Debate in Science: The Case of Acculturation

The acculturation paradigm of measuring assimilation, separation, integration and marginalization confuses dimensional and categorical conceptions of its constructs, fails to produce ipsative data from mutually exclusive scales, misoperationalizes marginalization as distress, mismeasures biculturalism using double-barreled questions instead of computing it from unicultural measures, and then tends to misinterpret and miscite this faulty science. Extensive published but widely uncited data cast doubt on claims that integration is preferred by minority groups or is beneficial for them. Such salient but unseen problems suggest that the community of acculturation researchers is biased and blinded by an ideology, probably the commendable ideology of liberalism, which advocates freedom of choice, tolerance, plurality, and redress of harm. Phenomenological observations that challenge the paradigm include the absence of studies of majority group acculturation, the well-replicated fact that minorities never prefer pure uniculturalism, the indistinctiveness of cultures, and the predominance of researchers, theory and data from similar Anglo-Saxon settler societies (USA, Australia, Canada). ACKNOWLEDGMENTS Helpful criticisms of early drafts came from Vali Ahmadzadeh, John Berry, Michael Bond, Friedrich Funke, Peter Guarnaccia, Uichol Kim, Maria Lewicka, Leonard Martin, Charles Negy, Yutaka Okura, Senel Poyrazli, Joachim Reimann, Ronald Taft, Eugene Tartakovsky, Joseph Trimble, Mehran Vali-Nouri, Ramón Valle, Rolf Van Dick, and Jie Weiss. Acknowledgement of helpful criticisms does not imply endorsement of the arguments and evidence presented here. Debate about Acculturation 2 But with far more subtlety does this mischief insinuate itself into philosophy and the sciences; in which the first conclusion colors and brings into conformity with itself all that come after though far sounder and better. Besides, independently of that delight and vanity which I have described, it is the peculiar and perpetual error of the human intellect to be more moved and excited by affirmatives than by negatives; whereas it ought properly to hold itself indifferently disposed toward both alike. Indeed, in the establishment of any true axiom, the negative instance is the more forcible of the two. Francis Bacon, New Organon, XLVI, 1620 Science advances by creating theories, then criticizing those theories, and then correcting or discarding them. This is a normal, necessary and inevitable aspect of science. Applied topics, like the psychology of intercultural contact, may influence public policy and thus effect the lives of millions of people. Those of us criticizing this kind of acculturation research have professional and ethical obligations to effectively inform the research community, especially considering that acculturation is a serious experience for many people, and that acculturation contexts can cause conflict, oppression, ethnic war, and genocide. Our contemporary world is in an acculturative crisis. Aboriginal peoples, migrants and other minorities suffer difficulties in almost all economically developed nations, and many millions have suffered or died in acculturative conflicts in Aceh, Afghanistan, Ambon, Angola, Armenia, Bosnia, Brazil, Cambodia, Cameroon, Chechnya, Chiapas, Congo, Corsica, Croatia, Cyprus, East Timor, Eritrea, Euskadi, Fiji, Georgia, Guatemala, Gujarat, Iraq, Israel, Ivory Coast, Kalimantan, Kashmir, Kazakhstan, Kosovo, Kurdistan, Lebanon, Liberia, Macedonia, Mali, Mynmar, Debate about Acculturation 3 Nagorno-Karabakh, Niger, Nigeria, Northern Ireland, Palestine, Peru, Philippines, Rwanda, Sri Lanka, Sudan, Tibet, Thailand, Tonga, Uzbekistan, West Papua, Xinjiang, Zambia, and Zimbabwe, plus the events of our current "clash of civilizations.” It is thus imperative that doubts about acculturation theories, methods, or findings be allowed into our discourse forum, even if that discomforts us. For the past century, the predominant conception of acculturation has presumed that minorities react to prolonged intercultural contact by assimilating to the dominant society, by separating from it, or by becoming bicultural, either successfully as bicultural integration or unsuccessfully as bicultural marginalization (Rudmin, 2003a; b). John Berry has been the most prolific and high profile contemporary scholar promoting this kind of conceptualization and has been a leader in developing a corresponding quantitative research paradigm, often referred to as “the Berry model.” This paradigm, though popular, is not without faults. In 1997, the journal Applied Psychology arranged a keynote article by Berry (1997a) followed by critical commentaries (Horenczyk, 1997; Kagitçibasi, 1997; Lazarus, 1997; Pick, 1997; Schönpflug, 1997; Triandis, 1997; Ward, 1997). For example, Triandis (1997, p. 56) argued that "the model is so complex that it is not testable," and others criticized the paradigm for lack of useful application. In 1998, the International Association for Cross-Cultural Psychology arranged a symposium on "A Critical Appreciation of Berry's Model" (Boski & Kwast-Welfeld, 1998; Schmitz, 1998; Weinreich, 1998). These critiques thus far focused on macro-model issues, and avoided problems of methodology, including constructs, confounds, and mistaken conclusions. Debate about Acculturation 4 In 2001, Rudmin and Ahmadzadeh criticized the psychometrics of the paradigm and raised ethical issues about faulty science on applied topics. In 2003, Berry and Sam replied and tried to justify the constructs, methods and findings. The present paper continues these critical discussions about faults in acculturation research and is written as a reply to Berry and Sam (2003). They are to be lauded for well articulating the fourfold acculturation theory, for developing a research paradigm which has produced a very large volume of data-driven literature which invites critical examination, and for now engaging in discussions with critics of their research. The present paper replies to their arguments in the spirit of scientific debate. Hopefully the outcome will be clearer understanding of the research issues that surround the topic of acculturation, to the ultimate end that our science will improve. Responses to personalized comments that deflect attention from essential matters of acculturation theory are presented as end notes. Acculturation research 1 that has not used the fourfold paradigm is omitted from the present discussion, though much of this has been critically examined elsewhere. For example, the large literature on Latino-USA acculturation has been criticized by Rogler, Cortes and Malgady (1991), by Negy and Woods (1992b) and by Hunt, Schneider and Comer (2004). The intent here is to broaden the debate by explaining the sequential observations and reasoning that led to the discovery of the faults and by showing how these lead to issues of ideology, research design, and perhaps new research paradigms. The goals are to promote better psychometrics, to provoke new thinking about acculturation, and to excite critical responses. Debate about Acculturation 5 As Kuhn (1962) would have predicted, a strong critique of a popular paradigm's methods and conclusions has had trouble passing peer review processes and coming to press. Researchers and editors experience such critique as personal criticism and may dislike that decades of research might need to be dismissed. But the faults in acculturation research are many, and once pointed out, are obvious and undeniable. The line of argument here, that these failings arise from ideological biases that blind the entire research community, is the most plausible explanation and the most exculpatory, compared to alternative explanations that question the competence or integrity of individual researchers and editors. Origins of the Psychometric Critique Rudmin's entry into acculturation research began as a late, stand-in supervisor for Merametdjian's (1995) thesis study of Somali refugees in Norway. She had used measures of acculturation developed by Sam under advisement by Berry (Sam & Berry, 1995) in which YES, NO answers to the issue of maintaining heritage culture and YES, NO answers to the issue of participating in the larger society define, respectively, the constructs of integration (YES, YES), assimilation (NO, YES), separation (YES, NO), and marginalization (NO, NO). Because the four constructs are mutually exclusive, agreement to items about one construct should impede agreement to the corresponding items about the other three. However, many of Merametdjian's respondents agreed to two or more scales about these mutually exclusive constructs, which is evidence of psychometric problems. This anomaly in a student's data led to the search and discovery of similar problems in other studies (Rudmin, 1996). For example, Kim (1988, p. 97) tabled mean scale scores for an acculturating group and two non-acculturating control Debate about Acculturation 6 groups. The three matched samples were: 1) Korean immigrants experiencing acculturation in Canada, 2) Koreans in Korea who had self-selected to emigrate to Canada but not yet done so, and 3) Koreans in Korea who had not sought emigration to Canada. Rudmin (1996) gave the English version of Kim's scales to another control group: 4) Norwegian students instructed to guess how they imagine Koreans in Canada might answer. On a five-point Likert scale, all four samples had similar mean scores: integration 4.1 ±0.2, separation 2.8 ±0.2, marginalization 2.7 ±0.3, and assimilation 2.1 ±0.1. The correlations between mean scale scores for each pairing of samples were .98, .96, .98, .99, .94 and .91, which average to r = +.97 (n = 6, p < .05), showing nearly perfect concordance among these four samples in their collective answers about Koreans acculturating in Canada. The scale standard deviations were also very similar, again showing concordance of r = +.97. When an acculturating group answers like non-acculturating control groups and like far away students with little knowledge and no experience of the two cultures in question, that is evidence of serious psychometric problems. Another unnoticed psychometric problem was evident in four studies reported in Applied Psychology by Berry, Kim, Power, Young and Bujaki (1989, p.199). Each study showed significant positive correlations between measures of mutually exclusive constructs, for example, between assimilation and separation for the Portuguese-Canadians (r = +.33, n = 117, p < .001). That is, when asked if they have positive attitudes towards Portuguese and Canadian cultures, respectively, respondents agreeing to the NO, YES construct also tended to agree to the YES, NO construct. This kind of problem was most dramatically displayed in the Journal of Cross-Cultural Psychology by Montreuil and Bourhis (2001, p. 709) who Debate about Acculturation 7 reported that measures of assimilationism (NO, YES) and segregationism (YES, NO) were correlated at r = +.60 (n = 637, p < .001). These five studies were all student thesis projects, each approved by a supervisor, a research committee, an examination committee, and then by a team of journal review readers and an editor. Thus, each study was critically appraised by 7 or 8 or more senior scholars, apparently none of whom noticed that significant positive correlations between measures of contrary constructs are evidence of psychometric problems. To have divergent validity, such measures should show strong negative correlations. But Berry et al. (1989, p.199) reported in Applied Psychology that assimilation (NO, YES) and separation (YES, NO) were almost perfectly uncorrelated at r = -.01 (n = 150, p > .05) for Koreans in Canada. Using acculturation scales modeled on this study, Safdar, Lay and Struthers (2003, p. 570) reported 14 years later, also in Applied Psychology, that assimilation (NO, YES) and separation (YES, NO) were again perfectly uncorrelated r = -.01 (n = 166, p > .05) for Iranians in Canada. Two studies reported in the International Journal of Intercultural Relations that used the same type of scales again found assimilation (NO, YES) and separation (YES, NO) to have near-zero correlations: JasinskajaLahti and Liebkind's (2000, p. 510) study of Russo-Finnish adolescents reported r = -.06 (n = 170, p > .05), and Pham and Harris's (2001, p. 289) study of Vietnamese-Americans reported r = -.01 (n = 138, p > .05). In a large study of 42 samples from 13 countries, Phinney, Berry, Vedder and Liebkind (2006, p. 98) again found assimilation and separation to have near-zero correlations (r = +.06, n = 5366, p < .05). Such well replicated lack of divergent validity for acculturation scales is evidence of systematic psychometric problems. Debate about Acculturation 8 The research community had not noticed in 30 years of using these acculturation constructs that they are mutually exclusive such that the resulting data should be ipsative rather than independent. This means that the scores generated by these scales should be negatively intercorrelated because agreement to one scale should reduce agreement to the other scales. It is not plausible that a respondent could validly give maximum agreement to all four scales, as would be possible if the constructs and the scales were independent, i.e., not mutually exclusive. Hicks' (1970) calculated that four ipsative measures have null intercorrelations of r = -.33 rather than r = .00 as happens with independent scales, and this was to be empirically demonstrated by Rudmin and Ahmadzadeh (2001). One consequence of ipsativeness is that it is invalid to analyze fourfold acculturation measures using multivariate methods such as factor analysis (Cornwell & Dunlap, 1994; Guilford, 1952; Johnson, Wood & Blinkhorn, 1988). Comparing Acculturation Measures The psychometric problems discovered and reported by Rudmin (1996) motivated the design of a new study proposed by Ahmadzadeh on the acculturation of Iranians in Norway (Rudmin & Ahmadzadeh, 2001). Four methods of measuring acculturation attitudes were used: 1) the established fourfold scales measuring a respondent's Likert agreement to double-barreled items that simultaneously ask about attitudes toward minority culture and majority culture, as exemplified by this integration item, "It is important to continue using my mother tongue even though I am learning Norwegian;" 2) standardized scoring of these scales, using each respondent's mean answer and standard deviation to transform item responses before scale summation such that acquiescence effects are neutralized and the Debate about Acculturation 9 scales become fully ipsative; 3) independent Likert measures of attitude toward the minority culture and attitude toward the dominant culture, as illustrated by the two items, "To know Iranian literature is part of my identity" and "Reading Norwegian literature gives me a new identity;" and 4) forced-choice measures of cultural preference which yield acquiescence-free and fully ipsative scale scores and which allow a multicultural option, as exemplified by an item about food, "I prefer: a) Iranian, b) Norwegian, c) both, d) from the whole world" (Rudmin & Ahmadzadeh, 2001, p. 48). These four methods of measuring the same acculturation attitudes of the same respondents produced very different results. Only one finding was replicated, namely, the uncorrected and the standardized marginalization scores both showed significant negative correlation with the Satisfaction With Life Scale (Diener, Emmons, Larsen & Griffin, 1985) and significant positive correlation with the Zung Self-Rating Depression Scale (Zung, 1965). However, a closer look at the marginalization items shows they had not been operationalized as the construct of rejecting both cultures (NO, NO). This conclusion was based on four observations. First, the literal meanings of the items were about social distress and indecision, not about a decided dislike of two cultures (Rudmin & Ahmadzadeh, 2001, p. 51): "Norwegians make us responsible for most of this country's problems." "It is difficult at present to find a friend you can trust." "I have difficulties in deciding whether to live like a Norwegian or like my co-nationals." "It is difficult to choose between my traditional way of living and the Norwegian way of living." Second, all 13 respondents who gave Likert agreement to this marginalization scale also gave Likert agreement to the integration scale, which is antithetical to Debate about Acculturation 10 marginalization. This should not be possible if marginalization were operationalized as saying NO, NO to minority culture maintenance and dominant culture participation and if integration were operationalized as saying YES, YES. Third, to show divergent validity, each of the marginalization items should have been negatively correlated with the integration, assimilation and separation scales. They were not. Fourth, if the marginalization items expressed rejection of the two cultures in question, then each should have been negatively correlated with the independent measures of attitudes toward the minority Iranian culture and toward the majority Norwegian culture. They were not. Presumptions, Constructs and Ideology These doubts about the marginalization items led to a closer examination of the origins and the historical evolution of the acculturation constructs and their operationalization. An explanation of why the marginalization items inquire about distress rather than about decisions to reject two cultures cannot be found by statistical methods. It was found in a 1989 footnote in Applied Psychology stating that the marginalization scale "was approximated by the scale of Marginality constructed by Mann (1958)" (Berry et al.,1989, p. 187). An explanation of that decision requires a search into the history of the paradigm. Thus, the critical force of Rudmin and Ahmadzadeh (2001) follows from trying to track the psychometric problems from the data, to the items, to the constructs, to the theory, to the history of the paradigm. Rudmin and Ahmadzadeh (2001) tried to articulate and criticize some of the presumptions of the paradigm. For example, to operationalize the marginalization construct (NO, NO) as distress caused by loss of cultural contact, it is necessary to Debate about Acculturation 11 impose three doubtful presumptions: 1) loss of cultural attachment followed from the respondents' decisions to reject the two cultures in question; 2) rejection of these two cultures implies rejection of all cultures; and 3) loss of cultural attachment is necessarily distressing. The misoperationalization of marginalization also requires an undeclared shift of perspective from the respondents’ points of view to the points of view of the two cultural groups. Respondents, from their own perspective, cannot be marginalized from groups they decided they do not want to belong to. It is only from the perspective of the groups that such people seem deprived of a cultural community and thus must suffer marginality. The NO, NO construct should not be conceived as "I am distressed because I decided to have no cultural community," but rather as "I prefer something other than those two cultures.” Whether or not that is distressing should be determined empirically. Distress is not a necessary aspect of disliking two cultures or of preferring something else. For example, Cohen (1956) presented theory and evidence that people who reject the minority and majority cultures are relatively free of the pathological traits that accompany ethnocentrism. Nash and Schaw (1963) argued that such people have secure self-identities in changing contexts and do not waste energy on psychological defense mechanisms. Mol (1963, p. 176) wrote that "rationality, objective observation, efficient management, [and] logical calculation require marginal attitudes.” Kim (1988, p. 170) found that Korean-Canadians who rejected both cultures were often successful, educated professionals who expressed multicultural attachment "to all humanity, regardless of culture and race.” Accordingly, Rudmin and Ahmadzadeh (2001, pp. 42-44) argued that the rejection of both cultural communities should be called "multiculturalism" or "cultural autonomy" Debate about Acculturation 12 since it implies a preference for a sub-culture, a third culture, self-actualization, or assertion of oneself as an autonomous individual in a liberal society. This kind of critical examination leads in sequence to the whole paradigm unraveling. If a preference for something other than the two cultures in question is "multiculturalism" or "cultural autonomy," then what is "marginalization?" It is failure to be accepted into the preferred cultural reference groups (Rudmin & Ahmadzadeh, 2001, pp. 43-44). Others have come to a similar conclusion (e.g., Campisi, 1947; Lewin, 1948; Marden & Meyer, 1968; Rothman, 1960; Slotkin, 1940; Taft, 1957; Voget, 1951). Berry has been engaged in a continuous scholarly effort, from 1972 to the present, trying to come to a satisfactory construct and scale for the NO, NO option (Rudmin, 2003b, pp. 36-37). Most recently, Berry (2003, p. 24) has himself come part way to accepting that marginalization is failure to enter the preferred reference group: "Although marginalization can be a strategy that people choose as a way of dealing with their acculturation situation, it can also result from failed attempts at assimilation." If marginalization is failure to enter preferred reference groups, then the person pursuing bicultural integration is most at risk of becoming marginalized because acceptance by two groups is more complicated and more doubtful than by one group. If so, then perhaps integration can be a stressful acculturation option. For example, Merametdjian (1995) and Rudmin and Ahmadzadeh (2001) found that all respondents reporting themselves to suffer marginalization also reported preference for integration. Sam (2000) found that immigrant adolescents in Norway who preferred integration suffered increased acculturative stress. Zajonc (1952) argued that the bicultural condition is distressing because the deep layers of the Debate about Acculturation 13 super-ego are threatened by the incompatibility of cultural norms, an argument echoed by Bochner (1982). Ichheiser (1949) argued that biculturalism is distressing because it sometimes requires inhibiting one's core cultural traits and because that can be misperceived by others as deception. Acculturation researchers have rarely noticed that biculturalism entails incompatibilities because the psychometric items usually ask only about surface behaviors which are amenable to code-switching, e.g., cuisine, language, and music (Berry & Sam, 2003). But many of the most important and defining aspects of culture, for example, religion, sexual norms, cleanliness, child-rearing, etc., are not open to code-switching because the norms of one culture preclude the practices of the other culture (Rudmin & Ahmadzadeh, 2001, p. 43). For example, it is not possible to integrate punitive child-rearing with non-punitive child-rearing, nor premarital virginity with premarital sexual license. Also, many aspects of culture are regulated by civil laws that criminalize cultural practices pertaining, for example, to dress, foods, child discipline, marriages, sexual behavior, drug use, weapons, gambling, etc. Biculturalism, in fact, is a very constrained concept and cannot be realistically practiced or promoted as a universal panacea for acculturation problems. Faulty acculturation theory is not limited to the community of contemporary researchers. Rudmin and Ahmadzadeh (2001) discovered several fourfold acculturation taxonomies developed prior to, or independently of, the contemporary fourfold paradigm. This led to the hypothesis that the faults in the contemporary paradigm might also be found in the work of earlier scholars. This led to the research and writing of a critical history of acculturation encompassing 68 taxonomies (Rudmin, 2003b), now expanded to 126 (Rudmin, 2003a). This Debate about Acculturation 14 historical research found that, yes, many of the faults in the contemporary paradigm are common to the extended, historic community of scholars, including such scholars of renown as Thomas and Znaniecki (1918), Robert Park (1928), and Kurt Lewin (1948). Thus the question arises of why the most prominent and competent scholars of acculturation, both past and present, have so many faults in their research, faults that violate well known norms of psychological science but that have nevertheless gone for decades unnoticed by a very large community of researchers, editors, teachers, and students. Such a situation would seem to have its explanation in a shared ideology that biases and blinds the entire community (Myrdal, 1969). Thus, the next step in this critical sequence is to see if the faults in acculturation research suggest an underlying ideological bias. Construct Universality Ideologies entail over-generalizations. That seems to be the case in the claim that all acculturating people experience themselves reacting to issues of 1) cultural maintenance and 2) positive intergroup relations. These two issues and the subsequent four constructs seem to have been imposed by researchers on to the phenomena of acculturation as argued by Bhatia (2003). But Berry (2003, p. 28) has argued that they arise universally from the phenomena: "During the course of this research, these two issues have moved from being an emic for only one group to being an emic for other groups and eventually to being a derived etic (perhaps a universal concept) for many groups during their intercultural contact." The claim of universality, that these are "two general issues facing all acculturating peoples," first appeared in 1989 in a summary of results from Debate about Acculturation 15 3 Australian Aboriginal samples, 9 Canadian Native samples, 3 Euro-Canadian samples, and 3 Canadian immigrant samples (Berry et al., 1989, p. 185). A total of 18 samples, mostly aboriginal people, in two similar Anglo-Saxon societies, is not a basis for claiming human universality. Furthermore, the origin of the two issues that define the four constructs was not reported as an emic discovery, but was declared as an etic assertion: "The model is based on the observation that in plural societies, individuals and groups must confront two important issues. One pertains to the maintenance and development of one's ethnic distinctiveness in society . . . The other issue involves the desirability of inter-ethnic contact" (Berry et al., 1989, pp. 186-187, bolding and underlining added). Second, acculturation research routinely generates data demonstrating that one or both of the defining issues are not fundamentally important to the people under study. As argued by Rudmin and Ahmadzadeh (2001, p. 46), positive correlations between any two of the acculturation scales are evidence that one of the fundamental issues is being disregarded by the respondents. For example, Berry et al. (1989, p. 199) reported in Applied Psychology that there were strong, positive correlations between integration (YES, YES) and separation (YES, NO) for the French-Canadian sample (r = +.48, n = 49, p < .001) and the Hungarian-Canadian sample (r = +.58, n = 50, p < .001). That means that respondents saying YES to intergroup relations also tended to say NO to intergroup relations, thus contradicting themselves on the defining issue of intergroup relations. Such self-contradiction implies that intergroup relations are not an important or fundamental issue for these respondents. Debate about Acculturation 16 Third, the paradigm does not fit the socio-historical context of the Black minority in the USA, as argued by Pettigrew (1988). Black acculturation is constrained by the historic racism of US society: "Save for West Indian migrants, blacks do not experience the traditional pull between another nation's culture and American culture, for which the assimilation and pluralism concepts were specifically fashioned. Nor do they face charges of being 'un-American' and not belonging. Their marginal position is more complex, and the use of theory and concepts developed for the immigrant experience obscures more than it enlightens. Blacks' marginalization is created by their being a long-term, integral component of American society, while, at the same time, being denied the privileges that otherwise accrue to such a central position" (Pettigrew, 1988, p. 24). Berry, Poortinga, Segall and Dasen (1992, p. 279) have cited this argument but apparently did not notice that it discounts claims that the fourfold acculturation constructs are universal. Finally, the presumption of universality cannot be correct if one considers people in the majority groups, which is most people, adopting foreign behaviors, for example, Russians playing ice hockey, Japanese eating curry rice, or Americans surfing the World Wide Web. For culturally secure majority groups, maintenance of cultural identity and participation in society are not issues when they acculturate. Although minority group acculturation has long been studied, the processes of majority group acculturation have only recently been studied psychometrically, e.g., Geschke & Mummendey (2005). Such a strong asymmetry in acculturation research is remarkable because it stands juxtaposed to often quoted definitions that acculturation is a two-way process of change (e.g., Redfield, Linton & Herskovits, Debate about Acculturation 17 1936, p.149) and juxtaposed to everyday evidence that the dominant majority do adopt practices from minorities and from foreign cultures. Consider such acculturatively acquired aspects of US culture as peanut butter, pizza, bagels, French fries, ketchup, tabasco, tacos, tobacco, wine, jazz, rock-n-roll, salsa, yoga, lacrosse, hockey, skiing, Boy Scouts, cars, autobahns, jet planes, Easter eggs, Thanksgiving turkey, Christmas trees, "Massachusetts," "Chicago," and "Los Angeles.” In the history of psychology, only Alexander Chamberlain in the 1890s had a research focus on majority group acculturation, documenting the impact of Native American and Black cultures on mainstream US society (Rudmin, 1990a; b; 1999). In addition to adopting foreign foods, words, technologies and recreations, there are many other types of majority group acculturation worthy of study: a) Mixed marriage entails individuals from the majority population taking spouses from a minority group or foreign country and adopting aspects of the spouses' cultures. b) Conversion entails individuals from the majority population adopting the religion of a minority group or foreign society. c) Foreign posting entails majority group individuals serving as overseas representatives for their company, government, or church, which may require them to learn new cultural skills even if they dislike the culture they are acquiring. d) Going-native entails individuals from the dominant group adopting minority culture behaviors and values, contrary to colonial expectations arising from cultural dominance and racism. e) The "-ophile" phenomena entail individuals becoming enchanted by often distant foreign cultures for their interesting qualities, as exemplified by Sinophiles, Francophiles, Russophiles, etc. f) Re-minoritization entails individuals from the majority group deciding to identify with, learn, and practice the forgotten minority heritage culture of Debate about Acculturation 18 their ancestors. g) Sub-culture creation entails individuals from the majority population rejecting their conventional culture and trying to create a new culture, as exemplified by utopians, communitarians, hippies, or punkers. That so many acculturative phenomena have been unnoticed and unstudied by psychologists suggests an unseen shadow of ideology controlling the entire research community. Mismeasure of Marginalism Ideology may also explain why the misconceptualization and the misoperationalization of marginalism has been so systematically unnoticed. Historically, the fourfold paradigm began with the concept of the Marginal Man first proposed by the American sociologist Park (1928) and further developed by Stonequist (1935) and Glaser (1958), all claiming that marginality describes the bicultural condition of minorities (Berry, 1970, pp. 239-241; Sommerlad & Berry, 1970, p. 24). For example, Glaser (1958, p. 34) wrote that bicultural competence results in a marginal person who "favors a pluralistic society in which he can feel identified with several ethnic groups." Berry explained in 1970 (pp. 240-241) that socio-cultural marginality arises from "a set of conditions characteristic of culture contact between two groups, one dominant over the other" and that psychological marginality is "characteristic of persons in the marginal situation. . . Traits thought to be included in this pattern are: aggression, suspicion, uncertainty, victimization-rejection, anxiety, and a lack of solidarity.” The four types of acculturation follow from this marginalizing bicultural situation: "Attitudes to modes of relating to the dominant society also enter into the complex set of traits claimed to characterize the Marginal Man. According to marginality theory, retention of a marginal relationship [now called Debate about Acculturation 19 marginalization], a 'passing' into the dominant society [now called assimilation], and a rejection of it (while reaffirming one's own society) [now called separation] are the three alternatives available as long-term adjustments to the marginal situation" (Berry, 1970, p. 241). Berry (1970, p. 242) added a fourth alternative: "moving, as a group, into the dominant society, while retaining a separate group identity" [now called integration]. Thus, questions about cultural identity and positive group relations are presumed to begin with, and to be caused by, the distressing state of minority group marginality. This causal chain is nicely illustrated in the Handbook of Cross-Cultural Psychology (Berry, 1980, p. 260), showing minority people of traditional culture coming into acculturative contact with the dominant culture, causing the minority to experience marginality, causing a crisis of choice between conflicting cultural identities and attitudes, causing the minority to choose between acculturation options of assimilation, rejection, or integration. Marginalization as the NO, NO mode of acculturation was not mentioned in 1970 and 1980 because it had not yet been conceived as an option arising from a decision, but instead marginality was conceived as the default consequence of continuing in the original situation of bicultural marginality. The American anthropologist Born (1970) had independently theorized that four modes of acculturation are adaptive responses to the acculturative stress of bicultural marginality. The measure of marginality is yet more complicated. In 1958, Mann proposed a measure of the marginal type of personality presumed to be caused by the bicultural context of minority individuals. None of Mann's marginality items has acculturative content, e.g., "Life is a strain for me," or "I regret the decisions I have Debate about Acculturation 20 made" (Berry, 1976, p. 178). This measure became one of two dependent measures of acculturative stress to be predicted by modes of acculturation (Berry, 1976, pp. 177-178). Thus, Mann's measure of marginality should not be confused with marginalization conceived as the NO, NO mode of acculturation. But these two constructs have been confused, as will be shown. In 1972, Berry, Evans and Rawlinson logically deduced that there should be a type of acculturation defined by minorities choosing to reject minority culture and to reject the dominant society, but it was called "inherently contradictory" (p. 29) for a cultural group to so choose its own demise. In 1976, this type of acculturation was called deculturation (NO, NO), but a proviso said: "Common sense and pilot work indicated that such an outcome was not chosen by anyone" (Berry, 1976, p. 180). In other words, "It should be noted that attitude items suitable for the 'deculturation' response are almost never accepted in a population; thus no scale has been developed to assess it" (Berry, Kalin & Taylor, 1977, p. 132). In 1983, the deculturation (NO, NO) construct was confounded by blending it with acculturative stress: " . . . it is accompanied by a good deal of collective and individual confusion and anxiety [and] is characterized by striking out against the larger society and by feelings of alienation, loss of identity, and what has been termed acculturative stress. This option is Deculturation, in which groups are out of cultural and psychological contact with either their traditional culture or the larger society. . . When stabilized in a non-dominant group, it constitutes the classical situation of 'marginality' (Stonequist, 1935)" (Berry, 1983, p. 69). In 1989 in Applied Psychology, this confounded deculturation construct was renamed "marginalization" and a Likert scale was devised to measure it (Berry et al., 1989, p. 188). A footnote explained that this marginalization scale was Debate about Acculturation 21 "approximated by the scale of Marginality constructed by Mann (1958)" (Berry et al., 1989, p. 187). Two example items for the new scale were presented, 1) "These days it is hard to find someone you can really relate to and share your inner feelings and thoughts;” 2) "Politicians use national pride to exploit and to deceive the public," showing that marginalization was not operationalized as a decision to reject both cultures but was instead operationalized like Mann's measure of marginal personality (Berry et al., 1989, p. 193). During this devolution of marginalization into a confounded construct, passive voice began to be used to describe the decision processes that are presumed to define the modes of acculturation. Passive voice creates ambiguity and thus serves ideological functions of keeping researchers unaware of confusions in their constructs. For example, passive voice was used in 1977: "(1) Is it of value to retain one's traditional culture? and (2) It is of value to work with the larger society in pursuit of common goals?" (Berry et al., 1977, p. 31). Twenty years later, the two defining issues were still presented in the passive voice: "Issue 1: Is it considered to be of value to maintain one's identity and characteristics? Issue two: Is it considered to be of value to maintain relationships with the larger society?" (Berry, 1997a, p. 10). If the two issues defining the acculturation constructs were written in the active voice, they would indicate who is choosing the acculturation modes. If active voice were used, researchers might have noticed the oxymoron of supposing that cultural minorities could prefer that their cultural preferences be denied or could prefer to become cultureless and marginalized. In sum, marginalization (NO, NO) is defined as a decision to reject both cultures, even though logic, common sense, and pilot studies had shown that minorities would not make such a decision. The use of passive voice leaves it Debate about Acculturation 22 unclear who, if anyone, decided both cultures should be rejected, and then the measure is not about rejection of both cultures but about the distress that is tautologically presumed to follow from culturelessness. Thus, marginalization is a confounded construct, and many citations, as will be shown, mistakenly claim that positive correlations between marginalization and acculturative stress are evidence that the marginalization (NO, NO) causes maladaptation. Such correlations should be interpreted as convergent validity for similarly operationalized measures. Faulty Psychometrics Another systematic bias in acculturation research is the wide spread use of survey questions that violate well known norms of psychometrics. The two founders of modern psychometrics, Thurstone (1928, p. 545) and Likert (1932, p. 45), both argued against double-barreled questions, as have virtually all subsequent psychometric textbooks. Rudmin and Ahmadzadeh (2001, pp. 44-45) focused criticism on the use of double-barreled questions. Asking about two cultures in one question results in items that are long, linguistically complex, and often requiring words of negation. These problems are well illustrated with Kim's (1984, pp.157167) acculturation items, which were recommended as model items by Berry et al. (1989, p. 193) and which have been widely copied by others (e.g., Abouguendia & Noels, 2001; Almyroudis, 1991; Ataca, 1998; Aycan & Kanungo, 1998; BenetMartínez & Haritatos, 2005; Berry, Phinney, Sam & Vedder, 2006; Bonifero, 1994; Eschel & Rosenthal-Sokolov, 2000; Fang, 1998; Hocoy, 1999; Inoue & Ito, 1997; Jasinskaja-Lahti & Liebkind, 2000; Jasinskaja-Lahti, Liebkind, Horenczyk & Schmitz, 2003; Krishnan & Berry, 1992; Kwak & Berry, 2001; Neto, 2002; Orr, Mana & Mana, 2003; Patridge, 1988; Pham & Harris, 2001; Roccas, Horenczyk & Schwartz, 2000; Debate about Acculturation 23 Safdar, 2002; Safdar et al., 2003; Sam, 2000; Sands & Berry, 1993; Schmitz, 1992;Tartakovsky, 2002; Virta, Sam & Westin, 2004; Young, 1984; Ziabakhsh, 2000).The worst of Kim's (1984, pp. 157-167) items are about history, herepresented with one proposition per line and with negation words underlined:1) Integration (YES, YES): For students who were raised in Canada,I would encourage them to take both Korean and Canadian historybecause it's important for them to know the history of both countries.2) Assimilation (NO, YES): For students who were raised in Canada,I would encourage them to take a course in Canadian history,but not in Korean historysince it has no utilityor value in Canada.3) Separation (YES, NO): For students who were raised in Canada, I would encourage them to take a course in Korean history, but not in Canadian history since there isn 't much worth learning about.4) Marginalization (NO, NO): Taking a course in history is a waste of time since it does not help you to learn anything practical or to get a job.These four items have a mean of 29 words and a mean of 3.75 barrels. Incomparison, Angleitner, John and Löhr (1986, p. 83) examined 1624 items from 10multi-scale personality inventories and found the worst sub-scales to have a mean of16 words and 2.5 barrels. Debate about Acculturation 24 Berry and Sam (2003, p. 66) now acknowledge that fourfold acculturationitems are double-barreled, but assert that that is necessary in order to capture thedouble-barreled aspect of acculturation. However, many acculturation studies havedemonstrated that double-barreled items are not necessary (e.g., in chronologicalorder, Campisi,1947; Szapocznik, Kurtines & Fernandez, 1980; Oetting & Beauvais,1990; Hutnik, 1991; Ward & Kennedy, 1992; Sayegh & Lasry, 1993; Birman, 1994;Donà & Berry, 1994; Ryder, Alden & Paulhus, 2000; Rudmin & Ahmadzadeh, 2001).One of the consequences of using double-barreled items is that respondentsmay focus on only one of the propositions, resulting in excessive agreement and inself-contradiction. Berry et al. (1989, p. 195) have explained that for each culturaltopic, like history, the corresponding questions for the four scales should bepresented together: "To reduce acquiescence tendencies, the four items werepresented in a single group where their contrasting meanings would be readilyapparent, as it would obviously be contradictory to agree to all four.” Rudmin andAhmadzadeh (2001) demonstrated several different ways that four-fold acculturationmeasures nevertheless show respondent self-contradiction and excessiveacquiescent agreement.Berry and Sam (2003, pp. 65-66) now argue that concerns about suchproblems are unjustified because "acculturation involves complexity, uncertainty,ambivalence, and many other psychological qualities that make such a Cartesianview of human behavior too simplistic for its proper study.” However, the fourfoldacculturation constructs are defined by a 2X2 Cartesian view of human behavior ashas been frequently illustrated in 2X2 figures defining the constructs. Furthermore,many phenomena measured in psychology are complex, uncertain, and ambivalent,and it is exactly those kinds of phenomena that require psychometrics to be most Debate about Acculturation 25 Figure 1: Dimensional depiction of the fourfold acculturation constructs based on positive or negative attitudes towards minority (M) and dominant (D) cultures: +M+D (biculturalism or integration), -M+D (assimilation), +M-D (separation) and -M-D ("marginalization" or multiculturalism). Hypothetical data show person Pi as an integrationist and person Ps as a