;.—. . My..._x A4.4.4A‘JA AAA-4A BIASEB 3599mm STYLES AS A YES? NR THE MERCY OF RACIAL {LABELS The“: for the Dog-rec a? M. A. WEEKS“ STATE UNEVEESITY Calvin Ray Matthews 1974 I" L W W 4 9639 1 1 t WI W W h 3; W NW NW ' 3 igan Sta liversity 1C1} ‘ 14......1RY ' M re 3 ABSTRACT BIASED REPORTING STYLES AS A TEST FOR THE POTENCY OF RACIAL LABELS By Calvin R. Matthews The present research examined the effects of race of subject, race of author and racial labels on subjects“ reSponses to journalis- tic products. Past research results concerning labels and stereo- types were reviewed. From these results it was shown that various labels and stereotypes may or may not be taken as offensive, depending on circumstances. The present research extends this issue to the third person. That is, the potency of a racial label may well de- pend upon who is using the label. Black and White subjects participated individually in this study by reading a fake article that described an incident in which a small group of Blacks or Whites were depicted as hijacking a plane. Each article, which referred to the hijackers as ”niggers,” “hunkies,” or no racial label, contained the picture of a White or Black author. To measure one's attitude toward the article, each subject was asked to rate an article according to the quality of the report. After handing in all material, each subject then filled out a question- naire which measured recall of what actually was depicted in the hijacking story. Calvin R. Matthews It was hypothesized that: (a) Subjects of both race would tend to respond differently to an author depending on the label used in an article. (b) Comprehension and ratings would also differ as a result of the two—way interactions of the independent variables (race of subject, race of author, and racial label used). Although the general hypothesis for comprehension was not supported, results provided some support for the hypothesis: (a) As for comprehension, both Black and White subjects tended to remember more from articles written by an author of their own race. (b) Over— all, comprehension was less when an author used the associated label of members of the opposite race. (c) Although just the opposite was predicted, it was found that subjects of both races tended to re- member more from articles containing their associated label. As for the ratings, Black subjects rated articles higher than White subjects and overall the Black author was rated higher than the White author. The ”nigger” label was rated lower than the “hunky" label suggesting that the ”nigger” label was a more negative stimulus. There were no significant two-way interactions for White subjects, but there was a tendency to rate articles lower when the hunky label was used by no author. Overall, results indicated that racial labels can be taken as offensive and some labels are more potent than others. In this study the ”nigger“ label tended to be the more negative stimulus and Calvin R. Matthews the race of the author and the race of subject were significant vari- bles in this determination. BIASED REPORTING STYLES AS A TEST FOR THE POTENCY OF RACIAL LABELS By Calvin Roy Matthews A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology T974 cS86>é DEDICATION To the greatest couple on earth, my parents, Mr. and Mrs. Washington Matthews Sr. ii ACKNOWLEDGMENTS The author expresses his deepest appreciation to Dr. Frank Schmidt, advisor and chairman of his thesis committee, for his sup- port, suggestions and assistance throughout the various stages of this thesis, and to Drs. Lawrence Messe and Andrew Barclay for their knowledgable comment and advice. Although not directly involved in this research, I would also like to thank Dr. David Wessel for his helpful advice and comments and for presenting me the options of various areas of research. A special thanks is also extended to Drs. William Hicks and Comradge Henton of Southern University and Drs. Irving Lane and Charles Burchell of Louisiana State University. LIST OF TABLES . LIST OF FIGURES Chapter I. INTRODUCTION 11. METHOD Subjects . Material Design Procedure III. RESULTS . Summary IV. DISCUSSION . APPENDIX TABLE OF CONTENTS A. Nine Different Combination of Articles 8. Rating Forms C. Questionnaire . BIBLIOGRAPHY iv Page vi ID ID ID 12 12 T4 23 29 36 46 48 50 Table LIST OF TABLES Labels for Identification of the Various Dependent and Independent Variables . Overall Analysis of Variance Table of the Q Scores with a Presentation of Eta2 Correlation Between the Independent and Dependent Variables Overall Analysis of Variance Table of the Rating Scores with a Presentation of Eta2 Correlation Between the Independent and Dependent Variables Rating Score Means for ROS X RDA X L Page 15 T7 20 21 LIST OF FIGURES Figure Page l. A 2 x 3 x 3 Factorial Design Whereby ROS x RDA x L . l2 2. ROS x RDA for Q Scores . . . . . . . . . . . 25 3. ROS x L for Q Scores . . . . . . . . . . . 26 4. RDA x L for Q Scores . . . . . . . . . . . 27 5. ROS x RDA x L for R Scores . . . . . . . . . 28 vi CHAPTER I INTRODUCTION One Sunday evening two black undergraduate ”brothers” were jiving about the past Sunday night in a conversation that went like this: Dig it man--the niggers really got down last night. I mean the niggers really partied hard. Man, you just can't beat niggers having a good time. Such lines are very common among young blacks and they are hardly ever. taken as offensive by other Blacks. Yet, the same words could have been stated by a White student, Chinese, Mexican or any member of another ethnic group attending the same party, and more then likely the response would have been entirely different. As a matter of fact, it would not In; very wise to test to see. This is not to suggest that all labels are taken as offensive because as this literature re— view indicates, there are labels with pleasant associations as well. Various labels are witnessed in all ethnic groups. Whether the subjects are Black, White, Italian, Jewish, Catholic, Irish-American, French—Canadian, or even White Anglo-Saxon Protestants, there are still labels like spook, hunky, wop, kike, papist, harp, cannuck, and wasp that would be taken as offensive if used in a similar fashion. Not only can these obvious labels be taken as offensive, but mere proper racial titles may act like potent symbols especially if they I arouse ethnic association. Such labels can be traced back to Bibli- cal time during the days of Samson. The Bible reveals that during Biblical time, just the name Philistine brought hatred from the Israelites, who had been in bond- age under the Philistines for 40 years. Thus the potency of the label was due to an avalanche of stereotypes associated with the label. The Philistines were considered foreign oppressors who worshipped idols. Therefore, the label was considered representative of something un- favorable to the Israelites. This same type of resentment also oc- curred when a Jew was called a Samaritan. Jews saw the Samaritans as being of a different racial stock and having a different manner of worship. So for reasons of this order, various labels took on stereotyped images that tend to block out concrete reality. That is, because of Unapotency of the labels, persons tend to overlook im- portant features, many of which could offer a sounder basis of judg- ment. Thus when Jesus met the woman of Samaria at the well and asked her for a drink, she answered, How is it that thou, being a Jew, askest drink of me, which am a woman of Samaria? For the Jews have no dealings with the Samaritans. Jesus did not let the label fade out her qualities as a human being in having the right to the tree of life. So whether the resentment was due to racial, religious, or ethnic differences the fact that they existed was reality and their continuation travels with time it- self. Even today, to some, the label Philistine symbolizes belonging to a despised class which is guided by material rather than intellec— tual values. It is an important problem for social psychologists to ex- plain how a single label can block out or destroy so much, that is, so many important qualities. In order for a nation to continue to grow and prosper, its people must stress and focus their attention on the more important things in our society. Thus, in the United States, where various racial labels are numerous, one can only imagine how many important thoughts and ideas are wasted as a result of a potent racial label. Thus, it is to questions relevant to this is- sue that the present research was addressed: Are there labels of- fensive to Blacks and Whites? If so, why and how potent are they? Are some labels considered more potent than others, that is, do some labels tend to mask the potential influence of additional information more than others? If so, do the potency of these same labels depend upon who is using the label? Allport (I958) refers to such labels as ”Labels of primary potency" that abstract from concrete reality some one feature and assemble different concrete realities only with respect to this one feature. Such labels may point to an irrelevant physical handicap such as cripple or blindness. Lee (l950) gives as example where a blindman had applied for a position as typist. I knew a man who had lost the use of both eyes. He was a ”blindman." He could also be called an expert typist, a con- scientous worker, a good student, a careful listener, a man who wanted a job. But he couldn't get a job in the department store order room where employees sat and typed orders which came over the telephone. The personnel man was impatient to get the interview over. ”But you're a blindman," he kept say- ing, and one could almost feel his silent assumption that somehow the incapability in one aspect made the man incapable in every other. So blinded by the label was the interviewer that he could not be persuaded to look beyond it. The label of primary potency is exceedingly salient here. Although the blind man meets job requirements, the fact that he is blind stands out in the personnel man's mind. Thus the label "blind" had distracted his attention from concrete reality. Therefore, the Blindman was discriminated against by a misclassified label that had power to bias his judgment negatively. Ethnic labels, typically, have the same effect, especially when they refer to some highly visible feature such as Oriental, Black or European. Thus, ethnic labels may act like labels of pri- mary potency in distracting attention from concrete reality. Not only is the potency of labels like ”nigger" and I'hunky," salient, but even proper names, which ought to invite us to look at the individual person, may act like ethnic symbols of primary potency if the names have ethnic associations. Razran (I936) did a study of ethnic dislikes and stereotypes in which (30) photographs of college girls were shown on a screen to 150 students. The ethnic descents of the 150 subjects were: 63 Anglo-Saxon, 24 German, IS Irish, 9 Scandinavian, 9 Slav, 6 French, 6 Jewish, 3 Dutch, 3 Spanish, 3 Hungarian, and 3 Greek. This group were so stratified to represent the ethnic composition of the white population in this country. The subjects rated the girls on a scale from one to five for beauty, intelligence, character, ambition, general likability. Two months later the same subjects were asked to rate the same photographs and fifteen additional ones. This time five of the original photo- graphs were given Jewish surnames, five Italian, and five Irish and the remaining girls were given names chosen from the signers of the Declaration of Independence and from the Social Register. When Jewish names were attached to photographs there occurred the following changes in ratings: decrease in liking; decrease in character; decrease in beauty; increase in intelligence; increase in ambition. For those photographs given Italian names there occurred: decrease in liking; decrease in character; decrease in beauty; decrease in intelligence. Thus a mere proper name leads to prejudgments of personal attributes. The individual is fitted to the prejudiced ethnic category, and not judged in his own right. The Irish names also brought about depreciated judgment, but not as great as the case of the Jews and Italians. The falling lika- bility of the ”Jewish girls” was twice as great for ”Italians” and five times as great as for "Irish.” It should also be noted that the "Jewish” photographs caused higher ratings in intelligence and in am- bition, findings which demonstrate that not all stereotypes of out- groups are unfavorable. There are also pleasant associations with the labels Black Velvet, Black Queen, and Sweet Chocolate, terms that could indicate a pleasant association for the subject using the label. For instance, the statement, “he is a fine nigger” made from one Black to another could make some Black men blush and would be taken as highly complimentary. Brigham (l97l) also presented a review of various conceptualiza- tions of stereotypes in social psychological research wherein theoreti- cal approaches linking ethnic stereotypes to negative attitudes and to discriminatory behavior are discussed. Some of the reasons given for the negative attitudes toward stereotypes are: incorrect gen- eralizations, categorizations, product of a ”faulty“ thought process, and their resistance to change. This idea of a stereotype is also a label that could very well encompass incorrect generalization, or categorization. Furthermore, this connotation of a label or stereo- type could fluctuate depending on who is using it. Brigham also re— views research on stereotyping and stereotypes; the reviewer is re- ferred to this work for further information on this related area of research. Hayakawa (l94l) states that some labels are more salient be- cause of their emotional tone. That is the label could have been used in a way that conditioned people to fear something associated with the label. Hayakawa gave the example of an introduction: ”John Doe is an ex-convict . . . he's a criminal!“ The impIication that he's a criminal is a much higher level of abstraction wherein this association sometimes causes emotional reaction. Even though John Doe may have undergone a complete reformation--or, for that matter, may have been unjustly imprisoned in the first place—-results show that, nevertheless, he may wander in vain, looking for a job. From Schoenfeld (I942) it should also be noted that the salience of the emotional toned label could possibly be explained by the ”kernel of truth” hypothesis which in essence here would assert that the probability of committing a criminal act is greater among ex-convicts than among the general population. Even though this may or may not be the case there are researchers like La Piere (I936) who would oppose this view and assert that a stereotype could exist with— out a ”kernel of truth.” Allport (I958) also asserted that “Just the mention of 'com- munist,’ 'Negro,’ 'Jew,‘ 'England,l 'Democrats,’ will send some people into a panic of fear or a frenzy of anger." This reminds the researcher of the 60's when the labels “Black~ Panthers” or “militant Blacks" would make some people shiver because of their associations with acts like riots and burning. The label “Klu Klux Klan” could also be considered emotional toned because of its association with acts like lynching and burning. At times, these labels would make people buck their eyes just to hear them. From this, it seems evident that some labels are more potent because of classical conditioning, that is along with the label came associated beliefs and meaning that was either pleasant, unpleasant, or neutral. The researcher became interested in the issue of the potency of racial label as a result of overhearing the conversation that was reported at the beginnign of this thesis, which actually took place between two Black undergraduates. Assuming that the labels, “nigger” and “hunky,' are highly unpleasant for Blacks and Whites, respectively, when used by a member of a different race, the researcher decided to study their potency in affecting responses to journalism. It was predicted that not only would the two labels be potent, but their potency would also depend upon who was using the label. Hovland and Weiss (l95l) conducted a study which investigated the extent to which the past behavior of a speaker was used by targets in making mand-tact1 discriminations of the speaker's present communica— tion. They showed a persuasive message was more effective in changing attitudes when the communicator was very credible than when his credi- bility was questionable. In this study, identical belief statements about atomic submarines were said to be made by J. Robert Oppenheimer (a communicator who tacts) and Prova (a communicator who mands). As expected, the audience altered their own beliefs in the direction of the persuasive message to a greater extent with the tacting communi— cator than with the manding communicator. Thus, Hovland's and Weiss's study has implications for the present research in that the most credible message was judged as the most favorable. Therefore, it follows that the extent to which a racial label is accepted as posi- tive can also depend on the credibility of the source. That is, in the present study, it was reasoned that the source credibility of a speaker would be determined by the extent to which the target of the 1Mand-Tact discriminations may be further explained in Hov- land's and Weiss's The influence of source credibility on communication effectiveness, Publ. Dpin, Quart., I95l, 16, 635-650. message (the subject) perceived that he and the communicator had common bonds-~e.g., a common racial heritage. In the present research three fake newspaper articles were constructed that used the labels, “nigger'I or “hunky" or no label. These articles--supposedly written by either a Black or a White, or an author whose race was not identified--were read and rated by either a Black or White subject. From the perspective outlined above, it was hypothesized that, in general, subjects of both races would tend to respond differently to an author depending on the label used in an article. It was expected that a subject who read an article that was written by an author of a different race would tend (a) to get most offended and (b) remember less about information in the article if the persons described in the story (by the slang racial label) were of the same race as the subject. Four other pre- dictions were also made. I. Comprehension will be higher when subjects read arti- cles written by authors of their own race. 2. In general, subjects of both races will tend to remember less from articles containing their associated label. 3. Comprehension will be less when authors use the associa- ted label of members of the opposite race. 4. Subjects of both races will tend to rate the quality of articles containing their associated label lower. CHAPTER II METHOD Mega: One hundred and eighty undergraduate males in Baton Rouge, Louisiana (90 Black subjects from Southern University and 90 White subjects from Louisiana State University) participated in the study in partial fulfillment of a research requirement in introductory psychology courses. Subjects of each race were assigned by chance to one of three groups (2 experimental and I control). The sixty sub— jects in each group were tested individually. Material Three fake articles, supposedly drawn from an underground newspaper, were used as stimulus essays in this study. These articles describe an incident in which a small group of Blacks or Whites were depicted as hijacking a plane. An underground newspaper was used as the source in an effort to make such an article seem more real. That is, it was assumed that these articles, which often contained slang racial labels, would arouse suspicion if they were attributed to more regular jounalism sources. The picture of a White person or a Black person—-supposedly the author--was attached to the stimuli of 2/3 of the subjects; the remainder read the same stimuli but the author of the article was not identified. IO II Also the article referred to people as ”nigger," ”hunky" or no racial label was used. These manipulations yielded a total of nine different comibnations of articles, as listed in Appendix A. Four rating scales (Appendix B) were developed to assess the subjects' perception of the quality of the report. The rating scales allowed responses on each item, that ranged from I to 5, with a low score representing a less favorable rating. Thus, the overall rating score (R score, Table l) which was the sum of the responses to the four items could yield a minimum of 4 points for the least possible favorable reaction and a maximum of 20 points for the most favorable reaction. The second test for the potency of the two labels “nigger” and “hunky,' was a questionnaire (see Appendix C), which consisted of ten questions that measured recall of what actually was depicted in the hijacking story, was constructed. Since three of the items were worth 2 points (since they asked for compound answers), the questionnaire score (Q score, see Table I), could vary from D to l3, with a 0 representing no recall of the story and I3 representing the highest amount of recall. Both the rating scale and quesionnaire were tested for reliability. The items on both scales were fairly homogeneous in terms of how the individuals responded; that is, the intercorrelations were quite high for both scales. The a coefficient was computed for the rating scale yielding a reliability coefficient of .87. The Kuder-Richardson and Hoyt coefficient yield a reliability of .8l for questionnaire. I2 Qfiéiflfl A 2 x 3 x 3 factorial design was employed in which race of subject (Black or White) x race of author (Black, White, or no author) x racial label in article (nigger, hunky, or no label) were the inde- pendent variables that were examined. Below (Figure l) is a graphic representation of the design with the three independent variables labeled as in Table I. There were ten subjects in each cell of the design. -.L__ NL HL NOL RDA I I I I I I -_. r_§A JIIIA %NDA_ BA % WA A+NDA on I WA IN0A_ ,5 I I I I I I | I I I I I 59$ l I I I i l BS I I I I I I I I I I I I I Figure l.--A 2 x 3 x 3 Factorial Design Whereby ROS x RDA x L. Eresefigze Each subject reported individually and was given one of the nine stimulus articles to read. He was told that this article was taken from a past issue of an underground newspaper and the researcher wanted him to read it. After reading the essay, the subject was asked to rate the articles on the quality of the report. He was told that he did not have to discuss his ratings and that all his responses I3 would be kept confidential. After handing in the essay, the subjects were asked to fill out a questionnaire which consisted of ten ques- tions that measured his understanding of what actually happened dur- ing the hijacking. They were then thanked for their participation and excused from the session. CHAPTER III RESULTS An information score (Q score, see Table l) and a rating score (R score see Table l) were the two dependent variables which were used to examine the hypotheses. The Q score was a measure of subject's comprehension of material in the various articles and the R score measured a subject's attitude toward an article. Since the initial step was to assess the effect of the three independent variables-- race of subject (Black and White), racial labels (nigger, hunky, and no label) and race of author (Black, White, or not identified)2--on subjects' responses to the dependent variables, a separate 2 x 3 x 3 analysis of variances (Tables 2 and 3) were performed. In addition a correlation coefficient of the overall Q scores with the R scores Iwas conmuted to examine the extent to which the amount of comprehension was related to the judgment of an article. Results yielded a correlation of .l2 which was not significant. Thus, it appears that a subject's likes and dislikes of an article did not relate to his amount of comprehension. The low correlation between R scores and Q scores made multi- variate analysis of variance unnecessary. That is, the potency of 2The various label notations, e 9., BA and WA etc., used to identify the independent variables are listed in Table I. I4 IS the two racial labels would not have been better determined by con- sidering both comprehension and attitude together, since a person's attitude toward an article told one very little about the ability to comprehend other material in an article. TABLE I --Labels for Identification of the Various Dependent and Independent Variables. Questionnaire Score Q Score Rating Score R Score Black Subjects BS White Subjects WS Black Author BA White Author WA No Author NOA Nigger Label NL Hunky Label HL No Label NOL Race of Subject ROS Race of Author RDA Labels L From the overall analysis of variance for Q scores, summarized in Table 2, a significant three-way interaction was predicted, since it was expected that a subject reading an article written by a member of the opposite race containing his associated label would tend to get offended and remember less in that condition. However, results did not support this hypothesis in that there were no significant three-way interaction for Q scores, but the three two—way interactions l6 were significant. The significant RDA x RDA interaction is pictured in Figure 2. Tests of simple effects (Winer, l97l, pp. 347-35l) were performed separately for the ROS at each level of RDA (i.e., BA, WA, and NOA). These tests supported the prediction that comprehension would be higher when articles were read by a member of the author's own race. For the ROS at BA (F = 5.85, df_= l/l62, p < .05), Black subjects tended to remember significantly more than White subjects in this condition, and for ROS at WA (F = 2.08, d: = l/l62, p < .05) the trend was just the opposite; White subjects tended to comprehend significantly more than Black subjects from the White author. For ROS at NOA, no significant difference was found. Simple effects for RDA at BS (F = 2.96, gf_— 2/l62, p < .l0) showed that comprehension did differ for Black subjects as a function of the race of the author. The Newman—Kuels test performed on the three means in this condition indicated that Black subjects remembered significantly more from the Black author (q = 3.49, p < .l0), while comprehension from articles written by a White author or no author was about the same for Black subjects. RDA at WS was not significant; that is, there was no sig- nificant difference in amount of comprehension for White subjects with regards to the race of the author. The interaction, ROS x L (depicted in Figure 3) VIEIdEd 519‘ nificance at the .002 level. Therefore, tests of simple effects were performed for the ROS at the HL, NL, and NOL. These tests yielded exactly the opposite of what was predicted. That is, for ROS at NL I7 TABLE 2.--Overall Analysis of Variance Table of the Q Scores with a Presentation of Eta? Correlations Between the Independent and Dependent Variables. ——_. _ --_ .___.___ .___. _.. Sources ss DF MS F p ETA2 ROS .05 1 .05 .01 ns RDA .84 2 .42 .11 ns L .74 2 .37 .09 ns Ros x RDA 38.93 2 19.47 5.00 .008 .05 ROS x L 50.70 2 25.35 6.5l .002 .06 RDA x L 85.22 4 21.31 5.47 .0005 .10 ROS x RDA x L 14.07 4 3.51 .90 .464 .02 ERROR 63l.l0 I62 3.90 .76 TOTAL 821.66 179 (F = 6.85, df = l/l62, p < .Dl), results show that Black subjects remembered significantly more than White subjects when the label ”nigger" was used. For ROS at HL (F = 6 l5, gf_= l/l62, p < .05) the trend was the same; that is, White subjects tended to rember sig- nificantly more than Black subjects when the label "hunky" was used. For ROS at NOL, there was no significant difference in amount of com- prehension. Simple effects were also computed across labels for each race of subject; that is, L at BS and L at WS. For L at BS (F - 3.50, df = 2/l62, p < .05) results showed that comprehension did differ for Black subjects as a function of the label used. The Newman-Kuels test on these means (q = 3.72, p < .05) revealed that Black subjects remem— bered more when the NL was used as opposed to the HL, while there were no significant difference between the use of the HL and NOL. Compre- hension also differed for L at WS (F — 3.l2, df = 2/l62, p < .05). l8 Thus, the Newman-Kuels test (q = 3.53, p < .05) showed that White sub- jects remembered significantly more wnen the HL was used as opposed to the NL. Again, no significant difference was found between the use of the NL or NOL. Also, in the 0 score analysis the RDA X L interaction (dis- played in Figure 4) yielded significance at .005 level. Simple ef- fects were performed for RDA at each labeland for L at each race of author. For the FWfllat NL (F = 6.85, gf_= 2/l62, p < .Dl), and ROA at HL (F = 4.92, df = 2/l62, p < .Dl), results indicate that compre- hension for both the NL and HL differed as a function of the race of the author. The Newman-Kuels test on the means showed that compre- hension was significantly higher when the BA used the NL as opposed to when the WA used the NL (q = 5.ll, p < .Dl). Also, interesting to note is that comprehension for the NOA using the NL was signifi- cantly greater than when a WA used the NL (q = 2.95, p < .05). This suggests that the NL is more potent when used by a WA. This test also showed that for the RDA at HL, comprehension was significantly greater when a WA used the HL than when a BA used the HL (q = 4.20, p < .Dl). For RDA at NOL, no significant difference was found (F = .03, df = 2/l62, ns). Analysis of simple effects for L at WA (F = 5.l3, df = 2/I62, p < .DI) and L at BA (F = 5.72, df = 2/l62, p < .Dl) indicated that for each race of author, comprehension did vary as a function of the label used. The Newman-Kuels test on the means of this simple effect showed that for L at BA, comprehension was signifi-