u; 2%.... . x l. v.2. . S215 1531‘. 5 5'. Y 3‘ I‘fi?.hw no» ‘ :3. .1 .5... :n Pt. .) fin, {:1.‘ .. ‘ is.“ mm LIBRARY 2 Michigan State 206 ‘2 _ University This is to certify that the dissertation entitled FACE AND FACEWORK IN “ELL-MEANING CLASHES: HOW AMERICANS MANAGE FACE THREATENING ACTS IN INTERCULTURAL COMMUNICATION . presented by XIAOWEN GUAN has been accepted towards fulfillment of the requirements for the Ph.D. degree in Communication W Major Professor's Signature 0'1 / (31% / 6L00 8 Date MSU is an dfinnative-action, oqud-opportmity employer PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 5/08 KIProj/Achres/ClRC/DateDue.indd FACE AND FACEWORK IN WELL-MEANING CLASHES: HOW AMERICANS MANAGE FACE THREATENING ACTS IN INTERCULTURAL COMMUNICATION By Xiaowen Guan A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Communication 2008 ABSTRACT FACE AND FACEWORK IN WELL-MEANING CLASHES: HOW AMERICANS MANAGE FACE THREATENING ACTS IN INTERCULTURAL COMMUNICATION By Xiaowen Guan This study incorporates both observational and self-reported measures to examine Ameri can communicators’ facework strategies in a problematic intercultural communication situation. Participants (N=103) completed two self-reported questionnaires and had an interaction with a confederate, who was supposedly an international student from China. The confederate made a comment regarding the subjects’ weight gain to create a well-meaning clash. Subjects’ strategies to manage the situation were measured with both the self-reported questionnaires including Likert-type scale items and open-ended questions, and the third-party observers who watched video- taped interactions. The main findings are: 1) avoidance is the primary facework strategy for Americans to manage a well-meaning clash caused by others; 2) negative face threat predicts avoidance faceowrk, and attributing the face threat to intentional offense leads to aggression facework; and 3) disagreements with the face threatening comment in a non- aggressive and non-avoiding manner, jokes and sarcasm are strategies in addition to avoiding and aggression facework. In view of the partially consistent results between the two measures, a post-hoe path model was proposed to further explore the relationship among general face needs, face threats, situational factors, and facework strategies. The final path model shows that negative face threat and situational factors mediate the relationship between other- oriented face needs and facework strategies. The results and the new variable relationship indicate that a mixed-method approach to study behavioral outcomes in how people manage problematic intercultural communication Situations. TOLM. F. iv ACKNOWLEDGEMENTS I would like to express my Sincere thanks for the great guidance, support and insights that my advisor Dr. Hee Sun Park has given me throughout the four years Of my doctoral study. She inspired me to step out Of regular research method boundaries with my dissertation and encouraged me every time when I thought it could not be done. I would also like to thank my committee members Dr. Tim Levine, Dr. Mary Bresnahan, and Dr. Yoonhyeung Choi for their valuable feedback, which improved my dissertation at each stage. I am very fortunate to have a committee that is so supportive. My sincere thanks also go to the research team members who worked with me, Wonsun Kim, Laura Cumbow, Shannon Kendall, and Jonathan Lopatin for the hours and hours of time they put into training, at the laboratory, and actual coding. Without their help, my dissertation could have never been done. Special thanks to the Graduate School and the College of Communication Arts & Sciences for awarding me the fellowships which enabled me to concentrate on my dissertation in the past few months. Lastly, I would like to thank my fiance' Ian Fitzpatrick for his unconditional support and encouragement throughout the whole process. This is for you, Ian. TABLE OF CONTENTS LIST OF TABLES ............................................................................. LIST OF FIGURES .......................................................................... INTRODUCTION ............................................................................. CHAPTER]: LITERATURE REVIEW .................................................. Well-Meaning Clashes ............................................................... Conceptualization of Face ............................................................. Facework ............................................................................... Aggression F acework ................................................................. Avoidance F acework .................................................................. Individual Level of Face ............................................................. Situational Level of Face ............................................................. Types of Attributed Intentions ....................................................... Attributional Confidence in Intercultural Communication ...................... CHAPTER 2: PRELIM STUDY ........................................................... Results of Face Needs and Face Threats ............................................ Results of Attribution and Attributional Confidence ............................. Additional Findings and Explanations ............................................. Research Questions ................................................................... Alternative F acework Strategies ...................................................... CHAPTER 3: METHOD AND ANALYSIS ............................................. Participants ............................................................................. Procedure ............................................................................... Scenario selection ............................................................... Detailed procedure .............................................................. Measures ................................................................................ Dependent variables ............................................................ Independent variables .......................................................... Control variables ................................................................ Analysis ................................................................................. CHAPTER 4: RESULTS .................................................................... Overview ............................................................................... Characteristics of Avoidance and Aggression Facework ........................ Research Question 1 .................................................................. Research Question 2 .................................................................. Additional Findings ................................................................... vi viii X 1 4 4 8 10 12 l3 14 16 20 23 26 26 27 29 36 37 41 41 41 41 43 47 47 49 51 53 55 55 55 6O 7O 72 CHAPTER 5: DISCUSSION AND ADDITIONAL ANALYSIS ..................... 75 Preference Of Avoidance ............................................................. 75 Use of Nonverbal Cues ............................................................... 78 Sex Difference ......................................................................... 79 Explanation of Insignificant Findings .............................................. 81 Additional Analysis ................................................................... 85 CHAPTER 6: GENERAL DISCUSSION ................................................ 89 Level of Agreement between Self-Reported and Observational Measures ....................................... 89 Mixed Method Approach ............................................................ 97 Practical Implications ................................................................. 98 Limitations .............................................................................. 101 Future Research ........................................................................ 102 Conclusion .............................................................................. 103 APPENDIX A ................................................................................ 105 APPENDD( B ................................................................................. 108 REFERENCES ............................................................................... 110 vii LIST OF TABLES Table 1. Means, standard deviations, correlations, and reliabilities .................. Table 2. Observed frequency of facework strategies coded by third-party coders .......................................................................................... Table 3. Observed frequency of detailed facework coded by third-party coders ......................................................................................................................... Table 4. OLS regression results (first set regression with face needs and face threat as predictors and self-reported scale avoidance as dependent variable) ...... Table 5. OLS regression results (second set regression with attribution types as predictors and self-reported scale avoidance as dependent variable) .................. Table 6. OLS regression results (first set regression with face needs and face threat as predictors and self-reported scale aggression as dependent variable) ...... Table 7. OLS regression results (second set regression with attribution types as predictors and self-reported scale aggression as dependent variable) ............... Table 8. Logistic regression results (first set regression with face needs and face threat as predictors and coder coded avoidance as dependent variable) ............ Table 9. Logistic regression results (second set regression with attribution types as predictors and coder coded avoidance as dependent variable) ..................... Table 10. Logistic regression results (first set regression with face needs and face threat as predictors and coder coded aggression as dependent variable) ............ Table 11. Logistic regression results (second set regression with attribution types as predictors and coder coded aggression as dependent variable)... . Table 12. Regression results comparison between prelim and current study with face needs and face threat as independent variables and avoidance as dependent variable... ... ............................................................................................................ Table 13. Regression results comparison between prelim and current study with attribution types as independent variables and avoidance as dependent variable... Table 14. Regression results comparison between prelim and current study with face needs and face threat as independent variables and aggression as dependent variable .......................................................................................... viii 48 55 57 62 63 64 65 66 67 68 69 9O 91 92 Table 15. Regression results comparison between prelim and current study with attribution types as independent variables aggression as dependent variable ...... 93 ix LIST OF FIGURES Figure 1. Original path model with path coefficients ....................................... 86 Figure 2. Revised path model with path coefficients ........................................ 87 INTRODUCTION A common problem in face-to-face intercultural communication is when an actor behaves competently and attempts to convey a well-meaning intent based on his or her own cultural norms, yet the communicative act or the message is perceived to be otherwise according to the other actor’s cultural norms (Brislin, 1993). For example, when a younger Chinese person calls someone who is much older old in a social occasion, usually he or she means to show respect, since senior citizens are generally respected for their age and seniority in China. The senior Chinese person would not be offended at being called old since he or she interprets the message as the younger Chinese intends. However, when a Chinese person communicates the same message to an American person, the well-meant intent might not be necessarily interpreted in the same way, because the word old can have negative connotations in American culture. The message may be considered to be not only impolite, but also ofl‘ensive by the American, in contrast to what the Chinese person intends. Brislin (1993) called this kind of communication problem in intercultural context a “well-meaning clash” (p.10). When social norms are violated in social interactions, undesired emotions might be felt by social actors, for whom the situation is a social predicament (Schlenker, 1980). Schlenker and Darby (1981) assumed that when actors are in social predicarnents, they usually engage in actions to address the undesirable consequence, repairing the damage to identities and images. These tactics that aim to be consistent with one’s image and impression needs are called facework (Goffman, 1967). In intercultural communication, using tactics to manage well-meaning clashes effectively is an important skill to ensure the quality of communication, because if those well-meaning clashes are not dealt with well, they might lead to misattributions, disputes, and conflicts, which can further reinforce negative stereotypes of other cultural groups and potentially obstruct inter-group understanding. Although maintaining a desired image in social interactions might be a common goal for social actors, it is not always easy to obtain. The difficulty for improving skill in managing well-meaning clashes, in particular, lies in that the party who violates the social norm remains unaware of his or her offense. In the above example, when a young Chinese person calls an American old, he or she attempts to convey a good intent, instead of insult or hurt, but his or her audience might feel insulted or hurt due to their cultural context. Although the Chinese person is making an innocent mistake, the American might not perceive the act as an innocent cultural mistake, but rather an example of the other person being ill-educated or socially inappropriate. Since the Chinese person does not recognize that any inappropriateness has happened, and so would not correct the mistake, it is up to the American to initiate actions to address his or her lost face after a well- meaning clash takes place. Therefore, important questions are raised regarding what the tactics are that people use to address unfulfilled image needs under such circumstance, and why the tactics are used in these situations. To this end, this study will apply face and facework concepts to I) explain Americans’ reactions towards well-meaning clashes caused by another communicator who does not share the American culture, 2) examine the Americans’ strategies in dealing with the well-meaning clashes, and 3) investigate the findings by using different research methods. This manuscript will begin with a review of the literature on well-meaning clashes in intercultural communication, the concepts of face and facework strategies in relation to manage well-meaning clashes, and attribution of problematic communication situations. Then, it will discuss the results from the prelim study on the same topic and delineate the rationale for conducting the study with both self-reported and observational measures. Two research questions will be presented. Last, the results from the observations and self-reports will be reviewed and discussed, along with a discussion of the implications, limitations and directions for future research. CHAPTER 1 LITERATURE REVIEW Well-Meaning Clashes According to Brislin (1993 ), well-meaning clashes take place when communicators from different cultures behave properly, and in a socially skilled manner according to the norms of their own culture, but these very behaviors are considered to be improper or inappropriate in other cultures. Previous intercultural communication studies have identified that behavioral norms which are not shared by the intercultural communicators are one of the fundamental causes of well-meaning clashes. A case in point is that, in US, a common reply to a sincere compliment is usually a simple sentence, such as “thank you”, displaying one’s acceptance of and appreciation for the compliment (Hu & Grove, 1991), whereas in Chinese culture, a more socially acceptable response to a compliment is a denial or rejection of the compliment. The different responses towards the same compliment between the Chinese person’s and that from an American lies in the Chinese peOple’s concern for humility and modesty in social interactions (Hu & Grove, 1991) and the Americans’ value of self-enhancement and self- affirmation (Gao & Ting-Toomey, 1998). To Chinese, deprecating oneself in public is a sign of good manners, respectability, and strength, while to Americans, people should be proud of their achievements, and it is appropriate to validate praise from others. The two different responses, which are both socially appropriate in their own cultures, may lead to a clash of meanings when used in the other’s culture. When an American hears a denial from a Chinese person in response to their compliment, he or she may be confused, possibly interpreting the reply as the communicator being ungrateful and unappreciative (Gao & Ting-Toomey, 1998). When a Chinese person does not receive a denial in retum for the compliment, he or she may attribute it to arrogance and lack of education in social manners of the other communicator. Difficulties and problems in intercultural communication thus arise when intercultural communicators who are unaware of the differences behave according to their own cultural norms, and unintentionally offend conversation partners who do not share the same cultural norms. In intercultural communication, it is common that the communicators do not share the same native language, and therefore, to engage in a meaningful conversation, one party has to speak the other party’s native tongue or both parties have to speak a second language, which is often English. Cross-cultural linguistic studies have found that when people speak in a second language, they tend to rely on the sociolinguistic knowledge of their native language to communicate their intentions (Blum-Kulka, House, & Kasper, 1989). Jung (1999) discovered that when communicating in English, advanced Korean learners of English use “sorry” instead of “thank you” as a response when they attempt to express appreciation for someone who has done a favor for them. The author explained that in Korean, the word sorry can function as an expression of the speaker’s recognition of and appreciation for others’ efforts in fulfilling the favor, which is how “thank you” functions in English (Jung, 1999). In other words, the Korean learners of English transfer their first language linguistic and pragmatic knowledge when using apology in a second language context. Clearly, the tendency of transferring the linguistic and programmatic knowledge of one’s native language into a second language context can lead to errors and failure in intercultural interactions, and further hinder the effectiveness of intercultural communication (Blum-Kulka, et al., 1989). Thus, speaking in a non-native tongue can increase the occurrence of well-meaning clashes, especially when the linguistic features do not match between the native languages of the intercultural communicators. Cultural differences do contribute to the occurrence of communication problems; however, this does not mean that well-meaning clashes are sure to take place whenever communicators are in intercultural contexts. Comparative cross-cultural researchers have endeavored to uncover cultural differences in various aspects of social interaction in order to educate intercultural communicators regarding cultural differences and prevent such problems; Cultural differences are found in many domains such as cognitive styles (e.g., Steward & Bennett, 1991), communication styles (e. g., Hall, 1978; Gudykunst & Ting-Toomey, 1996), and conflict styles (e.g., Ting-Toomey, 1988; Ting-Toomey & Oetzel, 2001). Increasing intercultural communicators’ knowledge and awareness about these differences, and training them to be mindfirl, sensitive, and attentive would undoubtedly facilitate the intercultural communication process. One limitation of the training programs aiming to improve intercultural communicators’ knowledge and awareness of cultural differences, however, is the one-sided emphasis on preventing misunderstanding from happening. That is, to train the communicators about the dos and don’ts so that they can avoid any miscommunication prior to actual encounters. The problem is that even if intercultural communicators are equipped with cognitive information, there is no guarantee that their communication process would become error—free in practice. In other words, no matter how competent a communicator might be, he or she cannot be immune to all communication problems. Yet, the training problems do not give much attention to the coping strategies after a communicative problem has taken place. Hence, in addition to learn how to prevent well-meaning clashes from happening, it is equally important for researchers and trainers, as well as intercultural communicators to learn about corrective strategies of managing problematic situations after they have taken place. Current intercultural communication research is limited in the number of studies focusing on the corrective strategies after a problem takes place. Therefore, the present study aims to explain communicators’ reactions to well-meaning clashes and to explore individuals’ strategies for managing the clashes. Intercultural communicators will benefit from the study through gaining more knowledge about how to manage problematic communication, which will help increase their communication competence. Two key concepts in well-meaning clashes are norms and norms violation. Sherif (1936) conceptualized social norms broadly to include “customs, traditions, standards, rules, values, fashions, and all other criteria of conduct which are standardized as a consequence of the contact of individuals, as specific cases of ‘social norms’” (p.3). Along the same lines, Goffman (1976) also believed that social norms are guidance for actions, regulating interactions between communicators across relationship types. Norms often function invisibly to social actors, as social actors may not be aware of norms until the guidance and rules are violated (Sherif, 1936). Well-meaning clashes are characterized by one party’s unintentional breaching of the social norms of the other party, in which case the first party creates a social predicament for the second. The person whose cultural norms are violated might experience negative emotions such as awkwardness, embarrassment, or even anger. This outcome is obviously undesirable in social interactions. When undesirable consequences are foreseeable, communicators can make efforts to prevent the conversation from leading to negative outcomes. For example, highly personal questions may engender discomfort for the person who is being questioned. Therefore, such questions are usually avoided in interactions. When social actors cannot foresee a norm violation or prevent it from happening, however, they will have to engage in actions to repair the social damage after the breach has taken place. This social phenomenon is essentially associated with the concepts of maintaining social actors’ positive image and managing a desirable impression in social interactions, which is where face and facework are concerned. To understand people’s reactions in well-meaning clashes, an examination of face and facework is needed. Conceptualization of F ace Face is a pervasive concept in social interactions. Ho (1976) claimed that there is not a facet of social life to which the question of face is irrelevant. Face can be used to explain people’s various communicative behaviors in social interactions, including why peOple apologize when they make a request, why they laugh when they are embarrassed in public, and why they change the topic during certain conversations (Tracy, 1990), all of which aim to maintain, save, or prevent damage to communicators’ faces. Empirical research has provided strong evidence for the powerful influence of face on communication behaviors, namely, politeness (e.g., Brown & Levinson, 1987), compliance gaining (e. g., Baxter, 1984; Tracy, Craig, Smith, & Spisak, 1984), emotional disclosure (e.g., Shimanoff, 1985, 1987), and conflict styles (e.g., Oetzel et al., 2001; Oetzel & Ting-Toomey, 2003). Face is directly associated with image and identity, which communicators in social interactions are assumed to desire. Goffman (1967) defined face as “the positive social value a person effectively claims for himself by the line others assume he has taken during a particular contact,” and “face (italic original) is an image of self delineated in terms of approved social attributes—albeit an image that others may share, as when a person makes a good showing for his profession or religion by making a good showing for himself” (p. 5). Extending Gof’fman’s (1967) conceptualization of face, Brown and Levinson (1987) connected politeness to face. According to Brown and Levinson (1987), face is “the public self-image that every member of a society wants to claim for himself/herself” (p.61), and face can be divided into two basic kinds: positive face (i.e., the desire to be appreciated and approved of by selected others) and negative face (i.e., the want to be unimpeded and free from imposition). Brown and Levinson (1987) asserted that to be polite is to satisfy social actors’ positive and negative face in interactions. To distinguish these two types of face from other face related concepts, positive face and negative face will be termed positive face needs and negative face needs in this study. Face needs are desirable image needs; however, they are not always satisfied. Social interactions in daily life can bring all kinds of threats to one’s face needs. Brown and Levinson (1987) called verbal or nonverbal acts that run contrary to the actor’s desired face needs face threatening acts (F TA). Requests, suggestions, threats, warnings, and reminders primarily threaten one’s negative face, since they impose on other person’s autonomy of action (Carson & Cupach, 2000). Disapproval, disagreement, complaints, and insults, on the other hand, are inherently threatening one’s positive face, as they attack the other person’s need to be approved of or appreciated for (Brown & Levinson, 1987). Some acts can threaten both positive and negative face. For example, when being criticized, one’s need to be approved of is not satisfied. At the same time, the directive and constraining nature of criticism also exposes one’s negative face to attack (Leichy & Applegate, 1991). In a well-meaning clash, an intercultural communicator’s message, although uttered out of good intent by the message sender, can be perceived to be a criticism or a complaint by the bearer and, consequently, threatening the hearer’s positive and negative face. For instance, when being referred to as old by a Chinese speaker, an American hearer would probably feel his or her need to have a positive image threatened, because the word old has such negative connotations in American culture as incapable, incompetent, and unattractive. At the same time, the American hearer might also feel that the Chinese person’s comment is intrusive, since looking old or not is a personal issue for Americans and should not be discussed in public. In other words, the American hearer would perceive both of his or her positive and negative face to be threatened when being addressed as old by a Chinese person. Facework When people’s desired images are threatened, attacked, or lost, people are motivated to engage in facework strategies to address or to repair the damage (Schlenker & Darby, 1981). Metts (1997) defined facework as “a variety of communicative devices available to interactants for preventing face loss (both their own and the others’), restoring face if lost, and facilitating the maintenance of poise in the advent of disrupted interactions” (p.374). In other words, facework is a set of communicative behaviors that aim to be consistent with one’s face needs (Goffman, 1967). Goffman (1967) described facework as a process that does not only involve one’s own face management but also that of the other social actors during interactions. Brown and Levinson (1987) provided a 10 detailed framework of the types of facework strategies that could be employed as a function of the levels of face threats. Empirically, facework has been examined in a variety of social interaction situations, including making and responding to requests (e. g., Holtgraves & Yang, 1990, 1992), coping with embarrassment (e. g., Cupach & Irnahori, 1993; Irnahori & Cupach, 1994), and managing complaints and criticism (e.g., Cupach & Carson, 2002; Trees & Mansuvo, 1998). F acework strategies are tactics that either prevent a face threatening act from taking place or manage situations after someone’s face is attacked. Studying facework strategies can help communicators take each other’ s face needs into consideration, and more effectively manage their own images as well as that of the others in social interactions. Ultimately, a greater store of knowledge of facework strategies would contribute to improving communication competence and creating harmonious interpersonal relationships. Goffman (1967) divided facework into two basic types, preventive and corrective. Preventive facework refers to strategies that social actors employ to avoid situations that would expose their face to threat, and a typical type of preventive facework is to avoid topics or activities that could lead to face threats. Corrective facework is used when face threatening acts have taken place, and it includes strategies that correct the wrongdoing in the incident as a means to address and restore the damaged face. In the current study, corrective facework strategies are of greater relevance. In the well-meaning clash where the Chinese actor unknowingly behaves in an impolite or even offensive way according to the American’s cultural standard, the Chinese actor’s acts pose face threats to the American person. The American actor cannot foresee the incident, and therefore, cannot engage in any preventive measures. That is, preventive measures are not available for the 11 American actor, and if the American chooses to repair his or her damaged face, he or she would be engaged in corrective facework strategies. It is assumed that corrective facework is usually initiated by the offender (Goffman, 1967). In well-meaning clashes, however, since the offender remains unaware of the face threat done to the other party, and there is no third-party who might remind the offender of the offense, it will be up to the offended party to initiate corrective facework strategies. That is, when the Chinese actor threatens the American’s face in a well-meaning clash, the American has to introduce corrective actions to repair the damage on his or her face. Otherwise, no actions will be taken to address the American’s threatened face. Aggression F acew ork For an offended party, to correct his or her face damage caused by the offender, one approach is to reclaim his or her desired face at the expense of the offender’s. This is called “the aggressive use of facework” (Goffrnan, 1967, p24). The initiator of aggressive facework aims to “introduce favorable facts about himself and unfavorable facts about the others” (Goffman, 1967, p.25). In other words, while putting the others in disadvantage, the aggressor demonstrates his or her advantages in the topic under discussion. In Hatch’s (1987) observation study of kindergarten children’s face management strategies with their peers, it is found that in order to improve their own status, children would engage in aggressive use of facework by pointing out others’ mistakes, weaknesses or inadequacies. By making their peers look bad, children address their own image needs (Hatch, 1987). Applying the aggressive use of facework strategies in how an American might manage a well-meaning clash, an American person might 12 attack the Chinese person’s face, including pointing out the offender’s inadequateness, confronting the offender, and demanding further facework from the offender, an apology, for example. Avoidance F acew ork An alternative approach to manage a face threatening act that already takes place is avoidance. Goffman (1967) called the avoidance process “making a gracious withdrawal” (p.15) from interactions. When avoiding, social actors choose to overlook the face threat, and pretend that no damage on their face has happened. Cupach and Metts (1994) explained that, in certain face threatening situation, without inviting any further attention to the face threatening act that already has taken place would be effective in saving one’s faces. In well-meaning clashes, by ignoring the face threatening act, pretending nothing serious has happened, and continuing the conversation, the offended person can minimize the negative consequences brought about by the face threatening act that has already occurred, and reduce its impact on interactions that follow. Another avoidance strategy, pointed out by Goffman (1967), is that instead of overlooking the offense, a social actor “openly acknowledges an incident as an event that has happened, but not as an event that contains a threatening expression” (p.18). Displaying a lack of importance of the face threat can help reduce the level of face damage. Further avoidance strategy includes sending subtle cues to warn the offender not to continue on the face threatening topic (Goffrnan, 1967). Such cues are usually conveyed nonverbally, especially when social actors fails to conceal his or her negative expressions (Goffman, 1967) Both avoidance and aggression under investigation are corrective facework 13 strategies. According to the prelim study conducted by the author, the two types of facework strategies were not significantly correlated with other, and thus are assumed to be independent of each other instead of on the opposite ends of the same dimension. A casein point is that a highly avoiding strategy can be viewed as highly aggressive at the same time. During an on-going interaction between person A and person B, for instance, when person A accidentally delivers a message which is perceived to be highly offensive by person B, person B may choOse to leave the scene as an extreme form of avoidance, which might be considered to be highly aggressive and face threatening by person A. This type of response is high on both the degree of avoidance in its physical form, and the degree of aggression in its implications. On the other hand, when social actors do not perceive the face threatening act as threatening at all, and therefore, choose to address it as any normal conversation topic, the facework might be neither avoidance nor aggression. Individual Level of F ace Research studies have been carried out to examine the factors that influence people’s facework strategies when they attempt to address their faces in face threatening situations. Individual differences have been identified as one of the most important factors that affect how people use facework strategies. Goffrnan (1967) claimed that everyone wants to be “in right face,” and avoid being “in wrong face” (p8). Brown and Levinson (1987) assumed that all humans not only want both positive face and negative face, they also choose means that will satisfy those ends. The theoretical assumption of these assertions is that face is a basic need of all social actors. However, not every social actor shares the exact same amount of need for their faces. In Snyder’s (1979) description 14 of the self-monitoring process, he argued that people who are highly sensitive to the situational and interpersonal appropriateness of behaviors tend to employ more verbal and nonverbal cues to control their self-presentation than those who are less sensitive to social surroundings. Although Snyder (1979) did not use the concept of face, the idea of self-presentation is very similar to face and facework, as facework concerns various actions one employs to present his or herself. Studies on self-esteem and cognitive complexity also showed that people who differ in their level of self-esteem and cognitive complexity vary their use of strategies to maintain their faces (e.g., Applegate, 1982; Schlenker, 1980; Schlenker, Weigold, & Halarn, 1990). In a word, the variations observed in people’s use of facework strategies are a function of individuals’ stable dispositions, such as self-esteem, self-monitoring, and cognitive complexity. Additional evidence that face needs are treated as a disposition can be found in Face Negotiation Theory (Ting-Toomey, 1988, 2005). The theory assumes that cultural and societal factors such as individualism-collectivism and power distance would determine the individuals’ orientation of face. Individualists are believed to have higher self-centered face needs than collectivists, while individuals from high power distance cultures are more oriented to others’ faces (Ting-Toomey, 1988). The cultural dimension on face affects individual’s conflict strategies. Those who are more concerned with self- face needs will be more likely to engage in dominating strategies in conflicts in order to address one’s own face needs, while those who are more concerned with other-face needs will be more inclined to employ avoiding strategies, in order to save the other’s face (Ting-Toomey, 1988). Face Negotiation Theory does not specify the context where face needs are 15 examined, or pinpoint situations where certain conflict styles are used. In other words, face needs are discussed in general terms, independent of any particular situations. This shows that face needs are treated as individual dispositions, remaining constant across situations. In a series of empirical studies using Face Negotiation Theory, Ting-Toomey and her associates measured face needs as an individual trait (e. g., Oetzel et al., 2001; Oetzel & Ting-Toomey, 2003; Ting-Toomey & Kurogi, 1998), and they found support for a significant association between individuals’ general face needs and their preferences for conflict strategies. This line of research further suggests that individual’s face needs are trait-like qualities residing in individuals, and can explain individual differences in their communicative behaviors. Situational Level of Face Individual difference in their face needs influence people’s reactions to various situations; however, such difference of outcome might also be affected by another important aspect of face, the situational level of face. The amount of face threat one perceives in a particular situation might also have impact on people’s reactions towards the situation. Tracy (1990) argued that “face is a social phenomenon; it comes into being when one person comes into the presence of another” (p.210). That is, one’s face is in relative terms to the other person in the interaction, and even for the same face threatening act, depending on the relationship with the other party, his or her reaction to the situation might differ. Brown and Levinson’s (1987) politeness theory provided a model of strategies for dealing with situations where their face needs are threatened, which are characterized with face threatening acts (F TA). According to Brown and Levinson (1987), the extent of 16 face threat would directly affect how individual reacts to the PTA and their choice of the politeness strategies, and the amount of face threat is derived from three situational variables: the social distance between the speaker and the hearer, the relative power of the speaker compared with the bearer, and the intrinsic degree of the imposition of the act. For example, asking someone for a ride is a face threatening act, as it threatens the person’s need to maintain the autonomy of his or her action, yet the act would be more face threatening if it were done by a superior rather than a peer, a stranger rather than a friend, and if the ride would take an hour rather than fifteen minutes. For the person who is requested, he or she might respond differently depending on if the person who requests is a superior or a peer, and whether the ride takes an hour or fifteen minutes. Clearly, the degree of the threat to one’s face might differ across situations given that the three variables carry different weight in different situations. The amount of face threat determines which type of politeness strategy should be used (Brown & Levinson, 1987). Empirically, the level of perceived face threat is found to be associated with a variety of communication outcomes, for example, communicative competence (e. g., Carson & Cupach, 2000), remedial and impression management strategies (e. g., Braaten, Cody, & DeTenne, 1993; Cupach & Carson, 2002; McLaughlin, Cody, & O’Hair, 1983). The impact of face threat on subsequent communicative behaviors can also vary depending on the individual’s general face needs. In other words, individual face needs might interact with situational face threat and influence individuals’ communication. For example, in the situation where someone requests for a ride to the airport, the perceived levels of face threat might moderate the relationship between face needs and facework strategies. For an individual who values his or her own negative face needs, he or she 17 may reject the request. However, if the perceived face threat level is low, for example, it only takes fifieen minutes, and requested by a close friend, he or she might not perceive the request as an imposition to his or her autonomy, and therefore, might grant the request. For the same individual, if he or she perceives a high level of face threat in the request, for example, it takes over an hour and requested by an acquaintance, he or she might be more likely to decline the request. In view of this, when examining how face affects communicative outcomes, not only the individual’s general face needs and the situational face threat must be examined, but the interaction between the two variables has to be taken into account. Prior face-related studies have been limited in their separate focus on either general face needs (e.g., Oetzel et al., 2001, Oetzel & Ting-Toomey, 2003) or situational face threats (e. g., Carson & Cupach, 2000). Few studies have combined the two factors or considering possible interaction effects between the two factors, which left the face literature lopsided. The current study aims to fill in the gap, examining how individual face needs, and situational face threats would separately and jointly affect individual’s facework strategies in intercultural well-meaning clashes. To capture the dynamics of how face firnctions in empirical studies, researchers divided the concept of face into more detailed types based on two broad dimensions: self versus other and positive versus negative. Combining these two dimensions results in four basic types of general face needs: self positive face (SPF), which is one’s own needs of personal self-image to be approved and appreciated; other positive face (OPF), which is the needs to protect others’ image to be approved of and appreciated; self negative face (SNF), which is the needs of maintaining one’s own freedom of action; and other 18 negative face (ONF), which is the needs of keeping others’ freedom of action. An individual can have all of the four types of face needs, though the degree of needs might not necessarily the same across the four types for each individual. Face Threatening Acts (F TA) run contrary to people’s face needs (Brown & Levinson, 1987). In view of the four types of face needs, F TA thus can be categorized into four types based on the primary threat to the type of face as well: threat to SPF, threat to OPF, threat to SNF, and threat to ONF. In the current study, threats to the offended person’s self-positive face and self-negative face are more relevant because of the nature of well-meaning clashes. Take the situation discussed at the beginning of the paper as an example, when a Chinese person who intends to show respect addresses the conversation partner, an American, as old, he or she poses a FTA to the American. The Chinese person is the naive offender who is responsible for the F TA but is not aware of his or her offense, while the American is the offended party. In the eyes of the American person, the FTA caused by the Chinese person constitutes a threat to his or her own face (i.e., threats to self positive face and threat to self negative face). One may argue that, the F TA can also be considered as a threat to the Chinese person’s face (i.e., threat to other’s face). By calling the American old, the Chinese person appears impolite or inconsiderate in front of the American. The FTA makes the Chinese person looks bad, hurting the other’s positive face from the perspective of the American person. This may be the case; however, the threats to either other positive or other negative face are projected from the Americans’ perceptions of the situation. Such projections are secondary to the threat to one’s own face needs, which are assumed to directly motivate the social actors to engage in facework. The nature of the well-meaning 19 clashes determines that the threat primarily targets the face of those whose cultural norms have been violated. Thus, the situational threats to the offended person’s (i.e., the American’s) self-positive and self-negative face are more relevant to the topic of interest. Types of A ttributed Intention According to Goffman (1967), the perceived responsibility of the face threatening act in the eyes of the offended person can vary his or her facework strategies. Goffman (1967) claimed that there are three different types of intentions an offended person could attribute when his or her face is under attack. The first type is called an innocent mistake, which refers to the idea that the offended person perceives the offender’s act to be an honest and unintended mistake. In other words, the offended person believes the offender lacks knowledge that his or her act would pose a threat. To the offended party, if the Offender has foreseen the consequences of his or her action, he or she would not have done it in the first place (Goffman, 1967). In the earlier example where the Chinese person addresses an American as old in order to show his or her respect, if the American believes that the Chinese person does not have sufficient knowledge about the negative connotations of the word old in social interactions, the American would view the incident as a faux pas. The American person probably thinks that if the Chinese person has known calling someone old in US. culture could pose a face threat to other Americans, the Chinese person would not have done it in the first place. The second type of attributed intention is called an intentional offense, which refers to the idea that the offended person perceives the offender’s act to be an intentional and malicious offense. When a face threat is perceived to be an intentional offense from the offended person’s perspective, the offender is assumed to be aware of his or her 20 action and has anticipated the negative consequences. The fact that the offender still chooses to conduct the face threatening act despite the knowledge shows that the offender neither cares about the offended person’s feelings, nor values the offended person’s face. The face threatening act is therefore considered to be out of meanness and to be intentionally hurtful. This type of attribution usually happens in well-meaning clashes, because the offended party might not be able to comprehend how the other party could utter such an offensive comment without knowing it and so attributes impoliteness and rudeness to the person. The third type of perceived intention is called incidental offense, which Goffman (1967) defined as “an unplanned but sometimes anticipated by-product of the action the offender performs in spite of its offensive consequences, although not of spite” (p.14). When the offended person attributes the act to be incidental, the offender act is believed to be able to foresee the consequences of the act, but perform the act without an intention to hurt. Incidental offense might be due to the offender’s carelessness or inexperience in social interactions. For example, when being teased, the person usually attributes teasing as incidental offense. Teasing usually takes place in a light-hearted and playful manner from the angle of the teaser; however, teasing might be perceived as hurtful and unpleasant by the person who is teased (Kowalski, 2000). Under such circumstances, the offended person, though experiencing negative emotions, is still aware that the offender does not conduct the face threatening act out of meanness or spite, and the offended person might believe that the offender could justify his or her actions, despite the undesirable outcome. 21 Assessing the, intention of others in social interactions is the individual making an effort to understand the situation in order to decide how to react. Allred (2005) claimed that understanding the causes of other’s behaviors can help people make appropriate and adaptive responses. Sillars (1980) found that in college roommate conflicts, attribution directly affects peeple’s conflict strategies. Bennett and Earwaker (1994) also discovered that offended person’s rejection of apology from the offender is strongly associated with offended person’s attributed intention of the offender, and the perceived degree of the offender’s responsibility. Kelley and Michela (1980) explained that depending on whether a person’s action is perceived as being intentional or unintentional, they will be treated differently. When being intentional rather than unintentional is attributed to negative behavior, the negative outcome elicits more blame (Kelly & Michela, 1980). In examining individual reactions towards well-meaning clashes, how the offended person attributes the intention of the face threatening act could also have an impact on his or her reactions to the well-meaning clashes. For example, if the American perceives the Chinese person to have been intentionally offensive or malicious when the Chinese person called the American old, the American is more likely to engage in aggressive strategies and directly confront the Chinese. Ifthe attribution is an incident or an innocent mistake, the American might not feel necessary to attack the Chinese person’s face, and may consider using less aggressive strategies to manage the threatened face. Believing that the Chinese did not mean to hurt, it is also possible that the American may remind the Chinese in a non-threatening way that his or her comments are inappropriate, and provide the Chinese with information on the differences of cultural 11011118. 22 Attributional Confidence in Intercultural Communication In addition to the perceived intention, another important factor that may affect an intercultural communicator’s reaction to the well-meaning clash is their level of certainty about the other person’s intention. Berger and Calabrese (1975) maintained that uncertainty involves peOple’s inability to explain and predict the other person’s behavior in interpersonal communication. Uncertainty was often operationalized as attributional confidence, which referred to the level of accuracy of interpreting one’s action in a retrospective manner and predicting one’s behavior in the future (Clatterbuck, 1979). The more confident one has about the attribution, the less uncertain one is. One factor that helps improve the attributional confidence is the similarities between communicators (e.g., Cannon-Bowers & Salas, 2001; Klimoski & Mohammed, 1994; McLeod & Chaffee, 1973; Triandis, 1959, 1960). In intercultural communication, such similarities are assumed to be significantly less compared to in intracultural communication. A lack of similarities in world-views, emotional displays, and behavioral norms between intercultural communicators not only decrease the amount of information shared by communicators but also reduce their accuracy in interpreting each other’s intention in their message (e.g., Armstrong & Kaplowitz, 2001; Lee & Boster, 1991; Li, 1999). Gudykunst (1988) concluded that people usually experience a higher level of uncertainty when communicating with an out-group member who does not share the same cultural background than with an in-group member. Being uncertain about understanding, explaining, or predicting the other person’s behavior in social interaction usually causes the social actors discomfort, increasing the difficulty to continue interactions, and thus should be avoided (Berger & Calabrese, 23 1975). In order to reduce uncertainty, people engage in actions to seek information about each other, for example, asking questions. When the offended party is uncertain about the other actor’s intention of delivering the offensive message, he or she may ask the other person to give an explanation of the message in order to clarify the situation. In a well- meaning clash, asking the offender questions would serve to reduce the offended person’s uncertainty, especially if they learn that the offended person is making an innocent mistake or being intentionally hurtful. At the same time, by asking questions, the offended person can also address his or her lost face, as the act might eventually lead the offended person to realize the F TA, and offer an apology on his or her own. Reducing uncertainty in social interactions is not the only consequence when uncertain arises. Kiesler (1973) noted that, under certain circumstances, increased uncertainty is actually preferred in social interactions. Eisenberg (1984) also found that maintaining a certain level of ambiguity can help empower the communicators in an organizational context. That is, instead of seeking more information, people may use uncertainty to their advantage. Alternatively, people may leave uncertainty unattended without any further actions. Berger (1979) stated that, when individuals do not expect any further interaction with the other party, they might not pursue any additional information to resolve their uncertainty. Mishel (1990) reported that when patients suspect that their illness might be terminal, they choose not to seek any information about the illness from their doctors as a means to cope with the high level of uncertainty. In interpersonal research, it is found that people would avoid discussing certain topics with their partner when they are uncertain about the relationship development (Affifi & Burgoon, 1998). In a well-meaning clash, the offended party might experience a high level of uncertainty of 24 attributing the offender’s act; however, the offended person might not perceive any chance to see the offender again. He or she might thus terminate the conversation, avoiding further interaction with the offender as a way to deal with the uncertainty. In short, uncertainty can moderate people’s reactions in well-meaning clashes. The offended party, who is experiencing uncertainty and a low level of attributional confidence, might handle the uncertainty by either directly asking for more information about the other party’s intention or indirectly terminating the conversation and avoiding further communication with the other person. 25 CHAPTER 2 PRELIM STUDY The prelim study was designed to examine the relationship among face needs, situational face threats and facework strategies. The research method was self-reported Likert-type scales. Three steps were carried out in order to examine the relationship among the variables. First, as a pre-study, eight vignettes that were assumed to have the well-meaning clash effect were tested with both American (N = 88) and Chinese participants (N = 76). The results showed that four out of the eight vignettes depicting intracultural communication scenarios were considered as polite or appropriate by Chinese participants whereas impolite or inappropriate by American participants. Second, the four vignettes were phrased with Chinese offenders, and tested with a different sample of American participants (N = 216). Multi-level analyses were used to parse out the impact of individual level of face needs and situational level of face threat on facework strategies across various well-meaning clashes. Multiple-regression analyses were also conducted in order to examine the effects of the three types of attributed intentions and the level of attributional confidence on facework strategies across different vignettes. Third, a follow-up study was conducted with another sample of American participants (N = 183) by using the same four vignettes with American offenders in order to compare and contrast the results from the intercultural context in the main study. The major findings and the implications are presented as follows. Results of F ace Needs and Face Threats The results of the multi-level analyses showed that both the dispositional face needs and the situational face threats have impact, separately and jointly, on the 26 individual’s use of avoidance and aggression facework strategies. Specifically, compared with individual level of face needs, situational level of face threats accounted for a large portion of the variance in the use of avoidance and aggression strategies. This means that when people encounter a face threatening act posed by an intercultural communicator, their intention to either avoid continuing the conversation with the intercultural communicator or directly confront the person is largely determined by the level of the face threat the person perceives in the particular situation. In addition to the influence of face threats on facework strategies, the joint impact between individual face needs and situational face threats on avoidance facework strategies was also observed. Other-oriented face needs moderate the relationship between positive face threat and avoidance. The magnitude of the relationship between positive face threat and avoidance differs when individuals have high levels of other- oriented face needs versus when they have low level of other-oriented face needs. More specifically, other-positive face needs accentuate the positive relationship between positive face threat and avoidance, whereas other-negative faced needs weaken the positive relationship between positive face threat and avoidance. Results of Attribution and Attributional Confidence In addition to the influence of the individual level of face needs and the situational level of face threats on facework strategies, the effects of attribution and attributional confidence were also examined. Multiple regression analyses were conducted for each of the four vignettes with the three types of attributed intention and attributional confidence as indicators, and facework strategies as outcome variables, respectively. The results showed that particular types of attribution have a main effect on aggression facework 27 across the four vignettes, and the confidence of the attribution has a main effect on avoidance facework across the four vignettes under investigation. Type 1 attributed intention, (i.e., innocent cultural mistake) received the highest ratings among the three types of attributed intention in all of the four vignettes. Participants believed that the intercultural communicator’s FTA is most likely due to a cultural mistake in all the four vignettes. Multiple regression analyses showed that type 2 attribution (i.e., intentional offense) functioned as a significant positive predictor of aggression facework. That is, the more the participants perceived the other person as being intentionally hurtful, the more likely they were to resort to aggression facework strategies. Attributing the FTA to the intercultural communicator’s malicious intentions seems to lead to confronting the offender directly. As indicated earlier, type 1 intention (i.e., innocent cultural mistake) received the highest ratings among the three types of attribution in all of the four vignettes; however, the level of confidence in this type of attribution was rather low. None of the attributional confidence scores in the four vignettes were significantly different from the scale mid- point. In other words, even though the participants attributed the incident as an innocent cultural mistake, they were not confident about the attribution. Attributional confidence was found to be a significant negative predictor of avoidance facework across the four vignettes. This means that the lack of confidence in attribution attribution leads the Americans to avoid further interaction with the Chinese person. Americans attempt to manage their threatened face by preventing further attention to the topic. Attributional confidence of attributed intention was further examined in the follow-up study which used the exact same scales as the main study except that the 28 characters’ cultural background in the vignette became the same as instead of different from that the participants (i.e., the characters who performed the face threatening act in the vignettes were Americans instead of Chinese). The results showed that, between an intercultural and intracultural offender, Americans have higher level of attributional confidence when the act took place in an intracultural. context than in an intercultural context. In other words, participants were more confident in their attribution of the character’s intention when the conversation partner was from the same culture. To sum up, the attribution process of FTA and the confidence level of the attribution have influence on people’s use of avoidance and aggression facework. The more people believe the FTA is out of intentional offense, the more likely they are to use aggressive facework, whereas the more uncertain they are about the attribution, the more likely they are to use avoidance facework. This pattern of results was observed in both intercultural and intracultural communication contexts, although participants reported higher attributional confidence in intracultural context compared to that of intercultural context. Additional Findings and Explanations The findings on face needs, face threats, attribution and attributional confidence showed that the four factors have significant influence on people’s choice of facework strategies in intercultural well-meaning clashes. There are a few additional findings revealed from the prelim study that are worthy of examination as well. First, Americans changed their evaluation of the same incident when they were in an intracultural context compared to they were in an intercultural context. More specifically, the participants coming from different samples perceived that the politeness 29 of each act depicted in the given scenarios was higher when their conversation partner was Chinese than when he or she was American. All of the four acts were considered as more impolite and inappropriate when they were initiated by an American than when the exact same message or act was performed by a Chinese. It appears that American participants gave their intercultural counterpart more benefit of the doubt compared to their intracultural communication partner when evaluating the politeness of acts they perform. The change of perception complicated the findings of the study, prompting speculation. One speculation of the changed perception of the same scenario is the increased level of uncertainty in intercultural communication compared to intracultural communication. Attributional confidence scored higher in the intracultural context than that in the intercultural context. It seems that, when the participants were unsure of the other person’s true intention in performing the FTA, they were more likely to give the benefit of the doubt to the other communicator. Uncertainty of attribution in the intercultural context led the Americans to withhold a negative evaluation of the act. Another explanation concerns the methodology issue, which is the social desirability bias in self-reported survey studies. Self-reported surveys have been used widely in behavioral research studies across disciplines; however, one of the major criticisms of using self-reported surveys alone is the social desirability response bias, which is assumed to directly threaten the internal validity of the study. Bradburn, Sudman, Blair and Stocking (1978) found that people tended to over-report socially desirable behaviors on self-reported surveys, and under-reported undesirable social behaviors. This tendency is mostly a result of people’s need to maintain a positive impression especially 30 concerning questions that are sensitive to self-image (Bradbum et al., 1978). When people answer questions in self-reported surveys, they might be concerned that honest answers which contain negative information about themselves would threaten their positive image or they might be evaluated negatively by researchers. As a consequence, rather than giving honest answers, they provide the answers that are more socially acceptable. For example, when asked about if they had been intoxicated before, some people disguised the fact that they had been intoxicated or under-reported the times they had been intoxicated (Bradbum et al., 1978). In a meta-analysis examining the accuracy of clinicians’ adherence to practice guidelines by self-reports, the researchers found that there was a 37% overestimation of guidance adherence reported by self-reported studies compared to the results from studies using objective measures (Adams, Soumerai, Lomas, & Ross-Degnan, 1999). Social desirability might be present in the prelim study. In the study, American participants were asked to answer questions regarding face threats caused by an intercultural communicator in the self-reported questionnaire. As the communication process concerns cultural differences, it is possible that the participants did not want to be viewed as discriminating against people from different cultures or appearing closed- minded about interacting with international peOple. As a result, the participants might have under-reported the degree of face threat they perceived in the given situation. It is also possible that they were concerned that, if they responded negatively, they might be evaluated by the international researcher who was administering the surveys on the spot. This speculation is supported by participants’ answers to an open-ended question regarding their demographic information. The question was phrased as “Please briefly 31 write down your overall experience regarding interactions with international people.” A large majority of the participants reported that their communication experience with international people had been positive. One of the most frequent answers was “All positive. It is no different from interactions with any other people.” The answer was in contrast to the intercultural communication research findings (e.g., Gudykunst & Shapiro, 1986; Li, 1999), where participants usually reported more difficulty and higher levels of anxiety in interacting with their intercultural conversation partners compared to intracultural communicators. The contrasting results between literature and the prelim study might be explained by the Americans’ de-emphasis of the impact of cultural differences or their purposeful overlooking of the difficulties of interacting with people of different social and linguistic backgrounds. It is also possible that the participants have not developed sophisticated cognitive frameworks to comprehend the impact of cultural differences. As a result, they attributed the difficulty in intercultural communication to individual differences rather than cultural differences. In a word, the American participants’ changed perception of the situation might be due to their concerns of being viewed as closed-minded or even racist. It might be explained by the participants’ conscious or unconscious unwillingness to acknowledge the difficulty in interacting with international peOple. Theoretical and empirical evidence shows that people may disguise their true behavior when they only need to report intentions, in other words, a discrepancy between behavioral intention and actual behavior (e.g., Corral-Verdugo, 1997; Fishbein & Ajzen, 1975; LaPiere, 1934; Pager & Quillian, 2005). Self-reported surveys can contain social desirability bias because people’s behaviors are usually measured with the report of their 32 own behavioral intentions. It is an empirical question that whether or not their reported intentions would match their actual behaviors, as is, whether or not they would perform the exact behavior in a given situation as they indicated with intentions. Trandis (1979) defined intentions as the self-instructions for one to behave or to obtain certain behavioral objectives. Intentions are thus assumed to be one of the major determinants of an individual’s actual behavior (e.g., Abrahm, Sheeran, & Johnston, 1998; Fishbein & Ajzen, 1975). Therefore measuring intentions with items such as “I intend to. . .”, “I will...” and “I plan to...” becomes the proxy of measuring actual behavior (Fishbein & Ajzen, 197 5). In the prelim study, facework strategies were measured with intention items such as “I will avoid interaction with him or her” and “I will point out the rudeness of his or her comments.” Participants were asked to rate their extent of agreement or disagreement with each statement by circling the corresponding number on a 5-point Likert-type scale (1=strongly disagree, and 5=strongly agree). Facework strategies were measured with participants’ intentions to use the facework strategies. It is a question whether people use the facework strategies that they reported in the questionnaires. The relationship between behavioral intention and actual behaviors has been in debate for decades. Fishbein and Ajzen (1975) claimed that behavioral intentions predict behaviors, and empirical evidence shows that there is a strong association between intention and behavior. A recent meta-analysis by Webb and Sheeran (2006), however, showed that, although intention and behavior are'strongly associated, the causal effect of intention onto behavior might be attenuated by moderators such as perceived behavior control and willingness to conduct the behavior. Ajzen and Fishbein (2004) also pointed 33 out that although behavioral intentions predict behaviors, how well intentions in self- reports reflect actual behavior is an empirical question. LaPiere (1934) conducted a study comparing the intention of accepting Chinese as guests in hotels and restaurants versus the actual acceptance of a Chinese couple at those hotels and restaurants throughout the US. While accompanying a Chinese couple traveling in the US, LaPiere noticed that, when asked if Chinese would be accommodated at their establishment, only one out of the 250 establishments directly refirsed the request, while 99% of the establishments agreed to accept the Chinese guests. Six months after the trip, LaPiere sent out a survey to each of the 250 establishments they had visited, and asked them “Would you accept a member of the Chinese race as guests in your establishment ” A total of 136 establishments returned the survey, and 92% of those establishments indicated that they would not accept the Chinese as guests. More recently, Pager and Quillian (2005) examined the employers’ attitude towards hiring ex- offenders, and the study showed that there is a large and robust disparity between employers’ self-reported intention of hiring stigmatized job applicants and their decisions in actual hiring situations. More specifically, out of the 89 employers who indicated that they were likely to hire an ex-offender as an entry-level, 92.7% of them did not give a call back to the applicant who had a criminal record in actual hiring situation. This discrepancy between intention and actual behavior is also found in recycling behaviors (Pager & Quillian, 2005). Corral-Verdugo (1997) also discovered a low correlation between housewives’ self-reported recycling behaviors with the direct observation of the recycled objects at their houses. These studies suggest that there is no assurance that subjects’ responses to behavioral questions would warrant their actual behaviors, and 34 therefore researchers should be more cautious in interpreting their participants’ verbal responses, which might not be the same as their actual behavior (LaPiere, 1934). The discussion so far addresses the speculations of an interesting finding from the prelim study that the American participants changed their perception of the same act when they are in different contexts. These speculations concern both theoretical and methodological issues. Before moving forward, it needs to be clarified here that the methodological aspect under discussion is pg to invalidate the results from the self- reported surveys. Rather, it attempts to answer the question that whether the relationship found among face needs, face threats, attribution, attributional confidence and facework strategies in the self-reported survey study will be replicated in a study with a different method. Replication itself is a way to extend the external validity of research (Kerlinger & Lee, 1999). Aside from the possible social desirability in self-reported surveys, using a paper- pencil instrument to examine the problematic situation of intercultural communication is limited in realism. The design misses out one important characteristic of intercultural communication: language. Lee and Boster (1991) claimed that lack in language competence is one of the most important barriers in intercultural communication because a lot of energy and effort has to be given to the clarification of meaning. In the prelim study, participants only read hypothetical vignettes. When they were reading the vignette, it was relatively difficult for them to imagine the language barrier factor in the interactions. Thus it is questionable whether the participants would respond similarly in a real-life well-meaning clash as to how they did in the questionnaires. Direct observation of people’s interaction behaviors in the naturalistic setting can 35 address the limitations of self-reported measurement. Nook and Kurtz (2005) claimed that direct observation can capture a wider range of behaviors in a more specific manner, and observational methods are free from response biases such as social desirability or halo effect. Bakerman (2000) also asserted that observations are especially usefirl when the behaviors under investigation are “social, involving interaction between two or more participants” (p.138). Kerlinger and Lee (1999) stressed that “observations must be used when the variables of research studies are interactive and interpersonal in nature, and when we wish to study the relations between actual behavior... and other behaviors” (p.741). In view of the interest of the study, and the strengths of observational methods, it seems to suggest that in order to firlly capture the dynamics of how Americans deal with well-meaning clashes in a given situation, direct observations of Americans’ behavioral reaction towards the F TA is needed. Again, conducting the study with observational methods is not designed to invalidate the results of the self-reported study, rather, it aims to address the limitations of the self-reported study, and extend the validity of the findings. A research question therefore is proposed: RQla: Will the same relationship among face needs, face threats, attribution type, attributional confidence, and facework strategies be found with observational measures? There are two major practical limitations to completely replicate the self-reported study with direct observations of people’s reactions to a well-meaning clash in naturalistic settings. First, it is difficult to control the occurrence of well-meaning clashes in naturalistic settings. Unlike other social behaviors, well-meaning clashes only take place in intercultural context, where polite behavior in one culture is considered to be 36 impolite in another’s culture. The particular characteristics of the situation will significantly reduce the chance of occurrence in the naturalistic setting. Second, in the prelim study, self-reports employed four different well-meaning clashes in order to capture the effect of situational variation on facework. In the direct observation study, it would be rather unrealistic to capture all of the four types of well-meaning clashes with substantial number of subjects in each type of scenario. Given the practical limitations, the current study will focus on one well-meaning clash and trigger the well-meaning clash in the laboratory. This way, participants’ reactions of the well-meaning clashes will be easily observed and properly recorded. Thus the first research question is modified as: Rle: Will the result pattern of both individual face needs and situational face threats affecting avoidance and aggression facework with self-reported Likert-type scale measures be found with observational measures? RQlc: Will the result pattern of the types of attribution and attributional confidence affecting avoidance and aggression facework with self- reported Likert-type scale measures be found with observational measures? Alternative F acework Strategies The study is essentially interested in finding out how people deal with well- meaning clashes, a problematic intercultural communication situation. There were two basic types of facework strategies, avoidance and aggression, in the prelim study. How well people apply the two types of strategies, and whether or not there are other strategies in practice are still questions that need to be answered. In the self-reported survey study, participants were asked to indicate their tendency to use avoidance and aggression facework strategies after they read each of the 37 four vignettes. The results showed that, across the four vignettes, there was a relatively low tendency of using both avoidance and aggression facework strategies, given that the mean scores of both avoidance and aggression facework strategies were significantly lower than the scale mid-point. Examining the score distribution of both avoidance and aggression facework revealed that, even though there were participants who had a higher tendency to use the facework strategies provided, the majority of the participants were unlikely to use either the avoidance or the aggressive facework strategies. This suggests that there might be other facework strategies that people would use in dealing with well- meaning clashes. In the prelim study, participants were asked to report the likelihood of using five additional facework strategies: no reply, using a joke, pointing out the rudeness of the other person’s act, demanding an explanation, and demanding an apology from the other person. Among the five types of replies, the strategy of using a joke received the highest mean score across the four vignettes, while the strategy of no reply was found to be correlated with avoidance, while pointing out the rudeness and demanding further action on the offended party were associated with aggression facework. Although the details of the joke were not available, the finding shows that participants might engage in facework strategies which are not proposed in the existing facework taxonomy. Therefore, it is worthwhile to find out that if using a joke is a common way for people to manage the FTA in intercultural communication, and if so, what the humorous remarks consist of. According to Graham, Papa and Brooks (1992), humor serves a variety of purposes in social interactions, and one of its function is to be a face-saving strategy (e. g., Cupach & Metts, 1992; Edelmann, 1985; Fink & Walker, 1977; Saunders, 1998). Edelmann (1985) claimed that, in embarrassing situations, the use of humor can help 38 distract outsiders’ attention from the difficult situation and allow the embarrassed social actor to regain composure. Cupach and Metts (1992) also asserted that the use of humor can indirectly acknowledge responsibility for the embarrassing situation and address the damage due to face loss. If used adeptly, humor would increase the perceived communication competence of the social actor (Cupach & Metts, 1992). Using appropriate humor was also found to be positively associated with communicators’ likability and influence (Graham, 1985; O’Quinn & Aronoff, 1981). As a face saving strategy, humor was used by memory clinic patients when they perceived their positive and negative faces were to be threatened (Saunders, 1998). The author explained the results as the patients’ endeavors to maintain their dignity and positive image by means of humor (Saunders, 1998). In a business negotiation setting, humor can diffuse tension and mitigate offense (Vourela, 2005). La Fave, Haddad, and Maesen (1976) conceptualized humor as a process that must involve a sudden happiness increment, such as a feeling of superiority, relief, or arousal on the bearer. Therefore, a hearer’s laughter is a usually a primary indicator of humor, and smiles, grins, or even exhalations may also represent a humorous experience (Meyer, 2000). Since the indicators of humor are highly bearer-centered, it would be inaccurate to determine humor from the speaker’s perspective alone. A sender’s humorous intent might not be interpreted with amusement by the bearer. Interpretation of humor is a particularly challenging task for intercultural communication (Ting-Toomey, 1999). Ifan American person engages in humorous remarks, it is questionable whether the intercultural partner would evaluate the American’s verbal response as so. To resolve the dilemma, in the current study, whether the American subject uses humor in the interaction was 39 determined by the third-party coders, who are native-speaking Americans. That is, the study only takes the Americans’ perspective regarding whether humor is used. Sarcasm can also be an alternative strategy people use to manage impressions in face threatening situations (e.g., Jorgensen, 1996). Sarcasm is used when a speaker’s intent is the opposite of his or her message’s literal meaning, as when a person says, “What a lovely day!” while it is pouring outside. In certain social interactions, sarcasm allows the speaker to state a positive message with a negative intent, which could be a non-violent way to express one’s negative emotions (Rockwell & Theriot, 2001). In other words, when a social actor experiences anger or is upset in a norm violation situation, he or she can rely on a positive message to vent his or her negative emotions towards the face threatening act without verbalizing the negativity directly and openly. Using sarcasm shows that the social actor chooses not to ignore or avoid the other party’s face threatening act, but address it indirectly. The means of addressing the face threatening act, however, is not necessarily aggressive, because the sarcastic message is characterized by the seemingly positive and polite meaning on the surface. Sarcasm is also culturally bounded. Therefore, its interpretation might pose difficulty for intercultural communicators. Again, the issue was resolved by taking the American coders’ perspective. A second research question therefore is formed: RQ2: What facework strategies in addition to avoidance and aggression will subjects use when dealing with the intercultural well-meaning clash with observational measures? 40 CHAPTER 3 METHOD AND ANALYSIS Participants A total of 109 students participated in the study, while five out of the 109 subjects were international students: one from Africa, one from Western Europe, and three from East Asia. The average duration of stay for those international students was approximately two years, ranging from eight months to four years. In order to ensure the well-meaning clash was between an American and a Chinese person, the international students were removed from the final analysis of the study. During one session, the confederate did not deliver the comment, therefore, the subject was also removed from the analysis, which resulted in a sample of 103 (67 women and 36 men). All participants were recruited from the Michigan State University Department of Communication research subject pool. The average age of the subjects (N = 103) was 19.80, SD = 1.42, ranging from 18 to 24. Among the subjects, 71% were Caucasian American, 18.4% African American, 4.9% Asian American, and 4.9% multi-racial. Procedure Scenario selection. In order to observe the subjects’ reactions towards a problematic situation, a well-meaning clash situation was selected and created in the department laboratory. The scenario was designed as a Chinese international student commented on an American student’s appearance change, in particular, gaining weight. The scenario was selected based on both theoretical and empirical evidence. Cultivation theory (Gerbner, Gross, & Morgan, 2002) posits that continuous exposure to thin-ideal image in the media would lead viewers to accept being thin as the 41 standard in society, and consider it as central to attractiveness. A recent meta-analysis by Grabe and her associates (2008) confirmed that exposure to media’s thin-ideal portrayal causes self-image dissatisfaction. In addition, Greenleaf and her colleagues (2004) examined US. college students’ interpretation of weight-related words, for example, over-weight, fat, and obese. The results showed that those words are more likely to associate with laziness, unintelligence, unattractiveness, and lack of self-control (Greenleaf et al., 2004). It is obvious that the implications of overweight are negative in the US. society, which causes topics related to weight, in particular weight gain, sensitive in social interactions. By contrast, in Chinese culture, the negative implications of being obese are not as strong as that is in the US. (Lee, Chiu, & Chen, 1989). Although the concern of weight has been increasing in China in the last few years (Chen, Gao, & Jackson, 2006), traditional beliefs such as “being able to eat is to have luck”, and “gaining weight is a sign of stress-free life style” are still well recognized in the Chinese society (Lee et al., 1989). The overall discrepancy in societal attitudes towards obesity may stipulate different social norms around the topic of weight-related issues. Given the general anti- fat attitude in the US, weight-related topics is very likely to be perceived as less than socially desirable, therefore, commenting on someone’s weight gain may directly pose a negative face threat to the bearer. The bearer is likely to feel an intrusion into his or her personal matters, disrupting his or her need for privacy. At the same time, commenting on one’s weight gain can be interpreted as implying the person as being lazy and unattractive, which constitutes a positive face threat for the bearer. In contrast, commenting on someone’s weight usually shows one’s care and concern for the other 42 person’s well-being in Chinese culture (Hu & Grove, 1991), rather than criticizing the person’s image or probing into his or her private matters. As the social norm of weight- related topics is more relaxed in Chinese culture, weight-related topic might not necessarily be surprising, unpleasant, or constituting face threats for the bearer. In the prelim study, the vignette of commenting on one’s weight gain had the strongest well-meaning clash effect. The perceptions of the weight gain comment between American and Chinese was the sharpest among the eight vignettes investigated. Further, the act of commenting on one’s weight gain was consistently viewed as both a positive and negative face threat across the three steps of the prelim study. The act received the highest scores on the positive face threat and negative face threat in the prelim study. Based on the theoretical conceptualization and empirical results, it was expected that the act of commenting on one’s weight gain would be most likely pose a face threat for the subjects, and correspond to the nature of well-meaning clashes in intercultural communication. Detailed procedure. The announcement of research participation was posted on the department research pool website (www.cxperimetrixcom/rrArsucom). Students’ participation was exchanged with course credits. The research announcement contained a brief description of the study, including what the participants needed to do, the amount of time it would take, and the number of credits to be awarded. The announcement restricted participants from only two classes in the department, both of which had approximately 200 students enrolled. The purpose of selecting those classes was to establish a context that the confederate might have seen the subject at some point, but the subject might not necessarily remember her due to the size of the mass lecture. 43 The confederate was a young East Asian female who was a graduate student from the department. She has been in the United States for about two years, and her English was conversationally fluent, although an accent can be easily detected. Three undergraduate students from the department took turns to act as research staff members, giving instructions to the subjects in the laboratory and recorded the subjects’ reactions on the spot. All of them (two females, one male) are Caucasian American. The observations of the conversation between the subject and the confederate were in the research department laboratory which was equipped with multiple rooms. In the room labeled as Waiting Room, a desk and two chairs were set up. One of the chairs was directly across the video camera. The interaction between the confederate and the subject was observed from the one-way mirror in the Waiting Room. Two other rooms were labeled as Interaction Room 1 and Interaction Room 2. The research staff monitored the progress of the interaction through the one-way mirror in the room next to the Waiting Room. Students who were interested in the study signed up through the website at a pre- set time slot and were reminded of the appointment the day before. On the experiment day, when a subject arrived at the laboratory at the scheduled time, the confederate came approximately the same time, pretending to be another participant. The research staff checked them in by taking down their names and the classes they wanted to assign the credits for. The research staff then directed the subject and the confederate into the Waiting Room. The confederate usually came in the room first, leaving the subject the seat directly across the video camera. There were a few subjects who entered the Waiting Room first, and took the seat that was off the camera. After both sat down, the research 44 staff explained the purpose of the study, duration of the study, and what the subjects were expected to do. The subject and the confederate were told that they needed to wait in the Waiting Room first, and they each would be taken to one of the Interaction Rooms where they would have an interaction with the researcher. Doing so was to give subjects an impression that their conversation with the researcher would count for the real study, while talking with the confederate was not. The subject and the confederate were presented with the consent form. They were informed that although the conversation with the researcher was the major part of the study, all of their interactions had to be video- taped. After the subject and the confederate signed the consent form, they were asked to fill out the first questionnaire that included measures of general face needs and demographic information. The research staff left the room and came back afier the subject and the confederate completed the first survey. The staff informed them that Interaction Rooms were being set up and he or she would come back once they were available. Then the staff member left the room. The conversation between the subject and the confederate started, which was mostly small-talk. Sometimes the confederate initiated the conversation. She usually greeted the subject, or simply asked whether the subject was taking the experiment for a certain class. Other times, the conversation was started by the subject. During the interaction, the confederate was instructed to reveal her name (i.e., Ming, a common Chinese name for both males and females), hometown (i.e., Beijing, China), and the duration in the States (i.e., one year). Subjects usually exchanged similar information. The confederate also said that she was in the same class as the subject, and had seen the subject at the beginning of the semester. She then commented “You look a little different 45 from the beginning of the semester. I think you have gained a few pounds.” At the time of the data collection, it was towards the end of the semester, which means three months have passed in the semester. Generally speaking, it is possible for people to gain weight over three months that is noticeable to others. Subjects’ reactions were recorded by camera and the research staff member through the one-way mirror. After subjects reacted to the comment and the conversation started to change direction, the research staff entered the Waiting Room, and instructed the subject and the confederate to go to one of the Interaction Rooms. In the Interaction Room where the subject was in, the research staff instructed the subject to fill out a second survey, and informed the subject that the researcher would interview them later, then left the room. The second survey included both open-ended questions regarding the subject’s reactions towards the confederate’s comment, and close-ended questions concerning his or her evaluation of the past interaction with the confederate, such as situational face threats, attribution of the comment, and level of confidence in the attribution. After the subject completed the second survey, the researcher came in the Interaction Room to debrief the study, explaining the real identity of the confederate, and the true purpose of the study. Each subject was then asked not to share the set-up of the study with those who might be potential participants in the study. The researcher also interviewed subjects for several questions, including whether they were suspicious of the confederate during their interaction, their reactions towards the comment, if they had similar experience with other international students or friends, and whether sex and relationship type of the person who made the comment would make a difference on their 46 reactions. The interview process was recorded with a video-camera in the room visible to subjects. Measures Dependent variables were measured with both Likert-type scales and third-party observations. Independent variables were measured with S-point Likert-type scales (1=strongly disagree, and 5=strongly agree) and open-ended questions on the self- reported questionnaires (see Appendix A). Reliability analysis and Confirrnatory Factor Analysis were performed to ensure the scale reliability and unidimensionality. Means, standard deviations, correlations and reliabilities of the variables are presented in Table 1. Dependent variables. Subjects’ facework strategies after the confederate’ 3 well- meaning comment were the outcome variables of the study. They were measured with both third-party coder observations and self-reported Likert-type scales. First, two independent coders coded the subject’s facework strategies by watching the video-taped conversation between the subject and the confederate. A coding scheme (see Appendix B) was developed based on the open-ended responses from the prelim study on the same topic. According to the coding scheme, the coders decided whether the subject was avoiding or not (1= yes, and 0= no), and aggressive or not (1= yes, and 0= no). The inter-coder reliability (Krippendorf’s alpha) was .86 for avoidance, and .81 for aggression. Discrepancies of coding were discussed and resolved among the coders and the researcher. Two subjects declined video-taping; therefore, no taping was used with the two subjects. Due to the unavailability of video-taped interaction, the research staff member’s live coding of the two subjects’ facework strategies was used in the final analysis. 47 8.vm..8.va: 5.8 3 :ar :8. .5.- 8r 8. 8. 8r 8. 8. 2. Sr :.- 2. 8r 8. :5 883...: 5.8 3 :.- :8. :.- Sr 8.- 8. :8. .8. 3. 8r :. 8r 8. 8.- 3:. 84.2.2 :38 $3 .8. 8. 2. :8. 8r 8 2. 2. m8. .8- .8.- 8.- :r a: 889...: 3 A88 95 8- 8. .2. 8r :8. .3. :8. 8.- .8. 8. 2. S. and 84.2.: 5.8 $3 3. :8. :2- :an- :3- :3- 8. 8r .3- 8.- 8r ”2 0888.2 8.8 $8 :2. :2r .3“. :8. :2. 8. 8 S.- Sr 2.- SN 8865.: 8.8 93 2.3- 8. 2. :8. .2- 8- ._~.- 8 2.- a3 25.8.2 $8 $3 2. 2. 8r :. .3. :3. .8. .8. .3 39.5... 3.8 - 83 :8. :8. :Sr .2. .8. 8. .3. A88 =5 82.” 8.: S3 :8. :8.- .8. :E. 8. E. «2 25-895 8.: £3 ..8.- .8. .8. S. S. a: 208m... 5.8 83 8o.- 8.- 2r 8.- NZ 3%.: 5.8 30¢ .159. EL? 2. nn.m .mzoe $.08 35v awn. econ. NBA.“ "EDA“ $8 5.8 :8. a; "Ema A88 83 8... .5 a 58 S 2 3 2 .2 : 2 a w e e m a m N _ =82 dog—58:2 new .maoufiotonc anon—~33 c.8283 £502 ._ 053. 48 In addition to coding subject’s facework categorically, the two coders also rated the degree of subject’s avoidance (M = 4.16, SD = 3.05) and aggression (M = 1.20, SD = 0.82) on a 10-point single-item Likert-type scale (1 = not at all, and 10= extremely). The inter- coder reliabilities were .92 for the magnitude of avoidance and.89 for aggression. These two dependent variables were only used in the path model analysis. Second, subject’s facework strategies on the self-report questionnaire were measured with multiple scale items. Avoidance strategies (M = 2.32, SD = 0.79) were measured with four items, including “After hearing what she said, I said no more on the topic,” “I ignored the comment,” “I changed the topic,” and “I ended the conversation.” The reliability (Cronbach’s a) was .77. Aggression strategies (M = 1.76, SD = 0.58) were measured with five items, including “1 confronted her about she said,” “I challenged about what she said,” “I demanded an apology from her,” “I demanded an explanation from her,” and “I pointed out the rudeness of her comment.” The reliability was .84. Independent variables. General face needs were measured with a modified 29- item self-reported scale from the prelim study prior to the interaction between the subject and the confederate. The four face needs included: self-positive face (SPF), self-negative face (SNF), other-positive face (OPF), and other-negative face (ONF). The reliabilities (Cronbach’s a) of the 6 items measuring SPF were .74, .76 for the 10 items measuring SNF, .76 for the 7 items measuring OPF, and .63 for the 6 items measuring ONF. Confirmative Factor Analysis (CFA) was conducted on the 29 items with four factors solution, and the fit indexes yielded an unacceptable fit: CFI (Comparative Fit Index) = .83, NFI (Normed Fit Index) = .69, NNFI (Non-Normed Fix Index) = .81, IF I (Incremental Fit Index) = .83, and RMSEA (Root Mean Square Error of Approximation) 49 = .09. After examining the reliabilities, factor loadings, and error variances, 9 items were deleted fi'om the scale, and the reduced 20-item model yielded a better fit: xz = 260.17 with df= 164, CFI = .90, NFI =. 78, NNFI = .89, [PI = .90, and RMSEA = .08. To ensure the four factors were distinct from each other, CPA with one-factor solution was conducted, and it yielded poorer fit: 1? = 534.89 with df= 17o, CFI = .70, NFI = .61, NNFI = .67, IFI = .71, RMSEA = .15, Ax’ = 249.35 with Adf= 11. The four factor solution was superior to the one-factor solution. The reliability for the 20-item scale was .80 for the 4-item SPF (M = 4.34, SD = 0.48) scale, .75 for the 6-item SNF (M = 3.79, SD = 0.53) scale, .76 for the 6-item OPF (M = 3.72, SD = 0.55), and was .66 for the 4- item ONF (M= 3.55, SD = 0.55) scale. Situational face threats were measured with the level of positive and negative face threat perceived in the confederate’s comment. Positive face threat (M = 2.78, SD = 1.08) was measured with six items, including “What she said made me feel bad about myself,” “What she said made me feel self-conscious,” and “What she said made me feel awkward,” and the reliability was .93. Negative face threat (M = 3.08, SD = 0.94) was also measured with six items, including “Her comments made me not to know what to say,” “What she said was intrusive,” and “Her comments made me uncomfortable,” and the reliability was .90. Attribution of the comment was measured with the level of agreement with three different types of attribution. Type 1 attribution, innocent mistake (M = 3.34, SD = 0.91), was measured with four items, which were “She was making an innocent cultural mistake,” “She did not know how her comment might be perceived,” “She had no idea that her comments might be inappropriate,” and “She would not have made the comment 50 if she knew it be seen as inappropriate.” The reliability was .88. Type 2 attribution, intentional offense (M = 1.69, SD = 0.66), was measured with four items, which were “She was trying to insult me,” “Her intentions were mean-spirited,” “She was trying to hurt my feelings,” and “She was trying to make me feel bad.” The reliability was .97. The third type of attribution, incidental offense (M = 2.67, SD = 0.66), was also measured with four items, including “She knew the comment might hurt me, but it was not her intention to hurt me,” “She might have anticipated that I would be offended by her comments, but she had a good reason to make the comment,” “It was not her intention to make me feel bad, but she might have anticipated that her comment would hurt me,” and “She was honest with me, even though she knew it might hurt my feelings.” The reliability was .68. Attributional confidence (M = 2.78, SD = 0.81) was measured with the degree of certainty of subjects’ attribution, including items such as “I am absolutely certain of her intentions in giving the comment,” “It is clear to me that why she gave me the comment,” “I am definite about her intentions, “I am not sure about her intention” (reversely coded), and “I am confident in my explanation as to why she gave the comment.” The reliability was .85. Control variables. In order to obtain a general understanding of subjects’ perception of the comment, subjects were asked to evaluate the level of politeness (M = 2.12, SD = 0.75) of the confederate’s comment. Sample items included “Her comments are appropriate,” “Her comments are considerate,” “Her comments show her good manners,” and the reliability was .89. In addition, the level of severity (M = 2.69, SD = 1.08) of the comment was also measured with subjects’ emotions after hearing the 51 comments. Sample items included “I am angry hearing what she said,” “I am upset by her comments,” “I’m left with negative feelings after hearing what she said,” and the reliability was .92. Subjects’ physical appearance was rated by the live coder on two 10-point (1 = not at all, and 10= extremely) Likert-type scale questions: one is concerned with the subject’s physical attractiveness (M = 6.46, SD = 1.24), and the other is the subject’s status of being overweight (M = 2.45, SD = 0.77). Prior to the study, live coders were trained to rate people’s physical attractiveness and overweight and reached reliabilities above .90. The two variables were correlated with each other r (101) = —.35, p < .01. During the debriefing session, subjects were asked that if they were suspicious of the confederate’s identity. Based on subjects’ answers, suspicion was coded categorically (O =not suspicious, and l=suspicious). Approximately17% of the subjects reported that they were suspicious of the confederate’s identity. Although suspicious subjects (M = 2.68, SD = 0.72) were less confident in their attribution than non-suspicious subjects (M = 3.25, SD = 1.03), t (101) = —2.80,p < .01, n2 = .07, they did not differ in the use of facework strategies. Specifically, non-suspicious and suspicious subjects did not differ in the use of avoidance facework measured with self-reported measures, t (101) = 0.14, p = .89, 112 = .00, nor in avoidance measured with observational measures, t (101) = 1.70, p = .09, n2 = .03. Suspicious and non-suspicious subjects did not differ in the use of aggression facework measured with observations either, t (101) = —0.07, p = .95, n2 = .00. Negligible difference was found between non-suspicious (M = 2.32, SD = 0.83) and suspicious subjects (M = 2.29, SD = 0.60) in the self-reported aggression facework, t (101) = 2.14, p < .05, oz: .04. 52 Analysis Two types of regression analyses using SPSS 16 were conducted in order to answer the research questions. First, Ordinary Least Square (OLS) regression analysis was conducted when the continuously measured facework strategies were the dependent variables, including avoidance and aggression facework reported by subjects on the Likert-type scales of the second self-reported survey. Binary logistic regression analyses were conducted when the dichotomously coded facework observed by third-party coders were the dependent variables. Considering the large number of independent variables, the independent variables were divided into two sets, and each regression analyses included only one set of independent variables. With one dependent variable, the first set of regression only included face related indicators, such as face needs and face threats, and the second set only included other situational factors, such as attribution and attributional confidence. Specifically, for the first set, block 1 of one regression analysis included, self-positive face (SPF), self-negative face (SNF), other positive face (OPF), other negative face (ONF), and negative face threat (N_threat), and block 2 of the regression included product terms of SPF x N_threat, SNF x N_threat, OPF x N_threat, and ONF x N_threat. The second set of independent variables were block 1 of another regression analysis included attribution type l-cultural mistake (Il), attribution type 2-intentional offense (I2), attribution type 3-incidental offense (13 ), attributional confidence (Confidence), subject’s sex, the degree of subject being overweight, and block 2 included 11 x confidence, 12 x confidence, 13 x confidence. A path model using LISREL 8.80 (JOreskog & SOrbom, 2006) was conducted as a 53 post-hoe explanation and further exploration of the variable relationships. 54 CHAPTER 4 RESULTS Overview The current study was designed to explore people’s facework in managing a face threatening situation in intercultural communication context. The main findings were: 1) avoidance is the primary facework strategy in managing a well-meaning clash caused by others; 2) negative face threat predicts avoidance faceowrk, and attributing the face threat to intentional Offense leads to aggression facework; and 3) a disagreement with the weight gain in a non-aggressive and non-avoiding manner, for example, “I still weighed the same as last year”, jokes and sarcasm are the alternative facework other than avoiding or aggression. Characteristics of A voidance and Aggression Facework According to the third-party coders’ observation, over half of the subjects employed non-aggressive and avoiding facework in response to the comment of weight gain, followed by non-avoiding and non-aggressive facework, aggressive and non- avoiding facework, and avoiding and aggressive facework . Table 2. Observed frequencies of facework strategies coded by third-party coders. Avoidance 69 (67%) Non-Avoidance 34 (33%) Aggression 11 (10.7%) 3 (2.9%) 8 (7.8%) Non-Aggression 92 (89.3%) 66 (64.1%) 26 (25.2%) Using sentences or words such as “I don’t know”, “perhaps” or “maybe” as the 55 only response was one of the most frequently used non-aggressive avoiding facework (see Table 3). Subjects did not continue on the topic of weight gain after those words by changing the topic or offering no more information, in which case, the confederate had to change the topic in order to maintain a conversation. For example: “(C buckling) I don’t know (looking down at the desk, them up at the confederate) Where do you live?” (Subject 28, Male) “1 mean maybe. Probably (maintaining eye-contact first, then looking away, coughing)” (Subject 48, Male) “Really? I don’t know (shrugging shoulders). I dunno (nodding head slightly, looking down)” (Subject 54, Female) “Maybe (hand putting on face), probably (moving away eye-contact)” (Subject 61, Female) In addition to those words, subjects also provided a brief explanation or justification of why they might have gained weight, but there was a clear intention to terminate the topic in discussion observed from their nonverbal cues after the explanation. For example: “I don’t know. Probably. My weight fluctuates around. That’s how campus is (moving away eye-contact)” (Subject 10, Female) “Yup, college. . .freshman 15 (moving away eye-contact)” (Subject 79, Female) Another non-aggressive avoiding facework was a simple agreement with the confederate’s comment, confirming the weight gain. Subjects simply stopped the conversation after the confirmation, leaving an impression that they were unwilling to continue the topic. For example: “Yeah, I have. (laughing)” (Subject 26, Female) 56 Table 3. Observed frequencies of detailed facework coded by third-party coders. Code Reply Freq Percent Facework Avoidance 13 Replied nonverbally only with apparent avoiding 5 4.9% cues l7 Replied with only filler words, trying to end the 16 15.5% topic 19 Replied with confirmation of the weight gain, 8 7.8% trying to end the topic 17, 18 Replied with fillers words, and brief justification 12 11.7% or explanation of weight gain 17, 19 Replied with filler words and confirmation of the 5 4.9% weight gain, trying to end the topic 18, 19 Replied with confirmation and explanation of the 3 2.9% weight gain, trying to end the topic Others (Different combinations) 20 19.4% Subtotal 69 67% Aggression 202 Replied with a challenge of message implication l 1% 205 Replied with pointing out the rudeness of message 1 1% 200, 1100 Replied with a clunge of message implication and 3 2.9% the statement that the comment was untnre 202, 1100 Replied with a challenge of message implication 2 1.9% and a statement that the comment was untrue 202, 205, Replied with a challenge of message implication, 1 1% 1100 pointing out the rudeness of message and the statement was untrue Subtotal 8 7 8% Neither 1100 Replied with a statement that the comment was 13 12.6% Avoidance untrue, no apparent avoiding/aggression cues nor 1101 Replied with positive remarks hearing the 3 2.9% Aggression comment 1103 Replied with jokes, with no apparent avoiding or 2 1.9% aggressive cues 1100, 1102 Replied with a statement that the comment was 2 1.9% untrue,andasareasticmarkwithnoapparent avoiding or aggressive cues Others (Different combinations) 10 9.7% Subtotal 26 25.2% Both 17, 200 Replied with filler words, and pointing out 2 1.9% Avoidance confederate was mistaken and 202, 1100 Replied with a challenge of message implication l 1% Aggression and a statement of the comment was untrue Subtotal 3 2.9% 57 “Probably. Since the beginning of the year, I know I have. Yeah... (looking down)” (Subject 51, Male) According to some male subjects, their weight gain was a result of intentional efforts. However, even they justified their weight gain due to a legitimate reason, their verbal (e. g., changing the topic) and nonverbal cues (e.g., moving away eye-contact) indicated an attempt to terminate the conversation or at least to avoid the tOpic in discussion. For example: “Hm. . .oh well, who knows. I have been trying to, so I ...work out a lot for the semester (smile disappearing)... .Do you go to the 1M?” (Subject 18, Male) “Really? Well, maybe. I’ve been working out a lot lately, I guess. So I’m trying to, but... (looking down at the desk)” (Subject 75, Male) Some subjects did not have any verbal response after hearing the comment, and replied with only nonverbal cues. No further verbal interaction on other topics was initiated on the subject’s side either, which signaled that the subjects’ attempt to avoid the further interaction with the confederate. For example: “(Shrugged her shoulders, and looked away)” (Subject 03, Female) “(Chuckled)” (Subject 08, Male) Nonverbal cues were also used by subjects in combination with the avoiding verbal statements to reinforce their discomfort of discussing the topic. Moving away eye- contact was the most frequently used sign of avoidance. For example: “That’s definitely possible (hanging head down with arms on knees). Definitely possible, I definitely put on. . .uh. . .the freshmen 15 (head kept down, no eye- contact)” (Subject 52, Male) 58 “Oh yeah? (smiling) well, I could have. . . It’s the muscles (smile gradually disappeared looked away) (Subject 14, Male) In summary, typical avoidance facework observed in the current study was characterized by 1) apparent nonverbal cues of avoiding firrther interaction with the confederate, such as withdrawing eye-contact and smile; and 2) a brief verbal reply to deflect the comment without firrther discussion of the topic. These verbal and nonverbal cues indicated subjects’ unwillingness or discomfort to continue the topic of weight gain. Non-avoiding and aggressive facework was characterized by subjects’ pointing out that the confederate was mistaken, or challenging the confederate for her implication of the comment. When using non-avoiding and aggressive facework, subjects did not attempt to change the topic of the weight gain, or avoiding firrther interaction with the confederate, instead, they addressed the issue directly, intending to convince the confederate that she was wrong. For example: “No, I lost weight. Maybe you were thinking I was someone else then.”(Subject 72, Female) “1 lost, yeah, I’ve lost a lot of weight. I’ve been really stressed out and I lost. . .Do I look fat?”(Subject 27, Female) “Gained a few pounds? (laughing) When did you think you saw me again?” (Subject 91, Male) In other circumstances, subjects attacked the confederate for making the comment in the first place. The nonverbal cues, however, seemed to contradict the seriousness of their verbal reply. For example: “Mean! Oh my God, that’ s crazy (sharp voice, animated facial expression)” 59 (Subject 102, Male) The most aggressive facework was when the subject directly confronted the confederate, pointing out the rudeness of her comment, and the nonverbal cues showed obvious negative emotions: “Are you being serious? (staring at the confederate, looking angry). That was extremely rude, but... (ignoring the confederate, looking into her bag)” (Subject 63, Female) The conversation was terminated, and the subject completed ignored the confederate, leaving the confederate in discomfort. Avoiding and aggressive facework strategies were characterized by the dual goals of the subjects. On one hand, subjects attempted to hint that the confederate was being mistaken, and on the other hand, they displayed clear tendency of avoiding the topic. For example: “I don’t know. I don’t think so (looking down, avoiding eye-contact). . .maybe someone else (mumbling, kept lowering head down, no eye-contact)” (Subject 38, Male) “Mmm. .. I don’t know. . .I don’t know. . .might not have been me (looking down, avoiding eye-contact)” (Subject 47, Male) Research Question 1 Research questions 1a, 1b and 1c were concerned with whether the results from observational measures would be consistent with that of the self-reported measures in the current study. The findings were mixed. When subjects’ self-reported facework strategies were the dependent variables, results showed that the only significant predictor for 60 avoidance facework was negative face threat, [3 = 0.39, t = 4.37, p < .001 (see Table 4). None of the individual general face needs or the interactions between general face needs and negative face threat was significant for self-reported avoidance, neither was the attribution types, attributional confidence, sex, overweight or any of the interaction effects (see Table 5). None of the individual general face needs, negative face threat, or the interactions between general face needs and negative face threat was significant for self-reported aggression (see Table 6). Attribution type 2-intentional offense was the only significant predictor for self-reported aggression facework, B = 0.41, t= 4.04, p < .001, (see Table 7). When the third-party coders’ coding were entered as the dependent variables, results showed that negative face threat was a significant predictor for coded avoidance facework, logistic coefficient (i.e., logit) = 0.76, odds ratio = 2.14, Wald’s statistic = 8.03, p < .01 (see Table 8). This means that increment in negative face threat was strongly associated with the increasing odds of using avoidance facework. None of the attribution types or other situational factors was significant for third-party coded avoidance (see Table 9). None of the face needs, negative face threats, or interaction effects was significant for third-party coded aggression (see Table 10). Attribution type 2 was not a significant predictor for aggression measured by observations (see Table 11). Overall, negative face threat was a significant predictor for avoidance facework with both measures. The higher of perceived negative face threat in the comment, the more likely were individuals to avoid the situation. Attributing the comment as an intentional offense was a significant predictor for aggression only for the self-reported measures. The stronger individuals perceive the comment as a result of mean- 61 Table 4. OLS regression results (first set regression with face needs and face threat as predictors and self-reported scale avoidance as dependent variable). B SE Beta t p—value Dependent Variable Avoidance Ssgéreport First-order effect Self-positive face (SPF) —0.05 0.18 —0.03 —0.30 .77 Self-negative face (SNF) 0.16 0.16 0.11 1.03 .31 Other-positiveface (OPF) —0.18 0.15 —0.13 —l.l9 .24 Other-negative face (ONF) 0.22 0.15 0.15 1.41 .16 Negativefacethreat 0.33 0.81 0.39 4.07 <.001 (N_threat) F(5, 97) = 4.98, p < .001, R2 = .20, adjustedR2= .16 Second-order effect SPFxN_threat 0.02 0.19 0.14 0.13 .90 SNFxN_threat 0.13 0.16 0.09 0.82 .41 OPFxN_threat —o.02 0.18 —0.05 —0.39 .70 ONFxN_threat —o.02 0.18 —0.01 —0.11 .91 Fchange (4, 93) = 0.22, p = .928, R’change = .007 Overall model: F (9, 93) = 2.77, p < .01, adjusted R2 = .14 62 Table 5. OLS regression results (second set regression with attribution types as predictors and self-reported scale avoidance as dependent variable). B SE Beta t p-value Dependent Variable Avoidance self-report scale F irst-order effect Cultural mistake (11) 0.04 0.10 0.05 0.42 .68 Intentional offense (12) 0.30 0.13 0.25 2.34 .02 Incidental offense (13) —0.06 0.13 —0.05 0.47 .64 Confidence -0.04 0.11 —0.04 —0.35 .73 Overweight 0.09 0.10 0.09 0.88 .38 Sex 0.22 0.18 0.13 1.27 .21 F(6, 96) = 1.57, p = .17, rtZ = .09, adjustedRZ = .03 Second-order effect 11 x Confidence —0.04 0.10 —0.05 —0.42 .68 12 x Confidence 0.02 0.19 0.01 0.11 .92 13 x Confidence 0.30 0.17 0.19 1.82 .07 Fchange (3, 93) = 1.24, p =30, R’chtmge = .04 Overall model: F (9, 93) = 1.46, p = .17, adjusted R2 = .04 63 Table 6. OLS regression results (first set regression with face needs and face threat as predictors and self-reported scale aggression as dependent variable). B SE Beta t p-value Dependent Variable Aggression Self-report First-order effect Scale Self-positive face (SPF) —0.26 0.14 —0.20 -1.82 .07 Self-negative face (SNF) 0.09 0.12 0.08 0.74 .46 Other-positive face (OPF) 0.02 0.12 0.02 0.14 .89 Other-negative face (ONF) —0.28 0.12 —0.26 —2.27 .03 Negativefacethreat 0.12 0.06 0.19 1.89 .06 (N_threat) F (5, 97) = 2. 10, p = .07, R2 = .10, adjusted R2 = .05 Second-order effect SPF x N_threat 0.14 0.14 0.11 0.99 .33 SNFxN_threat 0.15 0.13 0.13 1.16 .25 OFF x N_threat 0.11 0.14 0.10 0.81 .42 ONF x N_threat —O.27 O. 14 —0.23 -1.94 .06 Fchange (4, 93) = 1.45, p = .22, chhange = .05 Overall model: F(9, 93) = 1.83, p = .07, adjusted R2 = .07 64 Table 7. OLS regression results (second set regression with attribution types as predictors and self-reported scale aggression as dependent variable). B SE Beta t p-value Dependent Variable Aggression Ssgéreport First-order effect Cultural mistake (11) 0.10 0.07 0.15 1.44 .15 Intentional offense (12) 0.36 0.09 0.41 4.04 <.001 Incidental offense(I3) 0.11 0.09 0.12 1.19 .24 Confidence 0.03 0.08 0.05 0.45 .65 Overweight —0.003 0.07 —0.004 —0.04 .97 Sex -0.10 0.12 —0.08 —0.79 .43 F(6, 96) = 3.44,p < 01,18 = .18, adjustedR’ = .13 Second-order effect 11 x Confidence 0.05 0.07 0.09 0.83 .42 12 x Confidence —0.19 0.13 —0.15 -1.44 .15 13 x Confidence 0.18 0.12 0.16 1.56 .12 Fchange (3, 93) = 1.77, p =. 16, chhange = .04 Overall model: F (9, 93) -- 2.94, p < .01, adjusted R2 = .15 65 Table 8. Logistic regression results (first set regression with face needs and face threat as predictors and coder coded avoidance as dependent variable). B SE Wald Exp (B) p-value Dependent Variable Avoidance 00d“ First Block Self-positive face (SPF) —0.63 0.57 1.24 0.53 .27 Self-negative face (SNF) 0.28 0.49 0.32 1.32 .57 Other-positive face (OPF) —0.59 0.48 1.51 0.56 .22 Other-negative face (ONF) 0.33 0.47 0.48 1.39 .49 Negative face threat 0.76 0.27 8.03 2.14 <.01 (N_threat) Overall model evaluation: x2 (df = 5) = 12.08, p < .05 Second Block SPF x N_threat -0.30 0.62 0.23 0.74 .63 SNF x N_threat 0.31 0.54 0.34 1.37 .56 OFF x N_threat —0.45 0.60 0.56 0.64 .45 ONF x N_threat 0.23 0.59 O. 15 1.26 .70 Overall model evaluation: )8 (df = 9) = 13.49, p = .14 66 Table 9. Logistic regression results (second set regression with attribution types as predictors and coder coded avoidance as dependent variable). B SE Wald Exp (B) p-value Dependent Variable Avoidance coder First Block Cultural mistake (11) 0.07 0.28 0.06 1.07 .81 Intentional offense (12) —0. 12 0.36 0.11 0.89 .74 Incidental offense (13) —0. 10 0.36 0.07 0.91 .79 Confidence —0.14 0.31 0.20 0.87 .66 Overweight 0.51 0.32 2.56 1.66 .11 Sex -0.95 0.49 3.81 1.00 .05 Overall model evaluation: )8 (df = 6) = 8.32, p = .22 Second Block 11 x Confidence 0.28 0.28 0.96 1.32 .33 12 x Confidence —0.21 0.53 0.15 0.81 .70 13 x Confidence —0.03 0.49 0.004 0.97 .95 Overall model evaluation: x2 (df= 9) = 9.62, p = .38 67 Table 10. Logistic regression results (first set regression with face needs and face threat as predictors and coder coded aggression as dependent variable). SE Wald Exp (B) p-value 0.87 0.005 0.94 .95 0.71 0.25 0.70 .62 0.84 3.03 4.31 .08 0.67 2.76 0.33 .10 0.36 0.22 1.19 .64 0.96 2.28 4.25 .13 0.88 0.43 1.77 .51 1.21 0.97 0.31 .33 0.94 0.59 0.49 .44 B Dependent Variable Aggression coder First Block Self-positive face (SPF) -0.06 Self-negative face (SNF) —0.35 Other-positive face (OPF) 1.46 Other-negative face (ONF) —1 . 12 Negative face threat 0.17 (N_threat) Overall model evaluation: x2 (df = 5) = 5.53, p = .35 Second Block SPF x N_threat 1.45 SNF x N_threat 0.57 OPF x N_threat —1 . 19 ONF x N_threat —0.72 Overall model evaluation: )8 (df = 9) = 10.62, p = .30 68 Table 11. Logistic regression results (second set regression with attribution types as predictors and coder coded aggression as dependent variable). B SE Wald Exp (B) p-value Dependent Variable Aggression coder First Block Cultural mistake (11) —0.13 0.44 0.09 0.88 .76 Intentional offense (12) 0.37 0.52 0.50 1.45 .48 Incidental offense (13) 0.41 0.57 0.53 1.51 .47 Confidence —0.36 0.48 0.54 0.70 .46 Overweight —0. 14 0.43 0.12 0.87 .75 Sex -0.83 0.71 1.35 2.28 .25 Overall model evaluation: x2 (df = 6) = 2.86, p = .83 Second Block I] x Confidence 0.25 0.51 0.25 1.29 .62 I2 x Confidence —l.23 0.98 1.57 0.29 .21 13 x Confidence 0.13 0.80 0.03 1.14 .87 Overall model evaluation: )8 (df = 9) = 5.25, p = .81 69 spirited and intentional insult, the more likely were they to be aggressive in managing the situation. This predictor was not significant with the observational measures. Research Question 2 Research question 2 was concerned with additional facework strategies that individuals employ in managing the face threatening act. In the current study, approximately 25% of the subject’s reactions coded by the third-party coders were neither avoidance nor aggression (see Table 3), and among the 25% of the responses, the most commonly used strategy was a simple disagreement with the confederate’s comment about the weight gain in a non-aggressive and non-avoiding manner. Subjects provided facts to explain why the confederate’s comment might not be true in order to validate their disagreement. For example: “Really? (looking at the confederate) Uh. . .I still weigh 173 and that’s the same I weighed at the beginning of this year.” (Subject 20, Male) “Oh yeah? (rising tone). No, I think if anything, I probably have lost a couple, ‘coz I’ve been running, so. . .I duno” (Subject 35, Female) While disagreeing with the confederate, subjects did not change their facial expressions, eye-contact or the tone of voice. They continued the conversation in the way they previously had with the confederate. Some subjects adopted sarcasm and joke as a reply to the confederate’s comment. Based on their tone of voice and the facial expressions, the delivery of the message was not perceived to be either avoiding or aggressive. For example: “Oh thanks. Thanks (laughing loud moving her body farther fiom the confederate) Maybe it’s just ‘coz I am close up (hand gesturing her face, smiling)...1’m 70 kidding. Oh my gosh, this is so funny. I don’t know if I should take offense or not (smiling, and looking at the confederate).”(Subject 13, Female) “Really? Okay... (hand touching her face). . .Oh no! Maybe I should go work out (facial expression and voice were animated smiling while speaking then looking away). . .Um. . .yeah. . . .I don’t know, yeah. . .I try to work out a couple of times a week, but... (looking back at the confederate, smiling, and shrugging shoulders)” (Subject 78, Female) Approximately 12% of the subjects’ responses employed humorous and/or sarcastic remarks. While half of the humorous and/or sarcastic remarks were used with apparent avoiding cues to end the topic, the other half did not, which was coded as non- avoiding non-aggressive facework. In the self-reported survey, two five-point(1=strongly disagree, and 5=strongly agree) Likert-type scale questions asked the subjects to evaluate the level of their response as being humorous (M = 3.02, SD = 1.68) and sarcastic (M = 2.40, SD = 1.11). The two variables were found to be correlated significantly with each other, r (101) = .47, p < .01. However, they were not significantly correlated with avoidance facework, and were moderately correlated with aggression facework measured by self-reported scales, r (101) = .26, p < .01 between humor and aggression, and r (101) = .23, p < .01 between sarcasm and aggression. It seems to suggest that subjects tended to perceive their own use of humor and/or sarcasm as more of aggression than avoidance, although the third-party coders did not share the perception. In summary, in addition to avoidance and aggression, when managing a face threatening act, people employ disagreement with the comment, sarcasm, and jokes in a non-avoiding and non-aggressive manner to address their face. 71 Additional Findings The use of nonverbal cues was prevalent when individuals manage a well- meaning clash. Fifty-five subjects moved away their eye-contact after receiving the comment from the confederate, and 41 of these subjects were coded as using non- aggressive avoiding facework. Fifty-two subjects employed laughter, including nervous laughs and chuckles, after hearing the weight gain comment, and 35 of these participants were coded as using non-aggressive avoiding facework. Nine subjects withdrew their smile gradually after hearing the comment, and eight of the subjects were coded as using non-aggressive avoiding facework. In addition, 18 subjects shrugged their shoulders, and 15 subjects nodded their heads after hearing the comment. According to the confederate’s report, 19 subjects’ faces turned red after she made the comment about the weight gain. Sex difference was first found in the level of other-oriented face needs. Compared to male subjects (M = 3.50, SD = 0.49), female subjects (M = 3.84, SD = 0.55) had a slightly higher level of other positive face needs, t (101) = —3. 17, p < .01, n2 = .09. Female subjects (M = 3.65, SD = 0.51) also had a slightly higher level of other negative face needs than did male subjects (M= 3.37, SD = 0.60), r (101) = —2.58, p < .05, n2 = .06. No significant difference was found on the self positive face needs between male (M = 4.34, SD = 0.50) and female subjects (M= 4.33, SD = 0.52), t (101) = 0.10,p = .92, n2 = .00. Similarly, no difference was found on the self negative face needs between male (M= 3.71, SD = 0.50) and female subjects (M= 3.84, SD = 0.54), t(101) = —1.22,p = .23, n2 = .01. Male and female subjects also differed in their perceptions of the face threatening act. In terms of the politeness of the comment, though both male and females subjects 72 considered the comment was impolite, females (M = 2.00, SD = 0.71) perceived the comment to be even less appropriate than their male counterparts (M = 2.36, SD = 0.79), t (101) = 2.36, p < .05, n2 = .05. Females (M = 3.09, SD = 0.97) also reported higher level of negative emotions after hearing the comment compared to their male counterparts (M = 1.96, SD = 0.87), t (101) = —5.82,p < .001, n2 = .25. In addition, female subjects (M= 3.07, SD = 1.06) perceived the incident as a positive face threat more strongly than male subjects (M = 1.96, SD = 0.86), t (101) = —4.06, p < .001, n2 = .10. Similar pattern was observed as female subjects (M = 3.39, SD = 0.88) considered the incident as a negative face threat more strongly than male subjects (M= 2.51, SD = 0.78), t(101) = —5.01,p < .001, n2 = .20. However, male and female subjects did not differ in their attribution of the incident. For attribution type l-innocent cultural mistake, female subject (M = 3.37, SD = 0.91) and male subjects (M= 3.28, SD = 0.91) did not differ, t (101) = 0.43,p = .67, n2 = .00. For attribution type 2-intentional offense, female subjects (M = 1.73, SD = 0.66) and male subjects (M = 1.60. SD = 0.66) did not differ either, t(101) = 0.92, p = .36, 112 :00. No difference was found in attribution type 3-incidental offense between females (M= 2.76, SD = 0.62), and males (M= 2.51, SD = 0.71), t (101) = -1.81,p = .07, n2 = .03. Yet, female subjects (M = 2.58, SD = 0.64) had a lower level of confidence in their attribution than did male subjects (M= 3.14, SD = 0.95), t(101) = 3.56,p < .01, n2 = .11. Despite the sex differences in other-oriented face needs, and the perceptions of the face threatening act, sex was not a significant predictor of facework strategies with either the self-reported measures or the observational measures. Between perceived physical attractiveness and the rated level of overweight, overweight was entered into the regression analysis, as which was more relevant to the 73 comment on weight gain. This variable, however, did not predict the facework strategies with either the self-reported or the observational measures. 74 CHAPTER 5 DISCUSSION AND ADDITIONAL ANALYSIS The results showed that negative face threat is a significant predicator for avoidance facework between the observational and self-reported measures, while attribution for intentional offense is a significant predictor for aggression only when measured with self-reports. People’s avoidance facework is directly caused by perceived negative face threat in the face threatening act, and attribution of intentional offense is more likely to lead to manage a face threatening situation aggressively. Preference of A voidance It is not hard to understand why people want to avoid in a face threatening situation, since a face threatening act entails discomfort and unpleasantness. To escape from the undesirable situation, avoidance becomes a frequently used facework strategy in social predicaments (Cupach & Irnahori, 1993; Metts & Cupach, 1989). The motivation of avoidance is to maintain, what Goffman (1967) called, poise, “the capacity to suppress and conceal any tendency to become sharnefaced during encounters with others” (p.9). When individuals are caught in a face threatening situation, their faces are attacked, and their poise is at stake. The threat to one’s face is more obvious when the face threatening act is an unexpected violation of social norms and is caused by others. According to subjects’ explanations of their reactions in self-reports, many subjects experienced a loss of composure after hearing the comment. Approximately 40% of the subjects stated that the confederate’s comment was a surprise to them, and they did not know how to react. For example: 75 “I was shocked and didn't know how to reply. I felt like the comment was rude and didn't know how she could possibly know that about me or make that comment to a stranger.” (Subject 76, Female) “I didn't really say anything back, because I didn't know of how else to act or what to say. I was confused as to how she knew who I was.” (Subject 46, Female) Goffman (1967) stated that when a minor face threatening act takes place, social actors often choose to be blind to the incident, allowing those who are involved to recollect themselves. When individuals are faced with losing their poise, the most effective way to maintain or regain poise is to pretend that no face threat has happened, treating the face threat as non-exi sting or non-threatening. According to Cupach and Metts (1994), an absence of composure is usually associated with a lack of control and inability to cope with problematic situations. Losing control of a situation would accentuate the undesirable consequence of the face threat caused by others. Therefore, individuals want to make efforts to regain their composure as an attempt to take control of the situation. Such ability can demonstrate individuals’ competence in social skills of managing their own face (Cupach & Metts, 1994). Brown and Levinson (1987) described negative politeness strategies as avoidance-based, which are characterized by restraint and formality. By avoiding dealing with the face threat directly, individuals signal the other party about social distance, which prevents further damage on their own face. An alternative explanation of the positive relationship between negative face threat and avoidance facework is from the perspective of weighing social cost and benefit. When people’s negative face is threatened by others, they have the choice of addressing their threatened face needs; however, by doing so, they would have to risk threatening the 76 offender’s face. They have to tell the offender what he or she has done wrong or to warn him or her not to do it again, which would be considered as a direct threat to the other’s negative and/or positive face. In a situation like this, people are trapped in a dilemma of either addressing self-oriented face or attending the other’s face needs. It is possible that after weighing the benefit of addressing their own face needs, and the cost of threatening the other’s face, individuals draw the conclusion that it might be too costly to risk threatening the other’s face in order to address that of their own, and therefore, they sacrifice addressing their own face so as not to threaten that of the other’s. For example: “I was shocked she asked me the question, but I did not want to be rude and just come out and say that is none of your business.” (Subject 30, Female) “I didn't want to be mean and be like hey, many thanks for calling me fat.” (Subject 64, Female) “I didn't really want to make her feel bad so I brushed it off and she didn’t seem to scared to say it so I wasn't going to tell her she was being rude.” (Subject 36, Female) “1 just tried to avoid confrontation and/or emotions being revealed to her. I didn't want her to know that as a complete stranger she'd hurt my feelings.” (Subject 38, Male) It can also be understood that subjects’ choice of sacrificing their own face is to avoid further loss of their own face. Since the outcome of addressing their own face needs is relatively unknown to the subjects, they might be afraid that the situation would be exacerbated by giving it more attention (Cupach & Metts, 1994). In other words, the potential benefit of addressing their own damaged face is not justifiable to the efforts they 77 have to go through. Therefore, given that the costs outweigh the benefits, people choose to overlook what has happened, leaving their threatened face unattended. The dynamic of aggression facework is more complex than that of the avoidance. Negative face threat does not directly lead people to use aggression. For individuals to be aggressive in managing a face threatening situation, individuals must attribute the FTA which is imposed by the other party to an intentional offense. It seems that without attributing the FTA to the offender’s mean spirit, individuals do not feel a reason to sacrifice the other person’s face in order to repair that of their own. Use of Nonverbal Cues Nonverbal cues are frequently used both alone and in combination with verbal replies when people attempt to manage a social predicament. There are two major functions that nonverbal cues serve. First, nonverbal cues are used as expressing emotions. According to Burgoon, Buller, and Woodall (1996), people display emotional affect through nonverbal cues either intentionally or even when they are not conscious of it. The most common example is that a smile usually signals happiness, while a frown indicates confusion. In the current study, one subject stared at the confederate after she made the comment, and used a harsh tone to display anger after hearing the comment. However, nonverbal cues do not always reinforce verbal messages. In the current study, subjects sometimes maintained a smile throughout the conversation, leaving the confederate an impression that they had enjoyed the conversation. While during the debriefing session, subjects confessed that they had hidden their real emotions in their interaction with the confederate. Subjects consciously controlled the display of their nonverbal cues to prevent them from revealing their real emotions that they were 78 experiencing. Some subjects also laughed after hearing the comment. Obviously, the laughter is not a Sign of happiness. Instead, it serves as the subjects’ attempt to disguise their real feelings. Second, nonverbal cues function to regulate the direction of interaction (Patterson, 1988). Subjects used nonverbal cues to manage the flow of the conversation. For example, some subjects tried to create a physical or emotional distance from the confederate as a signal of their aversion to the comment, for example, shifting their body to the farther side of the chair, moving away their eye-contact, and withdrawing smile. Goffman (1967) suggested that in a social predicament, people use subtle signs to imply that the other is violating the face needs and to give him or her an opportunity to initiate remedial facework on their own. In the current study, revealing a moderate level of negative emotion could be interpreted as a hint from the subjects to terminate the tOpic. However, it is also possible that subjects were unaware of their use of the nonverbal cues, and displayed the expressions driven by how they felt. Sex Drflerences It is interesting that sex differences were observed only with independent variables, not with the dependent measures. Both males and females exhibit similar pattern of facework strategies in dealing with a well-meaning clash. Traditionally, men are believed to be more assertive, and potentially more aggressive, while women are more avoiding in managing conflicts (e.g., Brewer, Mitchell, & Weber, 2002; Papa & Natalle, 1989; Sorenson, Hawkins, & Sorenson, 1995). However, the empirical findings on sex difference of conflict style preference have been mixed. In a recent review of sex differences in conflict literature, Nichotera and Dorsey (2006) claimed that the observed 79 sex differences in conflict management are most likely driven by situational factors, such as status differentials and nature of the conflict, instead of biological sex alone. In the current study, men and women do not have the same level of other-oriented face needs, or perceive the face threatening situation exactly the same way; however, they do not differ in their attribution of the situation nor the facework strategy use. It is possible that the sex difference on face needs and perception of the situation might be a function of individual differences, which further confirms that different communicative behaviors are as a result of complex interplay of various individual as well as situational factors. In short, contrary to the stereotypes of Americans being direct and confrontational, when they encounter a problematic intercultural communication situation consisting of a negative face threat caused by a stranger, the majority of people choose to avoid further interaction with the offended party. Not giving further attention to the topic and pretending the comment are the major facework strategies Americans employ to cope with a well-meaning clash. A considerable number of subjects employed non-aggressive and non-avoiding facework strategies in the well-meaning clash. In most of the cases, the disagreement was based on the subjects’ awareness that they did not gain any weight. Offering the information and pointing out the fact serve individuals to deflect the face threatening comment. Without disclosing any apparent intent to challenge the offender, individuals’ disagreement disassociates themselves from both the comment and the negative implications of the comment. The disagreement thus firnctions as an indirect signal to the offender that he or she is mistaken. Certainly, one’s impression of another’s appearance might not be as accurate as how others see them or how they see themselves. Impression 80 by nature is subjective, and differs across individuals. Therefore, even though the comment might be wrong as a fact, there is still a possibility for someone to make a comment about one’s appearance change. Using a disagreement with the comment further illustrates that Americans tend to associate weight gain with negative implications. Disagreement with the comment obviously indicates one’s unwilling to be connected with connotations such as unattractiveness, lack of self-control or laziness. The overall anti-fat attitude in the US. is reinforced. Explanation of Insignificant Findings The vast non-significant results for aggression facework measured by the third- party observers are mainly due to a general lack of variance in the dependent variable. When avoidance and aggression strategies were observed by coders, each of them was coded with dichotomous categories, and analyzed by binary logistic regressions. When dichotomously coded aggression facework was the outcome variable, 90% of the subjects were coded as non-aggressive facework, which resulted in an overwhelmingly concentrated distribution in the dependent measure. This result pattern could have overshadowed the predictive capabilities of the predictors. The invariance of aggression facework strategies might be explained by the relationship type between the confederate and the subject. During the debriefing session after the interaction, a majority of subjects mentioned that they would have been more open or more confrontational with the confederate, if the person were a friend of theirs. For example: “If it’s a friend who said it, I’d probably be less upset. I probably have gone into more details, say something like ‘it’s not true, ‘coz I have lost weight’. I would 81 react a little more, explain a little more. I was completely caught off-guard. ‘Coz for strangers, they don’t know you, for someone to be so personal, to comment on something so personal is a lot different, but with friends, you can come out and be honest with them, but with strangers, you still worry about it.” (Subject 99, Female) “1 would probably react differently if it’s a friend who said it. I think I would be more offended by it. I don’t know. I mean. . .an example of that would be if my roommate said, ‘you gained weight’, I would be like ‘shut up!’ But it’s a stranger. . .I probably would be more polite to strangers.” (Subject 86, Male) Relationship type determines the level of closeness and familiarity. Brown and Levinson (1987) claimed that with increasing relational intimacy, the need for positive face would increase, and the need for negative face would decrease. Lim (1990) found that individuals make more efforts to avoid threatening a friend’s positive face than that of an acquaintance. Goldsmith (1994) also demonstrated that support messages between friends that acknowledge positive face needs receive the highest rating in helpfirlness. Few studies were conducted to directly compare the negative face needs between fiiends and strangers; however, it is more common for friends to threaten each other’s negative face than for strangers. Generally speaking, one would feel more comfortable to ask a friend than a stranger for a favor. In a study on apology intentions, it showed that when a social breach took place, Americans believe that a higher degree of severity would be felt by a stranger than by a friend imposed by the same face threatening act, and American offenders have a stronger intention to apologize to a stranger than to a friend (Guan, Park, & Lee, 2007). 82 According to Uncertainty Reduction Theory (Berger & Calabrese, 1975), uncertainty with strangers would motivate information seeking in order to reduce uncertainty. Berger (1979) modified the theory by pointing out that with strangers who people might not develop a future relationship, they would be more hesitant to risk crossing the stranger’s boundaries. In a face threatening situation caused by a stranger, individuals are concerned that the stranger’s reaction might worsen the situation, if they confront him or her. Not knowing exactly how a stranger would react prohibits individuals from using any aggressive facework strategies. A lack of relational development further reduces the need for doing so. However, with a fiiend who one is familiar with and values the relationship, such uncertainty will be low. Therefore, it might be possible that people vary their facework depending on the type of interpersonal relationship with the person involved in the conversation. Based on the explanations of individuals’ preference of avoidance over aggression facework, it suggests that both the relationship type and the need to regain composure are the motivation to choose to avoid in the given face threatening situation. According to Leary and Kowalski (1990), when managing one’s image, individuals have to decide which image they want to construct and how to go about making it. This process is influenced by both individual differences and situational factors. A primary situational factor would be the degree of face threat one perceives in the act (Hodgins, Liebeskind, & Schwartz, 1996). Hodgins and his colleagues (1996) argued that the degree of face threats is also influenced by individual differences. For example, people who have high need to seek external control might be more sensitive to the threat to their own face, 83 whereas people who are secure with the control of their own of actions tend to be more attuned to threats of others’ faces (Hodgins et al., 1996). Regression analyses with avoidance as the dependent variable also showed that neither the self- and other-oriented face predicts avoidance facework directly, nor moderates the effect between negative face threat and avoidance facework. The theoretical thread and the empirical results suggest that instead of an interaction relationship between face needs and face threats on facework, there might be a casual string from individual trait-like face needs, to the perceived face threats, and then to the avoidance facework. As for the aggression facework, even though logistic regression analyses with observational aggression facework did not reveal much information, OLS regression analysis showed that attribution type 2-intentional offense is a significant predictor for aggression facework when measured continuously by self-reported scales. Theoretically, different types of attribution of the face threatening act would affect social actor’ s facework (Goffman, 1967), although the corresponding relationship between the type of attribution and the facework strategies is not specified. In attribution literature, both offender’s responsibility and the severity of the offense have been shown to be strongly associated with the offender’s subsequent remedial strategies and the victim’s reception of the remedial strategy (e. g., Bennett & Earwaker, 1994; Han & Cai, 2006; Kuha, 2003). This seems to suggest that a casual string might also exist among the negative face threat, perceived severity of the comment on weight gain, attribution type 2-intentional offense and aggression facework. 84 Additional Analysis In view of the argument so far, a path model (see Figure 1) was proposed and tested with LISREL 8.80 (JOreskog & SOrbom, 2006). As LISREL cannot handle categorical dependent variables, instead of using the categorical avoidance and aggression facework coded by the third-party observers, in the path model, third-party observers’ evaluation of the magnitude of the avoidance and aggression facework were used. The two facework strategies were each measured by a ten-point Likert-type scale and were significantly correlated with their corresponding variables measured categorically. For avoidance, the correlation between the third-party observed magnitude and third-party observed coding was r (101) = .73, p < .001, and for aggression, the correlation between the two measures was r (101) = .72, p < .001. The fit indexes of the original path model indicated a marginal fit with x2 = 63.14, df= 39, p = .0085, RMSEA = .079, CFI = .88, IFI = .89, NF]: .78, NNFI = .84, and GFI = .90. However, the paths from each of self-positive face, self-negative face, and other positive face to negative face threat were not significant at p = .05 level. To develop a model that fits the data better, various modifications were made to the model. A model that removed self-oriented positive and negative face needs and connected other-oriented positive face with intentional offense provided the best fit with )8 = 47.70, df= 27, p = .0083, RMSEA = .088, CH = .89, IFI = .90, NF] = .80, NNFI = .85, and GFI = .91, and A12 = 15.44 was insignificant with Adf= 12 (see Figure 2). There are two major changes in the revised model. First, self-oriented face needs do not exert causal influence on face threats. Self-oriented face needs do not associate with the negative face threat, instead, only other negative face is a significant predictor 85 .8600 Deco lo>< ..de :82 mom-m3 mo. V a... .8er86 and a? sees ea .8an ._ seem nacho 3:32.85 comd e82 mom-es... eeocm ..vmé 3am wuz 350 088 mom 350 08m woz hem 08m mom hum 86 Deco Iww< 8. v a . .mEomoEooo £8 £3» .32: 53 @833“ .N uSmE .830 19?. ...de acmd emu-a2 8588 how ww< GQMAV —§_u=0a=_ / . 08:.— mOnm need new $1 550 828 ensue 8a.» 92 mom 22 .88 3882 .Re 87 for the negative face threat. A possible explanation is that individuals who are concerned about protecting others from being imposed or intruded may perceive higher degree of face threat in the F TAs that threaten their own negative face caused by others. Although it is seemingly contradictory to common sense, it is consistent with subjects’ explanations of their avoidance facework. Subjects who chose to “brush off” the comment were mainly concerned about not making the confederate uncomfortable. In other words, other negative face needs are a causal force for the avoidance facework. Second, other positive face has a causal impact on attribution. This can be explained that attribution could be influenced by personality variables. Mitchell (1987) found that a range of trait variables, for example, empathy, emotional stability, and extroversion, are strongly associated with attribution styles, and asserted that there is a casual relationship between the personality factors and attribution styles. The revised model indicates that the higher need to maintain other’s positive image one has, the less likely one would attribute the face threatening act as an intentional offense, which reduces the likelihood to the use of aggression facework in managing the face threatening situation. 88 CHAPTER 6 GENERAL DISCUSSION This study incorporates both observational and self-reported measures to examine people’s facework strategies in a problematic intercultural communication situation. A total of 103 American subjects participated in the study. They completed two self- reported questionnaires and had an interaction with an international student supposedly from China, who made a comment to create a well-meaning clash. Subjects’ strategies to manage the situation were recorded from both self-reported questionnaires including Likert-type scale items and open-ended questions, and third-party observers who watched both the live and the video-taped interactions. The results were partially consistent between the self-reported measures and the observational measures. Level of A greement between Self-Reported and Observational Measures To examine the level of agreement between self-reported and observational measures, the results of the current study were also compared with that from the prelim study. The comparison shows that they are partially consistent across measures and studies. Regression results were compared across the prelim self-reported study with the weight vignette, the self-report portion in the current study, and the observation portion in the current study (see Tables 12-15). Negative face threat is a significant predictor for avoidance facework across both measures and studies. With the level of negative face threat perceived in the well-meaning clash becomes higher, the likelihood of avoiding greatly increases. Attributing the comment as an intentional offense is a significant 89 3. u we Bases .8. v as: 8. n genes? .8. v 888 u A8 .8 a H.89: =85 .- EN .8 k “ seen. .826 Etudeénauafi sausagesomeameta Susanefimzwxua “nonsense .688 :85 .88 u as .8 888k .88 u Ex .8 .8888 8.. 8.. $828 3. Se- §887 2. m3- §.82.e- aae-zézo 8. 8.8 8e88,? 2. 8.8- 5888- 2. Se €828 see-2:8 8. e2 c.0888 S. 8.8 8.828 m. .85 A2888 aae-zfiE 8. e; @888- 8. 38 A2888 8. 82. £888 ease-Zenam 0.62m 888 8.va.8.§u$u§~a emumxeanaendaume etheanaenfitmx negas $88 :85 .88. v a .8... n Ea a: 48. v a .83 t c 3.8 a 825-28 8v in 588.8 83 83 388.8 83 88 8.82.8 8258a .3382 8.. e2 588.8 8. 28 A2888 2. Se- 8828- 82939889350 8. 88 58%.? em. «.3- 5828- 2. 88 58:8 $988328“...on R. E 5.88.8 :- :.e @888 8. 8.8 5.888 Ezmvooaosenwonesm 8. a... $88.8- R. 8.8.. €88.8- on 8.8- 5888- 588832.888 8.85 Ea snare 88am 58 m saga sum 88 m 628:. sum 88 m 8526.... cfioéoeoan c8728 canoe-cease 88qu confirm-ease E ~qu 28288-an 28.53 “severance 8 85203 e5 85st? 8033qu mm “8.5 08m 98 33: comm .23 .33» Hee-56 28 8:03 5953 cam—888 £38 888on .m. 2an 90 ”m. u @888 u a ".88 xx ”Gog—MK. Eve—c =80>O 3. 88 588 88: 2. $8 8388 :8. R. S: awe 28 .8. n .33 u G is «x ”cog—6‘5 385 =Eo>o :. of fine :8 3. 83 $8838. 8. $8 $88 :8. 88. 88 628 28. E. 38 Gas 28. S. 2: £88 88 3. n Q 883.88 .2. n «82 u 38 .88 k 8258 =eo>o 3. u mwéug 8m." a 82 u :8 d mmafiom 8. 28 E88 8M8 N8. 88 A288 88 we. 88: A288 38- 8. u mm 33.85 .88. "31:. ".52 u 68 8E an. 888 A288 888 a. 28 A288 28 Q. 38. A288 38. we. 88. 2 ~88 888. N8. 28 588 88 we. 38 A288 .88 2. "$388.88 .88. v.33. u $8 .3 k “ ESE =eo>o S. .I. wmfiqfig .om. u A .888 u 38 .9 umafié a. 888 5.828 mm. 888: 58:88- a. 88. 388888: 2. "$888838 "mm .88. vagmufqak 2. 888 588 28 :8. v 28. 3888381 8. 28 88.8 28 R. 88 $8.88 888 2. 88: A888 888: 28.3-8 88$ 68 m Amenzv 38848230 33.8 sum 69 m A8778 :82..=omua§o 83.8 sum 68 m 2865 383.5 @3qu 382.83 8:28 coco—Ego x m. 3:25:00 x N— oocouccoo x : goo—m vacuum E92230 xom 353200 AB 9.5% 38283 .2 an 9.8.88 38235-8 :< :8 93%: .8339. .2 x85 amt..— 3583‘ 633...? 33:33 8 35.86% 98 333...? 8288335 88 $90 .8955; .23 .338 80:3 23 8:05 5232. 83388 $388 :ommmohwom .2 058,—. 91 8. u h188. u a ".8 Na ”Gog—Syn 385 —_d._o>C 3. 8.8 $88 88: mm. :8 :28 8...: .m. 2.. 888 $8 2. 2... G88 2... 2.. u 2.3.“ n G ".8 ~x S. u E .822...8 .8. u a .2... n .8 .8 2 H.255 ..226 no. M 533.35% .8. u m .2... u .8 .8 85:852. 88. .28: 8.8 R8: N... 8.8 .38 ..8 2. 28 .28 28 mm. ..8 .38 :8 2.. "$823.58 8.. um... 2. "$523.58 ..8. v.33 u 98 .8...” .082 ..205 me. .I: mngug .3. u R .888 n EN .8 82.38... .8. -8 8.828 2. 8.8: 2.828: 88. .88: :88 .88: 2. N88 .2888 2.:EB§.§.2:E H88.82.25 .282 ..205 .8. u a. 8. .N n 98 .22 ..88. v a 82 u c 3.2... .8. 8... .38 :8 88. 28 $88 28 .88.v .28 688 38 8.. mm... 58 2.: 8. 38: .28 .28: .m. 88 8.8 8.8 .8. .m... .38 2... 8... 88 3.8 88 ..N. .88: ...8 ..8: N8. 888 2:8 28: 2.. .88 .28 888 8. 88 A. .8 2.8 8. .88 58 888: S. 88: 8.8 88: mm. .88 .28 .88 03.2... .89.... 68 m 252... 52. G8 m 25;... 52. .mm. m 8.8.03. 3.885 28.28 3580:2280 98.22. 288283835 .2 anz. 2822.285»... 825:2 2 ...zo 325:2 2 EC 325:2 x 82m 825:2 x .25 V.8... 288m 9.25:2. .82.... 38... 02832 $28 8». 35%...35 8.5. 82 05.88225 E28 85 252.823 £28 85 35.88208 U.8... .2... 2.23238. 033...; 252.328 3 5.88838 2.8 8038...? 28.85305... 2. 89.5 8.... 2... .68.. 88.. .23 .33.». 2.225 .23 8:9:— cooEon 88282.88 2.38 commmohwom .3 Bang. 92 2. u 22 8288...... ..8. v 2.83 2. u N2 828...... ..88. v 2.88.8 u .8 .88 2 U.8885 ..226 u ...8N .22 n .2.8:. ..285 .8. u 2 .82 n .8 :8... N8. ..8. n 88.6822 N8. n 825.8822 ..N. n 2 H8888......6 .282: ..226 .8. .u 2 .22.. u .8 .2 885.882 .N8.. : ...8N .2 8882.82 28. 8... .8888 28 N.. 8.8 .288 28 88. 88 .28. 888 882.880 x m. .N. 8N8 .8888 MN. .: 2. 28: .288 28: 88. 888 .288 888 8888888 x N. N8. 8N. ..888 8N8 N... 888 .88. 888 88. 8.8 .8888 28 852.88 x .. Jug—m vacuum 2.:2 .88.Nn.8.:..3N.. N..:N2822..88.....uN2 1:22.28? 85:22 H8888......5 .882: ..225 ..8. v 2 .888 n .88 .88 2 ..88. v 2 .88... u .. .N.2 2 2. $8 .288 ...8: 8. ..888: .2888 .888: - - - 28.2285 8N. 8N.N ..288 2.8: 2.. 888: .288 8.8: 88. N88 .. .88 88 .38 8... 828 .8888 888: 88. 888 .8888 88 NN. 888: .888 888: 8:28:88 2.. .8.. .2888 ...8 ..N. 28 .888. ...8 88. 88 .888. ..88 .28 282.8 382.652 5... 2.. 2... .N88. R8 .88.v ...8 .8888 888 .88.v N88 .8888 888 .28 2:2"... .88....25N 8... 82. 888 .388 28: 2. 28 .2888 8.8 .8. 28: .288 8.8: ...8 2.8.2:. .233... 8... 28... .2... 88.81. .692 .82 m 288>... 28m. .82 m 2281. 888m .m2 m 8.822.... .2223 282.83 .8728 3880:2258 .8 .28 28228.88:.5§u .8 . N:28 28828.88:E..22 638...? 85.8.8285 28 commmouwwm 2.8 8.83.2.2; 2.0.2.328... mm womb 2225528 .23 .838 82.80 new 8:22. 28:53 28.2888 2.22: 228833. .2 038,—. 93 predictor for aggression; however, it is only shared by self-reported measures between the two studies. Some findings in the prelim study are not shared by either the self-reported measures or the observational measures in the current study. First, attributional confidence was a significant predicator in the negative direction for avoidance in the prelim study. The more confident individuals were in their attribution of the F TA, the less likely were they intend to avoid the situation. Second, negative face threat was a significant predictor for aggression. The higher level of negative face perceived in the FTA, the more likely one was to use aggression facework. Third, Hierarchical Linear Modeling (HLM) analysis with multiple vignettes showed that other-oriented face needs moderated the relationship between face threats and avoidance facework. In the current study, however, no interaction was found between face needs and face threats in either of the self-reported measure or observational measure. It might be difficult for different research methods to completely replicate each other due to the limitations of each method. However, both self-reported measures between the prelim study and the current study also yielded different pattern of results. A possible explanation lies in the different designs of self-reports. The self-reported questionnaire in the current study asked subjects to recall how they reacted to the confederate’s comment, instead of indicating the intention that what they would do in a hypothetical situation, as it was measured in the prelim study. It is possible that the current self-reported measure might be more accurate in reflecting subjects’ actual behavior in managing the situation. However, the vast non-significant results with the self-reported measures in the current study do not provide a clear picture on the variable 94 relationship. Further, as the current study is only about one particular situation, the findings are limited in capturing the impact of situational variations on facework strategies. Between the self-reported and observational dependent measures in the current study, there is only a moderate level of correlation. Correlation determines the strength and direction of a linear relationship between the two variables (Cohen, Cohen, West, & Aiken, 2003). The moderate level of association between the two measures suggests that results from the two research methods do not completely converge, which questions the validity and reliability of the findings. Several factors could have attributed to the inconsistent results between measures. First, it is possible that the subjects’ own understanding of their own use of facework might not be consistent with how they carried out the facework in actual behavior. Subjects might believe they were trying to correct the mistake of the confederate, which was conceptualized as aggression, while their nonverbal cues indicated an obvious avoiding tendency. It is likely that subjects were not even aware of their nonverbal cues, which however, gave away their real intention in managing the problematic situation. In daily life, people’ facework strategies do not always match their intended outcome. Sometimes a not-so-funny joke that meant to case an embarrassing situation turns the situation even more uncomfortable. Second, the complexity of facework behaviors posed difficulty for third-party observers’ coding. The inter-coder reliability of the facework strategies was around .80, which was acceptable, though not satisfactory. The discussion of coding variations between the coders was often centered on the real intention of the subjects. The coders sometimes had different interpretations of the subjects’ facework 95 based on their observation of the subjects’ verbal and nonverbal cues. Additionally, most subjects did not rely on a single verbal reply to manage the situation. According to the frequencies of subjects’ replies, it is clear that the majority of subjects adopted a combination of verbal and nonverbal replies, which increased the difficulty of coders’ agreement. Third, the self-reported measures in the current study did not correspond to all possible avoidance and aggression facework strategies observed. The self-reported scale items of facework strategies were created and developed based on the self-report from the prelim study, which was a priori to the observations. Some strategies, for example nonverbal cues signaling avoidance and pointing out the mistake counted as aggression, were not discovered in the prelim study, and therefore were not included in the scale items of aggression and avoidance in the current study. The coding scheme was revised based on the observations from the study, and some categories had to be added to the original coding scheme in order to match the subjects’ reactions. The discrepancy between self-reports and the coding scheme might have contributed to the inconsistent results between the two measures. The medium level of agreement between the observational measures and the self- reported survey measures in the current study, and between the self-reported measures of the prelim and of the current study may suggest that the methodology could have influenced the findings. On one hand, social desirability might distort people’s answers on self-reports; however, self-reports also provide useful information on the relationship between predictors and outcome variables. On the other hand, observations are limited by contextual factors, lacking in generalizability (Denzin, 1989), yet they also yield rich content on the characteristics of facework strategies, especially in terms of the details of 96 facework strategies and use of nonverbal cues. In light of the weaknesses and the strengths of both measures, it would be too soon to draw the conclusion that one measure is necessarily superior to, or more valid and reliable than the other. Despite the discrepancies of the results between measures and across studies, the relationship between negative face threat and avoidance facework is consistent from the prelim study to the current study and from the self-reported measures to the observational measures. The consistent finding shows that the two methods do converge to some extent. The partially inconsistent results, although pose difficulty in understanding the facework dynamics of how people manage well-meaning clashes, provide a foundation for a new direction of the variable relationship. Specifically, instead of moderating the relationship between face needs and facework, face threat mediates face needs and facework. This new approach of variable relationship would have not been discovered without using different research methods. Therefore, rather than claiming one method is necessarily better than the other, research questions might be better answered by applying both approaches. Mixed-[Method Approach The prelim study and the current study demonstrate the importance of integrating different research methods to study behavioral outcomes in problematic communication situations. The mix-method approach, however, is not new in empirical research. Campbell and Fiske (1959) first introduced the concept of multi-trait multi-method (MTMM) approach, and they expected MTMM to help researchers evaluate research findings, and improve finding validity. Mathison (1988) believed that “good research practice obligates the researcher to triangulate, that is, to use multiple methods, data 97 sources, and researchers to enhance the validity of research findings” (p. 13). The major assumption of the proposition is that research methods are subjective due to the strengths as well as the limitations of each specific methodology approach (Mathison, 1988). Denzin (1989) argued that although there were risks of using multiple methods as it would be difficult to reconcile inconsistent findings, different findings would provide different understandings, reveal new information about the research, and force researchers to be more rigorous in their conclusions. Practical Implications One of the contributions of the current study is that it serves as a meaningful learning experience for the subjects. During the debriefing session, many subjects commented that they enjoyed participating in the study, because it made them aware of the discrepancy of the social norms in different cultures, especially for something that was considered as a unsaid social noun. The experience made them realize how social norms function, and how powerful of social norms are, especially when they are violated. For those subjects, the research participation experience helped them learn to be more conscious of the difficulty in intercultural communication, improve their accurate attribution of a problematic communication situation, and prepare them with potential facework strategies in their future intercultural encounters. All of these would eventually enhance their intercultural communication competence. A few subjects with ethnic backgrounds and some of the international students, who were removed from the analysis, confirmed that in their ethnic or national cultures, commenting on one’s weight gain is considered as an acceptable norm. Their family members, friends, or even they themselves had made similar comments like this before. 98 The information provided evidence on the validity of the well-meaning clash scenarios applied in the current study. American subjects’ unwillingness to correct the confederate’s cultural mistake provides insights on the difficulties faced by intercultural communicators. Although hiding one’s real emotions and pretending nothing serious happened might have helped the Americans regain poise during the face threatening act, the intercultural communicator left the interaction without knowing that she had committed a face threat. None of the subjects in the current study attempted to correct the confederate or point out that she was making a cultural mistake. When examining the interview transcripts and self-reported questionnaires, however, it becomes clear that subjects experienced negative emotions after hearing the comment. On one hand, intercultural communicators are completely unaware of their offense, and on the other hand, most of the Americans choose not to confront them, even though they believe that most likely the other party is making a cultural mistake. Then the questions become: how and when can intercultural communicators find out about their offense so they will not commit the same offense again? Is avoidance the optimum outcome for improving intercultural communication effectiveness in the long run? Is it possible that intercultural communicators might benefit from constructive criticism so they will not offend another American person in the future? Those questions would be useful to integrate into the current intercultural communication training programs. Nonverbal cues were observed for almost all subjects in the current study. However, the ambiguity of nonverbal cues could further increase the difficulty for intercultural communicators to accurately interpret the situation. When nonverbal cues 99 are used to disguise one’s real emotions, it would be unlikely for intercultural communicators to detect if there is anything wrong. Even when nonverbal cues are used as a signal to show one’s real emotions, the subtlety of the cues and the multiple meanings could be troublesome for intercultural communicators to catch the hint. Sarcasm is frequently used in interpersonal communication. An accurate interpretation of sarcastic remarks requires competence in both socio-linguistic and cultural knowledge, while a lack of which would easily mislead intercultural communicators. For example, one of the subjects used “thanks” as a sarcastic remark to imply his or her unhappiness when hearing the comment. The confederate later reported that the particular reacted positively to the comment because the person used “thanks” and was smiling. A misunderstanding like this could even filrther obstruct the effectiveness in intercultural communication. As many difficulties as intercultural communicators might face, there are channels that can be taken advantage of in order to facilitate the intercultural communication process. Many subjects expressed that they would have been more open and direct if it were a friend who commented on the weight gain. In other words, for intercultural communicators to improve their communication effectiveness, establishing friendship with native speakers and seeking their advice might be an efficient and non- threatening way. In addition, intercultural communicators fiom cultures outside the US. should be attuned to the subtle nonverbal cues during their interactions. This way, intercultural communicators could sense cultural norm violation on their part, which might lead them to initiate inquiry, and eventually correct their mistake, in case of a well- meaning clash. 100 Limitations A major limitation of the current study is a lack of variance in the aggressive facework. One of the primary reasons is that the confederate was a stranger and a female. Subjects indicated that they would have changed their reactions if the person were a friend of theirs or if the person were a male. With the relationship type and sex of the confederate being different, there may be more variations in individuals’ facework strategies. Another limitation of the study is that although observational data provide important insights to the research topic, the weaknesses of observations require a cautious interpretation of the findings. First, any observation is limited to specified contextual particularity (Denzin, 1989). In the current study, individuals were observed on a single occasion, on a particular day and in a laboratory. Whether or not and how well the facework individuals adopted during the observation can be translated and generalized in a different occasion, a different time or in naturalistic settings are unknown. Second, subjects’ reactivity might have influenced their reaction to the comment (Denzin, 1989). Subjects were informed that they would be video-taped the whole process of their participation in the study, including the conversation with the confederate. During the debriefing session, some subjects mentioned that they knew they were being watched when they were talking to the confederate. Being aware of the third-party observation could have hindered the subjects’ actual behavior, which might have affected the finding validity and reliability. Third, inter-coder reliability was not satisfactory in the study. One of the reasons was the original coding scheme was not comprehensive. The coding scheme was constantly revised based on new reply categories from observations, which 101 could have hurt accuracy of coding. During training, coders practiced on replies that were generated from the previous self-reported study due to the absence of actual behaviors. Compared to the observed replies, self-reported replies usually contained one type of the coded response. The complex replies observed from subjects in their interaction with the confederate increased the difficulty in coders’ agreement of coding. A third limitation of the study is that there is no control group with intracultural communication dyads to compare the results with. In the prelim study with intercultural and intracultural conversation partner, the cultural background of the conversation partner showed a main effect on the avoidance facework. Therefore, it is possible that the findings related to the avoidance and aggression facework could have resulted from a combination of negative face threat and intercultural communicator partner. Future Research In view of the limitations, one of the first directions for the future research is to vary the confederate’s identity, including sex, relationship type, and cultural background. Although practically, it might not be easy to arrange a fiiend of the subject to be the confederate, changing the confederate’s sex may be the first step. Observing intercultural friends interacting with each other unobtrusively or asking people to recall miscommunication with their intercultural friends due to cultural norm differences can also be an alternative. Recording how people react to the same comment when it is delivered by an intracultural partner can also help exclude the effect of cultural background of the conversation partner on facework strategies. Second, multiple observations in different occasions of the same type of well- meaning clash would benefit the reliability of the finding. Including different types of 102 well-meaning clash scenarios and observing subjects’ facework strategies across situations can help establish a general pattern of facework strategies in dealing with well- meaning clashes. Thus, the findings would not be limited by the contextual particularity. It would be interesting to study how people from other cultures manage a well- meaning clash. In the current study, a Chinese created a well-meaning clash and threatened an American person’s face. Intercultural communication literature will benefit with studies examining what acts are considered as face threatening to a Chinese person, but polite and appropriate in US. culture, and how Chinese react to a well-meaning clash caused by an American. Conclusion In summary, avoidance is the primary facework strategy for Americans to cope with a social predicament that threatens their negative face, and the perceived intentional offense leads Americans to use aggression facework in managing the situation. Other- oriented face needs, instead of self-oriented face needs, exert causal impact on the perceived level of face threat and attribution. Alternative facework strategies include a disagreement with the comment, jokes and sarcasm. Nonverbal cues are used in combination with verbal replies; however, the functions are contextual. The current study supports a mixed-method approach to study behavioral outcomes in intercultural communication. The decision of using multiple methods is determined by the nature of constructs under investigation, and the limitations as well as strengths of each research method. Researchers need to evaluate the findings considering the method effect, and be cautious in drawing conclusions. Face is a useful concept and possesses strong explanation power. Many empirical 103 studies have been done to examine face in social predicaments. The current study contributes to the general understanding of face. Researchers who are interested in face should combine both individuals’ general level of face needs and situational face threats when studying facework strategies in their research practice. Additionally, the current study advances in operationalizing and measuring face and facework. Future research should continue giving attention to the dynamic of face in social interactions and apply multiple research methods to answer research questions in order to generate valid as well as reliable research findings. 104 APPENDIX A Self-Reported Surveys Before the observation Self-positive face F1. It is important for me to look good in front of other people F2. Maintaining a positive image is important to me. F3. I feel good when people have a good impression of me. F4. Making a good impression is important to me. F5. Having my judgment valued is important to me. * F6. Having my opinions appreciated is important to me. * Self-negative face F7. I want my privacy to be respected. * F8. I don’t want my personal life discussed. F9. I want other people stay out of my business. F10. My boundaries should be respected. F11. I have clear boundaries for other people. F12. I prefer to keep people at distance. * F13. It’s important not to put myself in a situation where I would feel indebted to someone. F14. I don’t like to be obligated to do anything. * F15. I prefer to have control over my personal space. F16. I find the feeling of owing someone something to be undesirable. * Other-positive face F17. I try to keep my conversation partner feel good about themselves. F 18. Helping others maintain the positive image of themselves is important to me. F19. I make an effort to not let others embarrass themselves. F20. I feel bad for those who embarrass themselves in public. F21. It is important for me not to make my conversation partner look bad. F22. When in public, I hold bad my criticism of others so that they will not look bad. F23. Even if someone annoys me, I avoid hurting them by keeping my feelings to myself. * Other-negative face F24. I try to avoid telling other people what they should do. F25. It is important for everyone to allow others to choose how they act. * F26. It is important to respect other people’s boundaries. * F27. I try not to invade other people’s personal matters. F28. I withhold making personal comments about others. F29. It is important to me not to tell others how to behave. 105 After the observation Politeness of act Her comment was appropriate. Her comment was polite. Her comment was proper. Her comment was considerate. Her comment showed good manners. Her comment was uncalled for.(R) * Severity of act I was angry at what she said. I was upset by her comment. I was left with negative feelings afier hearing the comment. I was not happy with her comment. Positive face threat Her comment made me look bad. Her comment embarrassed me. What she said made me feel awkward. What she said made me feel ashamed about myself. What she said made me feel bad about myself. What she said made me feel self-conscious. Negative face threat Her comment made me not know what to say. Her comment made me speechless. What she said was intrusive. Her comment disturbed me. Her comment made me flustered. Her comment made me uncomfortable. Attribution type l-innocent cultural mistake She was making an innocent cultural mistake. She did not know how her comment might be perceived. She had no idea that her comment might be inappropriate. She would not have made the comment if she knew it would be seen as inappropriate. Attribution type 2-intentional offense She was trying to insult me. Her intentions were mean-spirited. She was trying to hurt my feelings. She was trying to make me feel bad. Attribution type 3—incidental offense She knew her comment might hurt me, but it was not her intention to hurt me. She might have anticipated that I would be offended by her comment, but she had a good reason to make the comment. It was not her intention to make me feel bad, but she might have anticipated that her comment would hurt me. * Even though what she said hurt me, I know she did not mean to do that. She was honest with me, even though she knew it might hurt my feelings. 106 Attributional confidence I am absolutely certain of her intentions in giving the comment. It is very clear to me why she gave me the comment. I am definite about her intentions. I am not sure of the intention of her comment. I am confident in my explanation as to why she gave the comment. Avoiding facework I said nothing more on the topic. I ignored the comments. I changed the topic. I ended the conversation. I made an excuse to leave the scene. * Aggressive facework I confronted her about what she said. I challenged her about what she said. I acted defensively. * I demanded an apology from her. I demanded an explanation from her. I pointed out the rudeness of her comments. Other facework After hearing what she said, I made a joke about her comments. After hearing what she said, I used sarcasm to reply to her. Open-ended questions Do you remember the student’s name? Do you remember where she is from? Please write down how you replied to her as detailed as you can after she made the comment “you have gained a few pounds”. Why did you reply this way (in other words, what goal did you want to accomplish by replying this way)? How did you she respond, and how did her response make you feel? *items were deleted in the analysis to improve reliabilities 107 Code 10 11 12 13 14 15 16 17 18 19 100 101 102 103 104 105 106 107 APPENDIX B Revised Coding Scheme Replies Subject replied with exclamation mark (e. g., “Oh my goodness”, or “Wow”) Subject did not say anything and left the scene Subject did not say anything and had no nonverbal reaction Subject responded nonverbally with apparent avoiding cues (e. g., moving away eye-contact) Subject said that he/she had to leave without further information (e.g., “You know I have to go now”) Subject said that he/she had to leave followed by an explanation (e.g., I have another meeting after this, I have to go now”) Subject did not respond to the comment at all, but started to talk about something different (e. g., “How about the class you took?”). Subject did not deny the comment, but tried to end the conversation with filler words (e. g., “Oh really?”, “Probably”, “I don’t know”). Subject did not deny the comment, and gave a realistic but short explanation of the weight gain, trying to end the conversation (e.g., “It’s stress”, “I haven’t been to the gym for a while”) Subject acknowledged the comment, trying to end the conversation (e. g., “Yeah, I have”) Subject denied the comment directly, and pointed out the confederate was wrong (e. g., “No, you must have mistaken me for someone else.”) Subject denied the comment completely with aggressive nonverbal cues (e. g., “No, I don’t think so”) Subject denied the comment, and questioned the confederate for implication of the comment (e. g., “No way! Are you saying I am fat?”) Subject denied the comment, and questioned the confederate for the intent (e. g., “It’s impossible. Why did you say that?”) Subject denied the comment, and directly expressed that his/her negative emotion after hearing the comment (e. g., “That’s not true. I’m not happy to hear that”) Subject denied the comment, and pointed out the inappropriateness of the comments (e. g., “You are wrong and that’s rude to say” “I don’t think so, and what you said is inconsiderate”, “That’s not true. I can’t believe you just said that”) Subject denied the comment, and warned the subject not to say similar things (e.g., “I don’t think so, and don’t say that again” “You should be careful”) Subject denied the comment, and demanded the confederate to apologize to 108 108 109 110 200 201 202 203 204 205 206 207 208 209 210 1100 1101 1102 1103 99 him or her (e. g., “That’s not true, and you should apologize to me”) Subject denied the comment, and attacked the confederate without using vulgar terms (e.g., “You are wrong, and you look fat”) Subject denied the comment, and started to use vulgar terms. SUbject denied the comment and replied with sarcasm to attack the confgggrggg (e.g., .‘It’s not true, but you just made my day”) Subject did not deny the comment directly, but pointed out that the confederate was mistaken (e.g., “You are thinking me as someone else”) Subject did not deny the comment directly, but denied the implication of the comment (e.g., “I am not fat”, “I don’t think I look fat”) Subject did not deny the comment directly, but questioned the confederate for the implication (e.g., “Are you saying I look fat?” “Do I look healthier?”) Subject did not deny the comment, but questioned the confederate for the intent (e.g., “Why did you say that?”) Subject did not deny the comment directly, but directly expressed negative emotions he/she felt (e.g., “I am not happy to hear that) Subject did not deny the comment directly, but pointed out the inappropriateness of the comment (e.g., “That’s a rude thing you just said”) Subject did not deny the comment directly, but warned the confederate not to say similar things again (e. g., “You got to be more careful”) Subject did not deny the comment directly, but demanded the confederate to apologize to him/her (e.g., “You should apologize for what you said”) Subject did not deny the comment directly, but started to attack the confederate without using vulgar terms (e.g., “You look like you gained more than a few pounds”, “You are no Brad Pitt either”) Subject did not deny the comment directly, but started to use vulgar terms. Subject did not deny the comment directly, but replied with sarcasm to attack the confederate (e. g., “Aren’t you polite?”, “What a great personality that you have”) Subject simply disagreed with the confederate’s comment, and/or gave explanation to support the disagreement. No apparent avoiding or aggressive cues (e.g., “I just weighed myself, it’s still the same”) Subject was positive to hear the comment and replied positively (e. g., “I am happy to hear that”) Subject replied with sarcasm, with no apparent avoiding or aggressive cues (e.g., “Well, thank you,” “Sweet.”) Subject responded with humor, with no apparent avoiding or aggressive cues (e.g., “I like food. . .a lot.”) None of the available codes applied. 109 REFERENCES Abraham, C, Sheeran, P., and Johnston, M. (1998). From heath beliefs to self-regulation: Theoretical advances in the psychology of action control. Ps h l H 13, 569-591. Adams, A. S., Soumerai, S. B., Lomas, J ., and Ross-Degnan, D. (1999). Evidence of self- report bias in assessing adherence of guidelines. International J Qumal fgr MM in Health Care 11, 187-192. Affifi, W. A., and Burgoon, J. K. (1998). “We never talk about that”: A comparison of cross-sex friendships and dating relationships on uncertainty and topic avoidance. Personal Relationships, 5, 255-272. Allred, K. G. (2000). Anger and retaliation in conflict: the role of attribution. In M. Deutsch, and P. T. Coleman (Eds). The k of C nfli R l 'on: Theo Practi (pp.236-255). San Francisco, CA: Jossey-Bass. Applegate, J. L. (1982). The impact of construct system development on communication and impression formation in persuasive context. Communication Mgnographs, 49, 277-289. Armstrong, C. B., and Kaplowitz, S. A. (1991). Sociolinguistic inference and intercultural coorientation: A Bayesian model of communicative competence in intercultural interaction, Human Communication Research, 27, 350-381. Bakerman, R. (2000). Behavioral observation and coding. In H. T. Reis and C. M. Judd (Eds) Handbggk of Researgh Methgis in Sggal and Personalig Psychology, (pp.138-159). New York, NY: Cambridge. Baxter, L. A. (1984). An investigation of compliance-gaining as politeness. Human Communication Regard}, 10, 427-456. Bennett, M., and Earwaker, D. (1994): Victims' responses to apologies: The effects of offender responsibility and offense severity, Journal of Social Psychology, 134, 457-464. Berger, C. R. (1979). Beyond initial interactions. In H. Giles and R. T. Clair (Eds). Lanmge and Social Psychology, (pp.122-l44). Oxford, UK: Blackwell. Berger, C. R., and Calabrese, C. (1975). Some explorations in initial interactions and beyond: Toward developmental theory of interpersonal communication. Human Communication Research 1, 99-112. 110 Blum-Kulka, 8., House, J ., and Kasper, G. (1989). Cross-cultural pragmatics. Norwood, N.J.: Ablex. Braaten, D. 0., Cody, M. J., and DeTienne, T. B. (1993). Account episodes in organizations: Remedial work and impression management. Management Communication @arterly, 6, 219-250. Bradburn, N. M., Sudman, S., Blair, E., and Stocking, C. (1978). Questions threat and response bias. Public Opinion Qparterly, 42, 221-234. Brewer, N., Mitchell, P., and Weber, N. (2002). Gender role, organizational status and conflict management styles. International qumg of Conflipt Management, 13, 78-94. Brislin, R. W. (1993). Understanding culture’s influence on behavior. Orlando, FL: Harcourt Brace Jovanovich. Brown, R, and Levinson, S. C. (1987). Pplitggss: Spme universgls in lapgpagp ugges, Cambridge, UK: Cambridge University Press. Burgoon, J. K., Buller, D. B., and Woodall, W. G. (1996). Nonverbal communication: The unspoken dialer (2“‘1 Ed.). New York, NY: McGraw-Hill. Campbell, D. T., and Fiske, D. W. (1959). Convergent and discriminant validation by the multitrait-multimethod maxtrix. Psycholpg'cal Bulletin, 56, 81-105. Cannon-Bowers, J. A, and Salas, E. (2001). Reflection on shared cognition. Journal of Orgag'zatipngl Behavipr, 22, 195-202. Carson, C. L., and Cupach, W. R. (2000). Facing corrections in the workplace: The influence of perceived face threat on the consequences of managerial reproaches. Journal pf Applied memunicatipn Rpsgpph, 28, 215-234. Chen, H., Gao, X., and Jackson, T. (2007). Predictive models for understanding body dissatisfaction among Chinese young males and females in China. Behavior Resgr'ch and Ther_apy, 45, 1345-1356. Clatterbuck, G. W. (1979). Attributional confidence and uncertainty in initial interaction. Human Communication Research 5, 147-157. Cohen, J ., Cohen, R, West, 8. G., and Aiken, L. S. (2003). Applied multiple regression/correlap' on anflysis for the behavifl sciences. (3r Ed). Mahwah, NJ: Lawrence Erlbaum. Corral-Verdugo, V. (1997). Dual “realities” of conservation behavior: Self-reports vs. observations of re-use and recycling behavior, Journal of Environmental 111 Cupach, W. R., and Carson, C. L. (2002). Characteristics and consequences of interpersonal complaints associated with perceived face threat. Journal of Social and Personal Releu'enehips. 19, 443-462. Cupach, W. R., and Irnahori, T. T. (1993). Managing social predicaments created by others: A comparison of Japanese and American facework. Western Journal of Communicatiop, 57, 431-444. Cupach, W. R., and Metts, S. (1992). The effects of type of predicament and embarrassability on remedial responses to embarrassing situation. Communication Quarterly, 40, 149-161. Cupbach, W. R., and Metts, S. (1994). Fecework. Thousand Oaks, CA: Sage. Denzin, N. K. (1989). The research get: A theoretical introduction to m’elogieu methods. New York, NY: McGraw-Hill. Edelmann, R. L. (1985). Social embarrassment: An analysis of the process. Journal of Social Es! Personal Reletienshipe, 2, 195-213. Eisenberg, E. (1984). Ambiguity as a strategy in organizational communication. Communication Monomphs, 51, 227-242. Fink, E. L., and Walker, B. A. (1977). Humorous response to embarrassment. Psychologig Repog, 40, 475-485. Fishbein, M., and Ajzen, I. (1975). Belief, atu'tuee, intenu'en epd behevier; An introduction to theogg and research. Reading, MA: Addison Wesley. Gao, G., and Ting-Toomey, S. (1998). Communicating effectively with the Chinese. Thousand Oaks, CA: Sage. Gerbner, G., Gross, L., and Morgan, M. (2002). Growing up with television: Cultivation processes. In J. Bryant and D. Zillman (Eds), Media effects: Advances in theory and researeh (2nd Ed.). (pp. 43—67). Mahwah, NJ :Erlbaum. Goffman, E. (1967). Ingaction ritual; Eeseye on face-to-feg behavier. Garden City, NY: Doubleday. Goffinan, E. (1976). Relations in public. Garden City, NY: Doubleday. Goldsmith, D. J. (1994). The role of facework in supportive communication. In B. R. Burleson, T. L. Albrecht, and I. G. Sarason (Eds) Communieeuon of Sflifl support: Mesgges, interactions, relationships and communipy. (pp.29-49). Newbury Park, CA: Sage. 112 Grabe, 8., Ward, L. M., and Hyde, J. S. (2008). The role of media in body image concerns among women: A meta-analysis of experimental and correlational studies. Psychological Bulletiu, 134, 460-476. Graham, E. E. (1985). The involvement of sense of humor in the development of social relationships. Communication Reports, 8, 158-169. Graham, E. E., Papa, M. J., and Brooks, G. P. (1992). Functions of humor in conversation: Conceptualization and measurement. Western Jmmal of Cemmgieetion, 56, 161-181. Greenleaf, C., Starks, M., Gomez, L., Chambliss, H., and Martin, S. (2004). Weight- related words associated with figure silhouettes. Body Imege, 1, 373-3 84. Guan, X., Park, H. S., and Lee, H. Y. (2007). Cross-cultural differences in apology. Paper submitted for publication in International Journal of Intercultural Relations. Gudykunst, W. B. (1988). Uncertainty and anxiety. In Y. Y. Kim & W. B. Gudykunst (Eds) Theories in interguturfl Qmmguu'eetion. (pp. 123-156). Newbury Park, CA: Sage. Gudykunst, W. B., and Shapiro, R (1996). Communication in everyday interpersonal and intergroup encounters. Int rn ti 11 al of Inter al R lati n 20, l9-45. Gudykunst, W. B., and Ting-Toomey, S. (1996). Communication in personal relationships across cultures: An introduction. In W. B. Gudykunst, S. Ting- Toomey, and T. Nishida (Eds). Commm’ Q1210!) in firsonal relatienshipe moss cultures (pp. 3-18). Thousand Oaks, CA: Sage. Hall, E. T. (1978). Beyond Culture. New York, NJ : Doubleday. Han, B., and Cai, D. A. (2005, May). The effect of responsibility, severity, and relationship on face concerns and apolegy: A egoss-cultural pgepe_c’g've. Paper presented at the 55 annual conference of the International Communication Association in New York, NY. Hatch, A. J. (1987). Impression management in kindergarten classrooms: An analysis of children’s face-work with peers. An 01 Education rl 18, 100- 1 15. Ho, D. Y. (1976). On the concept of face. American Journal of Sociology, 81, 867-884. Hodgins, H. S., Liebeskind, E., and Schwartz, W. (1996). Getting out of hot water: Facework in social predicament. Joumfl of Personalig and Sg'al Psyehology, 71, 300-3 14. 113 Holtgraves, T., and Yang, J. N. (1990). Politeness as universal: Cross-cultural perceptions of request strategies and inferences based on their use. Journal of Personaliug and Social Ps cholo 59, 719-729. Holtgraves, T., and Yang, J. N. (1992). Interpersonal underpinnings of request strategies: General principles and differences due to culture and gender. Journal of Pereenalig and Social Pychelegy, 62, 246-256. Hu, W., and Grove, CL. (1991). Encomtering the Chinese. Yarmouth, ME: Intercultural Press. Irnahori, T. T., and Cupach, W. R. (1994).A cross-cultural comparison of the interpretation and management of face: US American and Japanese responses to embarrassing predicaments. International Journal of Intercultural Relations 18, 183-219. Joreskog, K. G., and Sorbom, D. (2001). L_ierel 8: User’s referenfiguide. Lincolnwood, IL: Scientific Software International. Jorgensen, J. (1996). The functions of sarcastic irony in speech. Journal of Pregmatice, 26, 61 3-634. Jung, E. H. (1999). The acquisition of communicative competence in a second language. Journal of Pan-Pacific association of epplied linguistics, 3, 13-37. Kelley, H. H., and Michela, J. L. (1990). Attribution theory and research, Annqu Review of Psychology, 31, 457-501. Kerlinger, F. N., and Lee, H. B. (1999). Feuueeu'ons of pehavieral reeflch. New York, NY: Wadsworth Kiesler, S. B. (1973). Preference for predictability or unpredictability as a mediator as a reaction to norm violations, Journal of Pegeenalig and Sgu'fl Psychelegy, 27, 354-359. Klimoski, R, and Mohammed, S. (1994). Team mental model: The construct or metaphor? Journal of Megement, 20, 403-437. Kowalski, R. (2003). I was only kidding: Vrctims and perpetrator’s perception of teasing. Personalig and Social Psychology Bulletip, 26, 231-241. Kuha, M. (2000). Perceived seriousness of offense: the ignored extraneous variable. Journal of Prematics, 35, 1803-1821. La Fave, L., Haddad, J., and Maesen, W. A. (1976). Superiority, enhanced self-esteem, 114 and perceived incongruity humor theory. In A. J. Chapman and H. C. Foot (Eds), Humour and laughter: Thmpy, resegeh, ane epplications (pp. 63-91). New York, NY: John Wiley and Sons. LaPiere, R. T. (1934). Attitudes vs. actions. Seg'al Forces, 13, 230-237. Leary, M. R., and Kowalski, R. M. (1990). Impression management: A literature review and two-component model. Psyehelogical Bulletip, 107, 34-47. Lee, H. O., and Boster, F. J. (1991). Social information for uncertainty reduction during initial interaction. International and Intercultural Communication Annual 15, 189-212. Lee, S., Chui, H. F. K., and Chen. C-N. (1989). Anorexia nervosa in Hong Kong: Why not more in Chinese? British Journal of Psychiatry, 154, 683-688. Leichy, G., and Applegate, J. L. (1991). Social-cognitive and situational influences on the use of face-saving persuasive strategies. Human Communicatien Reseflh, 17, 451-487. Li, H. Z. (1999). Communicating in conversations: A cross-cultural comparison. International J oumal pf Intereul_uual Relatiens, 23, 387-409. Lim, T. (1990). Politeness behavior in social influence situations. In J. P. Dillard (Ed), Seeking compliance: The production of intemrsonal influence messeges (pp. 7 5- 86). Scottsdale, AZ: Gorsuch Scarisbrick. Mathison, S. (1988). Why triangulate? Edueeu'onel Reeeatcher, 17, 13-17. McLaughlin, M. L., Cody, M. J., and O’Hair, H. D. (1983). The management of failure events: Some contextual determinants of accounting behavior. Human Communication Regen, 9, 208-224. McLeod, J. M., and Chaffee, S. H. (1973). Interpersonal approaches to communication research. American Behavioral Scientist, 16, 469-500. Metts, S. (1997). Face and facework: Implications for the study of personal relationships. In S. Duck (Ed), Handbmk of pegenal relatienships (2nd. Ed) (pp.373-3 90). New York, NY: John Wiley & Sons. Meyer, J. C. (2000). Humor as double-edged sword: Four functions of humor in communication. Communicau'en Thm, 10, 310-331. Mishel, M. H. (1990) Reconceptualization of the uncertainty in illness theory. Imege: Jeurnfl of Nursing Scholarship, 22, 256—262. 115 Mitchell, J. V. (1989). Personality correlates with attributional style. Journal of Psychology, 123, 443-467. Nichotera, A. M., and Dorsey, L. K. (2006). Individual and interactive processes in organizational conflict. In J .G. Oetzel and S. Ting-Toomey (Eds). The Sage hauehmk of conflict communication: Intgtating theog, research and practice. (pp.293-326). Thousand Oaks, CA: Sage. Noclg M. K., and Kurtz, S. M. S. (2005). Direct behavioral observation in school settings: Bring science to practice. Cogm'tive and Behavioral Practice, 12, 359-370. Oetzel, J., and Ting-Toomey, S. (2003). Face concerns in interpersonal conflict: A cross- cultural empirical test of the face negotiation theory. Communieau'en Research, 30, 599-624. Oetzel, J ., Ting-Toomey, S., Masumoto, T., Yokochi, Y., Pan, X., Takai, J ., and Wilcox, R. (2001). Face and facework in conflict: A cross-cultural comparison of China, Germany, Japan, and the United States. Communication Monogtaphs, 68, 23 5- 258. O’Quinn, K., and Aronoff, J. (1981). Humor as a technique of social influence, Sexual Psychology gra_rteriy, 44, 349-357. Pager, D., and Quillian, L. (2005). Walking the talk? What employers say versus what they do. Amm'can Sociolegical Review, 70, 355-380. Papa, M. J ., and Natalle, E. J. (1989). Gender, strategy selection, and satisfaction in interpersonal conflict. Western Journal of Speech Communication, 53, 260-272. Patterson, M. L. (1988). Functions of nonverbal behavior in close relationships. In. S. W. Duck (Ed). Handbook pf pepsonu relatt'onehips (pp.4l-56). New York, NY: Wiley. Rockwell, P. and Theriot, E. M. (2001). Culture, gender, and gender mix in encoders of sarcasm: A self-assessment analysis. Communication Resmch Repeuts, 18, 44-52. Saunders, P. A. (1998). “You are out of your mind!”: Humor as a face-saving strategy during neuropsychological examinations, Health Communication 10, 357-372. Schimanoff, S. B. (1985). Rules for governing the verbal expression of emotions between married couples. Western Journal ef Spflh Communicatien, 49, 147-165. Schimanoff, S. B. (1987). Types of emotional disclosures and request compliance between spouses. Communication Monogtaphs, 54, 85-100. Schlenker, B. R (1980). Impression management: The self-concept, social identity, and intetpfional relatiens. Monterey, CA: Brooks and Cole. 116 Schlenker, B. R., and Darby, B. W. (1981). The use of apologies in social predicaments. Social Psychology Quarterly, 44, 271-278. Schlenker, B. R., Weigold, M. G., and Hallam, J. R. (1990). Self-serving attributions in social context: Effects of self-esteem and social pressure. Journal of Personaliug and Social Psychology, 58, 855-863. Sherif, M. (1936) The Psychology of social norms. New York, NY: Harper and Brothers. Sillars, A. L. (1980). Attributions and communication in roommate conflicts. Communication Monogtuphe, 47, 180-200. Snyder, M. (1979). Self-monitoring processes. In L. Berkowitz (Ed), Advgces in experimental social psychology (Vol. 12, pp. 85-125). New York, NY: Academic Press. Sorenson, P. S., Hawkins, K., and Sorenson, R. L. (1995). Gender, psychological type of conflict style preference. Management Communication Qtarterly, 9, 115-126. Steward, E. C., and Bennett, M. J. (1991). American cultural patterns: A cross-culu_rru perspective. Yarmouth, ME: Intercultural Press. Ting-Toomey, S. (1988). Intercultural conflict styles: A face-negotiation theory. In Y. Kim and W. Gudykunst (Eds), Theories in intercultural communication. Newbury Park, CA: Sage. Ting-Toomey, S. (1999). Qommunicatjng acreee gluuea New York, NY: Guilford Press. Ting-Toomey, S. (2005). The matrix of face: An updated face-negotiation theory. In W. B. Gudykunst (Ed). Theorizing about Intercultural communication. (pp.7l-92). Thousand Oaks, CA: Sage. Ting-Toomey, S., and Kurogi, A. (1998). Facework competence in intercultural conflict: An updated face-negotiation theory. International Journal of Intercultural Relations, 22, 187-225. Ting-Toomey, S., and Oetzel, J. G. (2001). Managing intercultural conflict efferfively. Thousand Oaks, CA: Sage. Tracy, K. (1990). The many faces of facework. In H. Giles and W. P. Robinson (Eds). Handhook of Language and Social Psycholegy. (pp.209-226). New York, NY: John Wiley & Sons. Tracy, K., Craig, R., Smith, M., and Spisak, F. (1984). The discourse of requests: Assessment of a compliance-gaining approach. Human Communieatien Resear_eh, 117 10, 513-538. Trees, A. R, and Manusov, V. (1998). Managing face concerns in criticism: Integrating nonverbal behaviors as a dimension of politeness in female friendship dyads. Human Communication Research, 24, 564-583. Triandis, H. C. (1959). Cognitive similarity and interpersonal communication in industry. Journal ef Applied Psychelegy, 43, 321-326. Triandis, H. C. ( 1960). Cognitive similarity and communication in a dyad. Hill—11E Relations 13, 175-183. Trandis, H. C. (1979). Values, attitudes, and interpersonal behavior. In H. E. Howe Jr., and M. Page (Eds), Nebraska sympeeium en metivatien (pp. 195-260). Lincoln, NA: University of Nebraska Press. Webb, T. L., and Sheeran, P. (2006). Does changing behavioral intentions engender behavior change? A meta-analysis of the experimental evidence. Psychelogieat Bulletip, 132, 249-268. Vourela, T. (2005). Laughing matters: A case study of humor in multicultural business negotiations. Negotiau'on tMu, 49, 105-130. 118 uarrtrrrrrrrrrrrtl 9