Need for approval from others and face concerns as predictors of interpersonal conflict outcome in 29 cultural groups

T he extent to which culture moderates the effects of need for approval from others on a person’s handling of interpersonal conflict was investigated. Students from 24 nations rated how they handled a recent interpersonal conflict, using measures derived from face-negotiation theory. Samples varied in the extent to which they were perceived as characterised by the cultural logics of dignity, honour, or face. It was hypothesised that the emphasis on harmony within face cultures would reduce the relevance of need for approval from others to face-negotiation concerns. Respondents rated their need for approval from others and how much they sought to preserve their own face and the face of the other party during the conflict. Need for approval was associated with concerns for both self-face and other-face. However, as predicted, the association between need for approval from others and concern for self-face was weaker where face logic was prevalent. Favourable conflict outcome was positively related to other-face and negatively related to self-face and to need for approval from others, but there were no significant interactions related to prevailing cultural logics. The results illustrate how particular face-threatening factors can moderate the distinctive face-concerns earlier found to characterise individualistic and collectivistic cultural groups.

Determinants of the successful outcome of conflicts have been explored extensively in contexts ranging from personal relationships to business negotiation and international diplomacy (Fisher et al., 2011;Stohl et al., 2013).However, the extent to which the predictors of successful outcome vary between cultural contexts has been less frequently addressed.The ways in which persons think about themselves in cultures emphasising individualism or collectivism (Hofstede, 1980;Markus & Kitayama, 1991) is likely to affect their approach to conflict resolution and what they would consider to be a satisfactory outcome to a conflict.In this paper, we sample a broad range of cultural contexts in order to examine the extent to which outcomes are dependent on relevant personal attributes and on the cultural context in which a given conflict occurs.In particular, we focus on conceptualizations of face, both in East Asian cultures (Hwang & Han, 2010) and more generally (Brown & Levinson, 1987;Goffman, 1959;Ting-Toomey, 1988, 2005).
A person's face is usually understood in terms of their reputation or public image in relation to performance in a given social setting.Face concerns become particularly important in contexts where there is some risk of embarrassment or conflict.Most theorists distinguish two aspects of face, dependent on the salience given to self or others.For instance, Brown and Levinson (1987) distinguish positive face in terms of gaining recognition or praise for one's actions from negative face, which is accomplished through retaining one's freedom of action.Chinese concepts of face distinguish lian, which is a reputation integral to the individual that is based on adherence to personal morality, and mianzi or social reputation which derives from actions in particular contexts involving others (Hwang & Han, 2010).

FACE-NEGOTIATION THEORY
The differing definitions of these concepts of face reflect the individualistic and collectivistic contexts within which they have been formulated.Ting-Toomey (1988) addressed this cultural contrast directly through her development of face-negotiation theory, in which face concerns are distinguished on the basis of whether one is more concerned with saving or enhancing one's own face, or with giving face to relevant others.As first formulated, the theory proposes that, while concern for face is universal, there will be greater concern for self-face in individualistic contexts and greater concern for other-face in collectivistic contexts.These differing concerns are predicted to guide actors in their choices as to how to negotiate in conflict situations.
Face-negotiation theory thus gave a central role to the contrast between individualistic and collectivistic cultures, within which persons with independent or interdependent self-construals (Markus & Kitayama, 1991;Singelis, 1994) respectively are likely to be more salient.Members of individualistic cultures tend to be socialised in a manner that encourages persons to construe themselves as relatively independent agents.In contrast, members of interdependent cultures tend to be socialised in a manner that encourages them to construe themselves in terms of enduring group memberships.In an initial study of how students reported handling interpersonal conflicts, concern for self-face was found to be more frequent in LUN ET AL. the USA and Germany, while concern for other-face was more frequent in Japan and China (Oetzel et al., 2001;Oetzel & Ting-Toomey, 2003).Furthermore, independent self-construal was more strongly associated with concern for self-face, while interdependent self-construal was more strongly associated with concern for other-face.
In a major revision of her theory, Ting-Toomey (2005) proposed that this simple cultural contrast would be moderated by a variety of contextual factors that may pose threats to the face of one or another negotiator.For instance, there may be status differences between negotiators, they may be from outgroups rather than ingroups, the conflict may be about a major rather than a minor issue, and the negotiators may be from differing cultural backgrounds.All such face-threatening processes are predicted cumulatively to enhance reliance on self-face in preference to other-face, and thereby imperil the outcomes of conflicts.Subsequent studies of interpersonal conflict have supported the view that concern for self-face and other-face is influenced by contextual factors and not solely by cultural contrasts.For instance, concern for self-face was found to predict expression of anger while concern for other-face predicted compassion in both China and the USA (Zhang et al., 2014).In both Hong Kong and the USA, concern for self-face predicted forgiveness negatively while concern for other-face predicted forgiveness positively (Zhang et al., 2019).
The present study considers circumstances within which a particular aspect of personality may generate a face-threatening process.Crocker et al. (2003) identified seven different bases upon which US students evaluated their self-worth.Among these, one is distinctively relevant to the handling of interpersonal conflicts, namely the need for approval from others.Their specific measure of this need was found to be associated with low self-esteem and high neuroticism (Crocker & Luhtanen, 2003).Using a different but related measure of the need for approval from others, self-image goals (example item: "Get this person to notice your positive qualities") were found to be associated with validation-seeking and defensive responses to conflict in both the USA and Japan (Niiya et al., 2013).In contrast, compassionate goals (example item: "Be supportive of this person") predicted belief in positive outcomes to conflict.
A person's high need for approval from others can clearly generate a face-threatening process in circumstances of conflict with others.The question at issue is whether it will do so equally in all types of cultural context.Do more collectivistic contexts provide some protection against a high need for approval from others?

THE PRESENT STUDY
In this study, we examine the interrelationship of three types of measure: (a) An aspect of personality, namely need for approval from others; (b) Respondents' reported concerns for self-face and other-face within a past interpersonal conflict; (c) A sample-level measure of the cultural context within which the reported conflict occurred.In order to provide adequate tests of variance between cultural contexts, we include data from a much broader range of samples than those employed in earlier studies.We also include a measure of the perceived outcome of the conflict experienced by respondents.Specifically, we test whether the relationships between these measures vary in predictable ways between groups that give particular emphasis to the importance of face and those that do not.

Interdependence versus need for approval from others
The conceptualization of interdependence first formulated by Markus and Kitayama (1991) was built upon the proposition that individuals in collectivistic cultural contexts focus their attention upon an intimate circle of family and close friends and define themselves in terms of membership in these groups.Awareness of their relations with these others is a continuing priority.In contrast, group membership is less exclusively defined in individualistic contexts, and attention towards others is focused more broadly and varies between persons.
In settings where conflict occurs, we may expect that all persons will focus their attention upon the other party to the conflict, but this focus will be constrained by the nature of the pre-existing commitments to that other party.Independent individuals are likely to prioritise protecting their freedom to act in ways that preserve their autonomy and dignity: this will require attention to self-face while also seeking to resolve the conflict through giving face to the other party.This dual set of priorities has been formalised as dual-concern theory in analyses of conflict within Western cultures (Rahim, 1986;Thomas, 1976).
In contrast, interdependent individuals involved in conflict, especially when it is within their in-groups, will need to act in ways that enhance and preserve their group memberships.This will involve giving face to the other party to the conflict.Key priorities will be animosity reduction and the avoidance of any threat of the disintegration of the group (Leung, 1997).Consistent with this view, Ohbuchi et al. (1999) found that when students were asked to analyse what had been their approach to past interpersonal conflicts, those from a relatively collectivistic culture (Japan) emphasised avoidance, whereas those from a relatively individualistic culture (USA) emphasised assertion.Thus, while need for approval from others will predispose towards concern for other-face in all contexts, it will be likely to elicit greater reliance on self-face in individualistic contexts and lesser reliance on self-face in collectivistic contexts.In other words, in relation to need for approval from others, there will be a main effect for concern for other-face but an interaction with cultural context for concern for self-face.In order to formulate these expectations as hypotheses, we must first specify our measures of cultural context.

Dignity, honour and face cultures
While the attributes of individualistic cultural groups have been much examined in studies of conflicts, it has been proposed in recent years that it is useful to distinguish two different ways of orienting oneself towards settings within collectivistic contexts (Leung & Cohen, 2011).Within this perspective, members of particular cultural groups are seen as making use of three distinctive cultural logics emphasising dignity, honour or face as a basis for understanding and guiding events around them.As Leung and Cohen state: "A particular cultural logic weaves together various scripts, behaviours, practices and cultural patterns around this central theme, giving them a meaning and a certain logical consistency and coherence for the people of a culture" (p.508).Where a given logic is frequently employed it is more likely to be seen as normative.Thus, in individualistic contexts persons are more likely to see the logic of personal dignity as normative.In collectivistic contexts, persons may emphasise as a logic either the honour of one's group, or the preservation of harmony within one's group through the maintenance of face.Leung and Cohen (2011) note that individuals may interpret specific events in their everyday life in terms of a variety of cultural logics.However, the coherence of cultural groups rests upon employment of a relatively similar array of logics.Group culture is therefore best assessed by asking respondents to rate which logics they perceive as normative in their context, rather than to rate their individual values or interpretive frameworks, Using this type of measurement, differential endorsement of dignity, honour and face norms between samples has been reported across a broad range of nations (Aslani et al., 2016;Smith et al., 2017;Uskul et al., 2018;Yao et al., 2017).
Prior analyses of the present data confirm that there is significantly greater concern for both self-face and other-face where face logic is more prevalent and significantly less concern for self-and other-face where dignity logic is more prevalent (Smith et al., 2021).There was no significant effect where honour logic was prevalent.Following our earlier reasoning, we can now consider how respondents' individual need for approval from others will moderate these overall sample-level effects.
H1: Need for approval from others will be positively related both to concern for self-face and to concern for other-face.

H2:
The relationship of need for approval from others to concern for self-face will be weaker within groups where cultural logic emphasises face.

Conflict outcome
Prior studies testing hypotheses relating to concerns for self-face and other-face have not included a measure of conflict outcome.However, research relating to the dual-concern theory (Rahim, 1986;Thomas, 1976) hinted at different possible conflict outcomes associated with the two face concerns.In a meta-analysis, De Dreu et al. (2000) noted that negotiators tended to achieve higher joint outcomes when they had a prosocial rather than egoistic motive.This observation is consistent with Leung et al.'s (2002) proposition that a genuine concern for harmonious relationship and mutual goal attainment by both negotiators in a conflict enables positive conflict outcome.Thus, concern for self-face is likely to relate to less desirable conflict outcome, whereas concern for other-face may associate with more positive outcome.
In addition, a cultural context which emphasises face logic may further moderate the impact of the concerns for self-face and other-face on conflict outcome.Aslani et al. (2016) studied simulated intracultural business negotiations among students from USA, Qatar and China.These authors found support for their predictions that negotiators from Qatar and China, as representing cultures favouring honour and face logics respectively, would be more competitive than US negotiators and would achieve less joint gain.Importantly, they were careful to instruct participants that they were negotiating with an outgroup party and noted that in-group negotiation in cultures relying on face logic would focus on preservation of harmony rather than competition.Illustrating this difference between in-group and out-group strategies among Chinese students, Leung and Bond (1984) used a reward allocation task.In-group allocations were found to be based on equality, but out-group allocations were based on equity.In the context of interpersonal conflicts which involve non-strangers, the nature of the conflict is essentially an in-group negotiation.Accordingly, we develop the following hypotheses: H3: Positive outcome will be related negatively to concern for self-face, and positively to concern for other-face.

H4:
The relation of concern for self-face to positive conflict outcome will be weaker, but the relation of concern for other-face to positive conflict outcome will be stronger, in groups where cultural logic emphasises face.

METHOD Participants
The sampling frame was intended to include a range of groups within which one or other of the three cultural logics is widely believed to be more strongly endorsed.Selection of a small number of exemplars of each type risks confounding differing cultural logics with other aspect of cultural difference.Although nations are complex entities with numerous subcultures, continuities of for instance language, religion, educational systems and wealth have served to create and maintain a degree of distinctiveness between nations (Smith et al., 2013).
Participants were 5064 students from 24 nations who completed the survey either online, or in the classroom.1Most respondents were students of psychology or social science aged 18-30.Respondents either received course credit or were thanked for their participation.Ethical consent for the research project was obtained from each university that was sampled.In a small number of cases, ethical approval was based upon the ethical scrutiny that had been conducted at the University of Sussex.Respondents provided details of their age, gender, country of birth, nationality, ethnicity and religion.The survey included additional items not reported in the present analyses.It was originally constructed in English and was then translated into the language for use at each location.After independent back-translation, any corrections made were based on discussions between relevant bilinguals (van de Vijver & Leung, 1997).Respondents who were not nationals of the location sampled were excluded from the data analysis.In Brazil, Greece, Mexico, Russia and the USA, data were collected from cities or regions with known cultural differences, so these samples were further separated according to the location of data collection.Details of all samples are provided in Table 1.Within each of the three national samples that included diverse response modes, there were no significant differences in perceived cultural logics between paper responses and online responses, after controlling for age and gender.

Need for approval from others
To measure need for approval from others, we used the five items defining this source of social worth, as developed by Crocker et al. (2003).Three of the five items are reversed (example: "I don't care what others think of me"), while two were not ("My self-esteem depends on the opinions others hold of me").Responses were made on 4-point Likert scales, with anchors from strongly disagree to strongly agree.Cronbach alpha was above .70for 22 samples and above .60for four further samples.There were no negative item-whole correlations in any sample.
In this study, this variable is treated as an individual-level measure of a self-reported aspect of personality.

Face concerns
Respondents were asked to think of a recent interpersonal conflict that they had experienced, using the same measure that has been employed in earlier tests of face negotiation theory (Oetzel et al., 2001;Oetzel et al., 2008;Oetzel & Ting-Toomey, 2003).Respondents first identified the gender of the other party and indicated whether the conflict involved their romantic partner, a family member, a friend, or someone from work or college.They then rated four items referring to concern for Self-Face (example: "I was concerned with protecting my self-image"), and six items referring to concern for Other-Face ("I tried to be sensitive to the other person's self-worth").Responses were on 7-point Likert scales, with anchors from strongly disagree to strongly agree.After within-sample standardisation, factor analysis of the 10 face items for the total sample yielded a two-factor solution with oblimin rotation explaining 56.5% of variance, as shown in Table 2.
When sampling across cultures, it becomes increasingly difficult to satisfy the criteria employed to evaluate measurement equivalence through confirmatory factor analysis (Marsh et al., 2010).Fletcher et al. (2014) reported partial measurement equivalence of the concern for self-and other-face items in samples across five nations.Within the present much larger number and wider range of samples, we sought to determine the adequacy of the two face concern scales within each of the sampled nations by computing Tucker-Lewis phi coefficients (van de Vijver & Leung, 1997).As samples from some nations were small, data from nations that were judged culturally similar or geographically adjacent were grouped together, yielding ten clusters. 2The items defining the factor structure obtained within each of the clusters were compared in turn with the structure shown in Table 2.For nine of the ten clusters, all 20 coefficients for factors exceeded .96.For the Southeast Asian cluster, comprising the Malay and Thai data, the coefficients were .89for concern for Other-Face and .69for concern for Self-Face.Adequate phi coefficients provide evidence of scale reliability both within samples and between samples.Thus, there is overall evidence for consistent structure for the face scales employed in the present analysis, but attention is required to the results when the Malay and Thai data are included. 3n this study, mean scores for concern for self-face and concern for other-face are treated as individual-level reports of how a specific conflict was handled.

Conflict outcome
Positive conflict outcome was measured based on a single item, tapping successful outcome ("The conflict was satisfactorily resolved").As our present focus is on respondents' perceived outcome of a recalled conflict, the use of this generic, single-item measure is considered appropriate.Past research has shown that even more complex concepts such as self-esteem can be adequately captured by a single-item measure (Robins et al., 2001), and multi-item measures do not necessarily perform better than their single-item counterparts (Gardner et al., 1998).Thus we consider this item sufficient for indicating the effects of different face-concerns on the conflict.This item had a 7-point response scale, anchored from "Strongly Disagree" to "Strongly Agree."Sample frequencies for descriptions of each type of conflict are shown in Table 3.

Perceived cultural logics
We next constructed sample-level measures of the perceived cultural logics of dignity, honour and logic.Respondents were first asked to rate how well each of three statements described " … the people around you (your school, workplace, town, neighbourhood, etc.)." The items selected from those with the highest factor loadings reported by Yao et al. (2017).These were: "These people think that they should be true to themselves regardless of what others think" (Dignity); "These people feel that they should uphold and defend their family's reputation" (Honour); "These people think they should be extremely careful not to embarrass others" (Face).The three 6-point response scales were keyed from strongly disagree to strongly agree.Following the procedures employed by Kashima et al. (1995), individual-level acquiescence was first discounted through within-subject standardisation across the three items.These scores were then made comparable by standardisation across samples, and mean scores were computed for each cultural logic for each sample.In this study, although these measures are derived from individual-level data, the mean scores are treated as sample-level characterisations of each group's culture, because the scores incorporate the ratings made by all respondents within that sample.
The use of just three items as the basis for these measures does not permit tests of measurement equivalence.Smith et al. (2021) reported the mean scores for each sample, indicating that the scores obtained do accord in most instances with the ways in which emphasis upon dignity, honour and face logics is understood to vary across cultural groups (Uskul et al., 2018;Yao et al., 2017), thus providing some plausible evidence for measurement validity.Because these scores are based on within-subject standardisation, they are not independent of one another.Sample-level face logic correlates with sample-level dignity logic at −.72 (p < .001)and with sample-level honour logic at −.15 (ns).Sample-level dignity logic correlates with sample-level honour logic at −.58 (p < .001).In the present analysis, only the face logic measure is employed.

RESULTS
Preliminary analyses were required in order to determine the extent to which mean scores for conflict outcome differed in terms of the type of conflict that each respondent chose to describe and in terms of the genders of the parties involved.Conflict type varied by gender (F = 9.90; p < .001),with women more frequently describing family conflicts, and no differences by gender in the frequency of other conflict types.There was no significant relation between conflict type and reliance on perceived face logic.Mean outcome scores are shown in Table 4.After controlling for differences between samples in mean rated conflict outcome, univariate ANOVA for the total sample showed a significant effect for the four combinations of gender (F = 3.05; p < .05),for the four conflict types (F = 8.89; p < .001)and for the interaction between gender and conflict type (F = 4.16; p < .001).Positive outcome was significantly lower for work conflicts, and for conflicts where women rated conflicts with men.In testing the hypotheses, these sources of variance are controlled.We adopted the five-step procedure for multilevel analyses recommended by Vauclair (2013).We first tested the null model (Model 1) of our target variables (i.e., face concerns and conflict outcome) to examine their intraclass correlations (ICC).In Model 2, we added individual-level control variables (i.e., age, relationship type and gender) and predictors, with only the intercept of the target variable varying randomly.In Model 3, we allowed the slope of the individual-level predictors to vary randomly, in order to examine whether there was significant sample-level variability of the individual-level predictors using a likelihood-ratio test.Then we introduced the sample-level predictor (i.e., perceived face logic) in Model 4 to examine its effect on the dependent variables.In Model 5, we further tested whether the sample-level predictor would account for the variability in the slopes of the individual-level predictors by including cross-level interactions.
Multilevel analyses were conducted using the statistical packages for multilevel modelling, namely, lme4, RLRsim and lmerTest (Bates et al., 2015;Kuznetsova et al., 2017;Scheipl et al., 2008) in the R environment.Following the recommended procedure in multilevel modelling, group-mean centring was applied to all Level-1 predictors, except for dummy-coded variables, and grand-mean centring was applied to Level-2 predictor (Aguinis et al., 2013;Enders & Tofighi, 2007).We used full information maximum likelihood (FIML) in the estimation, so that we can compare the relative fit of the models based on the −2 log likelihood ratio (i.e., deviance; see Bliese, 2002;Peugh, 2010).A significant deviance test between Models 2 and 3 would indicate that the model involving random slope of the individual-level predictors statistically fits the data better than the fixed-slope model.
The results for concern for self-face and concern for other-face are summarised in Tables 5 and 6, respectively.The ICCs for self-face and other-face were 0.107 and 0.096, respectively, indicating that around 10% of the variance of these variables was at the sample level and the remaining variance was at the individual level.The likelihood-ratio test indicates significant improvement between Models 2 and 3 for self-face concern but not for other-face concern.However, here we follow Nezlek (2011), who notes that moderation may be stronger or weaker in different samples and that it is therefore possible to test for cross-level interactions even when the overall random slope term is not significant.
Results in Tables 5 and 6 show that, consistent with Hypothesis H1, need for approval from others was associated with greater concern for both self-face and other-face.As one might expect, the effect is much stronger for concern for self-face than it is for concern for other-face.Concern for self-face was also reported more frequently in work conflicts, whereas concern for other-face was reported more frequently in non-work conflict contexts and by women.Model 4 in these tables shows that sample-level perceived face logic does not predict higher concern for self-face, whereas it does predict concern for other-face.Furthermore, consistent with Hypothesis H2, Model 5 in Table 5 shows that the association between need for approval from others and concern for self-face is significantly weaker in cultural contexts emphasising face logic more strongly.This effect is shown in Figure 1.In contrast and consistent with Hypothesis H1, Table 6 shows that the relation of concern for other-face with need for approval from others does not vary between cultural contexts.
Table 7 shows the results of the multilevel analysis for conflict outcome.The ICC for conflict outcome was 0.043, meaning that around 4.3% of the variance was at the sample level.Consistent with Hypothesis H3, conflict outcome is positively associated with reliance on other-face and negatively associated with reliance on self-face.Need for approval from others is also negatively associated with outcome.Contrary to Hypothesis H4, Models 4 and 5 in Table 7 indicate that were no significant main or moderation effects for cultural context.Outcomes were also reported as significantly more positive for romantic conflicts and for conflicts between males.
We next explored whether variations in outcome between conflict context and the genders of the parties involved could be explained by the frequency of face concerns.Concern for self-face varied between conflict types (F = 18.27; p < .001),being higher in work conflicts than in all other conflict types (p < .001).Concern for self-face was also lower in romantic and in family conflicts than in conflicts with friends (p < .05).Concern for self-face also varied between genders (F = 3.99; p < .01),being higher among women reporting conflicts with men (p < .001).
LUN ET AL.Note: FIML = full information maximum likelihood estimation; OA = need for approval from others.Conflict type is coded with three dummy variables with reference to work relationship.Gender type is coded with three dummy variables with reference to female-male.* p < .05. * * p < .01. * * * p < .001.Note: FIML = full information maximum likelihood estimation; OA = need for approval from others.Conflict type is coded with three dummy variables with reference to work relationship.Gender type is coded with three dummy variables with reference to female-male.* p < .05. * * p < .01. * * * p < .001.Concern for other-face varied between genders (F = 4.22; p < .01),being higher for women in conflict with other women than for women in conflict with men (p < .05).
Concern for other-face varied by conflict type (F = 26.44;p < .001),being higher in romantic conflicts than in all other conflict types, and higher in friend conflicts than in those with family or at work (p < .001).Thus, outcome was consistently rated most positively in conflict types that were highest on other-face concern and lowest on self-face concern.
In addition to the analysis of outcome within the overall dataset, we documented how conflict outcome was predicted by the target predictors in each sample.Table 8 reports the extent to which positive outcome is predicted by need for approval from others, concern for self-face and concern for other-face within each sample.It is notable that while consistent effects for concern for other-face are found, the results for concern for self-face are more diverse, with significantly positive effects from four samples and significantly negative effects from six samples.

DISCUSSION
In this research we examined aspects of face-negotiation theory in relation to interpersonal conflict across a wide range of cultural groups.The results provide evidence for both culture-general and culture-specific effects.Across a much broader range of samples than previously available, the predictors of reported interpersonal conflict outcome are found to be similar.This finding is important, but it leaves open questions as to whether these predictors of outcome are elicited in similar ways in different national-cultural contexts.
In relation to previous studies, the inclusion of a measure of need for approval from others provides a basis for understanding how face concerns towards interpersonal conflict may be elicited in cultural contexts that emphasise independence rather than interdependence.While Ting-Toomey's (1988) theory of face negotiation initially emphasised contrasting face concerns and self-construals between individualistic and collectivistic cultures, this theory has been tested principally in relation to facework behaviours that are consequent upon concerns for self-face and other-face.In contrast, the present results relate more closely to Ting-Toomey's (2005) identification of face-threatening processes that may interact with cultural determinants of face concerns.Need for approval from others is one such factor, underlining the ways in which aspects of an individual's personality may interact with cultural context in eliciting self-face concern.We find that cultural context does moderate the link between need for approval from others and self-face concern, but that no such effect is found in relation to other-face concern.Why should this be so?
We interpret this result as highlighting the role conflict that is more strongly present for persons engaged in conflicts within individualistic rather than collectivistic contexts.Preserving one's dignity can conflict with reaching agreement with the other party.Dual-concern theory (Rahim, 1986) has long emphasised the importance of addressing both one's own and the other party's priorities in a conflict, but this balance is by no means always easy to strike.In contrast, reaching agreement with close others in interdependent contexts will serve to preserve one's face.Any stress involved in achieving this satisfactory outcome will be less contingent on variations in approval from others, since lack of such approval is less likely to be a threat to one's identity or group membership.The results obtained by Crocker et al. (2003) indicate that within the US need for approval from others may be a relatively aversive state, linked with low self-esteem.In interdependent contexts, approval from others is more likely to be an aspect of collective rather than individual self-esteem (Krys et al., 2020).It would be important to determine whether this is so in a future study.
Sample-level perceived face logic was positively related to face-concerns, but its effect did not extend to conflict outcome.The ICC of conflict outcome also showed that much of the variance of conflict outcome was at the individual level rather than the sample level.This suggests that although cultural context may influence a person's face concerns in a conflict, the relation of these concerns to outcome does not differ between contexts.These results are consistent with the findings of Oetzel and Ting-Toome (2003) that face concerns did differ between four cultural groups and with those of Oetzel et al. (2008), who showed that the types of facework associated with concern for self-face and with concern for other-face did not differ across four groups.Thus, both studies concur with the view that while cultural factors LUN ET AL.Note: FIML = full information maximum likelihood estimation; OA = need for approval from others.Conflict type is coded with three dummy variables with reference to work relationship.Gender type is coded with three dummy variables with reference to female-male.* p < .05. * * p < .01. * * * p < .001.
may differentially elicit concerns for face, these concerns are then addressed in similar ways in all contexts.These findings also show that eliciting the other-face concern of individuals from different cultural backgrounds would be a useful focus for achieving satisfactory outcome in a conflict.How this concern may be elicited during inter cultural conflicts and how tactics for doing so may vary across persons and cultures varying in concern for self-face can be more fully explored though drawing on the extensive body of research into aspects of cross-cultural intelligence (Earley & Ang, 2003;Liao & Thomas, 2020).

Limitations
In this study, respondents were provided with a choice as to which type of conflict to report.Providing choice enhanced the likelihood that a given respondent would be able to recall a suitable personally relevant event.However, if the conflict types selected covaried with prevalent cultural logics, controlling for conflict type could have eliminated variance that would be relevant to the hypotheses concerning moderation.For instance, as shown in Table 3, conflicts with friends were reported more frequently in Pakistan, Saudi Arabia and Iraq.Differentiating such sources of variance requires a substantial number of cases, and although the number of samples included in the present study greatly exceeds those in the prior literature the possibility of Type 2 errors remains.
The validity of the measure defining perceived face logic at the sample level rests on the wording of a single item.Evidence for the nomological net for cultural logics with a broad range of measures obtained from the present sample has been provided by Smith et al. (2021), but use of multiple items referring to dignity, honour and face logics in future studies is preferable.The single-item measure of perceived conflict outcome has provided an initial estimate, but multiple items will also be required in order to ensure adequate tests of measurement equivalence.Ratings on Likert scales show differential variations in response style across cultural groups (Smith, 2004), and these may well have affected the outcome ratings obtained in the present data.However, the use of opposed items in the need for approval from others and perceived cultural logic measures, and the contrast in effects between different samples shown in Table 8 mean that this concern is not likely to have been a major threat to the validity of the effects obtained.
A further limitation of this study is that no measure was included of the degree of threat to face experienced by respondents in the conflict that they described.A measure of this type for each of the face threatening processes specified in Ting-Toomey's (2005) reformulation of face negotiation theory would be necessary in order to evaluate their relative effect sizes.

Future research
This study has focused on the cultural logic of face, and the present measure of the cultural logic of dignity was substantially correlated with it.Reliance on face concerns within cultures where honour logic is predominant requires attention, especially with cultural logic measures that are fully independent of one another.We also require much richer information as to the ecology of interpersonal conflicts within each type of cultural logic, considering for instance time parameters, the nature of acceptable concessions, apologies and other forms of facework, the role of third parties and so forth.
Variations in need for approval from others have been little explored across cultures.In exploring the utility of this construct in explaining conflict outcomes and other phenomena of practical importance, it will also be necessary to clarify its relation to relevant dimensions of personality and to variables that have been more fully explored cross-culturally such as self-esteem (Schmitt & Allik, 2005) and self-construal (Vignoles et al., 2016).

CONCLUSION
We find support for Ting-Toomey's (2005) incorporation of contextual threats to face within her face-negotiation theory.The threat to face posed by an aspect of personality-need for approval from others-has specific consequences for the face concerns that are elicited, with consequential implications for achieving a positive outcome.However, we find that the association between each type of face concern and conflict outcome are culture-general within the present samples.

Figure 1 .
Figure 1.Interaction effect between need for approval from others and perceived face logic on self-face.

TABLE 1
Details of samples

TABLE 2
Factor analysis of concern for face items

TABLE 3
Distribution of conflict types between samples Outcome is standardised within sample; means with the same superscript do not differ from one another.FF = female in relation to female; FM = female in relation to male; MF = male in relation to female; MM = male in relation to male.

TABLE 5
Multilevel analyses for concern for self-face

TABLE 6
Multilevel analyses for concern for other-face

TABLE 7
Multilevel analyses for conflict outcome

TABLE 8
Approval of others and face measures as predictors of conflict outcome across samples Note: OA = need for approval from others; OFACE = other-face; SFACE = self-face.Standardised regression coefficients are presented.Controlled for conflict type and gender type.* p < .05. * * p < .01. * * * p < .001.