'L'HE FOSWEGN OF WOMEN 3N HUMGR Thesis fer the Degree of M. A. - Mitzi-ABM STATE UNNE‘ASITY AWARD MARY BALL 1976 L AAAAAAAA a ”apazgm W 041105: ABSTRACT THE POSITION OF WOMEN IN HUMOR BY Howard Harry Ball The position of women in society with regard to domination-subordination was examined by using jokes as an indicator of cultural patterns. A content analysis was performed on jokes from two anthologies reflecting two time periods, 1936 and 1972, respectively. From within these, a random sample of 300 jokes was drawn. Two modes of dominance were formulated: the personal mode and the institutional mode. In the personal mode the findings were that dominance of males over females exceeded dominance of females over males in 1936, while in 1972 dominance of females over males exceeded dominance of males over females. The change was accounted for by a decrease in males dominating females. Recently there is less dominance of either kind. Dominance is princi- pally a same-sex phenomenon. The institutional mode findings are that males have higher occupational rankings overall, higher occupational and higher status rankings even when women are dominant Howard Harry Ball in the personal mode, and greater frequency of occupa— tional mention. The inequality is larger for 1972 than 1936. THE POSITION OF WOMEN IN HUMOR BY Howard Harry Ball A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Sociology 1976 ACKNOWLEDGEMENTS I wish to acknowledge the keypunching assistance of Mary L. Marshall. My committee members, John Gullahorn, PhD, Harry Perlstadt, PhD, and Barrie Thorne, PhD, showed unusual patience with the pace of develop— ment of this thesis. They were always supportive. John Gullahorn deserves special mention for his ready accessi- bility,without which I surely would have succeeded in trying the patience beyond endurance of each of the committee members. My parents, Hyman and Dorothy Ball, and my brother, Richard Ball, were steadfast in their confidence of me, without which I would have lost patience with myself. ii TABLE OF CONTENTS INTRODUCTION. . . . . FUNCTIONS OF HUMOR. BASIC ASSUMPTIONS SAMPLE. PROBLEM . RESULTS SUMMARY OF RESULTS. INTERPRETATION AND DISCUSSION . . The Personal Mode. The Institutional Mode Other Data . CONCLUSION. APPENDIX. REFERENCES. iii Page 10 12 14 22 35 38 38 39 41 43 44 48 Table 10 11 12 13 14 LIST OF TABLES RELATIVE FREQUENCIES OF MALE-DOMINATING- FEMALE AND FEMALE-DOMINATING-MALE JOKES BY TIME PERIOD . . . . . . . . . . RELATIVE FREQUENCY OF MALE-DOMINATING- MALE JOKES BY TIME PERIOD. . . RELATIVE FREQUENCIES OF FEMALE-DOMINATING- MALE JOKES COMPARED TO FEMALE-DOMINATING' FEMALE JOKES BY TIME PERIOD. . . . FREQUENCY OF APPEARANCE IN JOKES BY SEX AND TIME PERIOD. . . . . . . . . FREQUENCIES OF MALE AND FEMALE PATTERNS COMPARED O O O O O O O C O O O O O O O O 0 NUMBER OF JOKES WITH SUPERIOR OCCUPA- TIONAL RANK BY SEX . . . . . . . AVERAGE OCCUPATIONAL SCORE BY SEX AND TIME PERIOD (EXCLUDING NO MENTIONS). FREQUENCY OF MENTIONS OF OCCUPATION BY SEX AND TIME PERIOD. . . . . . . . FREQUENCY OF NO MENTIONS OF OCCUPATIONS BY SEX AND TIME PERIOD . . . . . . OCCUPATIONAL SCORE AVERAGES FOR MALES AND FEMALES IN JOKES WHERE A FEMALE DOMINATES A MALE . . . . . . . . . . . . . . OVERALL STATUS OF MALES AND FEMALES COMPARED FOR FEMALE-DOMINATING-MALE JOKES. . . . . . . . . . . . HYPOTHESES BY TEST RESULTS . RANKS AND FREQUENCIES OF THE MOST OFTEN USED METHODS OF DOMINATION BY TIME PERIOD. SEX DISTRIBUTION . iv Page 24 25 26 27 27 29 3O 30 31 32 33 37 45 46 INTRODUCTION At the time this is being written a Women's Libera- tion Movement has been ongoing for a number of years in the society. Among the premises of this movement are that women are an oppressed group, dominated by men, that women have been treated unequally and discriminately in jobs, that women receive less of the rewards of the society than men. Already social science has been brought to bear on this "social problem." For example, studies show woman's image in advertising (Komisar, 1971) and textbooks (U'ren, 1971) is invidious. Similarly their position in the economy (see Knudsen, 1971) has been shown to be an unfavorable position relative to men. The case for the existence of differentials of economic rewards also seems to be beyond doubt. These differentials are obviously due to discrimination. Discrimination usually occurs in forms prescribed by institutional and cultural patterns. For its beneficiaries its immediate effects, at least, may often be benign; and women often are its beneficiaries. The effect of such cultural patterns requires some inferences that often cannot be demonstrated directly. 1 2 Therefore, the more studies that indicate bias in cul- tural patterns, the stronger the case that differentials are due to sexism and not chance or arbitrary distinc- tions. This is more or less a basic assumption of content analysis (Berelson, 1952: 18; Berelson, 1948: 20-24). Studies which illustrate by example that the image of woman is that of a limited and/or inferior person are useful, but fail to show how much more often this is the case than the reverse image, for which examples of capable, equal or superior may also be found (that is, such studies do not meet the "requirement of system" Berelson describes--1952: 17. Studies which illustrate by example do not lend themselves to hypothesis testing oriented toward establishing scientific propositions. They do not allow statements of relative emphases and omissions)- The use of content analysis on products of the culture, especially its mass media, has been a traditional approach in ethnic studies (e.g., Berelson, 1946; Barron, 1950). It therefore seems that application of this technique in an additional study might be a fruitful contribution towards strengthening or weakening arguments about the nature of cultural patterns in the instance of women. The aim is to be somewhat more quantitative than some earlier studies (such as Komisar, 1971). Barron's (1950) study on humor is impressively novel; most content analyses use serious material of 3 substantial length. The assumption associated with serious material is that behind the theme, or phrases or key words lies an intention, or at least a habit of thinking, that relates more or less straightforwardly to the material. If positively evaluated adjectives always appear with male descriptions and the opposite with female descriptions, then a pro-male bias may be indicated. But jokes and anecdotes are generally much shorter than stories; they rely even more on stereotypes and the adjective may be used simply to make the joke funnier. Thus, straightforward assumptions may be less relevant. However, this is a question of degree, not kind. We might find that males are described with posi- tive adjectives and females with negative ones in jokes and draw the same conclusion. On the other hand, much in humor depends on the meanings the reader/listener supplies. Some degree of subtlety seems to be incor- porated in jokes. We doubt that the straightforward application of many techniques such as adjective counts can tap all that a joke conveys. Whereas, for example, a short story might indicate insult by derogatory words, jokes may insult by the inference or the conclusion one makes. The emotional force associated with a joke's punchline can add strength that might not show up with conventional techniques. Read the following selection: The young lady nodded her head at what the psychiatrist was telling her, and said, "Yes, I see, Dr. Schmidt. At least, I see every- thing but one point. The one thing I'm hazy about is this phallic symbol you men- tioned. What's a phallic symbol? "A phallic symbol," said the psychia- trist, "is anything that can be used to represent or symbolize a phallus." "But what is a phallus, doctor?" The psychiatrist said, "I think I can explain that most clearly by a demonstration." He stood up, unzipped, and said, "This my dear young lady, is a phallus." "Oh," said the girl, suddenly compre— hending, "I see. You mean it's like a prick, only smaller." (Asimov, 1971: 378) Somehow it seems to us that a woman accusing a man of having a small penis, while insulting, is far more force- ful as an insult in the punchline to a joke! We are obliquely pointing to the problem of the unit of analysis. Whether to take the word, the sentence or the theme as the unit of analysis may be more a matter of choice in cases other than jokes. With jokes, using the theme or entire joke as the unit of analysis may be virtually dictated. For example, who is dominant in the joke above? We might say there is a doctor-patient relation or, more generally, an expert-layman relation, indicating the man is dominant. The ability to insult is another indication of dominance, one which both male and female here achieve (the male by his act, which was unprofessional, the woman by her comment, which was degrading). But getting in the last insult (by virtue of the joke ending) is, in our Opinion, the most important. S It sets the joke apart, as the editor (Asimov) says, an anti-psychiatrist joke; the woman is the unquestion- able victor. FUNCTIONS OF HUMOR While humor lends itself to study as data on cul- tural patterns generally, there is an additional reason why humor may be useful for studying male-female rela- tions. The reason is that Humor lends itself particularly well to use as a conflict device because of its almost boundless limits in subject matter, and because its nature is such that it often contains more or less well concealed malice. (Burma, 1946: 710) The utility of humor for control and expression of conflict has received notice by several authors. Burma (1946: 710) in his article goes on to say, Throughout the history of minority—majority relations in this country the set of tech- niques which we may denominate by the general term humor has played a definite role in inter-personal and inter-group relationships. Apparently all minority groups suffer in this manner, and apparently all use the same weapon in return. While the question of the theoretical status of women as a minority group is interesting and one we hope to address ourselves to in another paper, here we wish merely to acknowledge that women often seem to suffer the same fate as minorities and there is sufficient reason to examine humor as it pertains to women. 7 Whether humor is returned by women is an issue which seems to be a subject of debate for social scientists (Weisstein, 1973) and the popular press (see, for example, Harrington, 1973: 21). Berlyne (1954: 811) quotes Stephenson on the con- flict and control functions of humor. Humor may ”strengthen the morale of those who are present and undermine the morale of those against which it is aimed..." and it may "express approval or disapproval, develop common attitudes, indicate safety or friendship." Coser (according to Berlyne, 1954: 811), for example, found that in staff meetings at a psychiatric hospital More jocular remarks were made by senior par- ticipants than juniors....Seniors frequently used juniors as targets for their wit, but the converse never occurred...[while] the usual targets for juniors were either them- selves or patients and patients' associates. The interpretation of these observations was that humor expressed aggression on one hand, while under certain circumstances it served "to reconcile and affirm social values." Weisstein (1973: 51) succinctly summarizes some of the functions of humor: ...to establish, maintain or reinforce dif- ferences in power. Humor can serve as an expression of pleasure, affection, love, play, recreation, or an aesthetic, and be used as a vehicle for wit, argument, thought. But it can also be used as an expression of the exclusion of others; it can define normative behavior; it can be used as a signal that a situation is not at all serious, or that it 8 is so serious that we had better laugh if we are to be able to do battle with it; it can be used as a display of personal charm and attractiveness. Given the above uses of humor, it would seem neglectful not to study also its content, especially as a device of domination. Weisstein (1973) seems to think that the so-called lack of sense of humor among participants in the woman's movement is at least two-fold. First, women are not laughing any more at jokes that are insulting to them and, secondly, a women's subculture has yet to be developed. She notes that other oppressed groups have developed humor-~"it is a weapon or a technique of sur- vival used by the oppressed. It is the powerless fighting back" (p. 88). This author finds that Weisstein's observations are selective, focusing only on the humor of the oppressed group that ridicules the oppressor and ignoring ways in which a group laughs at itself-~as, for example, the section of Jewish jokes selected by the Jewish Asimov * (1971). However, it is not our purpose to join in this it It is this, I think, that is behind the charge of a lack of sense of humor in the women's movement--a charge that Weisstein fails to adequately answer. Women do suffer many things, I agree, and on the face of it this would seem to be why women don't laugh. But other suffering groups laugh at themselves; why are women different: Or are they? Any copy of Playgirl magazine will have funny cartoons involving women, and I presume women laugh at them, although I would not describe the involvement as derogatory to women. Barrie Thorne notes 9 debate at this time. Rather, analysis can reveal whether there are jokes where women get the better of men, how often they appear, and in what form; as well, of course, as the reverse situation. that many women's magazines have male editors. This is obviously an interesting path for inquiry but one beyond the scope of our present purpose. Jokes of this sort are, we would assume, most preva- lent in sources generally restricted to women, such as women's magazines. How often they appear in a general collection, therefore, might be an underestimate of the existence of these jokes. Yet appearance in general collections, especially those edited by males, may there- fore be a better reflection of the culture overall. BASIC ASSUMPTIONS The specific assumption of this study was stated earlier: the content of jokes with regard to sexism is a reflection of the culture of American society. The general assumption of content analyses at large, of which this specific assumption is a part, is as follows: Content analysis assumes that inferences about relationships between intent and content or between content and effect can validly be made or the actual relationships established. (Berelson, 1948: 6) ********** This assumption that knowledge of the content can legitimately support inferences about non-current events is basic to the central contribution of content analysis, namely to illuminate certain non-content areas.... Content analysis is done to reveal the pur- poses, motives, and other characteristics of the communicators as they are (presumably) "reflected" in the content.... (Berelson, 1948: 6) A second assumption that Berelson considers funda- mental to content analysis is that the study of manifest content is meaningful. That is, the "meanings" the analyst uses correspond to the "meanings" intended by the communicator or his audience. Berelson considers "manifest content" to vary along a continuum. That is, the more manifest the content, the 10 11 more different readers will assign the same meaning to that content. A news story of a train wreck is an example from the "highly manifest" end of the continuum, while an obscure modern poem is at the other end (where it is unlikely that two different readers will get iden- tical meanings from the content). The material utilized in this study departs from the ideal of perfect manifest content. However, it is more manifest than latent. The categories are chosen to tap such manifest content, e.g., occupations. A third assumption listed by Berelson is: Content analysis assumes that the quantitative description of communication content is mean- ingful. This assumption implies that fre- quency of occurrence of various characteristics of the content is itself an important factor in the communication process, under specific conditions. , (1948: 8) The particular use of content analysis in this study rests upon the above assumption. This use falls under one of three main groupings identified by Berelson: First, there are the analyses of content which are designed to illuminate the cultural or personal conditions under which the com- munication was produced, i.e. the precondi- tions which determine the nature of communica- tion. In this group are the following specific uses: To reflect attitudes and interests ("cultural patterns") of population groups.... (1848: 18) A second use of the content analysis here is from Berelson's second group--"To describe trends in communi- cation content" (1946: 18). SAMPLE The usual notion behind a sample is that it should be obtained in such a way as to be representative of the universe one is sampling. Most commonly this is done by selecting a random sample. Every unit should have an equal chance of being included in the sample. This requires a list of every unit in the universe or some mechanism for choosing in a genuinely random fashion. Thus, the first problem was to select a random sample of jokes from all jokes produced or being used in a particular time period. A practical way of doing so is impossible. Limiting the universe to all published jokes does not overcome this difficulty. In order to meet the requirements of practicality, the choice of a random sample at this level was abandoned. Instead two anthologies were selected on an availability basis; one bore the copyright date of 1936 (COpeland), and the other 1972 (Asimov). These available anthologies were thought to be good sources because anthologies tend to be inclusive of all the jokes repeated at a particular time and remembered from the recent past. They cover several themes and subjects and tend to be less typed and selective than the 12 13 jokes in a particular magazine. This is far more efficient than trying to sample from many issues of many different magazines. It also avoids excessive redundancy. Within this now limited "universe" of jokes in each anthology 3 random sample was chosen by using a table of random numbers.* Since one anthology has its jokes numbered consecutively, the table could be applied directly. The other anthology had only pages numbered; therefore, the table was used to choose pages. Once a page was chosen, a joke was selected on that page by the roll of a die, counting down the page the same number of jokes as the number shown on the die. Because of the way the sample was selected, inferen- tial statistics can be validly used only for generalizing to the level of the anthologies. Inferences to jokes at large and cultural patterns generally can, of course, be made but cannot be supported with inferential statistics. This is no greater limitation than that of many similar studies, but one not to be forgotten. it This is the difference between the target universe and the sampled universe (Mueller, 1961: 341). PROBLEM Thus far, we have given the reasons why we expect a content analysis of jokes to be a worthwhile investi- gation and have stated the problem in a general way. Now hypotheses will be stated explicitly and specifically. To review, the general statement of the problem is that there is subordination of women in society and its associated cultural patterns. Jokes were the chosen indicator of cultural patterns. The major hypothesis states that in jokes men are dominant over women. This hypothesis can be formulated in detailed operational terms both generally and with specific sub- hypotheses (to follow later). The concern for domination-subordination led to a personalistic focus. After reading several jokes at the ’beginning of analysis, domination seemed most forcibly expressed in personal terms, one individual pitted against another. In America, jokes tend to portray personal relationships. Even jokes about groups such as religious groups express their points through the vehicle of indi- viduals. Thus, a joke that would show some difference between three religious groups will have Reverend Brown, perhaps, Rabbi Levine, and Father Sweeny. 14 15 For example, this joke from Ms. Magazine (November 1973): MAN: Do you know the Women's Movement has no sense of humor? WOMAN: No ..... but hum a few bars and I'll fake it! was analyzed as having the woman dominate by use of a (quasi) minority stereotype by the (quasi) minority group member against the dominant;* that is, turning tables. Additionally, it is the woman who changes the subject and who gets in the last word. Another example is the anti-psychiatrist joke quoted earlier. There the woman dominates by insult. This personalistic focus led to various "domination methods" including, for example, such methods as insult- ing, implying ignorance, having the last word, possessing special privilege, and changing the subject. Some colleagues have argued that in the psychiatrist joke, the man's occupational status made him dominant. More properly speaking, this is a role relationship of professional-client. The woman's occupation, if any, is not revealed. These colleagues are, of course, correct about this role relationship. It is also significant that the woman's occupation is not revealed while the man's is revealed. But who is dominant? Within the context of this joke the woman clearly comes out ahead. * See Hacker (1951). 16 To avoid ambiguity in coding such situations, an occupational category was tabulated separately, as was certain additional status information. Two modes of dominance were postulated: the personal mode, which included the sorts of methods indicated above (see also Table 13), and the institutional mode, which included status attributes such as occupational role. Although sought, no special formula for integrating both modes was found. Recording of the method of domination was retained and understood to apply to the personal mode only. Hypotheses l to 7 are personal mode hypotheses. Hypotheses 8 and 9, while more institutional in nature, are provided for descriptive purposes and the hypotheses that follow test facets of the institutional mode. “IA: The relative frequency of jokes with men dominant over women is greater than the relative frequency of jokes with women dominant over men. Operationally, this is simply a counting procedure using as the unit of analysis the item. The items in this case are the jokes in the sample. Elsewhere in the analysis we will use as the unit of analysis the characters in the jokes. Dominance was determined by judgmental decisions made by the author. Part of the basis of these judgments was identifying a domination method (e.g., insult, or other methods listed in Table 13). 17 While the major hypothesis stated that men are dominant over women, this raises the question whether this has always been so. It is presumed that in many areas society has moved away from clear, fixed roles for males and females to a society with unclear, overlapping roles. Therefore, we have included two time periods in the sample to provide some trend data and can add the hypothesis that men are dominant over women but less so now than in the past. Operationally H1A will cover the combined sample so that H1B can be added. HIB: The percentage difference between jokes with men dominant over women and jokes with women dominant over men is smaller for 1972 than 1936. What is being said here is that we expect the advan— tage males have over females is becoming smaller over time. The male's advantage is the amount that male over female dominance exceeds female over male dominance (female advantage is the amount that female over male dominance exceeds male over female dominance). On assumption behind hypotheses H1A and H is that 1B the percentage of jokes with men dominant over women is reciprocal with the percentage of jokes with women domi- nant over men. However, there may be a large "neutral" segment from which increases or decreases are drawn. Therefore, the following hypotheses were formulated as guides to a more careful analysis: 18 H2: The relative frequency of jokes with women dominant over men has increased over time. H3: The relative frequency of jokes with men dominant over women has decreased over time. Along the same line of thought we realized that men might still be dominant in jokes but their domination might not be over women but be an increased domination over other men. That is, male domination whether over males or females is likely to be a constant. Similarly with women. Therefore, two additional hypotheses were formulated: H4: The relative frequency of men dominant over men has increased over time. H5: The relative frequency of women dominant over women has decreased over time. The presumption here is that if men have a constant need to dominate and if they are dominating women less, they must consequently be dominating others (men) more. The presumption with regard to women is that if there is a constant need to dominate, and if they are dominating males more often, then they must be dominating women less often. l9 Berelson says Content analysis stands or falls by its cate- gories. Particular studies have been pro- ductive to the extent that the hypotheses have been insightful and the categories clearly formulated and well adapted to the communication content. (1948: 88) The categories for the above hypotheses form a group which was labeled "pattern of dominance": (1) male over male (M/M), (2) male over female (M/F), (3) female over female (F/F), (4) female over male (F/M), (5) not applicable (NA), (6) other. The "other" category gave way upon analysis to additional categories for jokes with indeterminate sex members: male over indeterminate (M/I), female over indeterminate sex (F/I), indeterminate sex over male (I/M), and indeterminate over female (I/F). The hypotheses thus far focus on the concern for power in the form of personal dominance. But there is a subtler form of favoritism: attention. Do men or women receive more attention? Because more attention is given to more highly valued objects, more characters in jokes will come from more highly valued categories of people. Evidence indicates that qualities of maleness are more highly valued in society than qualities of femaleness. Therefore: H8: Men will appear more often in jokes than women. This is a simple count of males and females (the categories) regardless of the number of jokes. This is 20 dependent on the selection of jokes--male names often being used as a convention when sex doesn't matter. More telling and useful for comparison may be: H9: There will be more male characters in each joke than female characters. In the extreme case attention is simply given to one sex in virtually all activities, the other being seen as unessential or a group of non-persons.* Three hypotheses were formulated concerning the institutional mode. Occupation being of high importance to one's overall status, the expection is: H10: In each joke in which men's and women's occupations can be identified, men's occupations will be superior in rank to women's occupations. Because of the duality of domination modes, we hypothesized that one might be a compensation, of sorts, for the other. When women bested men, which we expected would be via the personal mode, men might have superior social positions. Thus: * See Goffman's discussion (pp. 151-153) of non- person, especially "...the role of non-person usually carries with it some subordination..." (p. 152). 21 H When women are personally dominant over men, 11' the men will have suprior status attributes such as A. occupation B. majority-minority or minority rankings (e.g., ethnic minority woman, ethnic majority male) C. other status attributes. This is in some sense a test of "compensation." The two different modes may have a sex-role association and, if they occur in a pattern, it would suggest a weighting scheme, that "really" the institutional mode counts for more. If the two modes could be independently ranked and if the hypothesis is found true, it would be a further confirmation that we live in a male dominated society. The hypothesis is also a test of status inconsistency; our expectation is that people's status will be incon- sistent in a way that results in male dominance, or ameliorates a male's otherwise poor status. Thus, the hypothesis is phrased in the above terms rather than vice-versa. This would allow us to see if women ever become "fully dominant" over men-~the presumably rare event. RESULTS The Copeland anthology yielded a sample of 146 jokes while the Asimov anthology yielded a sample of 154 jokes for a combined total of 300 jokes. For each joke sampled a coding sheet indicated the sex of each character for up to four characters, the domination pattern (e.g., male dominating a male was recorded M/M), the method of dominance (e.g., insult), the number of the dominant character (i.e., l, 2, 3, or 4), the number of the sub- ordinated character, the occupation and its two digit code for each of the four characters, the overall status (low or high) rating for each character and other infor- mation. This other information included a primary status rating that was relevant to the relationship in the joke; for example, mother, daughter, son, foreigner, aged (low), aged (high), young (lo), young (hi), Jew, Black, Scotch, etc. Also, an identification number, the page number on which the joke appeared in the anthology, its place on the page, the source anthology, etc., chapter number and whether the character (1-4) spoke in dialect. The information from these sheets was punched into cards. Analysis was performed on the Michigan State 22 23 University CDC 6500 computer using the Statistical Package for Social Scientists. i: H can be tested by Table l. The total sample 1A shows weak support. Thirty-four percent of all the jokes had males dominant over females (M/F), while 33% had females dominant over males (F/M). This percentage was computed by using the 86 jokes which contained characters of both sexes as a base. However, this view of the overall sample masks what has occurred at the two different points in time. Look- ing again at Table 1, one sees moderate percentage dif- ferences in each subsample. In 1936, males were dominant over females more often than females were dominant over males, while in 1972 the reverse was true. Thus, the predicted trend of Hypothesis H1B is supported. Moreover, this reversal in dominance over time was because of simultaneous but unequal change in both patterns of dominance, M/F and F/M. Both of these moved in the directions predicted by H2 and H3, respectively. Note that the change in M/F was greater (13%) than that of F/M (3%). If the amount of dominance in jokes is constant over time, then the results above showing a decrease of M/F jokes suggests an increase in other patterns. F/M * A listing of hypotheses and a summary of results follows this section. TABLE 1 24 RELATIVE FREQUENCIES 0F MALE-DOMINATING-FEMALE AND FE LE-DOMI - LE T E D Domination pattern time 1936 1972 Combined Male dominating Female (M/F) 41% 28% 34% (16) (13) (29) Female dominating Male (F/M) 31% 34% 33% (12) (16) (28) Other* 28% 38% 33% (11) (18) (29) Total 2 sex jokes 100% 100% 100% (39) (47) (86) Percentage difference 10% -6% 1% * Jokes that had characters of both sexes but either had no dominant character or had M/I, I/M, F/I, or I/F. (I = indeterminate sex). showed only a small increase; therefore, one might expect an increase in the patterns such as M/M, as H4 predicts. However, Table 2 actually shows a decrease in the rela- tive frequency of M/M jokes, refuting the hypothesis. H5, on the other hand, is supported. Line 2 of Table 3 shows a decrease of F/F jokes over time. There is non-support for H6. Table 3 indicates that, in relative terms, women usually dominate other women rather than dominating men, although the 1972 data show a 14% decline from 1936. (Each percentage was calculated on the appropriate base, two or more females for F/F jokes, two sex jokes for F/M.) 25 TABLE 2 RELATIVE FREQUENCY OF MALE-DOMINATING-MALE JOKES BY TIME PERIOD Percent jokes Total jokes with Time Male-dominating-Male two or more males 1936 60% 98 (59) 1972 55% 113 (62) Total 57% 211 (121) Percentage difference 5% H7 also must be rejected. M/M jokes occur about 40% more frequently than M/F jokes. Eighty-one percent of 73 jokes from 1936 were M/M compared to 41% of 39 jokes that were M/F, and 65% of 95 jokes were M/M from 1972 compared to 28% of 47 jokes that were M/F. Men appear four times as frequently as women in the total sample, as the data in Table 4 show, supporting Hypothesis 8. This ratio of males to females was more pronounced in the 1936 sample, and declined in 1972. However, it may be in the world at large, jokes are a "man's world." Consistent with findings on Hypothesis 8 are the findings for Hypothesis 9. This hypothesis is supported, as can be ascertained by carefully examining the sex dis- tribution (in Appendix). 26 TABLE 3 RELATIVE FREgUENCIES OF FEMALE~DOMINATING~MALE JOKE M AR T E E-D IN T N - FE LE KE B E E D 1936 Domination pattern Percent Base N Female dominating Male (F/M) 31% 100%=39* Female dominating Female (F/F) 57% 100%=7** Percentage difference -26% 1972 Domination pattern Percent Base N Female dominating Male (F/M) 34% 100%=47* Female dominating Female (F/F) 46% 100%=l3** Percentage difference -12% Combined Domination pattern Percent Base N Female dominating Male (F/M) 33% 100%=86* Female dominating Female (F/F) 50% 100%=20** Percentage difference -17% * All in sample containing members of both sexes (see Table l). ** All jokes in sample with two or more females (see Table 14). 27 TABLE 4 FREQUENCY OF APPEARANCE IN JOKES BY SEX AND TIME PERIOD Sex 1936 1972 Combined Males 220 253 473 Females 50 68 118 Ratio of total males to total females = 4.0 Ratio of males to females in 1936 = 4.4 Ratio of males to females in 1972 = 3.7 To make this clear, the jokes that have more males than females (e.g., MMF), which for convenience will be called the male pattern, will be added up separately from those containing more females than males, which will be referred to as the female pattern. TABLE 5 See Table 5. FREQUENCIES OF MALE AND FEMALE PATTERNS COMPARED Pattern 1936 1972 Combined Male pattern 67% 70% 69% (98) (108) (206) Female pattern 5% 6% 5% (7) (9) (16) Equality pattern 28% 24% 26% (41) (37) (73) Total 100% 100% 100% (146) (154) (300) 28 As an indicator of equality of sorts, patterns MF and MMFF ("equality pattern") can be read from the table for comparison. Because the male population so vastly outnumbers the female population and therefore many of the jokes are MM, some may argue that Table 5 is not a true reflection of the situation. However, even if jokes where only both sexes appear are looked at (patterns MMF, MMMF, vs. MFF, MFF, MFFF), the hypothesis remains affirmed (7 to 3 in 1936 and 16 to 3 in 1972). Hypothesis 10 was tested by coding the occupations of characters with a two digit code--the "Socioeconomic Status Score for Categories of Occupation Component" as used in the 1960 Census (Miller, 1970: 179). The two digit code was punched on cards along with other data. The scale is termed one of status rankings and probably should be regarded as an ordinal scale. Even if not interval, there is a definite progression from lower to higher status in the ranking categories. For the gross comparisons made here, I have therefore used all jokes including both male and female in a routine that computed the average male occupational score for all the male characters in a particular joke and the average female occupational score for all the females in that same joke. Then a record was kept of the number of jokes with a higher average occupational score for males, the 29 number with a higher average for females, and the number where they were equal. Table 6 gives the results. TABLE 6 NUMBER OF JOKES WITH SUPERIOR OCCUPATIONAL RANK BY SEX 1936 1972 Combined Males superior rank 39% 49% 44% (15) (23) (38) Females superior rank 18% 6% 12% (7) (3) (10) Equal rank 44% 45% 44% (17) (21) (38) Total two sex jokes 101% 100% 100% (39) (47) (86) The table indicates support for the hypothesis (H10). This may be verified by comparing the average male and 9 average female scores for each subsample or whole sample as in Table 7. (These averages were computed by a trans- formation that placed each character's sex and occupational score on a single IBM card. Thus, a card for each char- acter. The routine cross-classified occupation by sex and gave mean scores as output.) Both of these tables suggest that inequality is becoming greater. Data in Table 7 offer further confirmation. For the total saMple, males averaged 66 points and females 48 points. There is a marked difference in male and female scores over time. 30 TABLE 7 AVERAGE OCCUPATIONAL SCORE BY SEX AND TIME PERIOD (EXCLUDING’NOAMENTIONS) Sex Time 1936 1972 Combined Males 56 77 66 Females ‘ 50 45 48 Score difference 6 32 18 In a complete survey such as a Census there are a few who do not report an occupation. The Census Bureau took account of this, apparently, for they gave a score of 33 to those not reporting an occupation. In this analysis, which is not a survey as such, one can expect many instances of no occupation mentioned. Table 9 gives information on the frequency of no occupational mention. This will be commented on shortly. TABLE 8 FREQUENCY OF MENTIONS OF OCCUPATION BY SEX AND TIME PERIOD 1936 1972 Combined (1) Male 125 111 236 (2) Female 13 9 22 Ratio of (1) to (2) 9.6 12.3 10.7 31 Not only are men's occupations superior in rank, but men's occupations are more frequently reported, as can be seen from Table 8. Men are listed with occupations about 11 times as often as women (236/22). It may be argued that this is an invalid comparison since men outnumber women in our sample by 4 to 1. This can be remedied by standardizing the female p0pulation by multiplying by four. Thus, the ratio of 236/88. But this is still a ratio of nearly 3 to l in favor of men. TABLE 9 FRE UENCY OF NO MENTIONS OF OCCUPATIONS EX ND M BR D 1936 1972 Combined Male 92 142 234 Female 37 57 94 Total 129 199 328 In light of the above, it is perhaps not surprising that while men have about as many non-mentions of occupa- tion as mentions, women have more than four times as many non-mentions as mentions (compare Tables 8 and 9). This is, perhaps, more impressive given the fact that women were arbitrarily given the score of 25 when they were clearly identified as housewives. This was based on the Census category of Housekeeper, private household, living in. 32 Table 8 also suggests that inequality is becoming greater over time. TABLE 10 OCCUPATIONAL SCORE AVERAGES FOR MALES AND FEMALES IN JOKE A L D N TE A M 1936 1972 Combined Males 49 88 73 (N=8) (N=l3) (N=21) Females 55 64 59 (N=S) (N=4) (N=9) Table 10 tests H11 for occupation. It indicates mixed results, in that the 1972 jokes would support the hypothesis, while the 1936 jokes as well as the figures for the overall sample are a basis for rejecting it. In any case, like much of the other data in the study, this finding should be regarded as tentative for two reasons. First, the selection for female-dominating- male jokes reduces the numbers such that cell frequencies are quite small. Secondly, the standard deviation for these occupational means are in the vicinity of 32 points. Such wide variation makes it difficult to attri- bute significance, in any sense of the word, to the results. 33 As a point of comparison, a similar table was examined for M/F jokes: it showed males with consistently higher scores than females. (This table does not appear.) TABLE 11 OVERALL STATUS OF MALES AND FEMALES COMPARED FOR FEMALE-DOMINATING-MALE JOKES Sex 1936 1972 Combined Status Low High Low High Low High Males 1 2 0 6 l 8 Females 3 1 7 l 10 2 N = 11 10 Total = 21 Parts B and C of H11 were not tested directly. Original recording of the data made this cumbersome. Low cell frequencies were expected, also, making a collapsed table attractive. Therefore, an overall status judgment was used. Dichotomous in conception, it con- tains merely Low and High categories. The resulting table in effect tests both B and C. Table 11 shows that this table too has low cell frequencies. The frequencies are clear enough so the table is not percentaged. The table indicates plainly for these few jokes that males have predominantly high overall status while females have low overall status. 34 This is true for the subsamples as well as the whole sample. On overall status the hypothesis is supported. SUMMARY OF RESULTS The following is a relisting of the hypotheses which will make reading Table 12, a summary of the results, easier: HlA: The relative frequency of jokes with men dominant over women is greater than the relative fre- quency of jokes with women dominant over men. HlB: The percentage difference between jokes with men dominant over women and jokes with women dominant over men is smaller for 1972 than 1936. H2: The relative frequency of jokes with women dominant over men has increased over time. H3: The relative frequency of jokes with men dominant over women has decreased over time. H4: The relative frequency of men dominant over men has increased over time. H5: The relative frequency of women dominant over women has decreased over time. 35 36 H6: Women will be dominant over males as often as over females. H Males will be dominant over females as often 7: as over males. H8: Men appear more often in jokes than do women. H9: There will be more male characters in a joke than female characters. H10: In each joke in which men's and women's occupations can be identified, men's occupations will be superior in rank to women's occupations. H When women are dominant over men, the men 11‘ will have superior status attributes such as: A. occupation B. majority-minority or minority rankings C. other status attributes 37 TABLE 12 HYPOTHESES BY TEST RESULTS Hypothesis Support or Rejection H1A Support-marginal H1B Support-weak H2 Support-weak H3 Support-moderate H4 Reject H5 Support H6 Reject H7 Reject H8 Support-strong H9 Support-strong H10 Support-strong H11 (A) Support (B)(C) Support-weak INTREPRETATION AND DISCUSSION The Personal Mode The weak support of Hypothesis H1A is in part explained by the change over time. That is, H1A is true for 1936 but false for 1972. Table 1 indicated an increase in F/M jokes and decrease in M/F jokes. This was expected and was the reason for formulating H2 and H3. However, it was expected that any change would be towards equality and not past it. Why this change? Perhaps authors are more conscious about women due to the Women's Movement, or perhaps there is simply less prejudice. Unfortunately, in the present study there are no data that can tell us the cause. It would be nice to think this result is due to a change in the attitudes of the members of society and we are about to have an era of sexual equality. The data on the institutional mode make one doubt that this is the cause. Rather, it would seem that there is response to criti- cism from the Women's Movement and there are some changes, but not to sexual equality. There is a clue to the nature of this change in that the M/F pattern has dropped considerably while the F/M pattern has changed little. This might be a response to 38 39 criticism rather than a significant change in the posi- tive image of women. One implication would thus be that males are acting so as to avoid punishment rather than to reassess their view of women. Further support for this line of thinking comes from the rejection of H4 and support of HS‘ That is, males are much less dominant over time in both the M/F and M/M categories, and women are only slightly more dominant over men and considerably less (down 11%) dominant over the females. If the new image were one of seeing women equal to men, one would expect sizable increases in female dominance--both of men and of women--but this is lacking. The personal mode data also indicate that dominance is much more a same-sex tool rather than a cross-sex device. Table 3 demonstrates this for women and the data for H7 reveal a much larger difference (40%) in the same-sex versus cross-sex dominance behavior of males. The data also suggest that there is not a zero sum game going on between men and women-—men's losses did not appreciably become women's gains. That is, the total amount of dominance does not appear constant. The Institutional Mode In examining the inStitutional mode results, it can be seen that a pattern almost opposite to that of the personal mode is found. Here males have been gaining. 40 There are more jokes where males have superior occupa- tional status. The inequality is greater in 1972 than in 1936. The average occupational score is much higher for males with the difference from females being rather considerable in 1972. Why the smaller difference in 1936? There were many Hobo jokes in this collection; Hobos received lower (0) occupational scores than housewives with whom they were usually interacting. It is not known how much this effect influenced scores-~but it certainly must have been a factor. Mention of Hobos and bums was rare in the Asimov collection. The point to be realized is that historical events influence social indicators. The results for occupation are fairly consistent. Not only are men's occupations superior in rank, but they are more often reported. The occupational superiority is evident even in jokes that in the personal mode have females dominant over males. However, in the 1936 sub- sample this was not the case. No other explanation besides the Hobo cases comes to mind. The major thrust of the results on the institutional mode is clear--women are more subordinate in occupational and other status rankings in 1972 than in 1936, and in 1936 they were already subordinate to men. The decrease in occupational mentions and the lower occupational score in the two time periods is consistent with data of actual events; for example, there has been 41 a decline of women faculty members at universities between 1930 and 1970. The Carnegie Commission (1973: 191) states that women faculty have substantially inferior remuneration compared to men. Other Data The much higher frequency of appearance (see Tables 4 and 5) of males shows that for jokes humor is a man's world. Certainly, many jokes use the masculine pronouns (he, his, etc.) as a convention. But doesn't such lopsided exposure further sexism? This is not the place to argue the causes of sexism. Rather, here is some evidence of uneven exposure for those who would join such arguments. Those who do enter this contro- versy should also be aware that the ratio of males to females in our sample is declining. Could this be because of a heightened awareness of women and their vocalness about such occurrences? ***** Should the superiority of women in the personal mode and of men in the institutional mode be viewed as merely two different types of dominance? Or should one mode be acknowledged as superior? Remember, males enjoyed superior rankings in both modes in 1936; and in 1972, when women were dominant in the personal mode, men retained higher occupational status. The answer to our two questions is not clear. But since institutional 42 characteristics are longer lasting, one may wish to characterize the entire situation as benefiting males. CONCLUSION The major thrust of the personal mode findings is that in this mode we have crossed the point of equality and females now dominate men more often than the reverse. This has been mostly due to a decrease in male dominance. Dominance overall has decreased showing it is not a constant. Also, one should remember that personal domi- nance is used more frequently as a same-sex device than a cross-sex device. The institutional mode findings are that males have higher occupational rankings overall, higher occupational and higher status rankings even when women are dominant in the personal mode, and greater frequency of occupa- tional mention. The inequality is larger for 1972 than 1936. Since women are increasingly dominant in the personal mode, and men increasingly dominant in the occupational sphere, contrasts should become more apparent in the years to come unless some change exercises counterinfluences. Continued disparity should make an interesting socio- logical case in the study of dominant-subordinate relations. 43 APPENDIX APPENDIX Other hypotheses were formulated but could not be tested for various reasons. Jokes were too brief to have major roles for characters in any but the longest jokes, which were too few to consider. Themes were found to be too disparate and non-recurring to be able to rank them. Associations between domination pattern and domina- tion method, domination method and sex distribution, types of joke as categorized by the editors in placing them in certain chapters with domination pattern and domination method were all tested but produced tables with too many empty cells. However, there are some observations that can be made from the data collected and analysis performed with regard to these unreported hypotheses. The methods of domination are many and diverse. Some 30 different methods were encountered. Table 14 gives the methods with the top seven ranks in each subsample. In the entire sample there were five jokes classi- fied as using the method of applying physical force. All of these five were male over male jokes. Half of the female—over-female jokes use the method of having 44 4S emH wooa qwuoo>cfi c we go mafiaamo mama m o>wuoo>ew A »m 90 mcflaamo can: u m we eofiumuomooo-:o: 4 OH *5 managesma oeo m ma wm ooemuoewm m.uo:po m HH we Hanan“ N ea woa gamma“ N mN smfl egos pmafi H ON sea who: umafl H .coum unmouom vogue: xzem .coum “awoken vogue: xemm whoa omma Qonmm mzHH >m ZOHH