National Stereotypes and Their Relationship to Self-Perceptions in Irish, Italian, and American College Students

This study examined self perceptions, and, their relationship to heterostereotypes and autostereotypes in three groups: Irish students from the University Colleges of Dublin, Galway, and St. Patrick's College in Dublin; Italian students from theUniversity of Palermo in Sicily; and American students at the Community College of Rhode Island, Bryant College, and the University of Rhode Island. It was predicted that, as a reflection of greater cultural homogeneity, a higher degree of within-group similarity would be found in the self perceptions of both Italian and Irish students than in the self perceptions of American students. It was further predicted that differences would be found across the three groups in the saliency of certain self-perceived "traits 11 • Finally, it was predicted that heterostereotypes would show a substantial degree of correspondence both with autostereotypes and self perceptions. The samples consisted of 186 Irish students, 89 males and 97 females; 179 Italian students, 83 males and 96 females; and 190 American students, 90 males and 100 females. Students in each group responded to two open-ended questions asking them to describe, in ten adjectives or fewer, their picture of the typical 11 member of the other two groups. Ethnic stereotypes of Irish, Italians, and Americans thus elicited were then compared to each group's picture of its own "typical member" (autostereotype), and to the self perceptions of group members. Autostereotypes were measured by asking students to select, from the 81-word checklist of the Activity Vector Analysis (AVA) all those adjectives which best describe the typical member 11 of their own cultural group. Self perceptions were measured in two ways: the subjects responded to the AVA adjective checklist a second time by choosing all those adjectives which best describe the person you really are 11 • Subjects also completed the 126-item, forced-choice Myers-Briggs Type Indicator (MBTI). The AVA yields scores on four vectors: aggressiveness, sociability, emotional control and social adaptability. Ipsative scores on these four vectors combine in different ways to form 258 possible personality profile 11 descriptions. The MBTI yields scores on four dichotomous preferences: introversion versus extraversion; sensing versus intuition; thinking versus feeling; and judging versus perceiving. A subject's preference within each dichotomy yields a four-letter type description 11 ; there are sixteen possible types. Very little correspondence was found among heterostereotypes and either autostereotypes or self-perceptions, except for a few isolated traits. However, distributions of MBTI scores for both the Irish and Italian samples showed greater homogeneity than either the American sample or the normative U.S. sample of 3860 college students used for comparison purposes. The Irish sampl~ was found to have a significantly greater proportion of introverted, intuitive, feeling, perceptive 11 types while the Italian sample had a significantly greater proportion of "extraverted, sensing, thinking judging 11 types, than either the American sample or the normative U.S. sample. American students showed a greater-than-expected frequency of sensing feeling 11 types than in the U.S. normative sample but overall the American sample was distributed more heterogeneously across the MBTI types than the Irish and Italian samples. Tight clusters for the self-perceptions of students as measured by the AVA profiles were also found for the Italian students, with 84% of the sample clustered tightly around either one of two AVA pattern shapes. The AVA distributions of both Irish and American students were more heterogeneous, with approximately 65% of each sample found in three clusters on the AVA pattern universe. ACKNOWLEDGMENTS This dissertation is dedicated, with the deepest gratitude, affection, and esteem, to my major professor, Dr. Peter F. Merenda. Words cannot adequately convey my thanks for his unfailing patience, wisdom, and support, not only duri~g the progress of this research, but in the eight years it has been my privilege to be his student. To each of the members of my committee, to Dr. Stanley Berger, Dr. Albert Lott, and Dr. Richard Purnell, for the time, effort, and excellent advice generously given during the successive phases of this research, I extend my deepest appreciation. Research is always a cooperative effort, but the implementation of this particular study would have been impossible without the extraordinary measure of help received from colleagues in Ireland and Italy who obtained samples and collected data for me in those two countries. I offer very special thanks to Dr. Patricia Fontes, research scientist and lecturer at the Educational Research Centre, St. Patrick's College, Dublin; Dr. Christopher Sims, lecturer at the University College of Dublin; Ms. Pamela Merenda, lecturer at the University of Palermo, Sicily; and Ms. Rosalia Sparacino, lecturer at the University of Palermo, Sicily. The translation into Italian of the test instruments used in this study was a crucial step in the research. For the time and painstaking care they generously devoted to this task, I am indebted to my husband, Dr. Santo Salvatore, and to Dr. Remo Trevelli, Professor of Italian in the Department of Languages, University of Rhode Island. vi I also extend my warmest thanks to Dr. William Pacitti, Ms. Jane Carey, Mr. Bob Heder, and Dr. Santo Salvatore for kindly allowing students in their classes at the Community College of Rhode Island and at Bryant College to participate in this research. I particularly wish to thank Mr. Katchadoor N. Kazarian, President of Walter V. Clarke Associates in Providence, Dr. Peter Merenda, and the office staff at Walter Clarke Associates for making available to me computer facilities for the scoring of the AVA. To Dr. Charles Collyer of the Psychology Department and Dr. Leo Carroll of the Sociology Department, University of Rhode Island, for generously giving their time to serve as additional outside mambers of my committee, and for their thoughtful comments on this dissertation, I give my thanks. Finally, I would like to thank my husband, Dr. Santo Salvatore. The never-failing help and support he has given me throughout every phase of this research is truly beyond measure, as is my abiding love and gratitude.

It was predicted that, as a reflection of greater cultural homogeneity, a higher degree of within-group similarity would be found in the self perceptions of both Italian and Irish students than in the self perceptions of American students.
It was further predicted that differences would be found across the three groups in the saliency of certain self-perceived "traits 11 • Finally, it was predicted that heterostereotypes would show a substantial degree of correspondence both with autostereotypes and self perceptions.
The samples consisted of 186 Irish students, 89 males and 97 females; 179 Italian students, 83 males and 96 females; and 190 American students, 90 males and 100 females.
Students in each group responded to two open-ended questions asking them to describe, in ten adjectives or fewer, their picture of the 11 typical 11 member of the other two groups. Ethnic stereotypes of Irish, Italians, and Americans thus elicited were then compared to each group's picture of its own "typical member" (autostereotype), and to the self perceptions of group members.
Autostereotypes were measured by asking students to select, from the 81-word checklist of the Activity Vector Analysis (AVA) all those adjectives which best describe the 11 typical member 11 of their own cultural group.
Self perceptions were measured in two ways: the subjects responded to the AVA adjective checklist a second time by choosing all those adjectives which 11 best describe the person you really are 11 • Subjects also completed the 126-item, forced-choice Myers-Briggs Type Indicator (MBTI). types while the Italian sample had a significantly greater proportion of "extraverted, sensing, thinking judging 11 types, than either the American sample or the normative U.S. sample. American students showed a greater-than-expected frequency of 11 sensing feeling 11 types than in the U.S. normative sample but overall the American sample was distributed more heterogeneously across the MBTI types than the Irish and Italian samples. Tight clusters for the self-perceptions of students as measured by the AVA profiles were also found for the Italian students, with 84% of the sample clustered tightly around either one of two AVA pattern shapes. The AVA distributions of both Irish and American students were more heterogeneous, with approximately 65% of each sample found in three clusters on the AVA pattern universe.

ACKNOWLEDGMENTS
This dissertation is dedicated, with the deepest gratitude, affection, and esteem, to my major professor, Dr. Peter F. Merenda  Broadly viewed, the proposed study focuses on several questions which can be stated in a general way as follows: In the face of increased communication, contact, and sophistication in contemporary society, do clear-cut ethnic stereotypes still exist?
If _ so, are there differences in the willingness of particular groups to stereotype other groups? For example, are American college students less willing to stereotype Irish or Italians now than they have been in the past?
If members of a cultural group do hold clear-cut stereotypes of either their own or other national groups, are there sex-related differences in these stereotypes?
For example, do Irish women have a sense of national identity which differs from that of Irish men? Do Italian men perceive the "typical" American differently from Italian women?
If individuals hold an ethnic as well as an individual identity, to what extent are these two perceptions congruent? That is, to what extent do individuals perceive themselves as being "typical'' of their own culture? Further, are there cross-national differences in the degree of congruence of the "real self" with the national autostereotype?
Are American college students, for example, more "alienated" from what they perceive as "typical" of their own culture than, say, Irish college students?
If individuals hold an ethnic as well as an individual identity, what is the effect of acculturation on this ethnic identity? For 2.
example, will Irish Americans of the third or fourth generation perceive the "typical" Irish person differently than the Irish national perceives himself?
If clear-cut national stereotypes do exist, to what extent do these stereotypes accurately reflect self-perceptions of group members?
The possibility that stereotypes may accurately reflect the selfperceptions of cultural group members carries with it the assumptions that there will be both within-group similarities and cross-group differ-enc_ es in such self-perceptions. This raises two further questions: To what extent will members of the same cultural group show agreement or congruence in their self-perceptions? To what extent will the self-perceptions of one cultural group differ in nature from the selfperceptions of another cultural group? That is, will cultural groups differ from one another in the saliency of certain subjectively-perceived "characteristics" or "traits''?
(For the purposes of this study, the perception of the members of a national group of either their own or another national group will be referred to as a national stereotype. Consistent with the terminology used by previous researchers (Triandis, 1972;Osgood, 1976) a group's perception of its own general characteristics is termed an autostereotype and a group ' s perception of the general characteristics of another national group is termed an heterostereotype.) 3.

CHAPTER I
Review of the Literature While stereotypes held by one culture or ethnic group with respect to another have been rather frequently studies by social psychologists, the degree of accuracy of such stereotypes has less often been examined.
This relative dearth of research interest is probably due, in part, to a bias on the part of social psychologists. As Middlebrook  observes: 11 Unti l quite recently, the statement that all stereotypes are false and irrational has almost been a truism in social psychology.
Some social psychologists have even defined the term as . . . an inaccurate, irrational over-generalization which, once formed, persists even in the face of contradictory evidence ... [p. 122]. 11 This persisting tendency on the part of many social psychologists to view ethnic stereotypes as generally negative, inaccurate over-generalizations and to treat the term as almost synonymous with the term 11 prejudice 11 can be traced back historically to the earliest delineation of the concept by Lippmann in 1922.
The noted newspaper commentator and author Walter Lippmann introduced and defined the concept of stereotype in his 1922 book, Public Opinion. Essentially, Lippmann characterized stereotyped thinking as rigid, factually incorrect over-generalizations applied to some aspect of the social environment and arrived at through faulty reasoning.
Lippmann's theoretical framework set the stage for a long series of empirical investigations of stereotypes which used all or part of his definition of a stereotype as their point of departure.
In 1933, Katz and Braly carried out the pioneer empirical study 4.
of ethnic stereotypes and, in the process, established the methodological model or paradigm which has since been followed by the majority of workers in the field.
In their now-classic study, Katz and Braly (1933) asked 100 Princeton students to describe 10 ethnic groups by selecting from a list of 84 traits "as many adjectives as you think are necessary to characterize these people adequately." Subjects were then asked to go back and star the five words in each list that "best described" the ethnic group in question. Only the starred words were used in the final analysis of data . Stereotypes of Blacks, Germans, Jews, Italians, English, Irish, Americans, Japanese, Chinese and Turks were thus elicited from Princeton students who, according to Katz and Braly, showed 11 a very high degree of agreement on the assignment of traits to ethnic groups." This research procedure has, incidentally, been repeated at Princeton twice since, in the ensuing years, and provides an interesting measure of changes, both in the willingness of students to stereotype, and in the content of the stereotypes themselves.
Thus Gilbert, in 1950, found that Princeton students showed considerably less agreement in the assignment of traits to ethnic groups, and interpreted this as evidence for a "fading out" of the tendency to stereotype (Gilbert, 1951).
In 1969, Karlins, Coffman & Walters, on the other hand, again using the identical original procedure used by Katz and Braly, found that Princeton students showed greater agreement in the attribution of traits 5.
to certain groups than Gilbert had found, but that students now assigned "new" traits which, in their opinion, indicated "more careful thinking about ethnic generalizations than their counterparts of the 1930 1 s [p. 14]. 11 In addition to this trilogy, a considerable number of other studies have been carried out in the ensuing years which have tended to focus on stereotypes of minority groups and have adopted both the Katz and Braly methodology and their definition of a stereotype as a "fixed impression which conforms very little to the facts it tends to represent and results from our defining first and observing second [p. 81]." More recently, however, Brigham (1971) has, in an extensive review of the concept of ethnic stereotypes, challenged this view. Among other objections, he feels that 11 characterizing something as invalid necessarily implies that a validity criterion is available. Yet, with reference to the actual distribution of traits within ethnic groups, empirical evidence is exceedingly scanty. Thus, in most cases, no criteria are available for assessing the factual validity of an ethnic generalization [p. 17].II 6.
traits [p. 285]. 11 The same point is made by the noted anthropologist Margaret Mead (]956) who, however, may have overstated the case somewhat when she remarked that "the elements which make up such stereotypes are accurate but incomplete descriptions of national character in particular crosscultural contexts [p. 222]. 11 In turning now to the few empirical studies which have actually compared heterostereotypes with autostereotypes or stereotypes with self-perceptions of a given group, some support for the 11 kernel of truth'' hypothesis is found. For example, people from the Noahkali region of Pakistan were stereotyped by Pakistanis from other regions as more pious (among other traits), and their frequency of prayer was found to be higher than that of other Pakistani sub-groups (Shuman, 1966). Triandis and Vassiliou (1967) studies the effect of frequency of contact on the autostereotypes and heterostereotypes of three groups of American subjects and three groups of Greek subjects _ . The groups were constituted on the basis of amount of contact with the other culture.
The two maximum contact groups, for example, consisted of Americans working in Greece in jobs which brought them into frequent direct contact with Greeks, and Greek students studying in America. There were two medium contact groups and two minimum contact groups also. Triandis and Vassiliou found that contact between the two groups had a differential effect on the autostereotypes held by Americans, and on the heterostereotypes held by each group with respect to the other. Amount of contact had no effect on the autostereotype held by the Greeks about themselvesall three groups showing essentially the same self-perception.

7.
Autostereotypes of Americans in the maximum-contact group, however, were generally more favorable than the autostereotypes of Americans in the other two contact groups. Apparently this was due to the positive feedback Americans received about themselves from Greeks with whom they were in a lot of contact. Further, the authors found substantial agreement between autostereotypes and heterostereotypes for both groups (one measure of validity). Finally, in an interesting finding, Greeks in maximum contact with Americans tended to hold more favorable stereotypes of -them than either the moderate or minimum-contact Greeks. On the other hand, Americans in maximum contact with Greeks tended to hold less favorable stereotypes of Greeks than their moderate-and minimum-contact American counterparts. The authors conclude: 11 The present data suggest that there is a 'kernal of truth' in most stereotype when they are elicited from people who have first-hand knowledge of the group being stereotyped [p. 324]. 11 Another study which found "relatively high 11 agreement between autostereotypes and heterostereotypes was reported by Abate and Berrien (1967). They had two large groups of Japanese and American college students rate behavior descriptions drawn from the Edwards Personal Preference Schedule (EPPS) on the basis of 11 how well these descriptions fitted Japanese and Americans in general". In addition, Abate and Berrien attempted to assess the actual characteristics of smaller samples of students, drawn from the larger groups, by having them complete the EPPS on the basis of the perception of the "real self 11 • Although they found relatively high, though differing, degrees of agreement between auto-and heterostereotypes for both groups, they found less agreement between 8.
stereotypes and vereotypes (actual characteristics), especially for the Japanese sample. Abate and Berrien interpret this discrepancy between national autostereotype and "actual characteristics" by suggesting the possibility that "respondents in both countries tended to give a stereotype of the socially accepted rather that the 11 typical 11 person of their own country [p. 437]. 11 This explanation seems a little weak, especially in view of the rather large difference between discrepancies for the American group and discrepancies for the Japanese group. It is possibl.e that Abate and Berrien failed to establish the validity of the Japanese translation of the EPPS.
Finally, Sue and Kitano (1973) found that American stereotypes of Asian Americans have changed quite dramatically over the years from the quite negative views held during the World War II period to the largely positive and, according to them, more valid views of the 1970 1 s. Aronson, editors). Reviews in the anthropological literature abound: Culture and Personality (Wallace, 1970); Personality in•Culture (Honigmann, 1967); and Culture and Personality (Barnouw, 1979), to mention only a bare few.
What is being done here is first, to briefly define the area of culture and personality -that is, the area where the interests of psychologists and anthropologists intersect; and second, (in what must necessarily be an oversimplification), to mention three major ways in which the perspectives of psychologists and anthropologists have generally differed somewhat from one another.
As Barnouw (1979) defines it, "culture and personality, or psychological anthropology, is an area of research where anthropology and psychology come together -more particularly, where the fields of cultural and social anthropology relate to the psychology of personality.
Ethnologists, or cultural anthropologists, are students of culture -of the different ways of living that have developed in human societies in differing parts of the world, while psychiatrists and (at least some) psychologists are analysts of human personality whose work involves an effort to understand why and how individuals differ from one another as they do. Serving as a bridge between ethnology and psychology, the field of culture-and-personality is concerned with the ways in which the culture of a society influences the persons who grow up within it The perspectives of anthropologists and psychologists studying personality in culture appeal to differ from each other somewhat in at least three major areas: explanatory emphasis; theoretical orientation; 11. and methodology.
In the first of these, explanatory emphasis, there are two prevalent emphases and these have been referred to as the 11 emic -etic'' distinction.
The terms "emic 11 and "etic" derive from the suffixes of the words "phonemics" and "phonetics". In linguistics, phonetics is a system for describing the units of sound found universally in all languages, whereas phonemics refers to specific sound units found only in particular languages. "In cross-cultural research, the emi c approach seeks to explain phenomena in terms of categories deemed meaningful within a specific culture, whereas the etic approach emphasizes the development of explanatory constructs that are applicable to all cultures.
In general, anthropological studies have used the emic approach, whereas psychologists have adopted an etic perspective [p. 1063[p. -64, Olmedo, 1979." To state this a little differently (and, again, in a necessarily simplified fashion), anthropologists have tended to emphasize the similarities in personality types within a given culture and, at the same time, therefore, the differences in (modal) personality types between one culture and another. Psychologists, on the other hand, have tended to emphasize both the greater variability of individual differences within a culture, and the universality of certain personality traits across cultures.
In fact, it was Freud 1 s insistence on the universality of the Oedipal conflict that historically provided part of the impetus for the pioneer studies of anthropologists like Benedict (1934), Mead (1939), and Malinowski (1927) who went into the field to study the effects, in given cultures, of societal constraints on sexual behavior and repression.

12.
Incidentally -and ironically -the impetus given by Freudian psychoanalytic theory to anthropological field studies of personality also appears to have given rise to the second major difference in perspective between psychologists and anthropologists: that of theoretical orienta- The foregoing material has been presented in an attempt to sketch the background for the present study and to bring together the two major lines of research that converge on the topic at hand.

13.
In summary, then, a review of the rather scantly literature dealing with the issue of validity of stereotypes tends to lend some support to the 11 kernal of truth 11 hypothesis: That is, at least in more recent times, ethnic stereotypes have sometimes corresponded in part to a group's actual characteristics and/or its perception of its own characteristics.
Further, there appears to be a movement in contemporary psychology away from Lippmann's original conceptualization of the stereotype as a product solely of faulty or illogical thinking. Hilgard et. al (1979) typify this new and softer view when, after defining a stereotype as a 11 belief about a group of people that, even if not totally false, has been overgeneralized and applied too broadly to every member of the group 11 , then go on to comment: 11 We emphasize again that the thinking process that gives rise to stereotypes is not in itself evil or pathological.
Generalizing from a set of experiences and treating individuals as members of a group are common and neces_ sary practices. It is simply not possible to deal with every new person as if he or s_ he were unique, and the formation of •working stereotypes' is inevitable until further experiences either refine or discredit them [p. 548]. 11 Finally, an extensive body of literature devoted to the study of personality in the context of culture forms an area of common interest to psychologists and anthropologists. This area has contributed strong and very extensive empirical support to the expectation that personality traits can and do reflect cultural influences.
These lines of research provide the theoretical background for the present study, which seeks to examine the degree to which autostereotypes of a given cultural group agree with heterostereotypes (that is, 14. views held about that culture by other cultural groups), and to selfperceptions.
The current trends noted and the studies just reviewed provide, as has been stated, the background and theoretical framework for the present study and give rise to the following assumptions and working defini- If cultures differ from one another in the attitudes, beliefs and behavioral expectations that are transmitted from generation to generation, it follows that the personality 11 types 11 of members of one culture should differ from those of another.
The definition of culture being used in this study is the one provided by Barnouw (1979). 11 A culture is the way of life of a group of people, the configuration of all the more or less stereotyped patterns of learned behavior which are handed down from one generation to the 16.
next through the means of language and imitation [p. 5]. 11 11 National character 11 refers to 11 the relatively enduring personality characteristics and patterns that are modal among the adult members of the society (p. 428 11 Inkeles and Levinson in Lindzey and Aronson, 1969).
The attempt to measure personality traits or states, or what a person "really is 11 raises formidable theoretical and methodological problems. Since the interest of this research is a comparison of certain cultural stereotypes with personality descriptions of individual members of those cultures, it is necessary to operationally define the term 11 personality 11 or "real characteristics".
For the purposes of this study, it is assumed that whenever summary descriptions of the self or others are given in terms of a few basic 11 traits 11 or 11 states 11 , it is likely that such descriptions are themselves stereotypes, to the extent that they act like statistical averages of trends or general tendencies, rather than predi~tions of the occurrences of specific behaviors in specific situations. As Triandis (1972) notes, 11~J e cannot know what we call reality in a direct fashion" but act instead upon our subjective perception of reality. Thus even the perception of the self may qualify as a stereotype.
To say: 11 ! am a quiet person" implies that in general, or as a whole, 11 I tend to be quiet".
Such a self description does not, thus, necessarily correspond with inherent, enduring II persona 1 i ty traits 11 ( a 1 though it may). For the purposes of this research, then, "actual characteristics 11 or 11 personality traits" are terms operationalized as a self-description or self-perception.

17.
It is conceivable that such a self-description may even act as a constraint upon the individual, a constraint which is accepted, however, in the interest of retaining a consistent and coherent sense of identity.
It is likely that such self-perceptions arise in part from values internalized by the individual in the process of being socialized into a given family and cultural group. Finally, then, an ethnic or cultural group will be defined as Greeley & McGready (1973) define it: "A large collectivity, based on presumed common origin, which is, at least on occasion, part of a self-definition of a person, and which also acts as a bearer of cultural traits [p. 210]." In this study, it is hypothesized that national stereotypes are related to "actual characteristics", that is, to self-perceptions, of individual subjects in the national group. Accordingly, it is predicted that when cultural and/or ethnic groups are compared, there will be measurable differences between them in both national stereotypes and in self-perceptions.
The major groups that were studies are southern Irish, southern Italian, and American college students. The term "southern Irish" is used to highlight the gee-political and cultural distinctions between the now independent Republic of Ireland, and northern Ireland. Similarly, the term "southern Italian" is used to highlight the widely recognized gee-cultural distinction between the largely agricultural 11 Mezzogiorno 11 and the more industrialized region of northern Italy. In order to trace possible interactions between ethnicity and acculturation, the attempt has been made to subdivide the American sample into !talc-Americans and  Maraspini (1968) points out that, in contrast, "the phenomenon of the 'citta contadina' or city of peasants is characteristic of the southern Italian agricultural region." These tightly-knit and isolated villages or "cities of peasants" typically organized around the local church with its bell tower. The Italian word for bell tower is "campanile" and has given rise to the word "campanilismo", which is used to describe the villagers' feeling of belonging.
"The sound of the be 11 was unique and familiar to each and every person in the village, and over the centuries the attachment the villagers felt for the bell metamorphosed into a sense of loyalty to the village itself and to one's neighbors. is to become simply American, is inhibited by strong elements in the social structure of the United States. It is inhibited by a subtle system of identifying which ranges from brutal discrimination and prejudice, to simply naming. It is inhibited by the simple unavailability of a simple "American identity". One is a New Englander or a Southerner or a mid-Westerner, and all of these things are too concrete for the ethnic to adopt completely, while excluding his ethnic identify.
In any case, whatever the underlying fault lines in American society that seem to maintain or permit the maintenance of ethnic identity beyond the point of cultural assimilation, the fact is ethnic identity (re- Since the researcher wished to examine not only cross-cultural differences, but also the possible effects of acculturation and frequency of contact on heterostereotypes, it was necessary to choose groups whose numbers would be sufficient in the Rhode Island population to make this possible. There is a sizeable population of both Italo-Americans and Irish Americans in the Rhode Island area and this fact influenced the selection of groups to be studied as well.
Finally, the practical question of the availability of subjects in particular cultural groups also, by necessity, influenced the selection. It happened that contacts were made available to the researcher in both Ireland and Sicily that would make possible collection of data from college students in those two countries.

Plan of the Study
National autostereotypes of the cultural groups involved were obtained by having the subjects of each group select all those adjectives from the 81 adjectives of the Activity Vector Analysis (AVA) which best describe "the typical member" of their respective cultures. Since scores on each of these four factors can combine in many different ways to form different profiles, the AVA is an instrument which yields at least 258 interpretable profiles or personality types. Using the AVA to measure national autostereotypes makes it possible to describe the sense of cultural identity held by members of a given cultural group in two ways. First, the frequency with which each adjective was selected by subjects in a given national group was tallied. If a given adjective were selected by at least 55% of the subjects in a particular group, it was assumed that this adjective formed part of a rather definite autostereotype for the group in question. All those adjectives, then, . which were selected by at least 55% of the group were considered as descriptive of the _ picture held by that group of its "general characteristics". (The use of 55% was an arbi-· · trary criterion used to ensure a clear majority.) Second, each subject's selection of adjectives were scored in the conventional manner to permit the determination of a profile or "personality type 11 • (Scoring procedures will be detailed in the Methods section). These profiles were then plotted on the AVA pattern universe. The AVA pattern universe is a Mercator-type projection of the 258 personality types in two dimensional space. The distribution of types for a normative sample of 1199 National heterostereotypes, on the other hand (the picture each group holds of the "typical" characteristics of the other two groups), were obtained by asking the three groups of subjects (Irish, Italian, and American) to describe, using ten adjectives or less, the "typical" member of the other two groups. For example, Irish students were asked to respond to this open-ended question by describing, in ten adjectives or less, their picture of the "typical" Italian, and then to describe their picture of the "typical" American in the same fashion. (Earlier, the point was made that the Katz and Braly method of eliciting ste~eotypes through checking off adjectives from a pre-existing list tends to introduce a certain degree of bias into the procedure. It is felt that opne-ended questions are less likely to do so).
Individual self-perceptions ("personality" or "actual" characteristics) were measured using two instruments measuring a total of eight Finally, subjects again selected adjectives from the 81-word list of the AVA which they felt would best describe their 11 real 11 or 11 basic 11 self. As described previously, this instrument also yields personality profiles or types in terms of basic psychological dimensions. Thus terms of basic temperamental characteristics such as "warm", "friendly", "outgoi ng 11 , "practical 11, etc. It was, therefore, necessary to choose measures of such basic traits, both to test hypotheses concerning the content of such stereotypes and also the hypothesis of "real differences" between the groups.
Second, both the AVA and the MBTI have highly satisfactory psychometric properties (reliability, validity, large normative samples, etc.) In addition, the AVA already had an existing form in Italian and had been used in cross-cultural research prior to this time. Finally, in order to make comparisons with previous empirical findings to be described shortly, it was necessary to choose comparable measures.
It will be recalled that subjects were to select adjectives from the AVA list twice: once as a measure of the cultural autostereotype and once as a measure of the "real" self. Thus discrepancies between these two ratings, as measured by the correlation between the two profiles, would also provide a measure of the degree to which the subject does, in fact, hold an autostereotype of his culture which is distinct from his self-perception, and also the degree to which the individual subject feels congruent with or alienated from, his own culture.

Predictions and Related Empirical Studies
In this study, the following specific predictions were made: With respect to within-group similarities, it was predicted that: With respect to between-group differences in self-perceptions, it was predicted that: (2a) Italian students, as a group, would show a greater frequency of "extraverted, sociable, "sensing" types than found in the Irish or American samples, or in the U.S. normative samples; (2b) Irish students, as a group, would show a greater frequency of "introverted, emotionally controlled, intuitive types" than found in the Italian or American samples, or in the U.S. normative samples; (2c) The distributions of -AVA and MBTI self-perception scores of the American students would not differ significantly from the more heterogeneous distributions of the U.S. normative samples on these two measures.
With respect to the relationship between ,self-perceptions and national stereotypes, it was predicted that: (3a) Some elements in the heterostereotypes for these three groups would show correspondence with their autostereotypes; and that (3b) Some elements in both the heterostereotypes and autostereotypes would show correspondence with the self-perceptions of group members. In regard to this last prediction it should be noted, however, that the sampling limitations inherent in this study may well restrict the degree to which such correspondence is found. The degree to which heterostereotypes of the "average" or "typical" member of a national group will correspond either with the autostereotypes or self-perceptions of groups of college students is problematical. 28.
The predictions made in this study were based on both theoretical considerations and on previous empirical findings. (Cf. Greeley and McCready, 1975;Merenda, Migliorino and Clarke, 1972;Abate and Berrien, 1967;.
Historically, one of the broad tasks that the science of psychology early defined for itself was the attempt to measure and explain individual differences. While many of the pioneer workers in this field assumed that such differences were largely due to innate, genetic factors, the prevailing theoretical bias has shifted. It is probably fair to say that the most widely accepted paradigm in contemporary American psychology has its philosophical roots in Lockean empiricism and its current expression in social learning theory.
Among other things, social learning theory is characterized by its emphasis on overt, measurable behaviors, rather than 11 intrapsychic conflicts, instincts, or dispositions ... Further, social learning theory assumes that ... personality is determined primarily through learning, which takes place in a social context." (Bourne and Ekstrand, 1976).
Elusive though the concept may be, the 11 culture 11 into which one is socialized must be counted as a vital and inextricable part of the social context or milieu in which learning takes place. The values, mores and traditions of a culture are, for the most part, immediately and proximally transmitted from generation to generation via the family. socialized into another culture. rt appears, from even a cursory examination of the representative empirical evidence, that cultural differences do exist across a broad spectruc of variables, and that these differences have important implications.
Glazer and Moynihan (1970) cite a study, for example, which found different patterns of cognitive skills, apparently in part as a function of ethnicity. This study (Stodolsky and Lesser, 1967)  0pler and Singer found ten areas in which there were significant differences between the two groups relating to their behavior and expression of psychopathology. Some of the most salient differences may be summarized as follows: Irish patients much more often had problems centering on a 11 dominating, rejecting mother 11 , while the fathers of Irish patients were generally perceived as weak or absent. Italian patients had conflicts of a less repressed kind centering instead on the father and male siblings. In addition, the Irish patients tended to exhibit greater degrees of repression and anxiety, while Italian patients were more overtly hostile and less repressed and anxious.
Irish patients appeared to utilize fantasy much more often and to appear quiet on the ward, while Italian patients were far more active and mobile. For example, 26 Irish patients showed no evidence of 11 acting-out11 behavior on the ward, while 23 Italian patients 11 showed marked and -repeated evidences of acting out. 11 31.
Twenty-four Irish patients were rated by various ward personnel as generally compliant to authority, whereas almost the same number of Italians were rated as non-compliant.
Delusional fixity was extremely prevalent in the Irish, while exactly two-thirds of the Italian cases showed no fixity whatever, "and the remainder had slightly scattered and changeable delusions ... " In addition to the diagnosis of schizophrenia, alcoholism was pre- differences between members of different cultural groups, it is worth noting that some authorities feel that the kind of behavior expressed in psychopathology tends to be an intensification, or extreme, of the basic personality traits that exist in the normal personality. Thus, Opler and Singer 1 s reported differences in the expression of psychopathology for Italian versus Irish patients may well reflect underlying normal personality differences in the two groups, carried to an extreme.
It also, incidentally, points up the importance of understanding differences in normative behavior for different cultural groups since it is also possible that normal differences in behavior may be misperceived by some as indicative of psychopathology when such may not be the case.
The possibility that expressions of psychopathology may be normative behaviors carried to extremes, and that these behaviors may differ for different cultural groups, receives some support from a study by Merenda, Migliorino, and Clarke (1972). In this study, the AVA was administered to 395 Sicilian males at a public technical school in Palermo.
As previously described, the AVA is an adjective check list of 81 words which yields a personality description in terms of profiles based on scores on the four dimensions measured by the instrument. The social perception and the perception of the real self of these Sicilian youths were compared to a normative sample of 7,732 Americans. It was found that, in contrast to the much more heterogeneous distribution of types in the American sample, the Sicilian sample clustered very tightly about two (somewhat similar) personality types or profiles.
In general, the sample showed a split between 11 those who perceive themselves as 33.
friendly, warm, and socially-outgoing personalities, and those who see themselves as being somewhat ruthless, agressive, hard-hitting egotists. 11 (The correlation between the two profiles is a moderately low r = .31). It is interesting to compare these findings with those of Opler and Singer for a sample of Italian male schizophrenics. It will be recalled that the Italians differed strikingly from the Irish in that the former were decidedly more extroverted, active, mobile and encouragement of independence in their children. They were also compared on such other broad behavioral and attitudinal dispositions as political participation, moral views, respect for the democratic process, and attitudes toward family structure.
With respect to the seven personality variables tested (and this is the aspect of the Greeley-Mccready study of the greatest interest here), the authors found only one variable on which both 11 the Irish and the Italians were significantly higher than Anglo-Saxon Americans 11 and this variable was fatalism. Irish and Italian-American students were significantly more fatalistic than Anglo-Americans.
Irish-Americans differed significantly from Anglo-Americans on four variables. They were more trusting and more fatalistic than Americans, but less authoritarian and less anxious.
Italians differed significantly from angle-Americans only on two variables. Italians were more fatalistic and less encouraging ~findependence in their children than Anglo-Americans.
Finally, in comparing Irish-American subjects with Italian-American subjects, they found Italians more authoritarian and more anxious than Irish, but less fatalistic, less conforming, less moralistic, and less inclined to encourage independence in their children.
In fact, on the seven personality variables measured, Italians and Irish differed significantly from each other on six out of the seven.
It is interesting to note that the Greeley-Mccready study found, in contrast to Opler and Singer, more anxiety in Italian-American than in Irish-American subjects. Although some of the differences, such as 35. this one, found by Greeley and Mccready are not consistent with previous findings (or, in fact, their own a priori predictions), the study is important in the principal finding that distinct differences were found as a function of membership in a particular ethnic group. It is felt that the particular measures of personality used in this study were not very refined. Apparently the personality instrument used by Greeley and Mccready was a seven-factor instrument "which emerged from a battery of 57 items. We (used) items which had factor loadings of over .200 for each scale." An item loading .20 on a factor is a very weak item.
Another disturbing aspect of the study is the absence of data on the subjects and sampling procedures. It is possible that methodological problems weaken the validity of the results.

Justification
It seems clear that there are both theoretical and practical reasons for pursuing the study of cross-cultural differences in stereotypes and self-perceptions, if these exist. As Rohner (1977) points out, knowledge gleaned from cross-cultural research helps to edit and extend" man's image of himself and to correct myths and stereotypes.  Based on the income level, education, and occupation of their parents as reported by these students, the sample can be described somewhat loosely as "middle class".

38.
The "typical subject" in this study, then, was a native-born, 19-year old single Catholic freshman or sophomore enrolled in a public liberal arts college in his or her respective country of birth. The subjects fell into one of three samples: the Irish sample (N = 186); the Italian sample (N = 179); and the American sample (N = 190). Each of these samples will be described individ~ally.
(See Table l  and the remainder (one each) in the U.S.A., China, and Belgium.
Ninety-nine percent of the sample were Irish citizens (one subject claimed dual U.S./Irish citizenship, and one was a citizen of Hong Kong). Ninety-seven percent of the sample was single. Ninety percent stated their religious affiliation was Roman .Catholic.
The overwhelming percentage of the parents of these students (92% of the mothers and 96% of the fathers) were also born in Ireland. Of the few who were not, most were born in one of the other parts of the British Isles.
Combining the reported education, income level, and occupation of parents as reported by students into a rough estimate of socioeconomic status (SES), it appears that the Irish students are sons and daughters of middle-to upper-middle class families. Details of the demographic data for these students will be reported in greater detail in Chapter 3 (Results).

41.
The Italian sample consisted of a total of 179 students, 83 males (46%) and 96 females (54%). All were students at the University of Palermo and were enrolled in Introductory Psychology classes and English language classes at the time data were collected in the Fall of 198.l .
The University of Palermo in Sicily is a national university, part of the highly centralized Italian educational system which includes at least 28 other universities located at centers throughout Italy.
Though the majority of students indicated they were freshman and sophomores (57%), a substantial proportion (especially of females) reported they were juniors and seniors. It should be pointed out, however, that this fourfold classification of students is not as meaningful for Italian as it is for American and Irish colleges, since the system of higher education in Italy is somewhat different. Students typically proceed more independently, studying until they are ready to present themselves for series of qualifying examinations and usually taking between four and six years to do so.
Ninety-seven percent of the students were between the ages of 17 and 25 with a modal age for the sample as a whole of 22 years. The mean age for females was 21.0 years, and for males, 20.52 years.

42.
Ninety percent of the sample stated that they were born in Italy.
However -in a pattern typical especially of the males in the Italian sample -a fair number of students failed to respond .to this (and other) questions. Since, however, only two subjects specifically gave birthplaces other than Italy, it is assumed that the true percentage of native birth probably approaches 99%.
The same reasoning may be applied to the question of citizenship.
Although 86% of the sample stated they were Italian citizens, it is felt that the remaining 14% who failed to answer the question probably are citizens as well, especially since no student reported citizenship in any other country. This means that the sample is very likely 99% to 100% Italian citizenship.
Ninety-three percent of the Italian students stated they were single and sixty-five percent declared their religious affiliation to be Roman Catholic, with a significant difference here between the 46%of the male subjects versus the 82% of the female subjects who stated they were Catholic.
It is likely that almost all of the parents of these students were also born in Italy, though only 75% of the subjects specifically stated this. The remaining 24% did not respond to the question although no subject reported any other country of birth for either parent.
The majority of students (83%) reported that the income level of their parents was average or above average and though the educational attainment and occupati onal status reported for their parents by the subjects was somewhat lower than for the Irish and American samples, the students can probably fairly be described as of middle or "working class". Ninety-five percent of the subjects reported they were born in the United States and ninety-seven percent stated they were U.S. citizens.
Ninety-one percent of the sample stated they were single. Seventytwo percent -80% of the females and 62% of the males -gave Roman Catholicism as their religious affiliation. Ninety-one percent of the parents of these students were born in the United States.
The educational attainment, income level, and occupational status of parents as reported by these students fell between the slighter lower levels for the Italian sample and the slightly higher levels for the Irish sample. Combining these three variables into an informal estimate of SES, the American sample can probably also be fairly described as "middle" or "working class" in background.

Instruments
The "Test Booklet" Each subject in this research received a "test booklet" consisting of materials to be completed in the following order. Part I: A single sheet requesting demographic information. Part II: A folded sheet consisting of three open-ended questions on the facing page eliciting heterostereotypes of the other two national groups and the amount of contact subject estimated s/he had with each; and, on the inside pages, a list of the 81 adjectives of the Activity Vector Analysis (AVA) repeated twice: once as a measure of the subject's autostereotype and once as a measure of the "real" or "basic" self. Part III: the Myers-Briggs Type Indicator (MBTI) and accompanying answer sheet. Eac.h measure will be discussed individually in the order in which subjects completed them.

Demographic Variables
For the Irish and Italian samples, the Demographic sheet consisted of the same set of questions, identical except that an Italian language version was provided for the latter group.
For the American sample, the Demographic sheet was identical except for the addition of a question used to determine ethnicity of the subject.
The demographic information requested from all three groups in- The Activity Vector Analysis (AVA): Measures of the National Autostereotype and of the "Real 11 or "Basic" Self In this study, the national autostereotype was measured by instructing the subject to: 11 Place an 1 X 1 before every word given below that you believe describes (your own) people as a whole or in general.
I believe the 1 typical 1 (member of my own national group) is ... 11 • The list of words following these instructions were the 81 adjectives of the AVA; In the same fashion, perception of the 11 real 11 or 11 basic 11 self was measured by instructing the subject to: 11 Now place an 1 X' before every word which you honestly believe is descriptive of you. 11 The 81 adjectives of the AVA were again listed below these instructions.

48.
The AVA was originally constructed by Walter V. Clarke (1956) as a personality test for use in industrial selection -that is, for the purposes of enhancing job placement by determining the degree to which an individual's temperament and behavioral tendencies are compatible with the demands of a particular job. Since its original development in the late 1940's, the use of the AVA has been expanded to include a considerable number of applications beyond this.
The conventional form of the AVA consists of a list of 81 adjectives, repeated twice. The subject is asked first to check "every word that has ever been used by anyone in describing you"; and then to select, from the second list of the same 81 adjectives, every work "you honestly believe is descriptive of you". In this way, two profiles are obtained: the "social self" or self the subject believes is perceived by others; and the basic or "real" self: the self the subject believes himself to "really" be. A third profile, the resultant of these two, is termed the "composite self" -or "how the person is apt to be in fact perceived by others" (Merenda and Clarke, 1965). These instructions may be altered to obtain other profiles -for example, profile of the "ideal self", profile of a public figure, or, as in the case of this research, profile of the "typical" member of a given group.
These profiles yield personality descriptions that are based on the integration of the scores on four factors or vectors measured by the AVA: aggressiveness (V-1); sociability (V-2); emotional control (V-3); and social adaptability (V-4). A fifth variable, activity level, determined by the total number of words checked, though not 49.
directly involved in determining the profile, is postulated to indicate the individual's energy level . and therefore to contribute psychologically significant information.
A profile shape (PS) is designated by a four-digit number which represents the subject's ipsative position on each of the four vectors.
A raw score on a given vector is determined by the number of adjectives checked by the subject which "belong to" that vector. Each vector of the AVA has approximately 20 words empirically determined, through factor analytic techniques, to be associated with its core meaning, (Merenda and Clarke, 1959).
Vector 1, or aggressiveness, has been determined, for example, to have a core meaning embodied in the AVA adjectives: stubborn, selfcentered, dominant, argumentative, opinionated, and impulsive. The more words of this type checked by a subject, the higher his score on vector one, or aggressiveness.
Similarly, V-2, or sociability, has a meaning embodied in the core words: Charming, appealing, admirable, smooth, and attractive.
Emotional control, or V-3, is signified by the subject's selection of words like: stable, calm, relaxed, quiet and patient.
Social adaptability, or V-4, is embodied by the core words: anxious, hesitant, submissive, tense, and dependent.
Raw scores on each vector are converted to standard scores having a mean of 50 and standard deviation of 10. An individual's profile is then determined ipsatively by assigning a value to each vector of from one (1) to nine (9), based on the deviation of the subject's score on 50.
that vector from his own mean. As an example, a subject's raw scores on vectors one through four, of: 26, 12, 18, and 8 respectively, when converted to standard scores, become: V-1 = 58; V-2 = 48; V=3 = 46; and V-4 = 46. The subject's own mean would thus become the sum of these four scores divided by four, or 50.
Pattern shapes are then derived by transforming standard scores to a nine-point scale, with one (l) representing the minimum value on a vector and nine (9) the maximum value. There are two constraints imposed on deriving the profile pattern: the four numbers of the profile must add up to 20; and at least one of the numbers must be either al or a 9, depending on whether the outstanding vector deviates from the mean in a negative (l) or positive (9) direction. For the illustration given above, since vector l = 58, it receives a value of 9.
Based on the relative deviations of the other vector scores from V-1, the profile designation in this case becomes: PS 9434. Scoring is thus ipsative and yields a four-digit number or pattern shape which is associated with a unique personality type. The AVA yields 258 profile or pattern shapes.
It should be emphasized that the personality type or profile derived in this manner is not simply a linear composite of the four vectors involved but, rather, a complex integration of four behavioral tendencies.
Clarke hypothesized that the observed behavioral tendency of an individual is the result of a complex interplay of psychological forces or variables acting upon or within the individual. He went on to postulate that an adequate, although not, of course, exhaustive, description of an individual ' s behavioral tendency could be derived by taking into account the interplay of the four independent vectors of aggressiveness, sociability, emotional control, and social adaptability.

51.
It was Clarke's belief that underlying these four variables are two basic kinds of responses (approach versus avoidance) to two basic kinds of social situations (favorable versus unfavorable).
Thus, aggressiveness (V-1) is conceived of as approach behavior in an ~nfavorable or antagonistic situation -or, in other words, the tendency on the part of an individual to meet difficulties head on.
Sociability (V-2) is conceived of as an approach behavior in a favorable situation -or the tendency of an individual to respond in an active, positive way to friendly situations.
Emotional control (V-3), on the other hand, is a basic tendency on the part of the individual to withdraw or retreat from a friendly or favorable situation and may be conceived of as similar to (though not identical with) traits li ke introversion and field independenceor the tendency to maintain a certain distance from the social environment.
Finally, social adaptability (V-4) is the tendency to withdraw or shrink from unfriendly or antagonistic social situations.
It will be recalled that the profile or pattern which results from the subject's scores on these four variables represents a unique personality type. It will further be recalled that, though in theory the numbe r is infinite, contraints imposed by the mathematical model underlying the AVA limit the actual number of discrete profiles to 258.

52.
These types taken together can be conceived of as a universe or sphere, It is now possible to conceptualize the universe of behavioral tendencies or possible profiles as a three-dimensional sphere, with a core or center, from which 258 radii project outward toward the surface in all possible directions.
(The four reference vectors, V-1, V-2, V-3, and V-4 lie within this sphere mutually equidistant from each other and separated by an angle of 110° (r = -.34).) Each of these 258 radii represents a uniquely determined behavioral tendency or pattern shape (that is, 11 the final single resultant vector 11 discussed above).
One of the outstanding advantages of this mathematical conceptualization, is that it permits the spatial representation of each type and allows relationships between types to be visualized in terms of 53.
angular separation between radii. 11 Since the cosine of the angular difference in direction between two pattern shapes is the coefficient of correlation 11 it further permits determination of the degree of Pearsonian relationship between any two types.
A method of projecting the three-dimensional space of this sphere onto a two-dimensional 11 map 11 known as the AVA Pattern Universe, has been accomplished through use of a Mercator-type projection technique.
This, in turn, permits the plotting on this universe of the distribution of profiles shapes in samples derived from various populations, a fact pertinent to the present research.
In summary, the AVA is a free-response list of 81 adjectives descriptive of human behavior which, although relatively simple and efficient to administer, nevertheless rests upon a psychologically sound theory and a sophisticated mathematical model. The subject's scores on four vectors provide the basis for determination of the profile. The psychometric properties of the test reflect its meticulous development and refinement that has been carried out over close to four decades.
The reliabilities of the word choices themselves, the vector scores and the profiles have been investigated extensively and found to be quite high. The validity of the test has been repeatedly demonstrated utilizing a wide variety of methods and samples. Construct validity has been established through, among other methods, factor analytic techniques of a measure of the social self (Merenda & Clarke, 1959a) and comparisons of AVA profile patterns with narrative self-descriptions of advanced graduate students (Merenda & Clarke, 1965).

54.
The predictive validity of the AVA has similarly been well-established by demonstrating, for example, that distinctive AVA profiles exist for successful versus unsuccessful life insurance agents (Merenda & Clarke, 1959b); management personnel versus line workers (Merenda and Clarke, 1959c); and self-made company presidents (Merenda & Clarke, 1959d Development of the test rests upon the fundamental assumption that 11 much apparently random variation in human behavior is actually quite orderly and consistent, being caused by certain basic differences in mental functioning (Myers, 1962, p. 51).
There 11 basic differences 11 , according to Carl Jung, arise in four areas of human functioning 11 perception, judgment, orientation with respect to the environment, and characteristic method of dealing with the environment.
Within each of these four basic areas there are, in turn, 11 two di sti net and sharply contrasting modes of functioning. 11 (Test Manua 1) · Thus Jung hypothesizes two kinds of preference in perceiving the world: (1) 11 sensing 11 , or reliance solely upon the direct evidence of the five senses; versus (2) 11 intuition 11 or 11 indirect perception 11 in which additional ideas or associations are 11 unconsciously tacked on to 11 the perceptions coming in from the external senses, thus leading to "hunches", "intuitions" or "insights" which go beyond the immediate evidence.

55.
A similar dichotomy of preferences exists in the use of judgment One last dichotomy of preferences is in the characteristic atti- 57.
In addition to items, the MBTI also contains some simple word pairs in which the subject must choose the preferred word. Some typical items from each of the four dichotomous scales of the MBTI, along with the scales on which they are scored, will be found in Table 2.
Sixteen possible personality types can be formed through the various combinations of the basic preferences. The Test Manual provides a qualitatively rich narrative description for each of the sixteen personality types. See Table 3 for the sixteen types.
These personality descriptions have apparently been developed, not only on the basis of theory, but also through the extensive empiri-  Table 2 Sample Items from the Myers-Briggs Type Indicator

I-E Scale: Introversion vs Extraversion
When you are with a group of people, would you usually rather E (a) join in the talk of the group, or I (b) talk with one person at a time.
Which word in each pair appeals to you more?

E S T P
Matter-of-fact. do not worry or hurry, enjoy whatever comes along. Tend. to like mechanical things and sports, with friends on the side. May be a bit blunt or insentive. Can do math or science when they see the need. Dislike long explanations. Are best with real things that can be worked, handled, taken apart or put together.

E S T J
Practical, realistic, matter-offact, with a natural head for business or mechanics. Not interested in subjects they see no use for, but can apply themselves with necessary. Like to organize and run activities.
May make good administrators, especially if they remember to consider others' feelings and points of view.

E S F P
Outgoing, easygoing, accepting, friendly, enjoy everything and ·make things more fun for others by their enjoyment. Like sports and making things. Know what's going on and join in eagerly. Find remembering facts easier than mastering theories.
Are best in situations that need sound common sense and practical ability with people as well as with things.  Likely to be honored and followed for their clear convictions as to how best to serve the common good. I N F P Full of enthusiasms and loyalties but seldom talk of these until they know you well. Care about learning, ideas, language, and independent projects of their own. Tend to undertake too much, then somehow get it done. Friendly, but often too absorbed in what they are doing to be sociable. Little concerned with possessions or physical surroundings.

I N T J
Usually have original minds and great drive for their own ideas and purposes. In fields that appeal to them, they have a fine power to organize a job and carry it through with or without help. Skeptical, critical, independent, determined, often stubborn. Must learn to yield less important points in order to win the most important.  Abl~ to do almost anything that interests them. Quick with a solution for any difficulty and ready to help anyone with a problem. Often rely on their ability to improvise instead of preparing in advance. Can usually find compelling reasons for whatever they want.
E N F J Responsive and responsible. Generally feel real concern for . what others think or want, and try to handle things with due regard for other people's feelings. Can present a proposal or lead a group discussion with ease and tact. Sociable, popular, active in school affairs, but put time enough on their studies to do good work.

E N T P
Quick, ingenious, good at many things. Stimulating company, alert and outspoken. May argue for fun on either side of a question. Resourceful in solving new and challenging problems, but may neglect routine assignments. Apt to turn to one new interest after another. Skillful in finding logical reasons for what they want · E N T J Hearty, frank, able in studies, leaders in activities.
Usually good in anything that requires reasoning and intelligent talk, such as public speaking. Are usually well-informed and enjoy adding to their fund of knowledge. May sometimes be more positive and confident than their experience in an area warrants. Sparacino. The translation was judged to be adequate.

Procedure
The instruments discussed in the preceding section were combined into a "test booklet" which was distributed to subjects during a 50minute lecture period. After briefly explaining that this was a crosscultural study, subjects were instructed to begin by filling out the demographic sheet and then proceeding at their own pace through the rest of the test material.
Printed instructions were provided so as to make the instrument virtually self-administering.

64.
The researcher administered the instrument to the American sample. Though the vast majority of students in all three groups were single, the slight differences in marital status that were found appear to reflect underlying differences in the populations from which the samples were drawn.

65.
The Irish students had the highest proportion of single students, 96%, with only three subjects who stated they were married, and five who did not respond to the question.
The Italian sample had the second highest proportion of single students, 93%, with seven students reporting they were married, and five who did not respond to the question.
The American sample had the lowest proportion of single students, 91%, with six married students, one widowed student, eight divorced students, and two who did not respond.
Though the majority of students in all three samples reported their religious affiliation as Roman Catholic, there were some 66.
unexpected differences in the proportions doing so.
The proportion of Irish students who stated they were Catholic was 88%. This falls somewhat short of the official census figures for the Irish population (94% as of 1978} but was still the highest proportion for the three groups, as was to be expected. Two students reported they were Protestants, 14 said they had no religious affiliation, and five did not answer. Responses across the three samples to the question regarding church attendance follow in general the same lines, and show, again, somewhat unexpected differences between Italian and American students. Irish students report the highest proportion of regular church attendance: 69% claiming to attend daily or weekly, 16% occasionally (monthly to several times a year); 11% 11 never 11 ; and 4% not responding.
In contrast, only 28% of the total Italian sample claim regular attendance: 18{ of the males and 36% of the females. Occasional attendance is reported by 32% of the females and 18% of the males.
Forty-four percent of the males and 25% of the females claim they 11 never attend 11 • Twenty-two subjects did not respond to the question.
Somewhat unexpectedly, a higher proportion, 34%, of American versus Italian students, report regular attendance. Thirty percent of the males and 37% of the females, state they attend weekly; occasional attendance is reported by 43%, and 20% state they never attend. Only four subjects did not respond to the question.
These data are summarized in Table 4. 68. Turning now to the data on parents of these students, the ethnicity of the American sample, and the education, income and occupation of parents of students in all three samples, will be discussed.
It is, first, not surprising to note that though the overwhelming majority of parents of students in a 11 three countries were born in their respective countries, a slightly higher number of U.S. students' parents were born elsewhere. There seems little question that for both This compares to 23% of the mothers and 32% of the fathers in the Italian sample, and 28% of the mothers and 39% of the fathers in -the American sample who had more than 12 years of formal schooling.
In terms of occupation, the same order appears to hold. Irish students claim the greatest proportion of professional parents, 20%, followed by .Americans, 14%, and Italians, 13%. Interestingly, Irish students had both the highest proportion of mothers whose sole occupation was listed as 11 housewife/mother 11 , and also the highest proportion of professional mothers. Sixty-three percent of the Irish mothers as opposed to 51% of Italian mothers and 26% of American mothers are fulltime housewives.
While the percentage of unskilled laborers among parents in the Irish sample was negligible, it rises to 7% in the Italian sample and 13% in the American sample.

71.
The modal occupational categories for fathers in each of the three countries are: Ireland -professional (followed closely by business, managerial level); Italy -clerical (followed not at all closely by business); and U.S.A. -business, managerial level (followed closely by skilled labor). See Table 5 for a list of the occupational categories used and the frequencies in each for the three countries.
In terms of approximate income level, the majority of students in all three countries report at least average income for their parents, but once again, the Irish report the highest percentage of 11 above 11 or 11 much above 11 average; 36.5% as opposed to 12% of Italians and 30% of Americans in these categories.
On the basis of these data, it is felt that the Irish students can be fairly described as sons and daughters of 11 middle 11 to, perhaps, 11 upper middle class 11 parents, while ·Italian and American students in these samples can be described as of 11 middle 11 or 11 working class 11 backgrounds.
Again, however, these conclusions are offered with the caution that the data derive from self-reports, not objective measures, and may well reflect response biases (e.g. exaggeration, understatement, etc.) as well. Furthermore, differences in the economic and educational systems in the three countries make strict comparisons impossi b 1 e.

Heterostereotypes
It will be recalled that heterostereotypes were measured in this study by having the students in each of the three groups respond to two open-ended questions asking them to describe in "ten adjectives or less their picture of the 11 typical 11 or average member of the other two groups.
A third, open-ended question asked each subject to estimate the amoung of personal contact with members of those groups.
The data were analyzed by compiling a list for each sample of all the words generated by the subjects in that sample, alphabetizing the words, and then tallying responses to each word. Three lists of varying lengths were obtained: the Irish students generated a list of approximately 525 words or word phrases; the Italian students a list of 417 words or word phrases; and the American students a list of 440 words or word phrases.
Once tallying was completed and response frequencies determined for each word, 11 frequently used 11 words were compiled into a shorter list and synonyms were combined into semantic categories -that is, clusters of two or more words conveying substantially the same meaning.

Response frequencies within clusters were then summed.
Given the open-ended nature of the questi0n, no a priori criterion was set for how frequently a given word would have to be used in order to qualify it as part of a group stereotype.
However, in the actual analysis of the data, it was found that almost any word used by even as few as four or five subjects almost always has synonyms in the list which also had appreciable response frequencies associated with them. This was particularly true in the Irish data.

74.
As an example, the word 11 conceited 11 was used by only three Irish subjects in describing Americans. Certainly a word used by less than 2% of the total sample would hardly qualify it as an element in a group stereotype.
However, when response frequencies (given in parentheses in the list that follows) to semantically similar words are summed, a stereotype does indeed emerge: 11 big-headed 11 (6); boastful (24); confident (28); egotistical (4); self-confident (3). A total of 90 subjects, or very nearly one half the sample, produced highly similar words in describing Americans, words which appear to convey the same meaning as the word 11 conceited 11 • The problem that arose in this approach to the data was deciding which words should be assigned to a given cluster. It may be argued, for example, that the word 11 confident 11 is not strictly synonymous with the 1t10rd •conceited 11 • The assignment to the word 11 confi dent 11 of the connotation 11 conceited 11 and its placement in the cluster of words just given is made on the basis of its relationship to other overlapping semantic clusters that were found to exist -that is, the connotation assigned to any word is influenced by the over-all context of which it is a part. Figure l should help make clearer the way in which the data were analyzed and interpreted.
In Figure 1,  ~ cluster. These clusters may be thought of as analogous to factors.
In the data to be presented it will be seen that such "factors" are often highly related to each other in such a way that a coherent heterostereotype based upon their interrelationships does emerge.

A. The Irish Heterostereotypes
The Irish, it will be recalled, produced the longest list of words or word phrases. In all, they used close to 525 adjectives. However, in contrast to both the Italians and the Americans, this list contained many synonyms. This finding is of interest in and of itself since both the number of words used and the words themselves reveal to a striking degree of fluency of the Irish students. Furthermore, most of the words produced were adjectives descriptive of psychological traits -in contrast, for example, to American subjects who tended to focus on physical attributes, and Italian subjects who, in a rather .diffident way, tended to describe superficial social traits.
More importantly, this finding is of interest because the presence of a great many synonyms in the list accounts for the fact that even though the Irish produced 100 or so more adjectives than either of the other samples, the strongest and clearest stereotypes of both Italians and Americans emerge from these data.
Irish males and females used a mean number of approximately seven adjectives in describing Americans and a mean number of approximately five to six adjectives in describing Italians. Inspection of Table 6 reveals that there were no significant differences between the sexes in the number of adjectives used, and, on the average, slightly more adjectives were used in describing Americans than in describing Italians.

78.
Irish students used a higher mean number of adjectives, in almost all cases, than used by the other two samples. Irish students as a group also reported more contact with Americans than with Italians, although many subjects spontaneously gave television and tourists as the source of this contact.
In the Irish stereotypes of Americans, five highly related clusters of synonyms reveal a picture of the typical American as loud, selfconfident, competitive, aggressive, and over-bearing. A sixth cluster, sociability, also emerges as a somewhat unrelated, or separate, 11 factor 11 , in which Americans are, in addition, pictured as friendly, extraverted and outgoing.
Four words having the highest response frequencies (ranging between 33 and 79) were used repeatedly by the Irish about Americans. Of these four words, three of them appear to be unrelated to any of the clusters.
The word 11 loud 11 had the highest freq·uency, being used in exactly that form by 79 subjects. This word, together with other highly related words like boisterous, articulate, vocal and vociferous, forms one of the clusters found in the data. Three other words, however, do not appear to relate to any of the clusters and, therefore, stand alone. These words are: 11 rich 11 (used 43 times); 11 fat 11 (used 36 times) and 11 talkative 11 (used 33 times).
Turning now to the Irish stereotype of the Italians, the same tendency to use a high number of synonyms, leading to clusters, is also found, but to a less striking degree. The Irish stereotype of the Italians is weaker and less clear. This may reflect, in part, the fact that the Irish, according to their own subjective estimate, have less contact with Italians. While nearly half the Irish sample (48%) report "some" or 11 a lot 11 of contact with Americans, only 17% report the same amount of contact with Italians. Amount of contact by itself, however, does not appear to account for all of this difference.
In the Irish stereotype of the ·Italians, presented in Figure 3, more 11 clusters 11 appear, they have smaller frequencies associated with them, and they are less closely related to each other. However, as in the stereotype of the Americans, a coherent picture emerges.
(Thirty-three percent of the males, for example, and 12% of the females failed to produce a stereotype of the Irish).
Italian students reported much more contact with Americans than with the Irish (many of them spontaneously indicating television, tourists, and American students as the nature of this contact) and they used, on the average, more adjectives in describing them. Forty percent of the Italian sample reported 11 some 11 to a 11 lot 11 of contact with Americans as opposed to only six percent reporting the same amount of contact with Irish. There were no significant sex differences in the number of adjectives used in describing a given group, but overall females used one or two more adjectives than males, and both groups used more adjectives in describing Americans. (See Table 6.) The Italian stereotype of the Americans does not include strong, interlocking word clusters as did the Irish data. Neither the few clusters that appear, nor the individual words used, have very high response frequencies associated with them. NATIOt~ALISTIC (9) EXTRAVERTED (7) LOUD (9) BOLD (6) HARD-WORKING (9)  The only words with appreciable frequencies are largely unrelated; 11 Allegro 11 (happy, cheerful), used by 24% of the sample is the most frequently used word, followed by "simpatico" (attractive, charming, genial), 20%; 11 pratico 11 (practical, experiences), 16%; 11 spendaccione 11 or 11 stravagante 11 (spendthrift, extravagant), 24%; and 11 megalomane 11 (megalomaniac or grandiose), 13%. The data are presented in Table 9.

87.
Finally, the Italian description of the Irish is notable for what it does not convey. There is a rather striking lack of words suggesting the trait of extraversion or sociability. The word "sociable" was used by Italians in describing Americans 23 times and only six times in describing the Irish. The word 11 extraverted 11 was used 14 times in reference to Americans and not at all in reference to the Irish. report approximately equal amounts of moderate to frequent contact with both groups. Forty-four percent of the total sample report "a lot" of contact with both Irish and Italians whereas only between three to seven percent report "no contact 11 with Irish and Italians, respectively.
Americans used approximately the same number of adjectives, on the average, in describing both Irish and Italians (refer to Table 6).
groups are those containing physically descriptive words such as dark, blonde, freckled, brown-eyed, etc.
In addition, clusters of synonyms appear to a greater extent than in the Italian data but less dramatically than in the Irish data, and the clusters are more independent of each other.
The American stereotype of the Irish, in addition to the predominating description of physical characteristics, emphasized their goodnatured and almost boisterous sociability, drinking habits, Catholicism, stu.bborness, and quick temper. See Table 11 for a summary of the data. Americans appear to be a more salient group for both Irish and Italian students than either of these latter groups are for each other.
Italian and Irish students report much more contact with Ame r icans, use more adjectives in describing them, and present a clearer, more coherent picture of them than they do of each other.

92.
Irish and Italian students show little agreement in their respective pictures of the 11 typical 11 American except that both groups perceive the Americans as somewhat boastful and self-aggrandizing. Americans are seen as aggressive and competitive by Irish students but as almost frivolous and outgoing by the Italian students.
Irish are perceived by both American and Italian students as drinkers, devout Catholics, traditional, and quick-tempered or irascible.
While the Irish are seen as jovial and friendly by a substantial proper-

93.
The existence of an autostereotype was defined in two ways. The frequency of selection of each word in the AVA list was tallied for the three groups and any word used by more than 55% of the sample was considered to be a significant element of their group self-description.
Second, the profile or personality "type" derived from scores on the four vectors of the AVA were plotted on the AVA pattern universe. The existence of distinct clusters, rather than a heterogeneous distribution, provided further evidence that an autostereotype existed. Further, thi -s autostereotype could now be described in terms of an integrative profile rather than in discrete words. An American normative sample of over l ,199 subjects was used as the reference group in defining departure from heterogeneity.
The results of the word frequency tabulations will be presented for each group first, and then the clustering (or lack of it) of personality types within each sample will be reported.
As will be seen in Table 13, American and Irish students used approximately the same number of adjectives in describing their autostereotypes. Italian students used significantly fewer adjectives than either of the other two groups. There were no within-group sex differences in the number of adjectives used for any of the samples.
There were also almost no within-group differences in the content of the autostereotypes themselves, either in terms of the most frequently chosen words or in terms of the least frequently chosen words.
Males and females agree strongly with each other in perceiving the typical Irish person as socially graceful -relaxed, friendly, easygoing, witty and amusing. Nine words were chosen by 55% or more of the 94.  Table 15.
American males and females agreed in their picture of the 11 typica l II American, as we 11. Fifty-five percent or more of each group selected 12 adjectives from the AVA list, ten of which were the same for both groups. The "top three 11 ( for both groups) were; 11 competiti ve 11 , "aggressive 11 , and 11 sociable 11 • The least frequently selected adjective *These words form part of the autostereotype of females but not males. The percentage for males is given for comparison. In addition to the ten words selected in common by both groups, 59% of the females saw Americans as "bold"

97.
and "interesting", whereas 59% of the males described the 11 typical 11 American as 11 skeptical 11 and 11 enterprising 11 • Results of the American autostereotypes are shown in Table 16.
AVA Profiles: Second Measure of the Autostereotype The second way in which the presence of an autostereotype was defined in this research was by deriving the patterns from the AVA vectors and then inspecting the distribution of these patterns for the presence of clustering. When profiles are plotted on the AVA pattern universe, highly similar profiles lie adjacent or nearly adjacent to each other. As discussed earlier, correlations between profiles are equivalent to correlations between resultant vectors.

98.
A Table of Correlations has been developed (Walter Clarke Associates, 1958) which provides these correlations between each profile and every other profile in the pattern universe. Profiles which correlate .69 or greater with each other are considered to be highly similar.
In addition, when profiles are plotted, clustering can be determined visually by inspecting the pattern universe for areas of especically dense distributions. A portion of the AVA pattern universe is reproduced in Figure 4 to illustrate the presence of clustering or a high degree of homogeneity in a distribution of scores. These data are taken from the actual distribution of American female autostereotypes on the AVA pattern universe, and show 72% of the total sample clustered in about 15% of the area of the pattern universe. Once clustering has been determined, a single profile is chosen from within the cluster to represent what is called the Group Centroid of that cluster. The group centroid is defined as the profile with which every other profile in the cluster will correlate .69 or greater, and may be considered to represent the "central tendency" of scores within the cluster. Thus, the personality description associated with the group centroid typifies all the persons falling within that cluster.
The selection of the group centroid is analogous to 11 the line of best fit 11 or the 11 line of least squares" in a scattergram. It is the selection of the profile or pattern that minimizes the distance between it and every other pattern in the cluster in terms of the correlation between them.
In the distribution illustrated in Figure 4, the group centroid is PS 7616, identified with an asterik in the diagram.
This discussion should help make clear the second way in which both the presence and nature of autostereotypes wer~ defined in this research.
Presence of an autostereotype was defined as the presence of clustering within the distribution, or the departure from heterogeneity.
The nature of the autostereotype -that is, the personality characteristics associated with it -was defined in terms of the profile description associated with the group centroid of a given cluster. With this approach to the data in mind, the autostereotypes of each of the three samples will now be discussed.

Irish Autostereotypes
The least amount of agreement as to what constitutes a "typical" member of one's own national group was found in the Irish sample.
Inspection of Table 17 shows that, while 84% of the total Irish sample can be accounted for by membership in one of these clusters, no one cluster includes more than 37% of the total sample.    Table   17 for a summary of these findings.) The distinctive nature of this very tight clustering reveals the high degree of consensus among Italian males and females about the 11 typical 11 member of their own culture.
The personality description associated 'with PS 5915 (held by 79% of the total sample) is that of a person who has acquired a sure sense of mastery over the social environment and is skillful at manipulating it, both in carrying out his/her own interests and promoting the interests of others. According to the description provided, this person 11 is constantly eager to seek out people and build an extensive network of personal contacts and social acquaintances which s/he accomplishes through the energetic use of social skills. 11 This is a person with good ability to mobilize others in the environment and to use situations to practical advantage.

106.
Although this is the autostereotype held by the majority of Italian students, a small part of the sample clustered, rather widely, around PS 5519. Their perception is of a person who lacks the confidence and sure mastery over the environment of the type just descrfbed but tends, instead, to be a worrier who anxiously regulates his/her life style, and goes to extremes to avoid trouble 11 or the perceived danger of such 11 • This person 11 exudes nervous tension; appears always to be 11 0n the 90 11 , usually to service the desires of others; and is gene.rally perceived as a very anxious type of person. 11

American Autostereotypes
Although not quite as dramatic as in the case of the Italian students, the American students also show a fairly high degree of consensus in describing their autostereotype. As can be seen in Table   17, 72% of the females and 66% of the males clustered rather tightly around PS 7616. A much smaller cluster, accounting for 22% of the total sample, formed around PS 5339. The two profiles correlate .44.
Only five males fell outside of these two clusters and their profiles were widely dispersed over the rest of the pattern universe. Eight females fell outside of these two clusters and all eight formed a loose cluster around a third centroid, PS 4952.
Pattern 7616 describes a tense, active, 11 high powered 11 person who is constantly 11 0n the go 11 • 11 S/he is a rather high-strung individual who approaches most demanding life situations with zeal, but invariably take on much more than s/he can handle ... 11 This person is prone to be irritable and impatient and tends to be impulsive in making decisions.

107.
A very different picture of the 11 typical 11 American is held by a minority of males and females in the sample who clustered around PS 5339. An almost identical profile also emerged as one of the autostereotypes held by the Irish (PS 3539, which correlates .83 with PS 5339).
Once again, this is the highly dependent person who 11 is more comfortable doing the bidding of others than in assuming a leadership role in any major enterprise in life -a humble, quiet, unassuming, 'yes' type of person." (It is possible that this description, found in both the Irish and American autostereotypes, is a projection of the 11 real 11 self.) The findings just described indicate that fairly clear autosterotypes, agreed upon by a substantial majority of the students in each of the two samples, are held by Italian and American students. Irish students, on the other hand, split into three groups in describing their differing perceptions of the 11 typical 11 Irish person.
An integration of the findings just reported with respect to the most frequently selected AVA adjectives, yields the following 11 thumb-nail11 sketches of the autostereotypes held by each group.
The Irish generally perceive themselves as a socially graceful p~ople, charming and persuasive in manner, and relaxed and at ease with the social environment. Two of the adjectives chosen most frequently by both Irish males and females in describing their autostereotype were 11 sociable 11 and "easy going". For Irish males, however, and to a lesser extent for Irish females, this social ease and grace masks an inner tension and intellectual drive.
(Recall that whereas Irish 108. females described their autostereotype with the words 11 leisurely 11 and 11 relaxed 11 , many fewer Irish males did so.) The Italian autostereotype held by the majority of both males and females emphasizes a dynamic control or mastery over the social environment, a mastery which enables them to put it to good advantage in carrying out their aims and interests. This is typified by the adjectives chosen most often by both males and females, 11 good mixer 11 and 11 ingenious 11 , for example, and is underscored by the profile of a type 11 constantly eager to seek out people and build up an extensive network of personal contacts. 11 American males and females agreed for the most part in the selection of adjectives like 11 competitive 11 "social" and 11 aggressive 11 in describing the 11 typical 11 member of their own culture, and this description was also reflected in the profile description of the tense, highpowered person who relates somewhat aggressively to the social environment chosen by 69% of the sample.

Self-Perceptions
In this study it was hypothesized that the process of being socialized into a given culture should give rise to a certain degree of within-group similarity in self-perceptions; and that, further, certain self-perceived 11 traits 11 would prove to be more salient in one cultural group than in another. The term II personality trait" was earlier defined in this study as the perception held by an individual of his/her 11 real 11 or 11 basic 11 self. The final portion of this section will present the results of the two measures used to determine this 109.
self-perception, the AVA 11 self 11 profile, and the Myers-Briggs Type Indicator. It will focus on the degree to which consensus was found to exist in the self-perceptions of members of the three groups studied, and the differences across the three groups with respect to the saliency of certain subjectively perceived 11 traits 11 • Irish Self-Perceptions With respect to the AVA profiles of the 11 real 11 or 11 basic 11 self, approximately 53% of the Irish sample fell into two clusters on the AVA 258-pattern universe, with exactly the same number of subjects (N=49) in each. The proportion of an American normative sample of N = 1199 cases, drawn from the general adult population, found in these two areas of the AVA pattern universe is, by contrast, 28.1%, or a little more than one half that of the Irish sample. · (This U.S. normative sample will be used as the reference sample in defining significant differences in distribution between the U.S. and the samples in this study. For a description of this normative sample see Clark (1973).
The chi-square test for the significance of the difference between the observed frequency in the Irish sample and the expected frequency For the second area, that around PS 5717, the difference between the observed frequency in the Irish sample and the expected frequency (derived from the normative U.S. sample) is equal to 47.37, (d.f. 1), p (.01. (See Table 18 for a comparison of the modal AVA self-perceptions for the three samples.) The clustering of the Irish sample was rather tight, with all the profiles correlating between .72 and .97 with their respective group centroids . No significant sex differences were found in group membership in either of these two clusters.
The remaining 47% of the Irish sample fell, for the most part, into two weaker clusters, within which the pattern types are dispersed more widely (.69 to .96 with their centroids). Only 10% of the sample was found outside of these four clusters.
The proportions of membership in these two latter and weaker clusters did not differ significantly from that found in the normative U.S. sample.
On this basis, it appears that there is a greater consensus in the self-perceptions of these Irish college students than among the general adult population of Americans whose patterns are distributed much more heterogeneously throughout the pattern universe.
AVA pattern shape 5717 was previously described as one of the autostereotypes held by these Irish students, particularly the males, about themselves. This is the "optimistic and socially outgoing person who possesses the enthusiasm and social poise that win friends easily ... " but who, behind this facade has an intellectual drive for perfection and truth that can never quite be satisfied." 111. AVA pattern 3449 is nearly identical to pattern 3539, which was also previously described as one of the three autostereotypes the Irish sample held about their 11 typical 11 countryman. This is the quiet, humble, unassuming person "driven by a strict sense of loyalty" and almost compulsively conscientious in carrying out tasks. This person tends to be dependent on others for direction.
A third, weaker cluster, representing eight females and seventeen males, formed around AVA pattern 5942 and the proportion in this cluster .did not differ significantly from the normative U.S. sample. This pattern, again, was also very close to one of the three Irish autosterotypes, PS 3944 (r = .88). This is the socially graceful, helpful, and facilitative individual always ready to promote ideas and projects, either their own or others ... a politically astute person with considerable charm and persuasive ability able to "influence others to go along with his/her point of view. 11 In the distribution of AVA profiles of the perceived "real self 11 , then, 66% of the Irish sample fell into three clusters, each of which can fairly be represented by a single personality description or "group centroid". can be made in the case of the MBTI, as well as the AVA.
The selection of a suitable normative sample, however, proved to be difficult because of the fact that the normative samples reported in the Test Manual are either exclusively male or exclusively female.
This problem was handled by averaging the percentages reported for two liberal arts normative samples, one male and one female, using a formula which weights the averages to take into account their differing N's.
The two normative samples selected for combination into a single reference sample were 3,676 male liberal arts college students drawn from Dartmough, Brown, Amherst, Stanford, Wesleyan Universities; and 184 Long Island University female liberal arts students. The sample sizes in this case are so vastly different that this solution is regarded as far from ideal. In addition, it should be pointed out that the colleges from which the normative sample is drawn are "ivy league" and probably quite atypical of the educational institutions from which the American sample in this study was drawn.
The procedure used in "averaging" the male and female normative samples was an application of the formula for obtaining weighted averages, applied to each of the sixteen cells of the MBTI type The normative U.S. sample distribution on the sixteen MBTI types is reported in Table 19. These normative samples are described in the MBTI Test Manual (1962).
Irish Self-Perceptions: MBTI types The distribution of MBTI types in the Irish sample is reported in Table 20. To facilitate comparison with the U.S. normative sample, which is also shown in Table 19, the percent of that type found in the U.S. sample will be shown in parentheses below the percentage of those types in the Irish sample.

Three cells of MBTI ty~es, INFP, INTP, and ENFP, appeared to draw
proportionately large percentages of the Irish sample, and the three together account for 43% of the total sample.
As seen in Table 21, which compares the modal MBTI types in all three samples with each other and with the U.S. normative sample, the frequencies of these three types differs significantly from that found in the normative U.S. sample. In addition, as will be noted in the         INTP types were also significantly more frequent in the Irish sample than in the normative U.S. sample. This type shares many of the characteristics of the INFP -depth, concentration, ingenuity, etc. -but are apt to have a greater capactiy for analysis and logic due to the 11 T 11 dimension.

ISTJ
As was mentioned, there is also a significantly much smaller proportion in the Irish sample of ESTJ types. The ESTJ 11 type 11 is the practical, down-to-earth, realistic and matter-of-fact kind of person who is not interested in theoretical topics. The low incidence of the 119.
ESTJ type further highlights and makes even more consistent the finding of INFP as the "modal" type.
Italian Self-Perceptions ~1ith respect to their AVA 11 self 11 profiles, 90% of the males and 79% of the females in the Italian sample fell into two fairly tight, somewhat overlapping clusters corresponding quite closely to the two clusters formed by their autostereotypes. The modal self-perception of both males and females was AVA PS 5717, and this is very similar to PS 5915, the mode for the autostereotype. Only seven males fell outside of these two clusters and their profiles were, for the most part, rather widely dispersed. Thus, for Italian males, 90% of the sample was accounted for in two clusters.
For the females, a third weak cluster, accounting for 12.5% of the female sample, formed around AVA PS 1667. The cluster also included three males. Of the eight females not included in the three clusters just described, six fell just outside PS 5717 and their profiles correlated .67 with it. 120.
Thus, for Italian females, three clusters accounted for 91 .5% of the sample, and the same three clusters for 94% of the males. (Refer   to Table 18 for a summary of these data.) AVA PS 5717 describes the active socially outgoing, basically optimistic person who is a "doer" and prefers action to inaction although beneath this facade of social poise this person is more deliberative than s/he may appear to be. There is also an intellectual drive to probe for logical answers and a desire for perfection that may make this person somewhat slow in making decisions and prone to changing them. Significantly more males than females shared this self-perception.
AVA PS 3539, with a much smaller proportion of Italian males and females clustered around it, describes the type of person who is highly dependent on others, finds it difficult to make independent decisions, and works best under supervision. This person tends to be an anxious worrier, to work with great conscientiousness and devotion to duty, and to be neat, organized, and accurate in approach to tasks.
This person appears to be the conservative type, approaching new situations with suspicion, and preferring traditional ways of doing things .
AVA PS 1667 drew a very small percentage of the total sample. This pattern is also described as a quiet and unassuming type, though one who lacks the anxiousness and tension of the preceding type. This individual exudes "peacefulness, tranquility, and serenity", is friendly, amiable, steady, patient, and empathic towards others. This individual also "resists change, preferring established systems and methods", as is 121.
the preceding type; and is good at carrying out routines with patience and dedication. This type is described as modest, cheerful, devoted and loyal.
The distribution of the Italian sample on the MBTI is shown in Hith respect to these four types (ISTJ, ISTP, ESTP, and ESTJ), the only significant sex difference in this distribution was for the ESTJ type, which has nearly three and one half times as many males (n = 24) as females (n = 7), a difference significant at the .01 level.
Owing partly to this skewed distribution, 11 of the sixteen cells of the MBTI type table showed a significant difference in the Italian versus U.S. normative, distributions. Outside of the 11 ST 11 cells mentioned, in which observed frequencies far exceeded expected frequencies, the significant differences in the other seven cells were in a negative direction -that is, contained far fewer-than-expected frequencies.
That this is not simply a function of the extremely skewed 11 crowding 11 of the Italian sample into the four 11 ST 11 types is evident from inspection of the distribution throughout the table. Rather than being dispersed rather evenly throughout the other 12 cells, one finds a marked lack of cases in the 11 NF 11 region of the table: that is, the area which 122. Table   Table 22 Distribution of MBTI Types in the Italian Sample  represents the polar opposite of the "ST". In this column of the type table, eight females and two males are distributed across four cells, and the lowest frequency in the entire table is found in the INFJ cell, which claimed one female and no males. It was just this case which was found in the Ir i sh data as well, but in the reverse direction.

Myers-Briggs Type Indicator Type
The Irish sample clustered in greater-than-expected proportions into the "NF" region, and in far-less-than-expected proportions in the 11 ST 11 region.
The ESTJ type was the most frequently occuring type in the Italian sample for males. The ESTJ is described as "an extraverted thinker .. . who has a great respect for impersonal truth, thought-out plans, and orderly efficiency . This type is at ease with the social environment, practical, observant, logical, decisive, critical, and organized. This kind of individual often makes a good executive.
The most frequently occuring personality types for the females were ISTJ and ISTP. The introversion of these two types adds depth and concentration to their capactiy for realism, observation, and logical analysis. The ISTJ type is better at organization while the ISTP type is better at adapting.
All of these types share a practical, orderly, matter-of-fact, realistic orientation to life and its problems. The introverted sensing types tend to remain more reserved, aloof, and detached from the environment , which may make them appear like "cool onlookers", or as serious and quiet, but which also give them an added capacity for concentration, depth, and i mpersonal analysis.
The extraverted sensing types tend to manage the environment very effectively though sometimes in a rather blunt way. 124.
There was, as noted, an extremely low incidence of 11 intuitive feeling types 11 in the Italian distribution, particularly INFJ. Only a single Italian (female) in the entire sample fell into this category. The low incidence of 11 NF" types appears to highlight a lack of a theoretical, as opposed to practical, orientation among these Italian students.
"Sensing" types tend to 11 see the facts 11 whereas 11 intuitive types" tend to 11 see the possibilities 11 • Thinking types tend to be good at impresonal analysis and organization and may make good admin-  Table 18 for a summary of these data.) Two of these clusters, PS 3539, and PS 6716, accounted for the self-perceptions of 51% of -the total sample, and showed a fairly high degree of similarity to two of the autostereotypes found earlier to account for 91% of the American autostereotypes . (Refer to Table 17).
The person typified by pattern 3539 has previously been described, then, as one of the American autostereotypes. This is the quiet, unassuming humble and dependent person who works best under direction from others. Such a person tends to be an anxious worrier~ meticulous, fastidious, and puntilious, and takes great precautions not to deviate from established routines and guidelines. Clinically, this person may sometimes be described as ''obssessive-compulsive'' (Merenda and Berger, 1978). This was the self-perception held by 38% of the females and 31% of the males in the American sample. (The sex difference is not 127. significant). It appears to describe a person rather low in selfconfidence.
AVA pattern 6716, which described 16% of the total sample, 22.5% of the males and 11% of the females, was also previously described as one of the American autostereotypes.
''This is a person who gives the appearance of being highly excitable, constantly irritable, and exceedingly impatient. S/he is a rather high-strung person who approaches most demanding life situations with zeal but who invariably takes on much more than s/he is capable of handling or accomplishing. S/he exudes nervous tension; is prone to impulsive actions and decisions; and is restless and uneasy in social situations -factors which often prevent him/her from being readily accepted by peers and other people toward whom personal contacts have been initiated." This pattern described a significantly greater proportion of males than females.
AVA pattern 1676, with 15% of the American sample clustering around it, 20% of the females and 9% of the males, described the "quiet and unassuming person whose presence among others exudes peacefulness, tranquility, and serenity. Being amiable and friendly, this person is well-liked by others. Steady and patient, s/he likes to be sought out by others in time of need, is modest in manner, and cheerful in spirit.
S/he resists change, performs duties with patience and dedication, and tends to be a loyal and devoted member of any groups/he belongs to. 11 This pattern described a significantly greater proportion of females than males.
Finally, AVA pattern 3944, with 14% of the total American sample clustered around it, also described a significantly greater proportion 128.
of females than males (18% as opposed to 9% of the latter).
This person is described as always ready to help others promote ideas and projects, both their own and others. This is a person who is at ease in social situations and initiating contacts, is attracted to a wide variety of people, and is usually a joiner of clubs and social groups. This is the politically astute individual who has the charm and persuasiveness to influence others to go along with his/her point of view.
(The fifth cluster centering around AVA pattern 2594 and representing 15% of the males and no females, can be described as a group, as peaceful persons, modest, serene, calm, and relaxed -a pattern not unlike PS 1676, described previously, with which it is correlated .79.
This was the only cluster in the American sample described thus far which did not show significant departure from the frequencies to be found in the U.S. normative sample . . It is included here, however, because, taken together with the four clusters described above, it accounts for 84% of the male sample. the American students in this sample, the 11 SF 11 type appeared to be mar~ frequent than in the general college student population, and the "NJ II types appeared to be less frequent.
The most frequent types for females in this sample were ESFJ and ENFP, both of which had nearly equal numbers of females, and taken together account for 35% of the female sample (versus 11% of the males).
The most frequent types for the males were ESTJ (13 males) and INFP (10 males) which, taken together, accounted for 29% of the male sample.
As was stated previously, however, only 11 SF 11 types showed a greater than expected frequency in the American sample, and only 11 NF 11 and "NJ" types less than expected frequencies.
11 SF 11 types are described by the MBTI manual (1962) as mainly interested in facts they can gather directly through the senses, but they approach their decisions regarding these facts with personal warmth. Their personalities tend to be sociable and friendly, their most successful fields tend to be in elementary teaching, nursing (especially pediatrics), social work, etc. where they can exercise their personal warmth effectively in concrete situations.  Type Table   131.

11
NF 11 types, which showed a low incidence in this sample, also possess the personal warmth of the 11 SF 11 types, but center their attention on possibilities and future projects rather than on concrete, here-and-now situations.
Thus, it appears that the American sample, with its high incidence of SF types, has, on the whole, a greater-thanexpected proportion of subjects who are rather more practical and realistic than theoretical, though they combine this with personal warmth, rather than impersonal analysis due to the 11 F 11 rather than 11 T 11 dimension.
It has already been noted that modal autostereotypes and self-perceptions frequently coincided in this study. nearly three-fourths of the sample having correlations between the 11 role 11 profile and the 11 self 11 profile of from .60 to .99. Furthermore, inspection of the frequency column shows that the frequencies go down in a highly orderly and consistent manner as the correlation declines.
Irish students showed the next highest degree of agreement, with nearly half the sample showing correlation between .60 and .99, of the autostereotype and the real self, though 27% of the Irish students had negative correlations.
American students had, overall, the least degree of agreement.
Of the total sample, 39% had correlations of between .60 and .99, whereas 31 .5% had negative correlations ranging from .00 to -.99. who is optimistic, outgoing, and socially poised but has an inner drive for perfection and truth that makes this type more inwardly tense than he appears; and PS 3449 -quiet, fearful, dependent, unassertive, and lacking in confidence. The four scales of the MBTI, taken as separate dimensions, showed the Irish students to be, as a group, introverted (65%), intuitive (69%), feeling (54%) and perceptive (60%). The most frequently occuring feeling (54%) and perceptive (60%). The most frequently occuring 11 type 11 for the Irish was INFP (18% of the sample), far more frequent than in the U.S. reference sample of American liberal arts colleqe students.

134.
The Italian students as a group had distributions which were much more homogeneous than the other two samples. The modal AVA 11 self 11 profile for the group, agreed upon by fully 56% of the sample, was PS 5717the outwardl.v active 11 doer 11 described above who performs we 11 socially but is inwardly more tense , discontented, and driven than appears outwardly. On the scales of the MBTI, the group as a whole were extraverted (52%though more males than females), sensing (78%), thinking (80%) and judging (51%). The most frequently occuring MBTI type for the sample as a whole was ISTJ (19%). However, there was a significant sex difference here. The mode for females was ISTJ and for males it was ESTJ. Thus there was good agreement in the Italian sample on the 11 S 11 and 11 T 11 scales, but females tended to be more introverted than males.
Thirty-five percent of the American students had an AVA self-perception represented by pattern 3539, which describes the "anxious, dependent, unassertive, individual who lacks confidence". This is in sharp contrast to the autostereotype, PS 7616, held bv 66% of the group, which described the 11 typical 11 American as excitable, irritable, impatient, and hiqh-strung Person always 11 On the qo 11 • This, taken together with the findings reported in Table 24, showed the group as a whole had the least correlation between self-perception and national identity, and suggests a real split in this samole. These students as a group seem to lack self-confidence and to perceive the typical member of the American culture as beinq quite different from them. On the MBTI scales taken individually, American students as a group were extraverted (63%), sensinq (57%), feeling (54% 0, and perceptive (58).

135.
The most frequentl.v occuri ng "tyoe" for the group reflects their preferences on these dimensions: The ESFP type was the most prevalent (12 %).
In the light of these findings an attempt can now be made to address the questions posed at the outset of this study and to examine the degree to which the hypothesis of cultural differences, and the specific predictions flowing from this hypothesis, received empirical support. Each of these questions and predictions will be briefly recapitulated preliminary to commenting upon them.

Heterostereotypes
(1) Were clear-cut national stereotypes found to exist? If so, were there differences among the groups, either in their willingness to stereotype, or the content of the stereotypes themselves?
For the purposes of simplification, groups are discussed as groups -but it should be kept in mind that each group was made up of nearly 200 individuals, and there was a good deal of variability across the individuals within each group. Thus, though the discussion will focus on 11 central tendencies 11 or trends of the groups taken as a whole, the fact of individual variability should not be forgotten. As was pointed out previously, for example, some persons -particularly among the Italian students -did not respond to the question regarding heterostereotypes at all. The researcher herself administered the test instruments to the American sample and found some students who were discomforted by the question and, in some cases, refused to respond to it.
It was apP.arent from the data, however, that some members of each of the three groups studies held stereotypic pictures of the other two groups. 137.
The Irish students appeared to have the greatest "enthusiasm" for the task of stereotyping, and they had the greatest degree of clarity and strength in describing these stereotypes. Their generally greater willingness to stereotype emerged in the larger number of total adjectives produced by the group as a whole, the overall greater mean number of adjectives used per subject, and the very low incidence (almost zero) of failure to answer the question. The clarity of the stereotypes produced by the Irish students was seen in the large number of synonyms occuring in the list of adjectives they generated.
The strength of their consensus was shown by the fact that, especially in stereotyping Americans, many of the individual adjectives or their synonyms, were used with relatively high frequency. The word "loud'' for example, was used by 79 Irish students in describing Americans.
No other single word was used with a frequency even remotely approaching this in either of the other samples. Also Irish students stereotyped Americans more strongly and clearly than they did Italians. In the stereotype of the former, Americans were seen as loud, dominant, aggressive, competitive, and friendly. Italians, on the other hand, were seen by these Irish students as gregarious, friendly, warm, temperamental, and emotionally expressive.
The Italian students appeared almost diffident in their approach to the stereotyping task. As a group they produced the shortest list of adjectives and there were fewer synonyms in this list. They used, 138. for the most part, a fewer mean number of adjectives per student and the frequency of use of any given word or its synonym was markedly lower than for the Irish sample. In addition, a substantial propor- groups agreed, however, in perceiving Americans as sociable and selfinflated.
Italians were seen as warm, friendly, gregarious, emotionally expressive and romantic by the Irish students whereas American students saw them as -in addition to dark-eyed, dark-haired, etc.aggressive, loud, hard-working and family-oriented.
Irish were perceived quite differently by American than by Italian students. Italian students saw them as conservative, traditional, emotionally cold, introverted, and stingy (though it should be emphasized again that this picture was held by only a small percentage of students) whereas Americans saw them as boisterous, happy, jovial, friendly, and 11 good partyers" -in addition, of course, to physical characteristics such as freckled, red-haired, etc.
In summary, then, stereotyping was found to exist in the three groups, although the willingness to stereotype varied, as did the content of the stereotypes themselves, across the three groups. Second, and related to this, the Irish students appeared to be the most fluent or literate of the ·three groups in terms of their language proficiency. Inspection of the list of adjectives generated by the Irish students is like looking at a lexicon of the English language. It would seem, in fact, that the Irish students took the task of producing ten adjectives almost as a personal challenge, and approached it with zest and zeal. This is especially interesting in light of the fact that the 11 INFP 11 type is described by the MBTI manual as having a "gift of language and expression". If richness of vocabulary is a valid measure of a facility for language, then it received some support in this research. It is entirely possible, however, that this finding is due, not to cultural, but to social class differences. It will be recalled that the socioeconomic status of 141.
the Irish students as a group was higher than that of either the Italian or the American students.
It was also previously noted that the nature or type of adjective used differed from group to group. The adjectives generated by the Irish students appeared to reveal, for the most part, a higher level of abstraction and depth of perceptiveness in their approach to personality trait description, a tendency also consonant with the higher prevalence of 11 introverted intuitive 11 types among the Irish students, a ·type characterized by their liking for the theoretical and abstract.
Italian students, a higher proportion of whom were 11 ST" types, and In addition, it appears that both in approach to the task itself, and in the nature of the answers, cultural differences were revealed.
It should also be mentioned in passing that there were no controls built into this study to assure that subjects produced their own 142.
privately-held stereotype, rather than~ stereotype -that is, the culturally held caricature of a given group. Although the intent of the question was to elicit the subject ' s own vie w of the target group, it appears that some students made the alternative interpretation and produced, instead, the "cultural stereotype". In spite of these reservations, however, it is apparent that some clear stereotypes did emerge that represented the actual views of the subject.
(2) Were there sex-related differences in the nature of the stereotypes produced?
This question could not be definitively answered due to the relative small frequencies of subjects who used exactly the same of a highly synonymous word. Sex comparisons on the basis of such small numbers would not be very valid. However, in the cases where words or their synonyms were produced in sufficiently large numbers to make legitimate comparisons possible, no sex differences were found in any of the groups. 143.
(3) (The question pertaining to ethnic differences in stereotyping or in self-perceptions -that is, in Irish-American students versus Italian-American students -was dropped from the study due to small numbers and very uneven sex proportions of subjects who met the criterion for ethnicity.)

Self-Perceptions
In this study it was predicted that: On the MBTI measure, the American sample showed a slight but significant tendency to cluster into the 11 SF 11 region of the type tended to describe themselves as the "quiet, unassuming type" of person on the AVA measure. To this extent, the prediction of heterogeneity in the American sample distributions was not supported.
The predictions with respect to within-group similarties in the self-perceptions of Irish and Italian students were made on the basis that the Irish and Italian cultures are more homogeneous with respect to nationality, and, to a lesser extent, socialization practices, than is the American culture.
Since it is assumed that, in addition to other factors, there are strong cultural influences on the formation of an individual's self-perception, socialization into a fairly cohesive culture should be reflected in a greater degree of within-group similarity in the self-perceptions of members of that culture. To the extent that, in addition, the agents of socialization within a given culture (e.g. family, church) are strongly organized and in consensus about the values to be transmitted, there is further reason to expect greater homogeneity in self-perceptions of cultural group members.
The results obtained in this research were, of course, based on samples of college students and the question of how representative these samples are of the three cultural populations from which they were drawn constitutes a severe limitation on the interpretations which can be made of these data. If it is assumed that these results do, in fact, reflect population characteristics, the high degree of homogeneity in the self-perceptions of Italian students and the greater proportion of "extraverted, sensing/thinking types" come as no surprise to those familiar with the literature on Italian 11 familism 11 147.
(Ban field, 1958; Gambino, 1974;. The conservative and highly structured nature of Italian family life or 11 l 1ordine della famiglia 11 could never be maintained without a high degree of planning, foresight, orderliness, and committment to preordained values that characterize the 11 judgi ng 11 and 11 thi nki ng 11 types. Further, family soci a 1 i zati on practices, with their high degree of consensus regarding the values to be transmitted tend, in themselves, to perpetuate similarities.
One of the values stressed is that of loyalty and committment to the family as a group rather than acting in behalf of one 1 s personal interests. This kind of socialization is not apt to encourage the development of individual differences or departure from group values.
The strength and cohesiveness of family life, as Gambino, Banfield, and others have pointed out, has had historical survival value for the Italian people in a culture marked by a long struggle against poverty, foreign invasions, political oppression, etc.
The Irish students in this study, although not as homogeneous in the distributions of their self-perceptions as were the Italian students, did show, however, a decidedly greater degree of homogeneity than either the American sample or the U.S. normative sample, on the MBTI measure. The prediction that there would be a greater incidence of "introverted intuitive" types in the Irish sample was rather well supported in this study. The prevalence in the Irish sample of introverted, intuitive, feeling, and perceptive types also appears to be consonant with certain traits widely attributed . to the Irish peoplee.g. their love of language and gift for expression, scholarship, and possibly, their celibacy and sexual repressiveness (cf. Scheper-Hugher, 1979).

148.
The greater homogeneity in the distribution of MBTI types for the Irish sample may possibly reflect, in part, the greater cohesiveness of the Irish culture as compared with the diverse, multiethnic culture of the United States. Although Irish family norms do not appear to be as strongly structured as are Italian family norms, they do appear to be a good deal more so than in the United States. Further, the strong identification of the Irish with Roman Catholicism probably provides another powerful agent of socialization within the culture which tends to promote a greater homogeneity in self-perceptions.
Finally, the American sample, while showing less homogeneity in distribution of self-perceptions than either the Italian or Irish samples, nevertheless departed significantly from the heterogeneity of the wider normative sample of U.S. college liberal arts college students.
It will be recalled that the modal AVA self-perception of the American sample depicted a rather quiet, unassuming, dependent, and non-assertive individual.
The distribution of the sample on the MBTI measure showed a greater-than expected frequency of 11 SF 11 types. This suggests that the American sample may have been atypical of the population from which the normative sample was drawn. It is likely, in fact, that students at the Community College of Rhode Island have a more practical orientation and more interest in direct preparation for a career in human services or business fields rather than in more abstract and theoretical fields, likely to attract "ivy league" college students.

149.
Relationship of Heterostereotypes to Autostereotypes and Self-Perceptions The final predictions made in this study concerned the relationship between national stereotypes and self-perceptions. It was specifically predicted that some elements in the heterostereotypes of these three groups would show correspondence with their autostereotypes; and that both would show some correspondence with self-perceptions.
Direct comparisons of heterostereotypes either with autostereotypes or self-perceptions was difficult because of the different ways in which the three variables were measured. In a general way, however, the specific content of the heterostereotypes produced showed a moderate to quite low degree of correspondence with either autostereotypes or self-perceptions.
The heterostereotypes of the Irish as "Catholic" and/or devout, drinkers, and hot-tempered were characteristics ascribed to the Irish by both the Italian and American samples. Data on religious preference and church attendance confirms this rather obvious perception of Irish Catholicism and/or devoutness. (As it happens, the stereotype of the Irish as drinkers is in agreement with what has been reported elsewhere -in the literature of the social sciences, for example -detailing the fact that though per capita consumption of alcohol is lower for this group than for other groups -e.g. Italians -the rate of alcoholism among the Irish is notably high.) The trait of "hot temper" was in agreement with the Irish autostereotype -59% of both Irish males and Irish females described the typical member of their cultural group as 150 . 11 argumentative 11 • A rather small percentage of Italian students held that the Irish were introverted, which, according to the MBTI scale measuring this dimension, as a group, these students tended to be.
The American students, on the one hand, and the Irish students on the other, produced stereotypes of the Italians which showed very little agreement either with each other, with the Italian autostereotype or with the self-perceptions of Italian students, with the possible exception of the word 11 romantic 11 which was used by all three groups in describing Italians.
In particular, the picture of the 11 typical 11 Italian as emotionally expressive, volatile, temperamental, etc. received scant suppoit from MBTI scores for these students showing a high preponderance (80%) of 11 thinking 11 and 11 judging 11 types in the Italian sample. The tendency to make judgments based on logic and facts, and to prefer an orderly and systematic life style is not consonant with the stereotype of an impulsive, spontaneous people. The trait of sociability was, however, reflected in the stereotype of the Italians and this is to be expected on the basis that stereotypes reflect more 11 public 11 and superficial behaviors.
Finally, the Irish heterostereotype of Americans as 1dominant aggressive, competitive, etc. was in fairly good agreement with the American students 1 autostereotype of themselves but in very poor agreement with their modal self-perception. The heterostereotypes of the Americans by the Italian students also showed fairly low agreement with either American autostereotypes or self-perceptions. There is very little in the data to suggest, for example, that these American 151.
students perceived themselves, either as a group or individually, as 11 happy 11 -the word used most frequently about them by Italian students.
In short, the heterostereotypes produced in this research by the three groups about each other do not appear to show much more than a very small degree of agreement either with group autostereotypes or self-perceptions of group members.
Three possible reasons for this lack of agreement suggest themselves. First and foremost, the stereotype elicited was of the 11 typical 11 or 11 average 11 member of the target group, whereas the autostereotypes and self-perceptions to which they were compared were those of young college students. In terms of age and educational status (and, in the case of the Irish students, socioeconomic class), these samples may well have been atypical of the larger populations from which they were drawn.
Second, Irish and Italian students apparently stereotyped each other and Americans on the basis of rather limited contacts with members of these cultures. In fact, in response to the open-ended question asking the subject to estimate the amoung of personal contact s/he had had with members of the groups to be stereotyped, many Irish and Italian students mentioned tourists, television, and "foreign students " as the nature of this contact. Thus the stereotypes of the Irish and Italians produced about each other and about Americans may well have been based on atypical members of the stereotyped cultures.
American students, on the other hand, may well have stereotyped Irish and Italians on the basis of contacts with Italo-Americans and Irish-Americans, rather than Italian and Irish nationals.

152.
Finally, as mentioned previously, heterostereotypes were measured with the use of open-ended questions whereas both autostereotypes and self-perceptions were measured with forced-choice instruments, and the differences in these formats created problems in direct comparison.
All of these limiting factors suggestt the need for greater refinements in the measurement of stereotypes in future research.
In addition to the predictions just examined this study also examined the relationship between autostereotypes and self-perceptions.
This was done in two ways: modal AVA autostereotypes were compared for each group with their modal self-perceptions, and the three groups were compared on the patterns which emerged when each subject's AVA stereotype was individually correlated with his/her self-perception.
The highest degree of correspondence between autostereotype and self-perception was found in the Italian sample. Seventy-three percent of the Italian sample selected AVA PS 5915 as their picture of the "typical Italian"; fifty-six percent of the Italian sample selected AVA PS 5717 as their self-perception. The correlation between these two patterns is r = .87. It is interesting to note that the Italian autostereotype (PS 5915) focuses on more superficial and public behaviors, whereas the modal self-perception (PS 5717) describes, in part, less visible or "inner traits". Both patterns describe the outgoing, extraverted person who is at ease with the social environment and well able to cope with it. The autostereotype, however, further stresses the ability to manipulate the environment in a forceful and effective way in furthering personal interests. The modal self-perception, 153.
though also describing a socially adept person, suggests the presence of less visible "traits" e.g. -the drive for perfection and for logical answers. This finding appears to support the validity of these measures since an autostereotype should, by definition, describe visible, public, superficial traits whereas the self-perception whould describe the 11 rea 1 inner person 11 • Further, one would expect more group consensus on the autostereotype than on self-perceptions, and this was, indeed, found.
The high degree of agreement between the modal autostereotype and the modal self-perception of the Italian students is also of interest since it suggests that valid cultural differences were reflected in these measures. Italian culture is, by all accounts, cohesive and highly structured. One of its central values is the ideal of effectively dealing with the social environment in furthering one 1 s personal interests (cf. Gambino, 1974). This high degree of consensus on the values transmitted, as well as the effectiveness of the socialization practices themselves, appears to be reflected in the correlations between the autostereotypes and the self-perceptions of these students.
Eighty-four percent of the Italian sample had correlations between the two measures which ranged from .40 to 1.00, and over half the group (52%) had correlations of .80 or greater (refer back to Table 24 for a summary of these data). This is to be expected in a cultural group in which socialization practices stress conformity to a public image and loyalty to group rather than individual values (Gambino, 1974).
A much more complicated pattern was found in the Irish sample.  Table 25 shows that the correlations between the three most frequently occuring autostereotypes and the three most frequently ocuring self-perceptions for the Irish sample were extremely high.
When, on the other hand, the pattern of individual correlations between autostereotype and self-perception was examined, the Irish students were found to be intermediate between the higher degree of congruence for Italian students and the lower degree of congruence for American students.
Sixty-two percent of the Irish sample had correlations between .40 and 1.00, compared to 84% of the Italian students and 49% of the American students.
These data suggest that the majority of Irish students, as individuals, hold self-perceptions which are in moderate to strong accord with their picture of the 11 typical member 11 of their culture, but the 155. students as a group are rather sharply split as to what this "typical member" is like.
It is possible, although of course strictly speculative, that the political history of the Irish people accounts for this finding of differing images. The political separation of Ireland from Great Britain occurred rather recently when the Republic of Ireland Act was passed in 1948 and "provided for Ireland's complete freedom from Great Britain. 11 In practice, however, as Dr. Leo Carroll 1 has suggested, Ireland's ties with Great Britain continue to be pervasive, complex, and ambivalent. It is possible, then, that a clear-cut and widelyheld sense of national identity has yet to emerge in this rather young Republic.
The lowers degree of agreement between autostereotypes and the self-perception was found in the American sample. This was true both for the modal patterns for the group as a whole, and for the pattern of correlations between individual profiles.
Although American students as a group showed moderately strong consensus in their picture of the "typical American" (66% chose AVA PS 7616 as the autostereotype), the modal self-perception for the group, PS 3539, correlates only r = .35 with this profile. AVA PS 7616 describes the "typical American" as a tense, highstrung, driven person -a picture somewhat congruent with the Irish stereotype of Americans as ambitious, aggressive, and hardhitting. AVA PS 3539, the modal American self-perception, 1 Dr. Leo Carroll, Department of Sociology, University of Rhode Island in a personal communication.

157.
describes an unassertive, anxious, insecure and meek type of personqualities certainly at variance with the autostereotype. The sharp discrepancy between the 11 typical American" and the self-perceptions of these college students is further underscored by the pattern which emerged when individual correlations were examined. Forty-nine percent of the students had correlations ranging from .40 to 1.00 between the AVA 11 role 11 and the AVA 11 self 11 profiles, but nearly a third of the sample (31.5 %) had correlations ranging from .00 to -1 .00.
Both the moderately high degree of consensus in the selection of th~ autostereotype and the diversity in self-perceptions is to be expected on theoretical grounds. The fact that sixty-six percent of the group agreed on an autostereotype as well as the content of the stereotype itself. suggests once again the validity of the measure. The diversity in self-perceptions is also to be expected in a multiethnic culture in which one of the few agreed-upon values is a stress on competitiveness and individualism .
Two alternative reasons suggest themselves in explaining the rather substantial proportion of students who saw themselves as quite different from the 11 typical American 11 • One possibility is that these junior college students are, as a group, insecure and low in selfesteem. Community College of Rhode Island, from which the majority of the sample was obtained, was formerly known as Rhode Island Junior College or, more popularly, as 11 Reject 11 -and, unfortunately, this image of a 11 reject 11 is one which quite a few of the students appear to have internalized. Thus a majority of these students see the typical American as aggressive and competitive, but see themselves as meek 158. and unassertive. A second, less likely, possibility is that the low correlations between the 11 typical American" and the "real self" represent a rejection of, or alienation from, ''cultural mores" as they are perceived by these students.
In conclusion, this study found some support from the prediction of cultural differences in self-perceptions. It cannot be too strongly emphasized, however, that neither the similarities within groups nor the differences between groups, should be overstressed.
In the face of the bewildering complexity of individual differences, psychology as a science attempts to simplify and generalize in order to provide useful explanatory and predictive constructs.
In this kind of nomothetic approach, specific instances and exceptions in general rules are minimized. It is most certainly not the case that the richness, complexity, and uniqueness of human personality can be captured by classification into a few 11 types". or explained on the basis of a handful of traits.
Just as the similarities within groups in this study should not be overemphasized, neither should the differences that were found between groups. Although the results of this study indicate that the concept of "modal self-perception" within cultures may be a fruitful one in terms of future research, the deeply pervasive similarities shared by all human beings should not be overlooked. The point was, perhaps, best made by one of the subjects in this research who, when asked to describe the "typical member" of the other two groups wrote, simply, 11 human 11 •