U UNIVERSITY OF CINCINNATI Date: May 18, 2009 Yan Yang I, , hereby submit this original work as part of the requirements for the degree of: Doctor of Philosophy in Educational Studies It is entitled: Causal Attribution and Culture--How Similar Are American and Chinese Thinking? Student Signature: This work and its defense approved by: Committee Chair: Ryan Adams Daniel D. Wheeler Rhonda D. Brown Approval of the electronic document: I have reviewed the Thesis/Dissertation in its final electronic format and certify that it is an accurate copy of the document reviewed and approved by the committee. Committee Chair signature: Causal Attribution and Culture— How Similar are American and Chinese Thinking? A dissertation submitted to the Graduate School of the University of Cincinnati in partial fulfillment of the requirements for the degree of Doctor of Philosophy (Ph.D.) in the Department of Educational Studies of the College of Education, Criminal Justice, and Human Services June 2009 by Yan Yang M.A., Shanghai Maritime University, 2000 B.A., Liaoning Normal University, 1997 Dissertation Committee: Ryan Adams, Ph.D., Chair Daniel D. Wheeler, Ph.D. Rhonda Douglas Brown, Ph.D. Abstract Recent psychological studies of culture have shown incompatible results about the differences in causal attributions between Americans and East Asians. Nisbett and his colleagues (Nisbett, 2003) found that Americans tend to make dispositional or internal attributions while East Asians tend to make situational or external attributions. Nisbett traced this pattern back to the differences between ancient Greek philosophy and Chinese Confucianism and Taoism. The opposite pattern of results was found by Crystal, Hui, and some other scholars (Crystal, 2000; Crystal, Parrott, Okazaki, & Watanabe, 2001; Hui, 2000, 2001). They theorized that Americans, as individualists, make external attribution as a self-enhancement mechanism and East Asians, as collectivists, make internal attribution to protect the group. The purposes of this study were to find out if those cultural differences in causal attribution could be replicated, if the content of attribution led to different responses, and if the question format (open-ended versus rating-scale questions) made a difference in participant’s attribution. This study used an online survey with one scenario from each of the two group’s representative studies, a campus-shooting scenario and a bullying schoolboy scenario. The two scenarios were rewritten to fit in both cultures and the two characters in the scenarios were given typical American names in the English questionnaire and Chinese names in the Chinese version. The open-ended questionnaires for both scenarios preceded the rating-scale questions. There were 76 American participants and 157 Chinese participants. There were separate web pages for the English and Chinese versions of the survey so that all participants took the survey in their native language. ii The results indicate that both scenario and question format had significant effects on causal attributions. The two scenarios led to opposite attributions on the rating-scale questionnaire: external attribution in response to the bullying scenario and internal attribution of in response to the campus-shooting scenario. Question format made significant difference in the bullying scenario: internal responses to the open-ended items and external responses to the rating-scale items. Both culture groups made similar causal attribution, either internal or external, of the same scenario in the same question format. American group held stronger opinions in rating-scale attributions and provided more responses in open-ended attributions compared to their Chinese counterparts. The results of this study do not follow the patterns predicted by either Nisbett (2003) or Crystal (20000). Causal attribution is more context-specific than culturespecific. iii Copyright iv Acknowledgement There are many people who have supported and helped me during my doctoral study. Without them, I would not have been able to attain this degree. First of all, I want to thank my advisor, Dr. Dan Wheeler, who has guided me with whole-hearted patience and support through the program and helped me grow intellectually. He is a mentor who showed me how to find my direction and steer my boat in academia. The rest of the committee has provided great advices and moral support during my study and dissertation as well. Dr. Rhonda Brown has been helping and providing advice since the preliminary hearing in the first year of my study. I appreciate the meetings during which she gave me advices, the detailed comments that she made on my work, and the opportunities that she provided for me to teach. I want to thank Dr. Ryan Adams for serving on the committee, for providing great feedback, for answering my questions, and for meeting me on a short notice. I also want to give my gratitude to Dr. Gulbahar Beckett who listened to my concerns, provided good advice for my problems, gave me support for my academic growth, and shared my Chinese connection. My husband, Robert, provided more support than I could ever have expected. He helped me with data collection and coding, took care of household chores and listened to my concerns. My parents always lavished me with their love and encouragement and they are the motivation for me to seek this degree. I also appreciate very much that my in-laws were supportive and helpful during my study. As a friend told me all the time, you cannot finish your dissertation on your own. It is a team work. There are many others who helped me during my research that I want to thank very much. My cousins, Hongli Xu, Di Wu; my friends, Dave Sacks, Lanbin Ren, Zhimin Liang, Jian v Xu, Shufang Zhang, Rachel Noll, Robert Kallmeyer, Karin Mendoza, and many other friends that I will have to fill another page up if I list their names. They helped me greatly with my pilot studies, data collection, and data coding. They helped me keep my moral up when I feel stressed out. They let me know that I was not alone. Finally, I want to thank Educational Studies program for providing me this excellent learning experience when I met so many wonderful people and took my first step in the academic world. vi TABLE OF CONTENTS List of Tables .................................................................................................................................. x List of Figures ................................................................................................................................ xi I. INTRODUCTION ....................................................................................................................... 1 Literature..................................................................................................................................... 1 The Psychology of Culture ..................................................................................................... 1 Western and Eastern Cultures................................................................................................. 5 Attribution............................................................................................................................... 6 Individualism-Collectivism Approach to Attribution........................................................... 10 Intellectual History Approach............................................................................................... 12 The Third Group: Individualism-Collectivism or Intellectual History? ............................... 16 Study Rationale and Purpose .................................................................................................... 18 Research Questions and Hypotheses ........................................................................................ 20 II. METHOD................................................................................................................................. 24 Participants................................................................................................................................ 24 Demographic Results ............................................................................................................ 25 Materials ................................................................................................................................... 25 Demographic questions......................................................................................................... 25 Scenarios ............................................................................................................................... 26 Open-ended Questionnaire.................................................................................................... 27 Rating-Scale Questionnaire .................................................................................................. 28 Procedure .................................................................................................................................. 29 Design ....................................................................................................................................... 32 vii III. RESULTS ............................................................................................................................... 33 Data Preparation........................................................................................................................ 33 Coding of Open-ended Responses ........................................................................................ 33 Factor Analysis of Rating-Scale Items ................................................................................. 34 Derivation of Dependent Measures from Open-ended Responses ....................................... 37 Derivation of Dependent Measures from Rating-scale Responses....................................... 37 Distribution and Transformation........................................................................................... 38 Data Analysis ............................................................................................................................ 38 Attribution Preference........................................................................................................... 39 Open-ended Responses ..................................................................................................... 39 Rating-Scale Responses .................................................................................................... 40 Internal Versus External Attributions ................................................................................... 42 Open-ended Responses to the Dave/De (Bullying) Scenario ........................................... 42 Open-ended Responses to the Michael/Ming (campus-shooting) Scenario ..................... 43 Rating-scale Responses to the Dave/De (Bullying) Scenario........................................... 44 Rating-scale Responses to the Michael/Ming (campus-shooting) Scenario..................... 45 Itemized Scores..................................................................................................................... 46 Analysis of Open-ended Category Endorsement of the Dave/De (Bullying) Scenario.... 46 Analysis of Open-ended Category Endorsement of the Michael/Ming (campus-shooting) Scenario............................................................................................................................. 47 Analysis of Rating-scale Items for the Dave/De (Bullying) Scenario.............................. 48 Analysis of Rating-scale Items for Michael/Ming (campus-shooting) Scenario.............. 50 Question Format Effect......................................................................................................... 53 viii IV. DISCUSSION......................................................................................................................... 55 Interpretation............................................................................................................................. 55 Causal Attribution and Question Format Effects.................................................................. 56 Causal Attribution and Cultural Effects ............................................................................... 57 Limitations ................................................................................................................................ 59 Conclusions and Implications ................................................................................................... 61 REFERENCES ............................................................................................................................. 64 APPENDIX A............................................................................................................................... 72 Recruitment Letters in English ..................................................................................................... 72 APPENDIX B ............................................................................................................................... 74 Recruitment Letters in Chinese..................................................................................................... 74 APPENDIX C ............................................................................................................................... 76 Cognition and Culture Survey in English ..................................................................................... 76 APPENDIX D............................................................................................................................... 83 Cognition and Culture Survey in Chinese .................................................................................... 83 APPENDIX E ............................................................................................................................... 90 Coding Rubric of Open-ended Responses .................................................................................... 90 ix List of Tables Table 1 Demographics of Participants.......................................................................................... 25 Table 2 Inter-coder Reliability for Open-Ended Responses ......................................................... 34 Table 3 Factor Analysis Summary of Dave/De (Bullying) Rating-Scale Items........................... 35 Table 4 Factor Analysis Summary of Michael/Ming (campus-shooting) Rating-Scale Items..... 36 Table 5 Dependent Measures........................................................................................................ 38 Table 6 Endorsement of Dave/De (Bullying) Open-ended Categories ........................................ 47 Table 7 Endorsement of Dave/De (Bullying) Open-ended Categories ........................................ 48 Table 8 Cultural Effects Among the Dave/De (Bullying) Rating-Scale Items............................. 50 Table 9 Cultural Effects in Michael/Ming (campus-shooting) Rating-Scale Items ..................... 52 Table 10 Participants’ Percentages for Each Scenario and Attribution Type............................... 54 x List of Figures Figure 1 Two Sequences of Online Questionnaire Pages............................................................. 31 Figure 3 Scenario and Cultural Effects on Causal Attributions of Rating-scale Questions ......... 41 Figure 4 Cultural Effect and Type-of-Cause Effect in Open-ended Attributions of the Dave/De (Bullying) Scenario............................................................................................................... 43 Figure 5 Cultural Effect and Type-of-Cause Effect in Open-ended Attributions of the Michael/Ming (campus-shooting) Scenario.......................................................................... 44 Figure 6 Cultural Effect and Type-of-Cause Effect in Rating Attributions of the Dave/De (Bullying) Scenario............................................................................................................... 45 Figure 7 Cultural Effect and Type-of-Cause Effect in Rating Attributions of the Michael/Ming (campus-shooting) Scenario.................................................................................................. 46 xi I. INTRODUCTION People usually want to make sense of the events in their lives, especially when these events are negative. The inference process by which people explain an event or behavior is referred to as causal attribution in psychology. People from various cultures differ in their explanations of why certain events have happened in their lives (Crystal, 2000; Crystal et al., 2001; Hui, 2000, 2001; Morris & Peng, 1994; Nisbett, 2003; Nisbett, Peng, Choi, & Norenzayan, 2001). This study is an effort to explore potential differences between American and Chinese explanations of two specific incidents—a campus shooting and a bullying behavior. The goal of this chapter is to provide background on psychological studies of culture and attribution theories, to review the literature on cross-cultural causal attribution, and to propose the research questions for this study. Literature The Psychology of Culture Culture has been an increasingly popular topic of psychological research during the past two decades. The scientific study of the link between psychology and culture can be traced back to the 19th century (Triandis, 1996). The psychological study of culture has greatly benefited from anthropologists’ efforts in defining culture, especially those of Tylor, Boas, and Sapir. British anthropologist Edward Tylor (1832-1917) was considered as a founder of cultural anthropology and he was widely recognized for an all-inclusive definition of culture. He proposed that “culture, or civilization, taken in its broad, ethnographic sense, is that complex whole which includes knowledge, belief, art, morals, law, custom, and any other capabilities and habits acquired by man as a member of society” (1871/1958, p. 1). There are two essential 1 components of culture in this definition—the human capability to create culture and the products of this capability. This capability comes from the human mind and originally develops inwardly; however, it is eventually affected by its social cultural context as well (Vygotsky, 1929). Tylor (1871/1958) also pointed out that the cultural conditions in different societies provided an apt subject for the study of human thought and action. Franz Boas (1858-1942) disagreed with Tylor’s cultural evolutionism and refused to accept that certain cultures and races were superior to others. He insisted that culture has to be understood in specific cultural contexts. Different cultures evolve in a non-linear way and are not parallel to each other. In addition, type, language, and culture are neither closely nor permanently related (Boas, 1911/1938). Type can be loosely understood as race in this case. This antievolutionist view along with Boas’ other thoughts were later summarized by his students as cultural relativism. “Culture may be defined as the totality of the mental and physical reactions and activities that characterize the behavior of the individuals composing a social group collectively and individually in relation to their natural environment, to other groups, to members of the group itself and of each individual to himself” (Boas, 1911/1938, p. 159). Though radical in his time, this definition remains central to anthropologists today (Holloway, 1997). Cultural relativism also gave rise to cultural psychology in the 1980s and indigenous psychology in the 1990s (Triandis, 2007) which will be covered later in this chapter. During the 1920s and 1930s, Boas’ students Ruth Benedict, Margaret Mead, and Edward Sapir significantly contributed to the theoretical positions for the field later known as culture and personality studies (LeVine, 2001, 2007). These studies are also referred to as the culture and personality movement in American social and behavioral science. These studies included many theories and approaches to the relationship between culture and the individual by anthropologists, 2 psychologists, and psychiatrists. An interesting remark regarding this relationship is the following: “Culture is to society what memory is to individuals” (Kluckhohn, 1954). This field is considered a precursor of contemporary cross-cultural studies on personality (LeVine, 2001). In the late 1950s psychologists started to shift away from behaviorism and began to communicate with linguists and computer scientists (Miller, 2003). Their efforts gave birth to cognitive science in the 1960s; in the meantime, dissonance theory and attribution theory came into being in social psychology. However, only a small group of psychologists, such as Harry Triandis, Michael Bond, John Berry, and Michael Cole, devoted themselves to cultural psychology (Nisbett, 2007, p.838). Interest in the psychology of culture did not revive until the 1980s; since then, more and more scholars have acknowledged the importance of culture in understanding the human mind and human behavior and have joined forces with researchers in cross- cultural, cultural, and indigenous psychology. Cross-cultural psychology is the systematic (explicit or implicit) comparison of thought and behavior in different cultural contexts. Cross-cultural research is usually conducted using quantitative method with culture as an antecedent or independent variable whereas cultural psychology uses ethnographic methods and treats culture as existing inside an individual (Greenfield, 2000; Triandis, 2000). Although cross-cultural psychology examines both similarities and differences in psychological functioning and behaviors of people from various cultures or ethnic groups, it is mainly concerned with understanding human diversity (Berry, 1997; Matsumoto, 2001). This has been criticized as conveniently labeling culture as the difference in behaviors and beliefs among different cultures (Hinde, 1987) and separating culture and individuals. Cross-cultural psychology has changed considerably over the past three decades with the emergence of the other two subfields—cultural psychology and indigenous psychology. 3 Cultural psychology usually avoids direct comparisons among cultures but endeavors to understand how mind and culture interact in specific contexts (Adamopoulos & Lonner, 2001). Cultural psychology honors the deep mentality differences among people from different cultures and aims at psychological pluralism without denying all universals or commonalities (Shweder, 2000). In cultural psychology, culture is the medium of learning about people and the world through which “the sources of development that underpin traditional developmental theories (nature-nurture, biology-environment, individual-society) interact to produce this development” (Cole, 1995, p. 31). Cultural psychology is still in its developing stage and is marked by its lack of a commonly accepted orientation and methodology. Indigenous psychology is similar to cultural psychology in the sense that it shares the emphasis on examining cognition and behavior in specific culture contexts. In addition, indigenous psychologists investigate “the knowledge, skills, and beliefs people have about themselves, and study these aspects in natural contexts” (Kim & Park, 2006). This approach is well-known for encouraging the development of a psychology of people in a specific cultural context. However, indigenous psychology has attracted criticism because of its cultural relativist views and difficulty in defining the indigenousness. Although there are heated debates among the advocates for each of the three approaches, differences among them are small and the approaches can complement one another. We cannot fully understand one culture without understanding other cultures, we cannot understand cultural differences without an understanding of cultural similarities, and we cannot understand cultural universals without an understanding of cultural peculiarities (N. Ross, 2004). It is very difficult to tease apart universals from specifics or to separate similarities from differences. When culture is looked at in terms of similarities, culture becomes defined as 4 integration of universal human knowledge, beliefs, and practices. When culture is treated in terms of differences, culture is defined as the unique social values and material traits of a racial, religious, or social group. There is general agreement that people from different cultures share similarities and have differences. There is little agreement, however, upon whether those differences are minor or major and whether they are stable. In order to study culture and psychology, it is necessary to create functionally equivalent and culturally sensitive instruments to tease apart the universal and specific aspects of culture (Cole, 2005). Cultural psychologists emphasize cultural differences among cultural groups because “cultural traditions and social practices regulate, express, and transform the human psyche, resulting less in psychic unity for human kind than in ethnic divergences in mind, self, and emotion” (Shweder, 1991, p. 72). Western and Eastern Cultures Western and Eastern cultures have been extensively investigated because of their different historical traditions, political ideologies, and social outlooks. Western cultures usually refer to European, North American, and Australian cultures. Eastern cultures usually refer to East Asia — China and the nations that were heavily influenced by Chinese traditional culture, such as Japan and Korea. Many differences in behavioral practices and beliefs have been found between these two cultures. For instance, parenting practices differ significantly between Americans and Asians in that American parents see great developmental value in play for their pre-school children whereas Asian parents see less developmental value in play but believe in early academic training (Parmar, Harkness, & Super, 2004). In addition, American college students were found to be better at creating artwork than their Chinese counterparts according to both American and Chinese judges (Niu & Sternberg, 2001). Western and Eastern people were also found to think differently. Studies show that East Asians focus on the field and rely more on 5 intuitive and dialectical reasoning whereas Americans focus on the object and rely more on formal and analytical thinking (Nisbett et al., 2001; Norenzayan, Smith, Kim, & Nisbett, 2002). Western and East Asian cultures have also been extensively compared on their different social outlooks — individualism and collectivism. Western society is generally accepted as the individualist culture within which people value independence, attend to the self, and seek to express unique individual attributes. East Asian nations are usually considered as typical collectivist cultures within which people emphasize interdependent relationships among individuals, attend to the goals of the group, and behave in a communal way (Markus & Kitayama, 1991; Mills & Clark, 1982; Triandis, 2001). The difference in independent and interdependent self-construal can influence emotion, cognition, and motivation. For instance, Chinese individuals were found to more readily infer personality traits based on group membership and tended to stereotype more than did their American counterparts (SpencerRodgers, Williams, Hamilton, & Peng, 2007). Attribution Attribution theories attempt to explain how people make causal explanations of behaviors and events around them and how they govern themselves in response (Crittenden, 1996; Kelley, 1973). The methods by which people make causal attributions has been used as one perspective to investigate the differences among cultures. People’s attribution styles or patterns can be classified by the dimensions of causes (e.g., locus, stability, controllability, and globality) and events (e.g., actor, domain, and valence) (Crittenden, 1996). The locus of control dimension refers to the degree to which a cause is internal or external to the actor of the behavior. The stability dimension indicates whether the cause will vary over time. Controllability refers to 6 whether the cause is subject to the actor’s choice. Globality indicates whether the cause is associated with multiple events or a certain event. The actor dimension refers to either the actor of the behavior or the observer, which is also known as self-other dimension. Examples for the domain of an event can include academic achievement, the relationship with one’s spouse, or a highway car crash. The valence of an event concerns whether the outcome is good or bad. Using the above dimensions, this study investigates the degree to which observers from America and China make internal or external attributions of two negative incidents, a campus shooting tragedy and bullying behavior at school. It is necessary to briefly review the history of attribution research in order to understand internal and external causal attribution. Psychological research on attribution started with the work of Fritz Heider (Malle, 2004). Based on Lewin’s field theory (1948/1997) that one’s behavior is a function of personal characteristics and social situation, Heider (1958) theorized that in common-sense psychology (also known as naïve psychology) the outcome of a behavior is dependent on two conditions, namely factors within the person and factors within the environment. These factors are also known as dispositional (or internal) and situational (or external) factors. Heider’s ideas did not arouse immediate interest until Jones and Davis (1965) published their acclaimed theory of correspondent inference (Malle, 2004). Jones and Davis (1965), building on Heider’s work, proposed that people tend to attribute an agent’s behavior to the agent’s disposition or personality but ignore situational factors. This was demonstrated by an experiment conducted by Jones and Harris in 1967.(1967) Participants read a short pro- or antiCastro essay and estimated the writer’s pro-Castro attitude. For writers who voluntarily chose their argument (position), participants rated the writers’ true attitudes to correspond to their essay 7 positions. For writers who were assigned an argument randomly, participants still rated writers of pro-Castro essays as having a more positive attitude towards Castro than those of anti-Castro essays. This tendency for observers to overemphasize dispositional explanations in another’s act but ignore situational explanations was later called the fundamental attribution error (FAE) by L. Ross (1977). This tendency to infer that people’s dispositions correspond to their behaviors can be demonstrated by an everyday example. People often like some movie stars because of the characters the stars played even though the actors are not the characters they play. FAE recently has been known as correspondence bias (Gilbert & Malone, 1995). Correspondence bias is less likely and is sometimes even reversed when people explain their own behavior. Loosely stated, people tend to attribute their own behaviors to situational factors. Jones and Nisbett (1972) called this effect actor-observer discrepancy, which is also known as the actor-observer effect (AOE). They proposed that there is a general tendency for actors to attribute their own actions to situational factors whereas for observers, the tendency is to attribute the same actions to actors’ stable personal dispositions. In Robins, Spranca, and Mendelsohn’s study (1996), external causes were further grouped into either general or immediate and concrete circumstances of the situation while internal causes were grouped into a stable dispositional factor and an unstable mood factor. Following dyadic interactions, most participants believed that their behaviors were influenced more by the immediate situation (the interaction partner) and their partner’s behaviors were more of an outcome of the stable dispositional factor. The actor-observer effect held true for most of the participants in the above study; however, there have also been studies indicating that this discrepancy between actors and observers can be absent and that there is a significant consistency between participants’ 8 explanations of their own and others’ behaviors. In Robins, Mendelsohn, Connell, and Kwan’s study (2004), participants were found to provide the same causal factors to explain their own behavior and the behavior of different partners. Participants who attributed their own behaviors to personality tended to attribute their partners’ behaviors to personality as well. Participants who favored situational factors for their own behaviors emphasized the same for their partners’ behaviors. This finding conflicts with AOE but has implications for the existence of attribution styles and generalized causal schemas. In the research of attribution styles or patterns, hypothetical negative events have been more frequently used than positive or neutral events as the material to trigger causal inference. The events or behaviors that have been used are diverse and include car crashes, gun-based violence, sexual victimization, domestic violence, disease, and many others. This line of research was inspired by Lerner’s Just-World Theory which stated that people need to believe in a just and orderly world where everyone gets what they deserve so that when they are confronted with injustice, they engage in a variety of causal inferences to maintain the sense of justice (Hafer & Begue, 2005; Lerner & Miller, 1978; Lerner & Simmons, 1966). Sometimes people blame innocent victims and rationalize that they deserve the unfortunate events that have happened to them in order to claim that the “injustice” does not exist and that justice has been maintained. Based on the internal-external attribution dichotomy, another attribution bias that has been identified is self-serving bias which denotes that people tend to attribute positive events to internal causes and attribute negative events to external causes. By taking credit for successes and making internal attributions, people achieve the purpose of self-enhancement. By avoiding blame for failures and making external attributions, people achieve the purpose of self-protection. This has been found by some studies and considered typical of individualist cultures, such as the 9 United States and Europe. This self-serving bias becomes group-serving bias in collectivist cultures, such as in China, Japan, and Korea, where people blame individuals and make internal attributions for individual failures in order to protect the group. One of the ways attribution theory has been applied to the study of cultures is to compare causal attributions made by people from individualist cultures and collectivist cultures. This line of research has been known as the individualism-collectivism approach. Individualism-Collectivism Approach to Attribution A group of psychologists have explained people’s cross-cultural differences in attribution with individualism-collectivism theory. In individualist cultures, people are autonomous and independent, giving priority to their personal goals; in collectivist cultures, people are interdependent and concerned with relationships, giving priority to their groups (Triandis, 2001). There is a group of scholars who emphasize individualism in Western cultures and collectivism in East Asian cultures. They utilize this approach for theoretical explanations of differences in causal attributions between the two cultures. Crystal’s study (2000) is representative of this line of research. He found that concerning peers’ deviant behaviors, American participants generally made external attributions for the purposes of individual protection or self-enhancement whereas Japanese participants usually made internal attributions to blame the individual and protect the group. In Crystal’s study (2000), the participants were fifth and eleventh graders from urban and rural areas in Tokyo and Michigan. All the participants listened to the same tape-recorded stories that each described one of four deviant behaviors of a hypothetical peer. The deviant behaviors included oppositional behavior, depression, school phobia, and aggression. After listening to the story, the students were asked three questions. The first question was whether the character had 10 done anything that seemed strange or unusual. The second question asked what was strange or unusual about the behavior or characteristic. The third question asked why the participants thought the character acted in that way. At the end of the interview, students were given a minute to describe themselves. Answers to the first three questions were analyzed for their content and classified into seven major categories. The three categories of the responses to the third question (why did the character act in a certain way) included external influences, psychological reasoning, and a “don’t know” category. External influences referred to family or friends. Examples of participant responses are “his father was too strict,” “his friends didn’t listen to him,” or “his parents didn’t give her enough attention.’ In contrast internal psychological reasoning involved the hypothetical peers’ feelings, cognitions, or emotional state. For example, “he took his anger out on the short kid” or “she felt worthless because of her mistakes”. The results show that, except for the scenario of the school phobic child, significantly more American students attributed the child’s behaviors to external influences in every scenario. In all four scenarios, more Japanese than American students explained the deviant behavior by references to the child’s psychological reasoning. Crystal discussed the cultural differences in attributions in terms of the individualism-collectivism paradigm and the self-serving bias. That is, Americans as individualists responded to the task in a self-enhancing manner to protect the self whereas the Japanese as collectivists responded in a group-enhancing manner to protect the group. Hui (2001) found a similar group-enhancing pattern in Chinese participants’ causal explanations of students’ concerns. Hong Kong students’ and teachers’ perceptions of students’ concerns and causal explanations of these concerns were investigated. Both students and teachers identified students’ ability and efforts (internal factors) as the top causal component for students’ 11 concerns. In an earlier study by Hui (2000), Hong Kong adolescent students took a survey on their concerns and the reasons/causes for their concerns. Results showed that even though students made references to both external and internal causes, they tended to give more weight to their own deficiencies (internal factors) rather than to others around them (external influences). Both students and teachers agreed upon the internal causes for student’s concerns. AOE is absent in this sample’s responses because actors and observers made similar attributions for the same events. The two studies provided evidence that the Chinese tend to make internal attributions with regard to school-related concerns. Intellectual History Approach Other research has come up with a very different analysis of the differences in causal attributions between Western and East Asian cultures. Nisbett and his colleagues propose that Western culture is more analytical and that Westerners make more internal attributions whereas East Asians are holistic thinkers and make more situational attributions. Nisbett (2003) argued that the differences in cognition between Western culture and East Asian culture can be traced back to the different philosophies and achievements of the ancient Greeks compared to the Chinese. Ancient Greek’s striking sense of freedom, individuality, and curiosity produced great philosophers, such as Aristotle, Heraclitus, Parmenides, and Plato. Due to the intellectual inheritance from ancient Greeks, Westerners today are analytic thinkers and pay attention primarily to the object, its categories, and formal logic. On the other hand, ancient Chinese philosophies, a blend of Taoism, Confucianism, and Buddhism, emphasized harmony, change, contradiction, simplicity, and holism. Therefore, East Asians are holistic thinkers, are more likely to attend to the entire field, and tend to rely on dialectical reasoning. 12 Nisbett and his colleagues have conducted a large number of comparative research studies in East Asians’ and Westerners’ perception and causal attributions. They argued that East Asians, as holistic thinkers, attend more to contextual or situational information and make more situational (or external) attributions whereas their Western counterparts (usually Americans) tend to make more dispositional (or internal) attributions (Nisbett et al., 2001). An early and representative study from this group of research is Morris and Peng (1994), motivated by several negative real-world events For example, in 1991, a Chinese physics student at the University of Iowa, Gang Lu, lost an award competition, appealed the decision unsuccessfully and subsequently failed to obtain a job. One day he entered the physics building and shot his advisor, the administrator who handled his appeal, a few bystanders, and then himself. Morris and Peng noticed remarkable differences between Chinese news reports and English news reports on this incident. In the American news reports, explanations for Gang Lu’s behavior focused on his psychological problems. In the Chinese newspaper, explanations focused on social pressure and aspects of the American context. Later in the same year, an American postal worker, Thomas McIlvane, killed his former supervisor, a few coworkers, and himself after he had lost his job, appealed unsuccessfully, and failed to find another full-time job. Morris and Peng performed another content analysis of American and Chinese newspaper reports on this incident and found results similar to those in Gang Lu’s case. They decided to design an experimental study to test their observations (Nisbett, 2003). Morris and Peng collected dispositional traits and situational factors from the news reports of both incidents and created a questionnaire. Half of the participants from each of two cultures (American versus Chinese) received the questionnaire from the campus-shooting 13 scenario and the other half received the questionnaire from the post-office shooting scenario. Each scenario preceded an open-ended question, which requested a brief explanation of the shooter’s behavior. The causal judgment task was composed of 28 hypothetical situational and dispositional causes of the murder on a 7-point rating scale (1 = “not a cause at all” and 7 = “the most important cause”). In addition, a counterfactual judgment task presented the participants with 16 hypothetical scenarios, each different from the actual scenario by the change of one causal factor (x). Participants were asked to decide whether the murder would have happened if only x had been different, where x stands for either a personal disposition factor or a situational factor in the actual case. The open-ended question served as a priming tool to trigger the participants’ spontaneous thoughts about the murder and these responses were not included in the analysis. The responses to the causal judgment task showed that personal dispositions were given greater weight overall by American participants while situational factors were given greater weight overall by Chinese participants. American participants particularly emphasized chronic psychological problems related to work and Chinese participants emphasized corruption and disruption of the society. American participants also committed the “ultimate attribution error” (or group-serving bias), which means they were more critical of the out-group (the Chinese graduate student Lu in this case). In addition, they gave more weight to personal dispositions and less weight to situational factors for the murder. Finally, the results of the counterfactual questions were inconclusive because Chinese participants believed the murder to be less likely than did the American participants when either dispositional or situational factors were altered. In addition to conducting verbal causality studies, intellectual history psychologists have conducted a considerable amount of cross-cultural research in perceptual and visual causality. 14 The assumption was that the social-economic environment of the two cultures described above, Greece and China, historically determined the different ways that we perceived the world and therefore the different ways we explained the world. Conversely, if we explain the world differently, we must see different worlds as well. Ji, Nisbett, and Peng’s art clips and rod-andframe study (2000) showed that Chinese participants reported stronger associations between events and were more field-dependent than their American counterparts. The researchers believed that East Asians attend to the field as a whole and that their cognition is relatively holistic whereas Westerners focus on objects and their cognition is more control-oriented. They proposed that East Asians believe that things are constantly changing, which leads them to “a fluid and open-minded view of people and events” resulting in situational attributions whereas “belief in no change may lead Americans to dispositionism that personal traits are fixed” (Ji, Nisbett, & Su, 2001). Ng and Zhu (2001) examined causality attribution and memory of events in situations involving both individual and group actions with Chinese and American participants. The Chinese participants, being more collectivist, showed greater preference for external attributions whereas the more individualistic participants, the Americans, showed a preference for internal attributions in situations involving individual actions.. The categorization of collectivists and individualists was measured by a self-construal questionnaire. Ng and Zhu argued that collectivism along with Confucianism and Taoism make the Chinese willing attributors but reluctant reductionists. The Chinese pay attention to both the actor and the target of an action, as well as the context in which the action occurs. In order to further explore the underlying mechanisms for the cultural differences in cognitive processing styles, Chua, Boland, and Nisbett (2005) tracked and measured American 15 and Chinese participants’ eye movements while they viewed photographs involving a focal object in a complex background. The researchers found that the Americans looked at the focal objects more quickly and fixated on them longer than did their Chinese counterparts. The researchers claimed that differences in attribution style and event memory may originate from differences in what people pay attention to. The intellectual history school of thought thus proposes that Americans make more dispositional attributions about an event whereas East Asians make more situational attributions. These differences are due to the influences of their intellectual inheritance and cultural backgrounds. The researchers of this school have provided empirical evidence to support the claim that people from Western and East Asian cultures make different causal attributions. They have proposed an interesting and provocative theory that links the cognitive differences to the intellectual history influences, although these connections can be hard to demonstrate. The Third Group: Individualism-Collectivism or Intellectual History? It is apparent that the individualism-collectivism and intellectual history theorists and researchers differ in their positions about the differences in causal attributions made by Westerners and East Asians. Nisbett and his colleagues believe that East Asians make situational and external attributions whereas Westerners make dispositional and internal attributions because each group inherits different thoughts and beliefs from our ancestors through education. Researchers such as Crystal and Hui argue that East Asians make internal attributions whereas Westerners make external attributions because collectivist and individualist cultures have different impacts on their members. To make the situation more confusing, there are other studies that adopted individualism-collectivism theories but that produced results consistent with those found by intellectual history researchers, as described below. 16 In a study conducted by Ho (2004), Chinese and Australian teachers were asked to rate the importance of causal factors for students’ problematic behaviors and the teachers differed significantly on the attribution of those behaviors. The Australian teachers emphasized students’ abilities as the main causal factor whereas the Chinese emphasized family background. Ho argued that in individualist cultures, individuals are held responsible for their behaviors; therefore, Australian teachers attributed students’ problematic behaviors to internal causes. In collectivist cultures, family factors were emphasized, thus helping to explain Chinese teachers’ emphasis on family factors. In a study conducted by Au, Hui, and Leung (2001), Chinese and Canadian participants were presented with a scenario in which a customer’s coat was stained with tea in a hotel and the hotel manager handled the customer’s complaint. The researchers found that Chinese participants attributed more responsibility to the hotel than did Canadian participants and that both groups were similar in the level of self-blame. The researchers argued that collectivists were more sensitive to in-group and out-group differences, which individualists are less sensitive to. Therefore, collectivists blamed the out-group service provider, the hotel, more than did the individualists. In Cheah and Rubin’s study (2004), American and Chinese (mainland) mothers were asked to make causal attributions on hypothetical vignettes of an aggressive child and a socially withdrawn child. The results revealed that Chinese mothers, compared to their counterparts, were more likely to attribute social withdrawal to external causes and less likely to believe that social withdrawal and aggression were stable or intentional. The authors believed that the Chinese mothers’ endorsement of external causal attributions is consistent with Confucian ideology and collectivist cultures that value unobtrusive and compliant behaviors in order to maintain social 17 harmony. It was thought that American mothers endorsed internal attributions because individualist cultures value assertive and independent behaviors. This third set of research shares the same theoretical classification of individualist collectivist cultures, classifying America as an individualist culture and China as a collectivist culture. However, the researchers produced different results from Crystal and Hui, who believed that Americans tend to make external attributions to blame others and protect the self whereas the Chinese tend to make internal attributions to blame individuals to protect the group. This third group believes that Americans favor internal attributions whereas the Chinese favor external attributions, similar to what Nisbett and his colleagues proposed. Study Rationale and Purpose Three groups of causal attribution studies have been presented with incompatible results. The individualism-collectivism studies show that Westerners, as individualists, tend to make external attributions to protect the self whereas East Asians, as collectivists, tend to make internal attributions to protect the group. On the other hand, the intellectual history group and results from several individualism-collectivism studies find that Westerners tend to make internal or dispositional attributions whereas East Asians tend to make external or situational attributions. The results of those studies are comparable because their participants were usually recruited from the same regions—North America and East Asia. In the individualism-collectivism studies, the individualist culture mainly refers to the U.S. and Canada, and the collectivist culture refers to China and Japan. In intellectual history studies, East Asian culture refers to China, Japan, and Korea, and Western culture mainly refers to the United States. The present study is an effort to explain why there are conflicting results from previous studies. On the one hand, the intellectual history group and some individualism-collectivism 18 scholars provided evidence that East Asian subjects were inclined to make stronger and more external attributions whereas American subjects were more likely to make internal attributions. On the other hand, another group of individualism-collectivism scholars came up with empirical evidence showing that East Asian subjects made more and stronger internal attributions while American subjects made more and stronger external attributions. Thus, the current study aims to address why there are conflicting results among these attribution and culture studies and how American and Chinese differ in their causal attributions. One of the differences among the previous research studies is that they tended to use different methods of soliciting responses. Most of the cultural studies of causal attribution requested participants to take paper-pencil questionnaires that usually included rating-scale questions, such as Morris and Peng’s study (1994) of the graduate student campus shooting. Some studies were carried out through in-person interviews where participants were requested to freely offer causal attributions, such as in Crystal’s study (2000). Methods of soliciting responses can have a large impact on participants’ responses. For example, in response to open-ended questions, participants are very likely to provide spontaneous causal explanations. Given multiple-choice or rating-scale questions, participants are more likely to provide responses constrained by the structure of the questionnaire. Thus, the researchers can impose their prior beliefs on the participants. To avoid this, the current study requested participants to offer spontaneous open-ended attributions first and then to respond to rating-scale survey questions. The rating-scale items were collected from pilot studies to minimize any researcher bias. Another difference among the previous studies is that they have used different scenarios, which may lead to different causal attributions; therefore, in this study, both scenarios were revised and used in an attempt to replicate the results of the two existing studies. 19 Research Questions and Hypotheses As mentioned in the above section, there are several possible explanations for the conflicting results in the above studies. First, different scenarios have been employed in previous studies. Second, different measurement methods can affect the outcome. Open-ended questions and rating-scale questions for the same material may solicit different answers from the same participant. Third, as noted by Markus and Kitayama (1991) a persistent issue in cultural and psychology studies involves the depth and pervasiveness of the cultural differences. The interpretation of cultural differences in causal attribution may contribute to the incompatibility among existing studies. Therefore, to test the validity of these explanations, three questions with corresponding hypotheses were developed for this study. Is Causal Attribution Content-Specific? The first question concerns whether causal attributions are content-specific. That is, does the specific scenario used in the survey lead participants to make certain type of attributions, either internal (dispositional) or external (situational)? People make causal attributions (either situational or dispositional) that correspond to the nature of the event or behavior. Studies show that valence of the behavior (Ybarra, 2002), attractiveness of the character (Forsterling, Preikschas, & Agthe, 2007), attributor’s cognitive load (Lieberman, Jarcho, & Obayashi, 2005), and many other factors have an impact on people’s causal attributions. It may be that there are not general differences in causal attributions between the American and Chinese groups but that there are culture-specific differences in their reactions to certain scenarios. For instance, in the shooting scenario, the murder weapon was a gun which is more accessible in the U.S. than in China. It might be that the cultural differences in causal attributions were due to the different reactions to guns by the two cultures. It is possible that if the weapon were a knife, there might 20 be a different pattern of causal attributions made by the two groups. Rather than there being general cultural differences, no matter what scenarios being used, there may well be cultural differences in the reactions to a specific scenario. Thus, the first hypothesis is that causal attributions are content-specific. It is predicted that the nature of the scenario is likely to interact with the culture. In addition, this hypothesis makes the common-sense assumption that people tend to hold adults responsible for their misdeeds and allow children excuses for their misbehaviors. Therefore, two predictions are made for the first hypothesis: I-1: For the aggressive schoolboy scenario, both groups will make stronger external attributions than internal attributions. I-2: For the graduate student’s murder scenario, both groups will make stronger internal attributions than external attributions. Is Causal Attribution Question-Format-Specific? The second research question asks whether causal attributions are question-formatspecific. That is, does question format (open-ended questions versus rating-scale questions) in the survey have an impact on participants’ responses? The rating-scale questionnaire is the conventional measure adopted by most attribution studies (Forsterling et al., 2007; Kenworthy & Miller, 2002; Kozak, Marsh, & Wegner, 2006; Schlenker, Weigold, & Hallam, 1990). This type of measure has the disadvantage that participants lack the option of not responding to those items that are unrelated to their prior thinking. However, open-ended questions provide the opportunity to elicit participants’ idiosyncratic responses (Eagly, Mladinic, & Otto, 1994). Therefore, the current study included both rating-scale and open-ended questions in the online survey in order 21 to compare the responses to each question format and find out if the two question formats will elicit different attribution patterns. Morris and Peng (1994) used open-ended questions to prime subjects prior to their responding to the rating-scale questions; however, the researchers discarded the open-ended responses which might provide richer information about the participants’ causal thinking patterns. Crystal (2000) used one-on-one interviews with his subjects. It is possible that people respond differently to questions presented in different formats; therefore, the incompatible findings from previous studies could result from the differences in instrumentation used. Rating-scale questions force participants to make a choice and are likely to provide participants with extra information that they would not have otherwise thought about. In these cases, participants will react to the items by retrieving previous attitudes toward the scenario (Eagly et al., 1994). However, openended questions allow participants spontaneous responses, which may be similar to responses that participants would give in a real-life situation. Therefore, as stated above, the second hypothesis is that causal attributions are question-format-specific; furthermore, the two predictions for the second hypothesis are: II-1: Participants from both cultural groups will endorse both internal and external causes in their responses to the open-ended questions so that, overall, neither of these causes will be favored. II-2: Participants from both cultural groups will favor one type of cause, either internal or external, in their responses to the rating-scale questions. Is Causal Attribution Culture-Specific? The purpose of the study is to find out if the two groups (American versus Chinese) will show different attribution patterns—either internal or external. The first two questions are potential explanations of the incompatible findings in the existing literature, and the first two 22 hypotheses are designed to detect sources rather than culture for the different attribution patterns between the American and Chinese groups. In contrast, the third research question asks whether causal attributions are culture-specific. This study adapted two scenarios from previous research—the bullying scenario from Crystal’s study (2000) and the campus shooting scenario from Morris and Peng’s study (1994). As stated above, in Crystal’s study, American participants were found to make external attributions, and East Asian participants were found to make internal attributions. In Morris and Peng’s study, Americans favored internal attributions and East Asians favored external attributions. Because there is considerable evidence to support both groups’ claims, it is hard to assume that one of these results is not valid. Therefore, it is possible that something else, such as the scenario or the question format, is causing the difference in the attribution patterns. It is also possible that American and Chinese perceive the causes differently from the internal-versus-external dimension. Again, the third hypothesis is that causal attributions are culture-specific. It is predicted that although there are differences in causal attributions between the two cultures, they will not appear in the form of opposite patterns as claimed in the existing literature. Culture affects causal attributions via interaction with the scenario and the question format. In other words, the main effect of culture may not be significant but the interaction effects will be. 23 II. METHOD Participants There were 209 Chinese participants and 138 American participants who accessed the online survey, Cognition and Culture. The Chinese participants were recruited from a university in a major city in China, and the American participants were from a Midwestern university in the U.S. The researcher emailed her colleagues in the U.S. and China to ask for their assistance in spreading the word about the study among their students. Potential participants (the students) were informed of the study by their teachers either via recruitment email or a flyer. Please refer to Appendix A for both the English and Chinese versions of the recruitment letter. In the letter, students were provided the URL to access the online survey. Please refer to Appendix B for the English version of the survey English and Appendix C for the Chinese version. Throughout the study, participants had no direct contact with anyone at the research site. Data collection lasted for one month. Participants were removed from data analysis based on three data-cleaning criteria. First, participants (N =46) who did not provide any response to the open-ended question were removed. Second, participants (N = 65) whose rating-scale responses were incomplete were removed. Third, international participants (N = 3) who stayed in the U.S. or China for academic purposes were removed because their opinions may not have been representative of the local people’s beliefs. After data cleaning, responses were saved from 76 American participants and 157 Chinese participants and used in the final analysis. 24 Demographic Results Participants were all well-educated, computer literate, and fluent in their native language (Chinese or English). American participants had an average age of 27 and were recruited from both graduate and undergraduate programs. Chinese participants had an average age of 21 and were mainly from undergraduate programs. American participants were somewhat older and had more years of education compared to their Chinese counterparts. The results are summarized in Table 1. Table 1 Demographics of Participants Culture N Mean Standard Minimum Maximum Deviation Age Education U.S. 76 27.05 10.81 18 61 China 157 21.48 1.88 17 29 U.S. 76 2.86 1.32 1 5 China 157 2.10 .45 1 5 Note: Education level 1 = High School Graduate, 2 = Some College, 3 = Completed College, 4 = Some Graduate School, 5 = Completed Graduate School. Materials Demographic questions Before participants took the questionnaire, they were asked four demographic questions—nationality, residency, age, and education level. Gender of the participants was not in the survey because existing research findings indicate that gender has little impact on the participants’ causal attributions of other’s behaviors (Crystal, 2000; Song, Smetana, & Kim, 1987). 25 Scenarios The scenarios of the survey were adopted from those used in the studies of Crystal (2000) and Morris and Peng (1994). In Crystal’s study, four scenarios of deviant behaviors— oppositional, depressed, school phobic, and aggressive—were presented to the participants. The aggressive scenario was about a schoolboy bullying his classmates and was used in the current study. Morris and Peng used two public shooting stories. The current study used the campusshooting scenario—an angry graduate student who committed murder and then suicide on campus. The two stories were rewritten to be applicable to both Chinese and American cultures, and the two main characters were each given a typical Chinese or American name. The bullying boy was named Dave in the English version and De in the Chinese version. The graduate student was named either Michael (English version) or Ming (Chinese version). Participants were expected to assume that the characters were from their own cultural backgrounds. Such approach was designed to avoid group-serving bias which was shown to exist among American participants in the literature (Morris & Peng, 1994). The scenarios and questions were written in English by the researcher and then translated into Chinese by a bilingual translator. To ensure the accuracy of the Chinese translation, it was back-translated into English by another bilingual translator and compared against the original English version. The English version of the schoolboy scenario used in the survey is as follows. Dave is an only child. His parents are doctors so they worked very hard and aren’t home very much. But they try to give him as much attention as possible. Dave doesn’t have much patience if things don’t go his way. One day, Dave’s teacher mentioned Dave’s low grade on a recent test in front of the class. After class, Dave was walking in the school 26 hallway when someone accidentally bumped into him. He grabbed the other person, pushed him up against the wall, and demanded he apologize. At lunch, a particularly short kid who was in Dave’s class stood in front of him on the food line. Dave began to tease him and call him names. That afternoon, during a chemistry lesson, Dave tripped this same kid as he was walking down the aisle carrying a tray of glass beakers. The English version of the graduate student scenario is: Michael was a graduate student and had been studying in a major university for five years. He had recently lost an award competition which was given to his advisor’s other graduate student. Michael suspected that his advisor had favored the other student and the other fellow student had cheated. He unsuccessfully appealed it. Later on he failed to get an academic job which he believed was due to his advisor's unenthusiastic reference letter. One day he entered the faculty meeting in his department and shot his advisor, the fellow student, and another two professors. He then proceeded two blocks to the administration building where he shot the associate vice president of academic affairs who was handling his case and a student helper. He then committed suicide by shooting himself in the head. Open-ended Questionnaire The open-ended questionnaire consisted of three parts: the scenario, the open-ended question, and five text boxes. The open-ended question asked participants to speculate beyond the limited information given in the short scenario and to provide as many causes for the event as they could think of. Below the instructions, there were five text boxes for participants to type in their speculated causes of the incident. Results in the pilot study proved that participants rarely gave more than five causes. The open-ended results were coded with a rubric that was created 27 after the data had been collected. More details regarding the coding rubric and the inter-rater reliability will be discussed in the next chapter. Rating-Scale Questionnaire The rating-scale items were developed in a series of two pilot studies. In the first pilot study, participants were recruited from the researcher’s social network in the U.S. and China. They were presented with the same two scenarios in their mother tongue, either English or Chinese, and asked to provide causal attributions freely for both scenarios and to comment on how to improve the wording of the two scenarios. The collected causal statements were summarized and rewritten to create the first draft of the rating-scale items, resulting in six internal attribution statements and seven external attribution statements for each scenario. This first draft was used in the second pilot study with another group of graduate students and some faculty members in a Midwestern university in the United States. There were both Americans and Chinese in this group of participants. They were asked to complete the ratingscale questionnaire, provide feedback on the rating-scale items, and offer any other causes attributions that they could think of. They were also asked to comment on the instructions of the questionnaire. After the participants turned in their responses, they were interviewed on the telephone by the researcher for further discussion of the instrument. Each item in the rating-scale questionnaire was either an internal or external attribution. One of the purposes for this second pilot study was to find out if participants interpreted the rating-scale items the same way as the researcher had intended, and this was confirmed. In the final version of the questionnaire, there were thirteen questions for each scenario. Each question consisted of two parts—one hypothetical causal statement for the incident in the scenario and a five-point scale (from “very important” = 5 to “not a factor” = 1) placed in front 28 of the causal statement. By selecting one of the five ratings, each participant indicated how important the cause was to the incident. Six internal causal statements and seven external causal statements were randomly ordered in the questionnaire. In the analysis each participant’s ratings of internal items were averaged separately from ratings of the external items. The difference between the two means indicated whether the participant favored internal or external explanations of the incident. Procedure In the recruitment email, the participants were provided with the URL of the online questionnaire. They accessed the questionnaire at a time and location convenient to themselves and had no contact with anyone at the research site. The online survey has one English version for the American participants and one mandarin Chinese version for the Chinese participants. In the recruitment email for American participants, the URL to access the English version was provided; in the recruitment email for Chinese group, the URL to access the Chinese version was provided. The online questionnaire included seven web pages. Page 1 consisted of the informed consent form, which told participants about the study in sufficient detail to allow them to make an informed decision about whether to participate. At the bottom of Page 1, there were two buttons to click on, “YES” and “NO.” Participants were informed that they had the option to click on “NO” to decline the survey. If they clicked on “YES” and agreed to participate, they would be led to Page 2, which consisted of the demographic questions. Since priming has been shown to have an impact on attribution outcomes (Morris & Peng, 1994), in the current study the open-ended questions for both scenarios were given before the rating-scale questions to ensure that the rating-scale items would not affect the answers of the 29 open-ended questions. Page 3 presented the first scenario followed by the open-ended question with five text boxes for answers. Page 4 presented the second scenario followed with the openended question and the text boxes. Page 5 presented the first scenario again followed with thirteen five-point rating-scale questions. Page 6 presented the second scenario, again followed with thirteen five-point rating-scale questions. The last page, Page 7, was to thank participants for their input. In order to control for order effects, the two scenarios were presented to the participants using one of two sequences (X or Y). In other words, from Page 3 to Page 6, the scenarios were presented as either Bullying-Campus-shooting-Bullying-Campus-shooting (sequence X) or Campus-shooting-Bullying-Campus-shooting-Bullying (Sequence Y). Half of the participants read the scenarios in the first order, and the other half read them in the second order. Once a participant accessed the website, he or she was assigned an ID number by the computer program. When the ID was an odd number, the participant received sequence X; when the ID was an even number, the participant received sequence Y. The order of the seven pages and the two sequences of presentation of the scenarios can be viewed in Figure 1. 30 Sequence X: Sequence Y: Page1. Welcome Page Page1. Welcome Page Page2. Demographic Questions Page2. Demographic Questions Page3. Dave/De Open-ended Questions Page3. Michael/Ming Open-ended Questions Page4. Michael/Ming Open-ended Questions Page4. Dave/De Open-ended Questions Page5. Dave/De Rating-scaleQuestions Page5. Michael/Ming Rating-scaleQuestions Page6. Michael/Ming Rating-scaleQuestions Page6. Dave/De Rating-scaleQuestions Page7. Thank-you Page7. Thank-you Figure 1 Two Sequences of Online Questionnaire Pages The website was made available in March 2007, and participants started to take the questionnaire in early March. On April 16, 2007, at Virginia Polytechnic Institute and State University (Virginia Tech), a student killed 33 people, wounded many others, and committed suicide (Hauser & O'Connor, 2007). This incident was similar to the Campus-shooting campusshooting scenario in the questionnaire. Because participants’ input may have been influenced by the publicity of the Virginia Tech incident, the website was immediately closed, and data collection was stopped. 31 Design To test for the scenario-specific hypothesis and culture-specific hypothesis, a series of 2×2 ANOVAs were conducted for the study. For the scenario-specific hypothesis, culture (American versus Chinese) was used as the between-subjects variable and the scenario (Bullying scenario versus Campus-shooting scenario) was used as the within-subjects variable. Since the rating-scale responses and open-ended responses constituted two different measures, one repeated-measured ANOVA was conducted for each. For the culture-specific hypothesis, the following analysis was run. For each scenario, a 2×2 ANOVA was conducted to test for a between-subjects cultural effect and a within-subjects type of cause (internal versus external, as determined by the use of factor analysis) effect on participants’ attributions, for the rating-scale and open-ended question formats examined separately. In addition, for each scenario, a MANOVA was conducted to test for the effect of culture on participants’ responses to the 13 rating-scale items, and another was conducted on participants’ responses to the categories of the open-ended coding rubric, which will be introduced in the next chapter. 32 III. RESULTS Data Preparation Coding of Open-ended Responses For the open-ended questionnaire of the Bullying scenario, 76 American participants offered 275 responses (in the form of causes), M = 3.79, SD = 1.42, and 157 Chinese participants provided 459 responses, M = 2.8, SD = 1.02. For the Campus-shooting scenario, there were 287 total responses from the American group, M = 3.6, SD = 1.56, and the total responses from the Chinese group numbered 457, M = 2.92, SD = .99. The first step in processing these open-ended responses was to create a coding rubric. All the open-ended responses were compiled into one file and were sorted into categories. The categories were used to create the coding rubrics to analyze the open-ended responses. For the campus shooting scenario, there were five categories for internal causes and five categories for external causes. For the bullying scenario rubric, there were four categories for internal causes and four categories for external causes. The coding rubric spelled out the content for each code number mainly in English. Some Chinese expressions or idioms were used, however, since they are absent in English or require extensive translation. In this case, the Chinese phrases remained in the rubric because the raters were Chinese who would have no problems understanding the rubric. Please refer to Appendix E for the coding rubrics of both scenarios. Though given five text boxes to fill in causes, some participants wrote a long paragraph in one or two boxes. In those cases, the researcher broke down the long paragraph into individual causes and re-numbered them. The researcher coded both American and Chinese participants’ open-ended responses with the coding rubrics. An English-speaking rater coded American participants’ open-ended responses and a Chinese-speaker fluent in English coded Chinese 33 participants’ open-ended responses. Both assistant raters were graduate students from the researcher’s institution. The researcher introduced them to the online survey and trained them to use the coding rubrics to code the open-ended responses. The inter-coder reliability was calculated on 100% of the sample, and Cohen’s Kappa indicated good inter-coder reliability. The average Kappa for the Dave scenario was .91, .87 for the De scenario, .88 for the Michael scenario, and .89 for the Ming scenario. The inter-coder agreement is summarized in Table 2. Table 2 Inter-coder Reliability for Open-Ended Responses Scenario Total Responses Agreement in Coding Agreement in Coding across for Coding across 8 or 10 Categories Internal-External Categories Dave (Am) 275 250 (90.91%) 269 (97.82%) De (Ch) 459 413 (89.98%) 449 (97.82%) Michael (Am) 287 255 (88.85%) 273 (95.12%) Ming (Ch) 457 415 (90.81%) 442 (96.72%) Note: Am=American, Ch=Chinese The differently coded English responses were recoded by another English-speaking research assistant, and the differently coded Chinese responses were recoded by another Chinese-speaking assistant fluent in English. This process determined the final coding for those responses which the researchers had formerly disagreed upon. Factor Analysis of Rating-Scale Items Factor analyses were conducted for the rating-scale responses of each cultural group in an attempt to confirm the internal and external components among the thirteen items. However, American responses and Chinese responses yielded different factors of both scenarios, indicating that the two cultural groups may perceive the rating-scale items differently. If factor analysis had confirmed the internal and external components, subsequent analyses should have been done on the factor scores. However, since the expected factors were not found, the subsequent analyses 34 necessarily assumed the face validity of the internal and external components. Rotated component matrices for each group of the same scenario are summarized into one table for comparison. Please view the results in Table 3 and 4. Table 3 Factor Analysis Summary of Dave/De (Bullying) Rating-Scale Items Rating-scale items of Dave/De (Bullying) U.S. Scenario Components Internal Causes 1 2 China Components 3 1 2 D1 He is self-centered and spoiled. .61 .47 D2 He was acting aggressive to gain attention .69 .35 from peers. D4 He was acting aggressive so that his peers .52 .50 .59 would not bully him. D5 He did not like school. .83 .57 D6 He did not like his schoolmates and teachers. .74 .71 D7 He did not like the classmate he bullied. .56 .60 .57 .44 .41 External Causes D3 His teacher mentioned his low grade in front of the whole class. D8 Movies, television, and other mass media .51 .43 glorify this type of behavior. D9 Most of his peers don't like him and make fun .53 .39 .68 of him. D10 He is friend with peers who behave like him. .72 .64 D11 Peers who had been bullied by him did not .78 .44 .47 report him. D12 Teachers and school did not attend to his bully .73 .68 behavior effectively and immediately. D13 His parents spoiled him but did not discipline him enough. 35 .38 .68 .74 Table 4 Factor Analysis Summary of Michael/Ming (campus-shooting) Rating-Scale Items Rating-scale items of Michael/Ming U.S. China (campus-shooting) Scenario Components Internal Causes M1 1 2 3 He is stressed out by school and 4 Components 5 1 2 .80 .38 .39 .59 .64 3 unpromising job hunting. M2 He did not know other ways to seek help .56 to release his frustration and anger. M4 He has some personality problems. M5 He was too obsessed with the award. M6 He was mentally instable and ill. M7 He had unrealistic fantasy of social .72 .56 .83 .61 .72 .77 .82 .36 .38 justice. External Causes M3 The vice president did not deal with the .82 .80 case fair and square. M8 Movies, television, and other mass media .63 .34 .42 glorify this type of behavior. M9 People around are individualistic and .81 .58 selfish. M10 His friends like to play with guns and .88 .67 .86 .67 solve problems by force. M11 The availability of gun and shooting training where he was living. M12 His advisor and fellow students were .70 -.40 .34 -.40 .80 cheating. M13 There is no third-party entity for appealing. 36 .59 .60 Derivation of Dependent Measures from Open-ended Responses Coded responses were summarized for each participant to produce the following dependent measures derived from the factor analyses (see Table 5). For each scenario, every participant received one external attribution score (OE-Ext), which was the sum of his or her responses categorized as external causes and one internal attribution score (OE-Int), which was the sum of his or her responses categorized as internal causes. Finally, an attribution proportion score (OE-Pref) for each participant was obtained by dividing the internal attribution score by the total attribution score. Proportion scores smaller than .5 indicate that the participant favors external attributions. Proportion scores larger than .5 indicate a favoring of internal attributions. Derivation of Dependent Measures from Rating-scale Responses There were five points on the rating-scale with two anchor points — 5 as “very important” and 1 as “not a factor.” For each scenario, every participant’s ratings of six internal causes were averaged to obtain one internal rating score (RS-Int) and his or her ratings of seven external causes were averaged to obtain one external rating score (RS-Ext). The final attribution proportion score (RS-Pref) for each participant was obtained by dividing the internal attribution score by the total attribution score. Proportion scores smaller than .5 indicate a favoring of external attributions. Proportion scores larger than .5 indicate a favoring of internal attributions. All of the above dependent measures are summarized in Table 5. Cronbach coefficient of the thirteen items of the bullying scenario is .742 and that of the campus-shooting scenario is .715, which means the internal reliabilities across the thirteen items of both scenarios are acceptable. 37 Table 5 Dependent Measures Variables for Each Scenario Question Dave-De Description of Each Variable Michael-Ming Format Open- OE-Ext-D OE-Ext-M Sum of causes coded as external causes ended OE-Int-D OE-Int-M Sum of causes coded as internal causes Questions OE-Pref-D OE-Pref-M Proportion of internal responses/total responses Rating- RS-Ext-D RS-Ext-M Mean of ratings of 7 external items scale RS-Int-D RS-Int-M Mean of ratings of 6 internal items RS-Pref-M Proportion of internal score/total score Questions RS-Pref-D Distribution and Transformation Tests for normality were conducted on the variables in Table 5, and results showed that the open-ended measures (OE-Pref-D and OE-Pref-M) were not distributed normally. Therefore, three transformations were conducted on the responses: a square-root transformation, a logarithmic transformation, and a reciprocal transformation. However, none of the transformations yielded a more normal distribution than did the original data. In addition, ANOVAs conducted on the three sets of transformed data resulted in the same findings as those from the original data. Therefore, results from the original data are presented in the following discussion of the data analysis. Data Analysis Data analysis results are presented by type of dependent variable—attribution preference scores, internal/external scores, and itemized scores. 38 Attribution Preference In order to test the hypotheses that causal attributions are scenario-specific and culturespecific, two repeated-measure ANOVAs were conducted to evaluate the scenario effect and cultural effect on participants’ causal attributions. Separate ANOVAs were conducted for the open-ended responses and the rating-scale responses. Scenario type at two levels (Bullying or Campus-shooting) was used as the within-subjects factor and culture at two levels (the Americans or Chinese) was used as the between-subjects factor. All statistical tests used an alpha level of .05. Open-ended Responses A repeated-measure ANOVA as described above were conducted to evaluate the openended questionnaire attributions (OE-Pref). A significant main effect was found for scenario, F (1, 231) = 51.67, p < .00, η2 = .18. No significant cultural effect was found, F (1, 231) = .9, p < .35, η2 = .00. A significant interaction effect was found between the two factors, F (1, 231) = 20.69, p < .00, η2 = .08. Follow-up t-tests were conducted to aid in the interpretation of the interaction effect. Independent t-tests were conducted to compare the two cultural groups’ causal responses for each scenario. A significant difference between the two groups’ responses was found for the bullying scenario, for which the American group provided more internal causes (M = .63, SD = .26) than did the Chinese group (M = .50, SD =.25), t (231) = 3.64, p < .00. There was no significant difference between the two groups’ responses of the campus-shooting scenario, t (231) = -1.93, p < .06. Repeated measure t-tests were conducted to compare the scenarios in terms of each group’s responses to them. Only the Chinese participants showed a significant difference in 39 their responses to the two scenarios. They provided more internal causes for the Ming scenario (M = .77, SD = .26) than for the De scenario (M = .50, SD = .25), t (156) =10.09, p < .00. In summary, both groups favored internal attributions for both scenarios; however, the Chinese group provided stronger internal attributions for the campus-shooting scenario than for the bullying school boy scenario. These results can be viewed in Figure 2. Preference for Internal Causes (OE-Pref) 1 USA 0.5 China 0 Dave/De Michael/Ming Scenarios Figure 2 Scenario and Cultural Effects on Causal Attributions of Open-ended Questions Rating-Scale Responses A repeated-measure ANOVA was conducted to evaluate the scenario and cultural effects on participants’ rating-scale attributions (RS-Pref). A significant main effect was found for the within-subjects scenario effect, F (1, 231) = 527.35, p < .00, η2 = .70. No significant cultural effect was found, F (1, 231) = .29, p < .59, η2 = .00. A significant interaction effect was found between the two factors, F (1, 231) = 15.39, p < .00, η2 = .06. To aid in the interpretation of the interaction effect, independent t-tests were conducted to compare the two cultural groups’ causal statement ratings of each scenario. For the bullying scenario, both groups’ internal preference scores are below .5, indicating the preference for external causes. However, the American group (M = .47, SD = .06) gave more weight to these 40 causes than did the Chinese group (M = .45, SD = .07), t (231) = 2.48, p < .01. For the campusshooting scenario, although both groups considered internal causes more important than external causes, as predicted, the Chinese group (M = .63, SD = .09) gave more weight to these causes than did the American group (M = .60, SD = .09), t (231) = -2.54, p < .01. Repeated measure t-tests were conducted to compare each cultural group’s ratings of both scenarios. The American group favored external explanation of the Dave scenario (M = .47, SD = .06) and internal explanation of the Michael scenario (M = .60, SD = .09), t (75) = -11.18, p < .00. The Chinese group also favored external attributions of the De scenario (M = .45, SD = .07) and internal attributions of the Ming scenario (M = .63, SD = .09), t (156) = -23.98, p < .00. In summary, both groups made external attributions for the bullying scenario and internal attributions for the campus-shooting scenario; however, for the bullying scenario, the Chinese group made stronger external attributions than did the American group. For the campus-shooting scenario, the Chinese group made stronger internal attributions than did the American group. Preference for Internal Causes (RS-Pref) These results can be viewed in Figure 3. 1 USA 0.5 China 0 Dave/De Michael/Ming Scenarios Figure 3 Scenario and Cultural Effects on Causal Attributions of Rating-scale Questions 41 Internal Versus External Attributions Morris and Peng (1994) compared participants’ ratings of external (situational) and internal (dispositional) items and found in the campus-shooting scenario that American participants had given internal causes greater weight than had Chinese participants whereas Chinese participants had given greater weight to external causes than had American participants. Therefore, in the current study, similar analyses were conducted to find out if the results would be replicated. A series of four repeated-measure ANOVAs were conducted to evaluate the cultural effect (Americans versus Chinese) and the type-of-cause effect (internal versus external) on participants’ attributions of the two scenarios for the rating-scale and open-ended responses. The alpha level used was .05. Open-ended Responses to the Dave/De (Bullying) Scenario A repeated-measures ANOVA was conducted to evaluate the cultural effect and type- ofcause effect on participants’ open-ended attributions to the Bullying scenario. A significant cultural effect was found, F (1, 231) = 35.33, p < .00, η2 = .13. A significant main effect was also found for the within-subjects factor, the type of cause, F (1, 231) = 27.63, p < .00, η2 = .11. Last, a significant interaction effect was found between the two factors, F (1, 231) = 20.87, p < .00, η2 = .09. Independent t-tests were conducted to aid in the interpretation of the interaction effect by comparing the two cultural groups’ open-ended responses of the Bullying scenario. For the internal categories, American participants (M = 2.39, SD = 1.2) provided significantly more causes than did the Chinese participants (M = 1.45, SD = .89), t (231) = 6.76, p < .00. Repeated ttests were conducted for each cultural group’s open-ended internal and external responses. The American group provided more internal causes (M = 2.39, SD = 1.2) than external causes (M = 42 1.39, SD = .97), t (75) = 5.29, p < .00. In summary, the American group gave many more internal causes than external causes; they also provided many more internal causes than did the Chinese group. These results can be viewed in Figure 4. Causal Category Endorsement 3 2 USA China 1 0 Dave/De Internal Dave/De External Figure 4 Cultural Effect and Type-of-Cause Effect in Open-ended Attributions of the Dave/De (Bullying) Scenario Open-ended Responses to the Michael/Ming (campus-shooting) Scenario A repeated-measures ANOVA was conducted to evaluate the cultural type-of-cause effect on participants’ open-ended responses to the Campus-shooting scenario. A significant cultural effect was found, F (1, 231) = 16.12, p < .00, η2 = .07. The American group provided more internal causes (M = 2.41, SD = 1.26) than did the Chinese group (M = 2.23, SD = 1.01); the Americans also gave more external causes (M = 1.18, SD = 1.10) than did the Chinese (M = .69, SD = .78). Also, a significant within-subjects effect was found for the type of cause, F (1, 231) = 151.71, p < .00, η2 = .4. Overall, for the Campus-shooting scenario, participants provided more internal causes (M = 2.29, SD = 1.1) than external causes (M = .85, 43 SD = .93). However, there was no significant interaction effect between the two factors, F (1, 231) = 2.00, p < .16, η2 = .01. These results can be viewed in Figure 5. Causal Category Endorsement 3 2 USA China 1 0 Michael/Ming Internal Michael/Ming External Figure 5 Cultural Effect and Type-of-Cause Effect in Open-ended Attributions of the Michael/Ming (campus-shooting) Scenario Rating-scale Responses to the Dave/De (Bullying) Scenario A repeated-measures ANOVA was conducted to evaluate the cultural and type-of-cause effect on participants’ rating-scale attributions of the Bullying scenario. A significant cultural effect was found, F (1, 231) = 33.72, p < .00, η2 = .13. The American group (M = 2.68, SD = .53) gave internal causal items more weight than did the Chinese group (M = 2.15, SD = .61); the Americans (M = 3.07, SD = .73) also gave external causal items more weight than did the Chinese group (M = 2.67, SD = .72). A significant main effect was found for the within-subjects factor, the type of cause, F (1, 231) = 99.68, p < .00, η2 = .30. Overall, participants rated external causal items (M =2.80, SD = .74) significantly more important than they did the internal items (M = 2.32, SD = .63). However, there was no significant interaction effect between the two factors, F (1, 231) = 2.17, p < .14, η2 = .01. These results can be viewed in Figure 6. 44 Weight of Rating-Scaled Items 5 4 USA 3 China 2 1 Dave/De Internal Dave/De External Figure 6 Cultural Effect and Type-of-Cause Effect in Rating Attributions of the Dave/De (Bullying) Scenario Rating-scale Responses to the Michael/Ming (campus-shooting) Scenario A repeated-measures ANOVA was conducted to evaluate the cultural and the type-ofcause effects on participants’ rating-scale attributions of the Campus-shooting scenario. A significant cultural effect was found, F (1, 231) = 19.20, p < .00, η2 = .08. The American group (M = 3.64, SD = .68) gave more weight to the internal items than did the Chinese group (M = 3.39, SD = .72); the Americans (M = 2.47, SD = .81) also gave more weight to the external items than did the Chinese (M = 2.02, SD = .75). A significant within-subjects effect was found for the type of cause, F (1, 231) = 380.93, p < .00, η2 = .623. Overall, participants rated internal cause items (M = 3.47, SD = .72) significantly more important than they did the external items (M = 2.16, SD = .80). However, there was no significant interaction effect between the two factors, F (1, 231) = 2.39, p < .12, η2 = .01. These results can be viewed in Figure 7. 45 Weight of Rating-Scaled Items 5 4 USA 3 China 2 1 Michael/Ming Internal Michael/Ming External Figure 7 Cultural Effect and Type-of-Cause Effect in Rating Attributions of the Michael/Ming (campus-shooting) Scenario Itemized Scores In order to determine if one cultural group favored any specific cause(s) significantly than the other group, repeated-measure MANOVAs were conducted. Analysis of Open-ended Category Endorsement of the Dave/De (Bullying) Scenario In order to determine if one group provided certain causes of the bullying behavior significantly more than the other group, a repeated-measures MANOVA was conducted for participants’ endorsement of the eight individual categories (four internal causes and four external causes) of the bullying scenario. Overall, analyses showed significant cultural effects, as the two cultural groups differed significantly on six categories. The American group endorsed significantly more often the categories of anger, self-esteem, and teacher’s responsibility whereas the Chinese group endorsed more often the categories of personality, parenting, and social-cultural influences. Each group’s endorsement rates for the eight categories showed that more than half of the American group believed that anger, self-esteem, and parenting were the important causes for 46 Dave’s bullying behavior. The majority of the Chinese group considered parenting as the major cause, and half of the Chinese group regarded personality as the most important cause. Table 6 is a summary of the percentage of each group’s participants who endorsed the eight categories and the F-values of the repeated-measure MANOVA. Table 6 Endorsement of Dave/De (Bullying) Open-ended Categories Dave/De (Bullying) OpenAmerican Chinese F ended Categories % of 76 % of 157 (1, 231) Internal Category D1 Angry at others; acts for self- 68.40% 16.60% 91.811* 51.30% 15.30% 43.478* enhancement or scapegoat. D2 Low self-esteem & needs attention. D3 Poor social & coping skills. 28.90% 32.50% .097 D4 Personality or mental problems. 34.20% 56.10% 9.428* External Category D5 Parenting problem. 57.90% 86.60% 7.776* D6 Teacher’s responsibility. 39.50% 7.60% 42.352* D7 Peer provoked him or not 9.20% 6.40% .607 1.30% 8.30% 4.343 reported to adults. D8 Social & cultural influences. Note: * p < .006 (adjusted error rate). F=4.343 was significant before adjusted error rate. Analysis of Open-ended Category Endorsement of the Michael/Ming (campus-shooting) Scenario In order to determine if one group provided certain causes of the shooting behavior significantly more than the other group, a repeated-measures MANOVA was conducted for the open-ended responses of the 10 categories (five internal causes and five external causes) of the Campus-shooting scenario. Cross-culturally, the American group provided significantly more 47 causes concerning emotion, revenge, school, advisor, peer, and job; the Chinese group gave significantly more causes concerning Ming’s social skills and personality problems. The majority of American participants regarded emotional cause and vindictive purpose as the major reason for Michael’s tragedy. Half of the Chinese groups considered mental health problems as the major cause of the campus shooting. The details can be viewed in Table 7. Table 7 Endorsement of Dave/De (Bullying) Open-ended Categories American Chinese F % of 76 % of 157 (1, 231) 73.70% 29.90% 89.172* 18.40% 42.70% 13.586* Internal Category M1 Emotional, vindictive, or selfenhancement. M2 Lack of social or communication skills M3 Personality problem 11.80% 40.10% 21.475* M4 Flawed belief or value system 27.60% 27.40% .002 M5 Mental health problem 46.10% 51.60% .568 32.90% 19.10% 4.656 External Category M6 School-related problem and failed job search M7 Social, cultural influence, esp. media 15.80% 14.60% .978 M8 Family background, parenting 14.50% 19.10% .732 22.40% 5.10% 18.906* 9.20% 6.40% 3.136 problem, traumatic childhood M9 Advisor, peer, or administration being unfair M10 Drug abuse, health issue, relationship issue Note: * p < .005 (adjusted error rate). F=4.656 was significant before adjusted error rate. Analysis of Rating-scale Items for the Dave/De (Bullying) Scenario 48 In order to determine if one group favored certain causes of the bullying behavior significantly than the other group, a repeated-measured MANOVA was conducted to evaluate the cultural differences in participants’ responses across the thirteen rating-scale items of the Bullying scenario. Overall, the American group gave more weight than did the Chinese group to 11 out of 13 items, among which 7 were significantly different between the groups. The Chinese group (M = 4.08, SD = 1.15) only gave significantly more weight to item 1 (“He is self-centered and spoiled”) compared to the American group (M = 3.29, SD = 1.24), F = 23.01, p < .00, η2 = .09. The Chinese group also gave more weight to item 13 (“His parents spoiled him but did not discipline him enough”) although this rating was not significantly higher than the American group’s rating. The details can be viewed in Table 8. 49 Table 8 Cultural Effects Among the Dave/De (Bullying) Rating-Scale Items Items in Rating-scale Questionnaire American Chinese (N=76) Internal Causes (N=157) M SD M SD F (1, 231) D1 He is self-centered and spoiled. 3.29 1.24 4.08 1.15 23.01** D2 He was acting aggressively to gain 3.99 1.05 2.11 1.29 122.01** 2.95 1.22 1.63 0.96 80.64** attention from peers. D4 He was acting aggressively so that his peers would not bully him. D5 He did not like school. 1.91 0.85 1.64 1.09 3.639 D6 He did not like his schoolmates and 1.99 0.97 1.93 1.2 .130 1.93 0.81 1.49 0.92 12.95* 3.7 1.08 2.7 1.28 34.46* 2.4 1.13 2.28 1.25 .509 2.79 1.27 2.17 1.28 12.30* 2.96 1.3 1.69 1.13 58.15* 2.49 1.22 2.37 1.32 .418 3.26 1.3 3.25 1.37 .003 3.87 1.12 4.26 1.19 5.63 teachers. D7 He did not like the classmate he bullied. External Causes D3 His teacher mentioned his low grade in front of the whole class. D8 Movies, television, and other mass media glorify this type of behavior. D9 Most of his peers don't like him and make fun of him. D10 He is friend with peers who behave like him. D11 Peers who had been bullied by him did not report him. D12 Teachers and school did not attend to his bullying behavior effectively and immediately. D13 His parents spoiled him and did not discipline him enough. Note. * p< .000 (adjusted error rate). Analysis of Rating-scale Items for Michael/Ming (campus-shooting) Scenario 50 In order to determine if one group favored certain causes of the shooting behavior significantly than the other group, a repeated-measures MANOVA was conducted for the responses to the 13 rating-scale items for the Campus-shooting scenario. The American group gave 11 out of 13 items more weight than did the Chinese group. Among the 11 items, 7 items were given significantly higher ratings by American participants. Chinese participants only gave slightly more weight than the American group to item 4 (“He has some personality problems”) and item 13 (“There is no third-party entity for appealing”). Results are shown in Table 9. 51 Table 9 Cultural Effects in Michael/Ming (campus-shooting) Rating-Scale Items Items in Rating-scale Questionnaire American Chinese (N=76) Internal Causes M1 He is stressed out by school and (N=157) F (1, 231) M SD M SD 4.03 1.03 3.59 1.19 7.45 3.82 1.17 3.66 1.28 .75 unpromising job hunting. M2 He did not know other ways to seek help to release his frustration and anger. M4 He has some personality problems. 3.74 1.14 3.86 1.22 .55 M5 He was too obsessed with the award. 3.37 1.09 2.96 1.19 6.28* M6 He was mentally instable and ill. 3.78 1.16 3.27 1.22 9.24* M7 He had unrealistic fantasies of social 3.13 1.28 3.00 1.42 .45 2.80 1.32 2.15 1.26 13.41* 2.47 1.29 1.87 1.12 13.64* 2.08 1.03 2.01 1.30 .18 2.16 1.22 1.58 1.04 14.11* 2.32 1.35 1.81 1.27 7.70* 3.14 1.48 2.04 1.32 33.20* 2.30 1.21 2.65 1.49 3.15 justice. External Causes M3 The vice president did not deal with the case fairly. M8 Movies, television, and other mass media glorify this type of behavior. M9 People around are individualistic and selfish. M10 His friends like to play with guns and solve problems by force. M11 The availability of gun and shooting training where he was living. M12 His advisor and fellow students were cheating. M13 There is no third-party entity for appealing. Note: * p < .003 (adjusted error rate), F=7.45 was significant before adjusted error rate. 52 Question Format Effect Descriptive statistics were used to detect question format effects due to the fact that rating-scale attribution scores (RS-Pref) and open-ended attribution scores (OE-Pref) were two measures on different scales and cannot be compared directly. Based on their RS-Pref and OEPref scores, participants from each cultural group were classified as an internal, neutral, or external attributor. Participants with RS-Pref and OE-Pref larger than .5 were classified as internal attributors and those with RS-Pref and OE-Pref smaller than .5 were classified as external attributors. The number and percentage of participants for each category can be viewed in Table 10. Results show that most participants made internal attributions of the campus-shooting scenario in both the rating-scale and open-ended questionnaires. This indicates that question format did not make much difference in the attributions of the campus-shooting scenario. However, question format made a difference in the bullying scenario to which most participants attributed external causes for the rating-scale questionnaire and internal causes for the openended questionnaire. Also, in the open-ended responses of the bullying scenario, Chinese participants’ responses were divided into three approximately equal attribution types . Finally, for both scenarios, there were more participants favoring a neutral attribution for the open-ended questionnaire than for the rating-scale questionnaire. Overall, question format made a difference in causal attribution in that rating-scale questions evoked participants’ favoring either internal or external causes depending on the scenario but open-ended questions inspired more diverse responses. 53 Table 10 Participants’ Percentages for Each Scenario and Attribution Type Type of Culture Dave/De (Bullying) Scenario Michael/Ming (campusAttribution Internal Neutral External shooting) Scenario Rating-scale Open-ended Rating-scale Open-ended Am 19 (25%) 45 (59.2%) 58 (76.3%) 53 (69.7%) Ch 30 (19.1%) 50 (31.8%) 140 (89.2%) 130 (82.8%) Am 6 (7.9%) 18 (23.7%) 1 (1.3%) 14 (18.4%) Ch 4 (2.5%) 58 (36.9%) 3 (1.9%) 32 (20.4%) Am 51 (67.1%) 13 (17.1%) 17 (22.4%) 9 (11.8%) Ch 123 (78.3%) 49 (31.2%) 14 (8.9%) 13 (8.3%) Note: Am=American, Ch=Chinese. American N=76, Chinese N=157. 54 IV. DISCUSSION Interpretation Previous studies showed incompatible attribution patterns of Americans and Chinese. Some studies found that Americans, being individualists, tend to make external attributions to achieve self-enhancement and protect the self, whereas the Chinese, being collectivists, tend to make internal attribution to blame individuals and protect the group. Other studies claimed that Americans, influenced by ancient Greek philosophy, tend to make more internal and dispositional attributions whereas the Chinese, influenced by Taoist and Confucian philosophy, tend to make more external and situational attributions. However, the results of the current study did not replicate the cultural differences claimed by any of these previous studies. Overall, American and Chinese participants favored the same type of attribution, whether it was internal or external, although they did show different extents to which they favored those attributions. The participants’ causal attributions and evocation of different type of causes differed significantly by scenario. Question format—either rating-scale or open-ended—was associated with different types of causal attributions as well. In this chapter, the interpretation of the analysis results will be discussed in the same order as the research questions and hypotheses. Causal Attribution and Scenario Effect The first research question of this study concerned whether the type of scenario used in the survey led the participants to make certain type of attributions, either internal (dispositional) or external (situational). The predictions for the first research hypothesis were that participants would make external attributions of the bullying schoolboy scenario but internal attributions of the campus-shooting scenario. These predictions were confirmed only by the rating-scale 55 responses. In the open-ended responses, participants made internal attributions of both scenarios, but participants made stronger internal attributions for the campus-shooting scenario than for the bullying scenario. The tentative explanation for this result is that people allow children external excuses for their misbehaviors but hold adults accountable for their own wrong-doings. The study findings are consistent with the view that the perpetrator will be attributed with greater responsibility under conditions of a serious nature (high severity) than those of a less serious nature (low severity) (Gleason & Harris, 1976). The campus-shooting scenario yielded stronger internal attributions than did the bullying scenario because the outcome of a campus shooting is much more tragic and severe than that of bullying; therefore, participants attributed more internal attributions to the shooter than they did to the bully. In summary, the content of a situation appears to play an important role in participants’ causal attributions and different events or behaviors may be interpreted quite differently due to the nature of the events or behaviors. Scenario type also interacted with culture in participants’ causal attribution, as will be revisited in the discussion of cultural effect. Causal Attribution and Question Format Effects The second research question of the study involved whether the question formats (openended versus rating-scale) had an impact on participants’ causal attributions. The results confirmed that question format made a difference in participants’ causal attribution patterns. It was predicted that both cultural groups would favor neither internal nor external causes in their responses to the open-ended questions. Therefore, participants should have provided more neutral attributions than internal or external attributions. However, the results cannot confirm this prediction because, for both scenarios, there were more people endorsing internal attributions 56 than neutral attributions. A second prediction was that both cultural groups would endorse either internal or external causes in their responses to the rating-scale items. This is confirmed by the results in that both groups favored external attributions for the bullying scenario and internal attributions for the campus shooting scenario. Overall, question format showed an impact on the bullying scenario in which participants attributed external causes for the rating-scale questionnaire but attributed internal causes for the open-ended questionnaire. The number of external attributions decreased dramatically and the number of internal attributions increased significantly when participants responded to the openended questions. In addition, the majority of the American participants responded with internal attributions to open-ended questions. Thus, some evidence exists to support the prediction that question format makes a difference in people’s causal attributions. Last, the majority of participants in both groups made internal attributions regarding the campus-shooting scenario based on their responses to the rating-scale and open-ended questions. A possible explanation for this result is that a campus shooting may be considered by most people to be more severe of an event, and therefore the perpetrator was held responsible for the consequence of his behavior. Causal Attribution and Cultural Effects The third research question investigated whether causal attributions are culture-specific. It was predicted that the two cultures would attribute internal and external causes to the scenarios to a different extent; however, they would not show opposite attribution patterns. In the previous chapter, cultural comparisons of attributions were presented using three sets of scores:overall attribution preference scores, internal versus external attribution scores, and the itemized attribution scores. Results confirmed the prediction and showed that cultural background had an 57 impact on causal attributions. However, the opposite attribution patterns claimed by previous studies was not found. Other interesting findings emerged from the results and are discussed below. This study aimed to find out whether American and Chinese groups favored different type of causal attributions (i.e., did Americans make more internal attributions and did the Chinese make more external attributions or vice versa). The results of the attribution preference analysis showed that in their open-ended responses, both cultural groups favored internal causes over external ones for both scenarios. Also, in their rating-scale responses, both groups favored external causes for the bullying scenario and internal causes for the campus-shooting scenario. Thus, the opposite attribution patterns claimed by previous studies were not duplicated in this study. However, a different magnitude of internal and external attributions was found between the two cultural groups. An interesting finding is that, compared to their Chinese counterparts, American participants were found to consistently provide more internal and external causes in the open-ended responses and give more weight to both internal and external causes in the rating-scale responses. This is validated by the itemized analysis of both open-ended categories and rating-scale items. Since the American group showed the tendency to rate most causes as important and provided more causes relative to the Chinese group, it will be more illuminating to view the few causes that the Chinese group provided more often and rated as more important than did the American group. Chinese participants emphasized personality as the main cause for both scenarios in both question formats. In the campus-shooting scenario, social skills and no third-party for appeal were the other two causes besides personality that received more weight from Chinese 58 participants. For the bullying scenario, besides the personality factor, the Chinese group considered being self centered and parenting as more important for the boy’s bullying behavior. This may be explained by China’s one-child policy. Since the 1980’s when this policy was adopted, most families in urban areas have only one child who are lavished with attention and love and usually get what they want. Trimetric Classic, the 13th century children’s book about Confucian principles, says that a father is at fault if he raises his children but doesn’t teach them; a teacher is lazy if he only teaches his students but doesn’t discipline them. It seems that the Chinese participants’ answers to the survey reflect both traditional beliefs and today’s realities. However, the bullying scenario happened to provide both the only child and busy parents as possible causes of the bullying behavior. Therefore, the Chinese emphasis on being self-centered and parenting as causes may be a product of both scenario and culture effects. In summary, culture does make a difference in people’s causal attributions. However, this difference is the relative extent to which the participants make either internal or external attributions. The pattern of an American preference for internal or dispositional attributions compared to the Chinese preference for external or situational attributions has not been found in this study. Limitations Some limitations of the study may have affected the conclusions drawn from the analysis. First, the samples of American and Chinese groups are not completely equivalent, although participants were recruited from university settings to help ensure equivalency. The data showed that all the participants were well-educated and computer literate. However, the American sample was a little more diverse in their education level and age compared to their Chinese counterparts. For example, 60.5% of the American group were from 17 to 26 years old, 26.3% 59 were from 27 to 36 years old, and 13.2% were from 37 to 61 years old. Also, 67.1% of the Americans were undergraduates, 7.9% were graduate students, and 24.0% had finished graduate school. The Chinese sample was more homogeneous; the majority of these participants were college students in their early 20s. Also, 97.5% of this group were from 17 to 26 years old, and 87.5% were college students. Gender and social-economic status can be influential in participants’ causal thinking as well. However, this information was not included in the survey; otherwise, the results may be better interpreted and understood. A second limitation is that both scenarios employed in the survey described anti-social behaviors. This was done for comparison purposes with previous studies that used such behaviors. However, valence of the behavior and severity of the consequences of anti-social behavior have been proven to affect people’s causal attributions of such behaviors. Based on the results of the current study, scenarios have an important impact on such causal attributions; therefore, scenarios using pro-social behaviors or events may evoke different attribution patterns from the participants. Therefore, the attribution patterns found in the current study might vary if the participants had been presented with scenarios of pro-social behaviors with positive consequences or with scenarios of daily events without any apparent consequences. Third, the psychological constructs are not consistent between the two cultural groups. That is, the American and Chinese groups perceived the internal and external causes in the rating-scale questionnaire differently. As was referenced in Table 3 and Table 4, factor analysis of each group extracted different factors and the two cultural groups’ ratings of the 13 causes for each scenario did not load on two factors—internal and external as intended by the researcher. Therefore, the analyses were done based on the face validity of the two factors which may affect the results and hence the conclusions, discussed below. 60 Conclusions and Implications Results suggest that the content of the scenarios evoked different attribution patterns in each culture. There is also some evidence to support the hypothesis that question format affects causal attributions. Overall, the results showed that American and Chinese participants made similar attributions for the same scenario, which does not support the occurrence of opposite attribution patterns as claimed by previous research. One explanation for the different findings is that culture is changing quickly due to the fast technology development. The studies by Morris and Peng was done in 1994 when personal computers were rarely seen in average household in China. Nowadays personal computers are quite usual in average households in China. Today’s internet technology and international transportation make global communication more convenient and efficient which provides both Americans and Chinese more exposures to the world on the other side of the Pacific Ocean. People learn more about and from other cultures and are influenced by other cultures as well. College and graduate students, this study’s major participants, are usually the most technology savvy population who benefit greatly from the internet communication. They also have more opportunities to mingle with international students on campus and learn from other cultures. They may represent the converging trend how people will think when the national boundaries become blurry and the true global village come into being. Different cultures may differentially affect people’s thinking and attribution patterns; however, these differences do not appear to be fundamental or absolute. Then why does so much research claim significant cultural differences in causal attributions? First, one possible explanation comes from the culture-as-similarities-or-differences debate. When culture is treated as differences between cultural groups, it is assumed that the differences are stable features that 61 distinguish one group from another. Some cultural heuristics might occur in certain instances, and cultural differences might appear and disappear under different circumstances (N. Ross, 2004). However, such differences have been overemphasized and universalized. Educators devote great amounts of attention to cultural diversity, and many schools have cultural activities to teach children about other cultures and people. What are we teaching children?—essentially, the difference between the U.S. culture and other cultures. It is true that certain cultural influences are internalized in the process of our cognitive development; however, we might not think as differently from each other as some researchers have proposed. Another explanation for the differences is that a variety of instruments and scenarios have been employed in previous studies (e.g., computer art clips, hypothetical scenarios, vignettes, video or audio clips, rod-and-frame, etc.). This variety may contribute to the different causal attributions made by participants. Third, the interpretation of the results may contribute significantly to the differences among the conclusions. Participants are usually presented with one or a few scenarios and requested to make causal attributions accordingly. The attribution differences for a particular event may not represent cultural differences in attribution styles. Finally, one last possibility is that “researchers in the field virtually ignore each other’s findings, explaining their data with radically different theories as if their individual data were completely isolated from other research” (N. Ross, 2004, p. p.6). Culture is still a growing area of research for psychologists; researchers are still working toward the convergence of different theories and approaches. It would be easier for educators and researchers if we could claim apparent or opposing cultural differences in people’s behaviors or reasoning. For instance, when we teach children different food made by different cultures, Chinese children learn that Americans like hamburgers and Italians make pizzas; American kids 62 learn that Chinese make dumplings and Japanese like sushi. Those concepts easily become stereotypes which could stop children or anyone from exploring more about other cultures. When American children are exposed to a Chinese context, they will realize that Chinese make dumplings occasionally and during their daily lives, they cook rice, steamed bun, noodles, and all sorts of food that they could imagine. Unless Chinese children come to the United States, they rarely will find out that Americans make all kinds of food as well and hamburger is only for convenience and picnic. Researchers have to claim cultural differences cautiously bearing in mind that a great deal of other variables that may contribute to the differences. As of today, the international cooperation is growing and the availability of internet is increasing; inter-cultural exchanges are more convenient and frequent due to the advanced traveling and communication technology; people learn from each other quickly and adapt to the new environment fast to keep up with the changes. Those factors may contribute to less and less cross-cultural differences and more and more individual differences when we eventually become a global village one day. Future study should address how changing culture influences people’s thinking and how people adapt to changing culture. 63 REFERENCES Adamopoulos, J., & Lonner, W. J. (2001). Culture and Psychology at a Crossroad: Historical Perspective and Theoretical Analysis. In D. Matsumoto (Ed.), The Handbook of Culture and Psychology. New York: Oxford University Press. Au, K., Hui, M. K., & Leung, K. (2001). Who should be responsible? Effects of voice and compensation on responsibility attribution, perceived justice, and post-complaint behaviors across cultures. The International Journal of Conflict Management, 12(4), 350364. Berry, J. W. (1997). Preface. In J. W. Berry, M. H. Segall & C. Kagitcibasi (Eds.), Handbook of Cross-Cultural Psychology: Social Behavior and Applications (2nd ed., Vol. 3). Boston: Allyn & Bacon. Boas, F. (1911/1938). The mind of primitive man. New York: Macmillan. Cheah, C. S. L., & Rubin, K. H. (2004). European American and mainland Chinese mothers' responses to aggression and social withdrawal in preschoolers. International Journal of Behavioral Development, 28(1), 83-94. Chua, F. H., Boland, J. E., & Nisbett, R. E. (2005). Cultural variation in eye movements during scene perception. Proceedings of the National Academy of Sciences, 102(35), 1262912633. Cole, M. (1995). Culture and cognitive development: from cross-cultural research to creating systems of cultural mediation. Culture and Psychology, 1, 29. Cole, M. (2005). Culture in Development. In M. H. Bornstein & M. E. Lamb (Eds.), Developmental Science: an Advanced Textbook (5th ed.). Mahwah, NJ: Lawrence Erlbaum Associates. 64 Crittenden, K. S. (1996). Causal attribution processes among the Chinese. In M. H. Bond (Ed.), Handbook of Chinese Psychology (pp. 263-279). New York: Oxford University Press. Crystal, D. S. (2000). Concepts of deviance and disturbance in children and adolescents: a comparison between the United States and Japan. International Journal of Psychology, 35(5), 207-218. Crystal, D. S., Parrott, W. G., Okazaki, Y., & Watanabe, H. (2001). Examining relations between shame and personality among university students in the United States and Japan: a developmental perspective. International Journal of Behavioral Development, 25(2), 113-123. Eagly, A. H., Mladinic, A., & Otto, S. (1994). Cognitive and affective bases of attitudes toward social groups and social policies. Journal of Experimental Social Psychology, 30, 113137. Forsterling, F., Preikschas, S., & Agthe, M. (2007). Ability, luck, and looks: an evolutionary look at achievement ascriptions and the sexual attribution bias. Journal of Personality and Social Psychology, 92(5), 775-788. Gilbert, D. T., & Malone, P. S. (1995). The correspondence bias. Psychological Bulletin, 117(1), 21-38. Gleason, J. M., & Harris, V. A. (1976). Perceived freedom, accident severity and empathic value as determinants of the attribution of responsibility. Social Behavior and Personality, 4(2), 171-176. Greenfield, P. M. (2000). Three approaches to the psychology of culture: where do they come from? Where can they go? Asian Journal of Social Psychology, 3, 223-240. 65 Hafer, C. L., & Begue, L. (2005). Experimental research on just-world theory: problems, developments, and future challenges. Psychological Bulletin, 131(1), 128-167. Hauser, C., & O'Connor, A. (2007). Virginia Tech Shooting Leaves 33 Dead Retrieved July 12, 2008, 2008, from http://www.nytimes.com/2007/04/16/us/16cnd-shooting.html Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley. Hinde, R. A. (1987). Individuals, relationships & culture: Links between ethology and the social sciences. New York: Cambridge University Press. Ho, I. T. (2004). A comparison of Australian and Chinese teachers' attributions for student problem behaviors. Educational Psychology, 24(3), 375-391. Holloway, M. (1997). The Paradoxical Legacy of Franz Boas - Father of American anthropology [Electronic Version]. Natural History. Retrieved January 28, 2009 from http://findarticles.com/p/articles/mi_m1134/is_10_106/ai_53479059. Hui, E. K. P. (2000). Personal concerns and their causes: perceptions of Hong Kong Chinese adolescent students. Journal of Adolescence, 23, 189-203. Hui, E. K. P. (2001). Hong Kong students' and teachers beliefs on students' concerns and their causal explanation. Educational Research, 43(3), 279-284. Ji, L., Nisbett, R. E., & Su, Y. (2001). Culture, change, and prediction. Psychological Science, 12(6), 450-456. Jones, E. E., & Davis, K. E. (1965). From acts to dispositions: the attribution process in person perception. In L. Berkowitz (Ed.), Advances in Experimental Social Psychology (Vol. 2). New York. Jones, E. E., & Harris, V. A. (1967). The attribution of attitudes. Journal of Experimental Social Psychology, 3, 1-24. 66 Jones, E. E., & Nisbett, R. E. (1972). The actor and the observer: divergent perceptions of the causes of behavior. In D. T. Gilbert (Ed.), The Selected Works of Edward E. Jones (pp. 179-240). Hoboken: Jonh Wiley & Sons. Kelley, H. H. (1973). The process of causal attribution. American Psychologist, 28, 107-128. Kenworthy, J. B., & Miller, N. (2002). Attributional biases about the origins of attitudes: externality, emotionality, and rationality. Journal of Personality and Social Psychology, 82(5), 693-707. Kim, U., & Park, Y.-S. (2006). The Scientific Foundation of Indigenous and Cultural Psychology: the Transactional Approach. In U. Kim, K.-S. Yang & K.-K. Hwang (Eds.), Indigenous and Cultural Psychology: Understanding People in Context. New York: Springer. Kluckhohn, C. K. M. (1954). Culture and behavior. In G. Lindzey (Ed.), Handbook of social psychology (Vol. 2, pp. 921-976). Cambridge, MA: Addison-Wesley. Kozak, M. N., Marsh, A. A., & Wegner, D. M. (2006). What do I think you're doing? Action identification and mind attribution. Journal of Personality and Social Psychology, 90(4), 543-555. Lerner, M. J., & Miller, D. T. (1978). Just world research and the attribution process: looking back and ahead. Psychological Bulletin, 85(5), 1030-1051. Lerner, M. J., & Simmons, C. H. (1966). Observers reaction to the "innocent victim": compassion or rejection? Journal of personality and Social Psychology, 4(2), 203-210. LeVine, R. A. (2001). Culture and Personality Studies, 1918-1960: Myth and History. Journal of Personality, 69(6), 803-818. 67 LeVine, R. A. (2007). Anthropological Foundations of Cultural Psychology. In S. Kitayama & D. Cohen (Eds.), Handbook of Cultural Psychology (pp. 40-58). New York: The Guilford Press. Lewin, K. (1948/1997). Resolving social conflicts and field theory in social science. Washington, DC: American Psychological Association. Lieberman, M. D., Jarcho, J. M., & Obayashi, J. (2005). Attributional inference across cultures: similar automatic attributions and different controlled corrections. Personality and Social Psychology Bulletin, 31(7), 889-901. Malle, B. F. (2004). How the Mind Explains Behavior: Folk Explanation, Meaning, and Social Interaction. Cambridge, MA: MIT Press. Markus, H. R., & Kitayama, S. (1991). Culture and the self: implications for cognition, emotion and motivation. Psychological Review, 98(2), 29. Matsumoto, D. (2001). Preface. In D. Matsumoto (Ed.), The Handbook of Culture and Psychology. New York: Oxford University Press. Miller, G. A. (2003). The cognitive revolution: a historical perspective. Mills, J., & Clark, M. S. (1982). Exchange and communal relationships. In L. Wheeler (Ed.), Review of personality and social psychology (Vol. 3, pp. 121-144). Beverly Hills, CA: Sage. Morris, M. W., & Peng, K. (1994). Culture and cause: American and Chinese attributions for social and physical events. Journal of Personality and Social Psychology, 67(6), 949-971. Ng, S. H., & Zhu, Y. (2001). Attributing causality and remembering events in individual- and group-acting situations: a Beijing, Hong Kong, and Wellington comparison. Asian Journal of Social Psychology, 4, 39-52. 68 Nisbett, R. E. (2003). The geography of thought: how Asians and Westerners think differently ... and why. New York: Free Press. Nisbett, R. E., Peng, K., Choi, I., & Norenzayan, A. (2001). Culture and systems of thought: holistic versus analytic cognition. Psychological Review, 108(2), 9. Niu, W., & Sternberg, R. J. (2001). Cultural influences on artistic creativity and its evaluation. International Journal of Psychology, 36(4), 225-241. Norenzayan, A., Smith, E. E., Kim, B. J., & Nisbett, R. E. (2002). Cultural preferences for formal versus intuitive reasoning. Cognitive Science, 26, 653-684. Parmar, P., Harkness, S., & Super, C. M. (2004). Asian and Euro-American parents' ethnotheories of play and learning: effects on preschool children's home routines and school behaviour International Journal of Behavioral Development, 28(2), 7. Robins, R. W., Mendelsohn, G. A., Connell, J. B., & Kwan, V. S. Y. (2004). Do people agree about the causes of behavior? A social relations analysis of behavior ratings and causal attributions. Journal of Personality and Social Psychology, 86(2), 334-344. Robins, R. W., Spranca, M. D., & Mendelsohn, C. A. (1996). The actor-observer effect revisited: Effects of individual differences and repeated social interactions on actor and observer attributions. Journal of Personality and Social Psychology, 71(2), 375-389. Ross, L. D. (1977). The intuitive psychologist and his shortcomings: distortions in the attribution process. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 10, pp. 173-220). New York: Academic Press. Ross, N. (2004). Culture and cognition: implications for theory and method Thousand Oaks, CA: Sage Publications. 69 Schlenker, B. R., Weigold, M. F., & Hallam, J. R. (1990). Self-serving attributions in social context: effects of self-esteem and social pressure. Journal of Personality and Social Psychology, 58(5), 855-863. Shweder, R. A. (1991). Thinking Through Cultures: Expeditions in Cultural Psychology. Cambridge, MA: Harvard University Press. Shweder, R. A. (2000). The psychology of practice and the practice of the three psychologies. Asian Journal of Social Psychology(3), 207-222. Song, M.-J., Smetana, J. G., & Kim, S. Y. (1987). Korean Children's Conceptions of Moral and Conventional Transgressions. Developmental Psychology, 23(4), 5. Spencer-Rodgers, J., Williams, M. J., Hamilton, D. L., & Peng, K. (2007). Culture and group perception: dispositional and stereotypic inferences about novel and national groups. Journal of Personality and Social Psychology, 93(4), 525-543. Triandis, H. C. (1996). Foreword. In J. W. Berry, M. H. Segall & C. Kagitcibasi (Eds.), Handbook of Cross-Cultural Psychology: Social Behavior and Applications (2nd ed., Vol. 3). Boston: Allyn & Bacon. Triandis, H. C. (2000). Dialectics between cultural and cross-cultural psychology. Asian Journal of Social Psychology, 3, 185-195. Triandis, H. C. (2001). Individualism-Collectivism and Personality. Journal of Personality, 69(6), 17. Triandis, H. C. (2007). Culture and Psychology: A History of the Study of Their Relationship. In S. Kitayama & D. Cohen (Eds.), Handbook of Cultural Psychology (pp. 59-76). New York: The Guilford Press. Tylor, E. B. (1871/1958). The Origins of Culture. New York: Harper & Brothers. 70 Vygotsky, L. (1929). The Problem of the Cultural Development of the Child II [Electronic Version]. Journal of Genetic Psychology, 36, 7. Retrieved 5/13/2005 from www.marxists.org/archive/vygotsky/works/1929/cultural_development.htm. Ybarra, O. (2002). Naive causal understand of valenced behaviors and its implications for social information processing. Psychological Bulletin, 128(3), 421-441. 71 APPENDIX A Recruitment Letters in English 72 Participants Needed for Psychology Study by a Ph.D. student at University of Cincinnati Don’t you wonder all the time why people do what they do? Help us in this psychology research and you will contribute to the explanation. Please visit the website to participate: http://homepages.uc.edu/cnc/2 00701/consent.php We’d be delighted to have your inputs. It’s absolutely confidential and we won’t ask for your name or any other identification. 73 APPENDIX B Recruitment Letters in Chinese 74 辛辛那提大学的心理学博士研究需要您的参与 你是否经常纳闷: 人们行为背后的动机和原因是什么呢? 参加我们的问卷调查,找到问题的答案。请访问: http://homepages.uc.edu/cnc/2007ch/ 我们真诚地感谢您的参与。您的观点将被完全保密; 我们不会在问卷里问及您的姓名和任何私人问题。 75 APPENDIX C Cognition and Culture Survey in English 76 Welcome to the Cognition and Culture Project! We really appreciate your taking time to consider participating in our study. This page explains the study and gives you the information you need to make a decision about whether to participate. If you agree to participate after reading this page, please click the YES button at the bottom of the page. University of Cincinnati Consent to Participate in a Research Study College of Education, Human Services, and Criminal Justice / Division of Educational Studies Yan Yang <[email protected]> 513-556-4596 Title of Study: Cognition and Culture Introduction: I am inviting 10 to 20 students to take part in a research study that I am doing as part of my doctoral degree program. Please read the following explanation carefully and then make a decision about whether you wish to participate. Purpose: The purpose of this study is to try out and improve a questionnaire created for my doctoral dissertation which will be focused on the cultural differences in cognition between Chinese and Americans. Duration: Completing the web survey for this study will probably take 10 to 15 minutes. Procedures: There are two web forms for this study. The first part of Form I presents a brief scenario describing someone's actions and asks you to give possible reasons for the person's behavior. The second half of Form-I asks you about whether the first half was clear and asks for your suggestions for improving them. The first half of Form II presents the same scenario again, lists possible reasons, and asks you to rate their importance. The second half of Form II asks you about whether the first half was clear and asks for your suggestions for improving them. Since I'm trying to improve the questionnaire, both your answers to the questions on the form and the comments about the form are very important to me. Risks/Discomforts: I do not believe there are any risks associated with your participation in this study. The scenario may include violent actions but I do not think that it will be upsetting to adults. Benefits: You will receive no direct benefit from your participation in this study, but your participation may help teachers, cognitive psychologists, and researchers better understand the cultural differences between Americans and Chinese. Confidentiality: The survey is anonymous so there will be no record of what answers you gave or even whether you participated. Offer to answer questions: If you have any other questions about this study, you may contact me at <[email protected]>/513-556-4596 or my faculty advisor, Dr. Daniel D. Wheeler at <[email protected]>/513-556-3607. If you have any questions about your rights as a research participant, you may contact the Chair of the Institutional Review Board—Social and Behavioral Sciences at <[email protected]>/513-558-5784. Voluntary Participation: I will have no way of knowing whether you have participated in this study. You may QUIT AT ANY TIME just by closing your web browser. Agreement: By clicking on the YES button below, you are indicating your understanding of this consent document and agreeing to participate in the study. If you wish to keep a copy of this consent form, you may print it from your web browser. I HAVE READ THE INFORMATION PROVIDED ABOVE AND I VOLUNTARILY AGREE TO PARTICIPATE IN THIS STUDY. NO YES 77 Before you proceed to the questionnaire, please answer the following questions by checking the button in front of your answer or fill in the blank: 1. What is your nationality? American Mainland Chinese Others, please specify 2. Where do you live? USA Mainland China Others, please specify 3. How old are you? 4. What is your education level? High school Some college Completed college Some graduate school Completed graduate school Please proceed to the survey by clicking Submit 78 Please read the following scenario: Dave was an only child and both his parents gave him a lot of attention. His parents were doctors so they worked very hard and weren’t home very much. Dave didn’t have much patience if things didn’t go his way. One day, Dave’s teacher mentioned Dave’s low grade on a recent test in front of the class. After class, Dave was walking in the school hallway when someone accidentally bumped into him. He grabbed the other person, pushed him up against the wall, and demanded he apologize. At lunch, a particularly short kid who was in Dave’s class stood in front of him on the food line. Dave began to tease him and call him names. That afternoon, during a chemistry lesson, Dave tripped this same kid as he was walking down the aisle carrying a tray of glass beakers. Now speculate about why Dave behaved the way he did and fill in your answers in the boxes below. Try to give at least two reasons. If you think of more than five reasons, you can put the rest of them in the "More" box. First Reason Second Third Fourth More Submit Answers 79 Please read the following scenario: Ming was a Chinese graduate student who was admitted to a major university in the USA and had been studying in the school for five years. He had recently lost an award competition which was given to his advisor’s other graduate student. Ming suspected that his advisor and the other fellow student cheated on him. He unsuccessfully appealed it and subsequently failed to get an academic job. One day he entered the faculty meeting in his department and shot his advisor, the fellow student, and another two professors. He then proceeded two blocks to the administration building where he shot the associate vice president of academic affairs who was handling his case and a student helper. He then committed suicide by shooting himself in the head. Now speculate about why Ming behaved the way he did and fill in your answers in the boxes below. Try to give at least two reasons. If you think of more than five reasons, you can put the rest of them in the "More" box. First Reason Second Third Fourth More Submit Answers 80 Here is the scenario again: Dave was an only child and both his parents gave him a lot of attention. His parents were doctors so they worked very hard and weren’t home very much. Dave didn’t have much patience if things didn’t go his way. One day, Dave’s teacher mentioned Dave’s low grade on a recent test in front of the class. After class, Dave was walking in the school hallway when someone accidentally bumped into him. He grabbed the other person, pushed him up against the wall, and demanded he apologize. At lunch, a particularly short kid who was in Dave’s class stood in front of him on the food line. Dave began to tease him and call him names. That afternoon, during a chemistry lesson, Dave tripped this same kid as he was walking down the aisle carrying a tray of glass beakers. For each listed cause, please click a button to indicate to what extent you think the cause contributed to Dave's behavior. Very important Important Not a factor 1. He is self-centered and spoiled. 2. He was acting aggressive to gain attention from peers. 3. His teacher mentioned his low grade in front of the whole class. 4. He was acting aggressive so that his peers would not bully him. 5. He did not like school. 6. He did not like his schoolmates and teachers. 7. He did not like the classmate he bullied. 8. Movies, television, and other mass media glorify this type of behavior. 9. Most of his peers don't like him and make fun of him. 10. He is friend with peers who behave like him. 11. Peers who had been bullied by him did not report him. 12. Teachers and school did not attend to his bully behavior effectively and immediately. 13. His parents spoiled him but did not discipline him enough. Submit Answers 81 Here is the scenario again: Ming was a Chinese graduate student who was admitted to a major university in the USA and had been studying in the school for five years. He had recently lost an award competition which was given to his advisor’s other graduate student. Ming suspected that his advisor and the other fellow student cheated on him. He unsuccessfully appealed it and subsequently failed to get an academic job. One day he entered the faculty meeting in his department and shot his advisor, the fellow student, and another two professors. He then proceeded two blocks to the administration building where he shot the associate vice president of academic affairs who was handling his case and a student helper. He then committed suicide by shooting himself in the head. For each listed cause, please click a button to indicate to what extent you think the cause contributed to Ming's behavior. Very important Important Not a factor 1. He is stressed out by school and unpromising job hunting. 2. He did not know other ways to seek help to release his frustration and anger. 3. The vice president did not deal with the case fair and square. 4. He has some personality problems. 5. He was too obsessed with the award. 6. He was mentally instable and ill. 7. He had unrealistic fantasy of social justice. 8. Movies, television, and other mass media glorify this type of behavior. 9. People around are individualistic and selfish. 10. His friends like to play with guns and solve problems by force. 11. The availability of gun and shooting training where he was living. 12. His advisor and fellow students were cheating. 13. There is no third-party entity for appealing. Submit Answers 82 APPENDIX D Cognition and Culture Survey in Chinese 83 欢迎访问“认知与文化"研究! 我们真诚地感谢您花时间来考虑参加我们的研究。本网页向您介绍我们的研究,并提供您在参加问卷之前 需要了解的信息。当您阅读之后并同意参加我们的问卷,请点击页底的按钮“是"。 辛辛那提大学 参加认知与文化研究的同意书 教育学院 教育研究系 杨彦 <[email protected]>(513)556-4596 论文题目: “认知与文化" 介绍: 我 邀请大约 100 名学生参加我的博士学位科研项目。在同意参加本研究之前,请仔细阅读所提议的 研究过程。 目的: 这份问卷的目的是了解不同文化背景的人对他人行为的解释。 历时: 完成问卷将需要您大约 10 至 15 分钟。 过程: 包括本页在内,问卷一共有 7 页。其中 4 页是研究内容的核心,在完成每页的问题之后,点击 “是"或者“提交"进入下一个页面。 风险: 参加这份问卷不会给您带来任何的风险。虽然问卷提及暴力行为,但应该不会给成年人造成困扰。 利益:本研究将不会让您直接受益。不过通过参加研究并和采访人讨论有关的话题将会有助于教师,认知 心理学家,和科研人员更好地了解人们如何解释行为。 机密性: 问卷是匿名的,所以没有记录显示您给出哪些回答或者您是否参加了问卷。 如果您对研究有疑问: 如果您对此还有疑问,您可以和我联系 <[email protected]>/513-556-4596 ,或者 联系研究人的导师 Daniel Wheeler 博士 <[email protected]>/513-556-3607。如果您就作为研究参加者的 权利有什么问题的话,您可以打电话给辛大社会与行为科学检查管理委员会的主席 Claudia Norman <[email protected]>/513-558-5784 。 自愿参加的权利: 我无法可知您是否 参加这份问卷。您可以随时关闭您的网络浏览窗口而退出问卷。 同意: 阅读了如上信息之后,如果您愿意参加这份问卷,请点击下面的按钮“是"。如果您想保留此同意 书,请打印本页。 我已经阅读了如上信息,并愿意回答这份问卷。 是 否 84 正式回答问卷之前,请回答以下问题;点击选择您的答案前面的按钮,或者将您的答案填 写在空格里。 1. 您的国籍是: 美国 中国大陆 其它 2. 如果您问题 1 的答案是“其它”,那么请将您的国籍填写到下面的空格里: 3. 您现在的居住地是: 美国 中国大陆 其它 4. 如果您问题 3 的答案是“其它”,那么请将您的居住地填写到下面的空格里: 5. 您的年龄,请在空格里填写数字: 6. 你接受的最高教育水平是: 高中 一些大学 大学毕业 一些研究生学习 研究生毕业 请点击“提交”按钮进入正式的问卷调查 提交 85 请仔细阅读下面的材料,并回答随后的问题。 德是家里的独子,他的父母都是医生, 工作很忙,在家里的时间有限。可是,他们尽可 能给了德许多的关注。如果事情不和德的意,他就发脾气。有一天,德的老师在全班同学 面前提到了他一次近期测验的成绩分数很低。下课后,在走廊里,一个同学不小心撞到德 身上。德揪住这个同学,把他推到墙上,要求他立刻道歉。午餐时间,大家排队等待就 餐。班里的一个小个子同学排在德前面。德戏弄他,辱骂他。下午化学课,这个小个子同 学端着一盘子的试管从德的身边走过时,德把他拌倒了。 请发挥想象,推测德为什么会有以上的行为,并把这些原因填写在下面的空格里。请给出 至少两个以上的原因。如果您可以想象得出五个以上的原因,请把第五个和更多的原因填 写在“更多"之后的空格里。 第一个原因 第二个 第三个 第四个 更多 提交 86 请仔细阅读下面的材料,并回答随后的问题。 明是一名在校五年的研究生。近期,他申请了一项奖学金,未获批准;他的一个同学得到 了这份奖学金。明怀疑是他的导师偏向这个同学,这位同窗作了弊,抢走了他该得的奖学 金。明向校方申诉未果。此后,明申请一所大学的教职,也没成功;对此他归咎于其导师 写的推荐信不力。不久后的某日,他走进本系的教授例会,开枪打死了他的导师,获奖的 那位同窗,和另外两位教授。然后,明来到两个街区以外的行政楼,找到当初处理他的申 诉的教务助理副校长。明开枪打死了这位副校长和一位在现场的学生;随后,明开枪自 尽。 请发挥想象,推测明为什么会有以上的行为,并把这些原因填写在下面的空格里。请给出 至少两个以上的原因。如果您可以想象得出五个以上的原因,请把第五个和更多的原因填 写在“更多"之后的空格里。 第一个原因 第二个 第三个 第四个 更多 提交 87 请仔细阅读下面的材料,并回答随后的问题。 德是家里的独子,他的父母都是医生, 工作很忙,在家里的时间有限。可是,他们尽可 能给了德许多的关注。如果事情不和德的意,他就发脾气。有一天,德的老师在全班同学 面前提到了他一次近期测验的成绩分数很低。下课后,在走廊里,一个同学不小心撞到德 身上。德揪住这个同学,把他推到墙上,要求他立刻道歉。午餐时间,大家排队等待就 餐。班里的一个小个子同学排在德前面。德戏弄他,辱骂他。下午化学课,这个小个子同 学端着一盘子的试管从德的身边走过时,德把他拌倒了。 以下是有可能导致德的行为的原因,请在每一项原因前选择您认为该原因对德的行为的决 定程度 非常重要 重 要 无 关 1. 他很自私,被宠坏了。 2. 他欺负别人,以获得同学们的注意。 3. 老师在全班面前提到他的成绩不好。 4. 他欺负别人,这样同学就不会欺负他。 5. 他讨厌上学。 6. 他讨厌他的同学和老师。 7. 他讨厌那个被他欺负的同学。 8. 电影,电视,和其他的媒体宣扬暴力行为。 9. 他的很多同学讨厌他,取笑他。 10. 他的朋友和他一样爱欺负别人。 11. 那些被他欺负了的同学没有告发他的行为。 12. 学校和老师没有及时并有效地处理 D 的不良行为。 13. 德的家长对他宠爱有加,管教不足。 提交 88 请仔细阅读下面的材料,并回答随后的问题。 明是一名在校五年的研究生。近期,他申请了一项奖学金,未获批准;他的一个同学得到 了这份奖学金。明怀疑是他的导师偏向这个同学,这位同窗作了弊,抢走了他该得的奖学 金。明向校方申诉未果。此后,明申请一所大学的教职,也没成功;对此他归咎于其导师 写的推荐信不力。不久后的某日,他走进本系的教授例会,开枪打死了他的导师,获奖的 那位同窗,和另外两位教授。然后,明来到两个街区以外的行政楼,找到当初处理他的申 诉的教务助理副校长。明开枪打死了这位副校长和一位在现场的学生;随后,明开枪自 尽。 以下是有可能导致明的行为的原因,请在每一项原因前选择您认为该原因对明的行为的决 定程度。 非常重要 重 要 无 关 1. 因为学校和就业的事情,明的精神压力很大。 2. 明没有寻找其它途径释放自己的挫折感和愤怒。 3. 教务副校长的处理不够公正,没有到位。 4. 明的性格有缺陷。 5. 明对奖学金一事太过斤斤计较。 6. 明的精神状态不稳定。 7. 明对社会公平抱有不切实际的幻想。 8. 电影,电视,和其他的媒体宣扬暴力行为。 9. 明身边的人都很自私,自我中心。 10. 明的朋友也舞枪弄棍,以暴力解决问题。 11. 明所处的环境允许拥有枪支,并有射击训练的场 所。 12. 明的导师和同窗确有作弊行为。 13. 明找不到可以公正裁决奖学金纠纷的第三方机构。 提交 89 APPENDIX E Coding Rubric of Open-ended Responses 90 1 Scape goat/selfenhance Anger, frustration, irritation @ parents/teacher, Make others feel bad, Embarrassed to revenge, humiliated, Selfenhancement, to make himself look better, Control Take out on others, pick the weak one Divert focus/attention Poor grade, ashamed of grades, consequence 讨厌同学,老 师,学校 面子,心情不好 Coding Rubric of Open-ended Responses to Dave/De (Bullying) Scenario 2 3 4 5 6 7 Parenting Teacher Peer Poor Personality, Low selfsocial/coping Mental esteem, skills health desire (brat) attention Get/need attention, Selfesteem, dignity, insecure at height, too high/low esteem, think himself is better than others Lonely, insecure, Feel abandoned/neglec ted, Unworthy, inadequate, inferior, loser 自闭,自尊心强, 自以为是 Social/coping skills, not get along with others, deal with problems/criticisms /anger, He does Poor self control/self discipline/patience, There is no one to talk to, no social network, no friends, No self reflection 认不清自己的缺点, 修养差, 缺少交流, 心理承受力差,不 会自我调节 Brat, have his way, Only-child, get his way, He is bully/mean, Selfish, selfcentered, inconsiderate, no respect to others, Psychological problems, bipolar, psy-disorder 心理不健全/扭曲/ 不平衡 孤僻,不乐观, 个人主义,小皇 帝,唯我独尊, 无 道德,无礼, 无心之失 虐待,暴力倾 向,恶作剧 脾气坏,钻牛角 尖,不知羞耻, 逆反心理 91 Parent/home: no love/affection, no attention/time, spoil, high expectation, wealthy, neglect/ignore, can’t rely on, no Ed./discipline/soc ialize Teacher mention grade/negative attention Mean peer, he was bullied, There is no correction from peers, Peers didn’t report him and let him get away with his misbehavior 同学放纵他 8 Social/cultur al 学校,社会(德 育)教育不足 Coding Rubric of Open-ended Responses to Michael/Ming (campus-shooting) Scenario 1 Self-enhance, Vindictive (Emotion, revenge, selfenhance) Anger, upset, frustrated, worked up, jealous(no one gets it if he doesn’t), pride hurt, feel cheated/ betrayed/ victimized/discriminate d/ignored/rejected/not appreciated Fear, avoid penalty, afraid of losing, regret killing, Hurt to revenge, to feel good, Blame others, take out on others, victims wronged him &responsible for his problems Prove self, empower self, no control/to take control, kill to fix problems, 找客观原因,推脱责 任,怨天尤人, 挫折 2 3 Social, communicat ion skills Personality, Belief/value Mental health School (award), Job Lack of skills/abilities to handle failure, No social/coping skills, bad relationship with people around, no friends, don’t get along, no self reflection, no self criticism, history of violence=a bully, not competent, isolated, no communication, 人际关系不好, 人情冷漠 读死书,书呆子 暴力倾向,爱欺 负人, 看不到自己的缺 点. 心理承受力差, 素质差,不能应 对失败,不会自 我调节/发泄 Pessimistic, hopeless, life goes nowhere, sense of loss, insecure, introvert, , skeptical, self-centered, selfish, rebellious, self esteem, feel like a loser, disappointed at self, Cloudy view/value of reality, social justice, antisocial, sociopath, false assumptions about justice, vanity, money, success, award, value school/award too much, Denial of loss/failure 计较得失/名利 不满社会,不 接受现实,只 看丑陋 自大,好胜, 期望过高,天 之骄子,过于 自信/自卑. hardworking but NO reward & NO recognition, Lose award/appe al, deserve it but lost Stress, fatigued, pressure, grad school, length of time at school, no job offer, 逆反心理,叛 逆 猜疑,不信任 别人,疑心, 执着,性格缺 陷,悲观,小 心眼,内向, 冲动,想不 开,主观,心 胸狭窄,脆 弱,心态不 好,计较, 梦想破灭 4 5 depressed, inhibited, psychologica l problems, emotional/m entally unstable, insane, schizophreni a, nuts, suicidal, 心理变态, 心理疾病, 一 时冲动/不平 衡,情绪不 稳定 6 不如人意, 学习压力, 无回报 92 7 8 9 10 Social, Cultural, Media Famil y, Parent childh ood Victim (advisor / fellow student/ adminis tration) Other s society/school /culture value grades but value neither morality nor personality ed., gun control Racism No other choice, no way out, no legal solution, no proper channel, no third party for further appeal Media/movie/ internet/mag azine 学校社会不 关心学生的 心理/德育教 育 childhoo d or past experien ce(he was ignored/ betrayed ; was cuddled too much, his past was too smooth & easy) Parents/ home: no commun ication, no trust, spoiled him & no disciplin e 没有阅 历,太 顺利 advisor: favoritism, no attention, no help, no communic ation, no recognition , unfair treatment Unfair treatment, misfortune, world against him, wronged him, cheated, Fellow student, cheater drug addict, offmeds, illness, girlfrien d, 打击,委 屈,自己 伸张正义
© Copyright 2026 Paperzz