scores, but cannot rely on comparisons of means.

Depression

Depression was measured with the 20-item version of the Centre for Epidemiological Studies Depression scale (CES-D) (Radloff, 1977). These items have 4-point response scales keyed in terms of frequency of symptom occurrence. Four items describing positive symptoms are reverse keyed. The depression factor was scaled by fixing one item loading to 1. After removing five items (including all the reversed items), adding two covariances between items across all national groups, and 13 covariances between items for specific nations, a multi-group CFA supported partial metric invariance: χ² (891) = 1791.973, p < .001, CFI = .90, RMSEA = .074 and SRMR = .088. Factor scores standardized within nations were saved from this model for use in analyses.

Life satisfaction

Satisfaction with life was measured with the five-item scale developed by Diener, Emmons, Larsen, and Griffin (1985). A sample item is ‘I am satisfied with life’. These items have 7-point response scales keyed from ‘strongly disagree’ to ‘strongly agree’. There are no reversed items. The average of Cronbach alphas by nation was 0.82 with the lowest score being 0.72. A multiple-group CFA with one covariance across all national groups supported full metric invariance, with good fit: χ² (68) = 142.693, p < .001, CFI = .978, RMSEA = .078, SRMR = .077. Factor scores were saved from this model for use in analyses.

Results

Although it was argued that there is no basis for expecting that perceived values will accord with prior characterizations of dignity, face, and honor cultures, mean scores standardized relative to Brazil for the present samples are of interest and are provided in Table 2. Brazil was arbitrarily chosen as the basis for comparison since its name come first alphabetically. The table shows that among the nine samples, Brazil scored sixth on dignity, sixth on face, and seventh on honor. Thus the Brazilian respondents did not see any of the three dimensions as describing their culture very strongly. Mexican respondents scored joint second on dignity, eighth on face, and fifth on honor, so that they saw their culture as approximating most closely to a dignity culture.

Table 2.

Standardized factor means for perceived beliefs.

Sample	Dignity	Face	Honor	Depression	Life satisfaction
Brazil	.00	.00	.00	.00	.00
China – Beijing	−.25	.16	.59	.36	−.56
China – Chengdu	.08	−.14	.61	.38	−.26
Finland	−.06	.17	−.95	.06	.44
Lebanon	−.44	−.40	.65	.10	−.31
Malaysia	.40	.61	.14	.48	.26
Mexico	.40	−.23	.27	−.12	.90
Turkey	.54	.41	−.12	.40	−.34
UK	.04	.45	.29	.21	.02

Table 2 also shows the standardized scores for depression and life satisfaction. Brazil scored second lowest on depression and fifth highest on life satisfaction, whereas Mexico scored lowest on depression and highest on life satisfaction.

Hypotheses 1–3 were tested by the correlations shown in Table 3. As predicted by Hypothesis 1, respondents in UK and Finland differentiated between the extent to which they perceived their culture as emphasizing dignity, face, and honor values. Five of the six correlation coefficients were significantly negative. As predicted by Hypothesis 2, respondents in China and Malaysia did not differentiate their perceptions of dignity and face values, although the magnitude of the association between the two measures did vary between samples. Perceptions of honor values in these samples did not consistently correlate with perceptions of dignity or face values. As predicted by Hypothesis 3, the respondents from Lebanon, Turkey, Brazil and Mexico perceived honor values as contrary to dignity and face values, with seven of eight correlation coefficients strongly and significantly negative. Furthermore, perceptions of face and dignity values in these samples were strongly and positively correlated.

Table 3.

Correlations between perceived culture by nation and by culture type.

Sample	Dignity–Face	Dignity–Honor	Face–Honor
Finland	−.28**	−.38**	−.17*
UK	−.27**	−.40**	.02
Dignity samples	−.28**	−.38**	−.09
China – Beijing	.22**	.10	.16*
China – Chengdu	.43**	−.17	.08
Malaysia	.82**	.18*	−.14*
Face samples	.41**	.08	.01
Lebanon	.71**	−.72**	−.63**
Turkey	.90**	−.27**	−.02
Mediterranean Honor samples	.81**	−.44**	−.23**
Brazil	.81**	−.63**	−.68**
Mexico	.96*	−.76**	−.71**
Latin American Honor samples	.89**	−.70**	−.70**

Note: * p < .05; ** p < .01.

Hypothesis 4 was tested by the correlation coefficients shown in Table 4. Evidence in support of the hypothesis is found only in the data from the UK, and the Beijing Chinese sample. UK respondents who perceived their nation as characterized by dignity values were less depressed and more satisfied with life. Furthermore, UK respondents who saw their nation as emphasizing honor values were more depressed and less satisfied with life. In the Beijing sample, those who perceived their nation as favoring face values were less depressed and more satisfied with life. There was also an effect in the Beijing data linking satisfaction with life and emphasis on dignity values, but as noted above perceptions of face and of dignity values were associated in this sample. In the remaining seven samples, there was only one significant effect: contrary to prediction, respondents in Lebanon were more satisfied with life when they perceived their nation as favoring dignity values. The table also indicates a significant pan-cultural correlation between perception that one's nation favors dignity values and higher satisfaction with life. Consistent with this overall effect, Mexican respondents reported the second lowest mean score for depression and the highest mean score for life satisfaction. As noted earlier, they also perceived Mexican culture as favoring dignity rather than face or honor values.

Table 4.

Correlations of perceived culture with depression and life satisfaction.

Sample	Dignity		Face		Honor
	CESD	LS	CESD	LS	CESD	LS
Finland	−.07	−.03	.05	.07	.02	.07
UK	−.22*	.21*	−.04	.03	.25**	−.18*
Dignity Samples	−.14**	.09	.01	.05	.11*	−.02
China – Beijing	−.09	.15*	−.13*	.18**	−.05	−.01
China – Chengdu	−.01	.14	.03	.12	−.14	.13
Malaysia	.05	−.11	−.04	−.07	.13	.00
Face samples	−.03	.07	−.07	.06	.00	.01
Lebanon	−.16	.16	.00	−.04	.11	−.09
Turkey	−.05	.18*	−.03	.11	.07	−.08
Mediterranean Honor samples	−.10	.17**	−.01	.05	.08	−.08
Brazil	.01	.06	.09	.03	.07	−.02
Mexico	.04	.01	.07	−.01	−.01	−.03
Latin American Honor samples	.03	.03	.08	.01	−.03	−.02
Total sample	−.04	.07**	.01	.04	.04	−.03

Note: CESD = depression; LS = life satisfaction.

* p < .05; ** p < .01.

Discussion

This study has explored the way in which members of distinctive types of cultural groups perceive the values of those around them. Consistent with the results of earlier studies showing variability of values and beliefs within nations (Fischer, 2006; Fischer & Schwartz, 2011), we have made no assumption that the perceptions of members of a cultural group will necessarily concur with one another in their estimates of prevailing values or beliefs. An averaging of their perceptions consequently may or may not provide a valid estimate of prevailing culture. Some respondents are likely to be better informed or more influential than others, and they may also differ in the reference points against which they choose to anchor their judgments (Heine, Lehmann, Peng, & Greenholz, 2002). The estimates of values provided by the present student samples may have been drawn more from knowledge of shared campus values than from knowledge of their national cultures as a whole. As Table 2 showed, there was in fact very little concordance between mean perceived values and prior characterizations of culture types. This may also be partly due to the fact that the values measures did not achieve full structural measurement equivalence.

Our emphasis here has been on the structure of respondents’ perceptions as an indirect indicator of cultural priorities, and the problems identified above do not affect within-subject data analyses. The substantial divergence found between the ratings made by respondents from dignity, face and honor cultures does lend support to the contrast made between them by previous researchers. Cultures are known to differ in their propensity for acquiescent responding (Smith, 2011), but the present correlational results cannot be attributed to a global tendency toward acquiescence because respondents from all samples did make distinctions between at least some of their ratings.

Little evidence was found for the predicted of association between the predominance of each type of perceived values and the indices of depression and life satisfaction. Rather than reflecting a differential response to each culture type, the results suggest an overall association between the dignity dimension and positive outcomes. Particularly in the UK, Turkey, Lebanon and China, the perception that one's nation is characterized by high dignity values and low honor values is associated with low depression and high life satisfaction. In the sample as a whole, only the association of perceived dignity and high life satisfaction achieves significance. While this is an individual-level association, it is reminiscent of the nation-level finding that mean life satisfaction is higher in dignity cultures (Diener, Diener, & Diener, 1995). However, the variance explained by the present association is modest, as might be expected given the greater salience of more proximal causal factors for depression and life satisfaction.

Conclusion

The use of a perceived values measure has provided evidence confirming the presence of contrasting cognitive styles among respondents in cultural groups assumed to exemplify dignity, face, and honor cultures. However, at least among the present samples of respondents, the measure does not itself provide a defensible basis for distinguishing directly between these three types of cultural groups. Fuller investigation of appropriate survey items and the contexts to which these items refer will be required in order that the required contrasts can be more adequately operationalized.

Funding

No financial support was provided.

Conflict of interest

Since the editor is a co-author of this paper, the review process and decision to accept it for publication were handled by Sofia Rivera-Aragón. The authors have no other conflicts of interest to declare.