WSI‘ANEIER REACTIG-Nfi T0 DISPLAYS OF AGGREng-QN: A MUDY’ OF SQCIALLY R‘EWONS I 8 LE EEI-IAVEOR Thesis In? {‘50 Beam 6? M: A; MICHIGAN STATE UNIVERSITY Gerald L‘ Borofsky 3%? THESIS IIIIIIIHIIINWIIIIWI I 293 011001421 ’ A LIJRARY I “:1 Chip?) 5ft. 1 I? [I‘ t J m é» * 1a raw-aim [1mm it} I; 3-W- «un- rréfi‘w 8.; ABSTRACT BYSTANDER REACTIONS T0 DISPLAYS OF AGGRESSION: A STUDY OF SOCIALLY RESPONSIBLE BEHAVIOR by Gerald L. Borofsky The purpose of this study was to observe the reactions of college men and women, when confronted with a situation in which one person was assaulting another. In a "psychodrama" situation, a "spontaneous" fight emerged between two apparent §§, 0f the six participants, only one was actually an §, The two fighters were actors, and the three additional group members were confederates. Data was used from thirty-five of forty—two §§J who were randomly assigned to one of four fight conditions: 1) male-male; 2) female-female; 3) male-female (male injuring female); and 4) female-male (female injuring male). The verbal and non-verbal behaviors of each §_were rated M“ H.“— by trained coders through one-way observation mirrors. The major findings were as follows: a) Combining all conditions, female §§_showed significantly fewer positive behaviors (attempts to interfere with the fight) than did male §g_(X2 - 4.11; df . 1; p <.OS). b) 1002 of male §§_interfered in the male-male condition, while 02 of male §§_interfered in the male-female condition (Fisher's Exact Test, p <.025). * c) Significantly more non-interferers laughed or smiled throughout the fight, than did interferers (Fisher's Gerald L. Borofsky Exact Test, p‘<.025). Conversely, interferers changed facial expressions from an initially nervous smile or laugh, to one of worry and concern, as the argument became a fight, significantly more frequently than did non—interferers (Fisher's Exact Test, p< .025). The above results suggest that college women are unlikely to directly interfere in a situation where one person is injuring another--regardless of the sex of the participants. This was attributed to the influence of cultural expectations for sex role behavior. The results also suggest that while college men are likely to directly interfere when two men are fighting, they are unlikely to interfere if a college man is beating up a college woman. It was speculated that college men failed to interfere in the latter case, because they were obtaining some kind of vicarious sexual or hostile gratification. This leads to the suggestion that men in general will not interfere when a women is being injured by another man, possibly because they are obtaining some kind of gratification which conflicts with socially responsible thoughts, feelings, and actions. If this is the case, and the presents results tend to suggest such a disturbing conclusion, then situations similar to the Kitty Genovese incident will probably continue to occur. A.woman being assaulted in the city streets may be functionally alone, even (or especially) if she is surrounded by many byst Approved ’IChairman, Thesis Committee Date :5: (5‘: Q77 BYSTANDER REACTIONS TO DISPLAYS OF AGGRESSION: A STUDY OF SOCIALLY RESPONSIBLE BEHAVIOR BY Gerald LEJBorofsky A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology 1969 @5835?“ I/cy/QBZ/Qb? To My Wife Jeanne "...because we are the gardeners of each other." ----Vincent Ferrini ii "I'll get by with a little help from my friends." ----John Lennon & Paul McCartney iii ACKNOWLEDGEMENT S I would like to express my most heart-felt apprecia- tion to Dr. Gary E. Stollak, an educator in the literal sense of the Latin verb educare: an elicitor of potential, and a humorous evocator of personal responsibility. I would also like to acknowledge the friendly support of his wife, Mary Alice, who despite being only six years old, provided a friendly refuge for two nomadic New Englanders. To both of them I am deeply appreciative. I would like to thank Dr. Lawrence Messé' for his aid in making statistical operations something less than a source of continued aggravation, and Dr. Andrew Barclay whose friendly encouragement was sustaining. Acknowledgement is due to the dramatis personae for their talent and enthusiasm. My thanks to Thomas Clark, Karen Grossman, Michael Oberfield, Vicki Sanchez, Bernie Tato, and Debbie Tomlinson. And finally I would like to express my gratitude to the "confederates" and coders, for their patience and enthusiasm: Frank Benison and his wife Marla, Sandy Bierowicz, Leon Brenner, Janet Keinath, Marlene Martin, Pat Southwell, Dennis Staulauskas, and Frank Winn. iv TABLE OF CONTENTS Chapter INTRODUCTION....... ........ . ................... General Overview... ................. . ..... General Hypotheses.. ................ ...... METHOD Design of the Experiment................... Subjects................................... Group Composition for Each Experimental Trial................................ Treatment of Subjects...................... Behavior of Confederates During Each Experimental Trial.............. Criterion for Plausibility of Experimental Manipulations........... Debriefing Procedure....................... Measures of Subject's Behavior............. Reliability of the Ratings................. HypotheseSOOOOOOOOOOOOOOOOOOOO0.. 000000000 0 RESULTS General Findings............. ....... ....... Comparison of Interferers And Non-Interferers...................... Comparisons Among Non-Interferers.......... Summary of Results......................... DISCUSSIONOOO0.0...OIOOOOOOOOOOOOOOOOOOO0...... Directions For Future Research............. Ethical Considerations................ ..... SUMMARYAND CONCLUSIONSOOOOOOOOOOO0.0.0.000...O REFERENCESOOOOOOOOOOOOIOOOOOOOIOOOOOOOOOOOOOOOO APPENDICES...0.0.00.0...OOOOOOOOOOOOOOOOOOOOOOO Page 1 1 5 ll 12 13 14 l7 18 20 37 38 41 45 58 59 61 67 70 Table l. 2. 10. LIST OF TABLES Page Distribution of SS_in each condition, by sex...OOOOOOOOOOOOOOOOOOOO ........ .0. 20 Distribution of Positive and Negative behaviors, by sex of S (all conditions COIr'bj-ned)... ..... OOOOOOOOOOOOOOCOOOOOOOO 21 Distribution of Intrapersonal, Interpersonal (peer), and Interpersonal (actor) behaviors, by sex of S (all conditions combined)............TL................. 22 Distribution of Verbal and Non-Verbal be- haviors, by sex of S (all conditions combined)..... ..... ................ ..... 23 Level of arousal by sex of S (all conditions combined).................... 24 Analysis of variance table, with sex of "victim", sex of "aggressor", and sex of S as main effects upon level of arousal................... 27 Mean arousal scores and standard deviations for each condition, by sex Of §OOOOOOOOOOOOOOOOOOOO ......... 28 Distribution of Positive and Negative behaviors, for male SE, by conditionOOOOOOO0.0.0.0.... 0000000000 29 Distribution of Positive and Negative behaviors, for male SS, in the male-male and male-female conditions........................ ..... 30 Distribution of Positive and Negative behaviors for male SS, in the female-female and male-female conditions................... ........ .. 31 vi Table Page 11. Distribution of Positive and Negative behaviors, for female SE, by condition............... ............ 32 12. Distribution of Intrapersonal, Inter- personal (peer), and Interpersonal (actor) behaviors, for male SS, by condition........................... 33 13. Distribution of Intrapersonal, Inter- personal (peer), and Interpersonal (actor) behaviors, for female SS, by condition........................... 34 14. Distribution of Verbal and Non-Verbal behaviors, for male 83, by condition................. ........ .. 35 15. Distribution of Verbal and Non-Verbal behaviors, for female SS, by condition.......... ......... .... 36 16. Characteristic facial expressions of interferers and non-interferers........ 37 1?. Characteristic facial expressions among male and female non-interferers........ 38 18. Most frequently used rationalizations for non-interference, among male and female non-interferers........... ...... 39 vii LIST OF APPENDICES Appendix Page A. Instructions to Subjects...................... 70 B. Basic Categories for Rating Subject's Behavior........................ 71 C. Score Sheet For Raters.............. .......... 73 D. Self Report Questionnaire Completed by -S—After DebriefingOCOOCO0......000O0...... 76 viii INTRODUCTION GENERAL OVERVIEW In recent years, increasing public attention has focused upon the issue of socially responsible behavior. Newspapers have reported numerous incidents in which by- standers have failed to come to the aid of a person who was in dire need of help. Kitty Genovese was attacked in front of her home in 1964. Amidst screams and desperate pleas for help, thirty-eight of her neighbors witnessed her murder with- out making any attempt to interfere. Not one of her neighbors even called the police, although the attack lasted for over half an hour. Darley and Latane' (1968b) have chronicled a number of similar episodes, and indicate in a personal communica- tion that "dozens of such incidents" have been brought to their attention in the recent past. In all of these inci— dents, individuals who were in desperate need of bystander assistance either died or suffered extreme pain because such assistance was not forthcoming. "Theories" to explain these disturbing events are as numerous as the "armchair psychologists" who incline toward such speculations. Explanations ranging from Apathy to Zeitgeist have been invoked, but until recently very little researCh had been undertaken on this pressing issue. Why do bystanders fail to act when another human being is in distress? Such a question has motivated an increasing flurry of research activity. The findings have been as dis- turbing as the actual incidents themselves. Research on the question of socially responsible behavior has followed two different, but related approaches. One approach has been the study of destructive obed- ience and conformity. The work of Milgram (1963; 1964; 1965a; 1965b) is an example of this type of approach to the study of socially responsible behavior. In part motivated by the behavior of the German people during World War II, this approach used the following paradigm: If a person is told by an authority, or representa- tive of authority, that he is to hurt a third person, when will he follow such instructions and when will he refuse to obey? The laboratory procedure in such experiments called for ordering a nai S_to hurt a "victim" by administering electric shocks with increasing levels of intensity. Mil- gram (1963) found that twenty-six out of forty §§ followed the experimenter's directions and administered levels of electric shock to the "victim", which would have been injur- ious and painful, had there actually been current flowing through the wires. Subsequent research has manipulated a number of situational variables in order to locate the criti- cal determinants of socially responsible behavior. Milgram (1965b) studied the influence of two variables- immediacy of the victim, and closeness of authority. He found that: "...something akin to fields of force, dimin- ishing in effectiveness with increasing psych- ological distance from their source, have a controlling effect on the subject's performance. As the victim is brought closer, the subject finds it harder to administer shocks to him. When the victim's position is held constant relative to the subject, and authority is made more remote, the subject finds it easier to break off the experiment. The effect is substantial in both cases, but manipulation of the experimenter's position yielded the more powerful results." In two other experiments, Milgram (1964; 1965a) studied the effects of group pressure upon S's behavior. When two confederates urged S to administer increasingly severe shocks to the "victim", S responded to this pressure from the confederates, and the mean shock level increased. When §§ first observed two confederates defy E's orders to increase the shock level, S§_themselves were more likely to defy E's orders to administer increasingly severe shocks. On the other hand, first observing two SS comply with E's orders, did not raise the mean shock level administered by S, over that administered in the baseline condition where no confederates were present. Using a similar method, Tilker (1967) studied the influence of two other variables upon bystander intervention. The degree of bystander responsibility for the "victim", and the nature of the feedback received from the "victim" were the independent variables. Tilker found that bystanders intervened most rapidly when given total responsibility for the"victim", and when supplied with both auditory and visual feedback from the "victim". Kaufman(l968) studied the influence of four variables (Status of S in the experiment; expectation of future partic- ipation as the person administering the shocks, cul- pability of the "victim"; and legitimacy of the authority for shocking the "victim") upon the tendency for S, the bystander, to interfere when a "victim" is being shocked. Kaufman found no significant differences between the condi- tions tested. In fact 176 out of 186 §§ failed to inter- fere, thus allowing the "victim" to be shocked a total of thirteen times, at increasing levels of intensity. Another line of attack has been utilized in the in- genious work of Darley and Latane' (1968a: 1968b; 1969). These experiments have sought the determinants of socially responsible behavior, but have utilized a design which is more representative of the real-life settings in which such emergencies have actually occurred. Utilizing a situation in which §§ overheard an epileptic seizure, Darley and Latane' (1968a) found that "the presence of other bystanders reduced the individual's (§L§) feelings of personal responsibility", and lowered the speed with which he reported the "emergency" to S, Popular folklore would lead one to speculateIthat the probability of bystander interVention would increase with the number of bystanders. The above findings directly contradict such a notion. It appears instead, that the probability of any one person in- tervening, decreases as the size of the bystander group in- creases. The authors have replicated this finding in a number of equally ingenious experiments of similar design. S§_who found themselves in a room which was filling up with smoke reported the "emergency" less quickly when in the presence of other bystanders, than when alone (Darley and Latane', 1968b). Using a situation in which S§_heard a woman fall from a chair and call out in pain, Latane and Rodin (1969) found that in groups of two (a naive S and a confederate), §§ offerred to help the woman less frequently than did §§ who were alone. This study also found that pairs of friends were more likely to interfere than were strangers. The present study was conceived, in order to study the influence of yet another set of variables, namely the sex of the people involved in the attack, and the sex of the by- stander. This study follows a design which is kindred in spirit similar to those utilized by Darley and Latane' in their research. GENERAL HYPOTHESES It is hypothesized that in a situation where one person is injuring another, bystanders will behave differentially, depending upon their sex, the sex of the attacker, and the sex of the victim. The specific techniques used to study this question are described in the Method section. Following a discussion of experimental procedures, a number of specific hypotheses are put forth. These are listed at the end of the method section. METHOD DESIGN OF THE EXPERIMENT A 2 x 4 design was used with sex of S and experimental conditions as the independent variables. The four experimental conditions were as follows: 1) Male-Male: One man beating up another man. 2) Female-Female: One woman beating up another woman. 3) Male-Female: A man beating up a woman. 4) Female-Male: A woman beating up a man. SUBJECTS Both male and female S§_were students in undergraduate sections of an Introductory Psychology class. Of the forty §§ originally obtained for the experiment, five male and five female S§_were randomly assigned to each of the four experi- mental conditions. Failure of §§ to appear for the experiment, plus the failure of seven §§ to satisfactorily meet the cri- terion for plausibility of the experimental manipulations (see discussion of this in a subsequent part of this section), necessitated obtaining additional §§' Difficulty in obtaining additional S§_limitéd the total number of §§ used in the data analysis to SS (17 male and 18 female). The distribution of these 35 SS_over the four experimental conditions is shown in Table l of the Results section. GROUP COMPOSITION FOR EACH EXPERIMENTAL TRIAL In each experimental trial, there were S_people in the room, in addition to S, The composition of this group, all of whom were ostensibly 85, was as follows: 1) One ACTUAL S, obtained in the above-described manner. 2) Three undergraduate assistants, who as far as S knew, were also §§ in the experiment, but who actually served as E's Confederates. The confederate group had a fixed comp- osition (f two males and one female. 3) Two Actors who were drama students. The actors engaged in what appeared to be a spontaneous fight sequence, arising out of the psychodrama exercise. The actors indicated that they were unacquainted with each other, before beginning the psychodrama. TREATMENT OF SUBJECTS Upon arrival, all participants (S, 3 confederates, and 2 actors) were seated (Figure 1). The participants were then asked to read the Instruction Sheet (Appendix A). O O O—Raters (3“ L~ one-way glass 0 {OT O}\ C) C) ~\\participants in C) the psychodrama ()1 exercise -p3rticipants Fig. 1. LOCATION OF PARTICIPANTS AND RATERS S then summarized the main points and answered any questions. The confederates asked a number of procedural questions, in order to encourage the actual S to ask any questions he might have. In all four conditions, S selected two confederates to act out the first psychodrama exercise. This was a male-female pair with the confederates cast in husband and wife roles, respectively. S then gave a brief outline of the situation which the two confederates were to act out. The situations were varied among four variations, in order to maintain some freshness and enthusiasm among the confederates. Each situation portrayed a strong disagreement between the husband and wife, a disagreement of sufficient intensity as to lead to an argument. The two confederates were asked to act out the assigned situation in the manner detailed by the instructions. S stressed the importance of being spontaneous and free in ex- pressing emotion. These preliminary exercises lasted approx- imately four to five minutes. They were stopped at the point where one or both of the confederates appeared to be getting angry. S then made some positive comments regarding the quality of their pefformance, and then suggested constructive improve- ments to be used by future pairs. Spontaneity and freedom of emotional expression were again stressed. S then asked for comments from the other participants, attempting to involve S more fully in the experiment. The next pair selected by S was the two actors. The sex of the actors was varied, according to the experimental condi- tion. The situation to be acted out was explained. Next S began to shuffle through his papers. He looked puzzled, and then explained that he had left some evaluation sheets in his office, and that he must leave the room for a short time in order to get these forms. S remarked that the first pair had done a fine job and that the present pair should try to act in similar fashion. S emphasized that all the other participants should pay close attention to what transpired, because S was interested in their evaluations of how the psychodrama was working out. As he departed, S reassured that he would return as soon as he got the forms from his office. At this point S went into the observation room. The actors then engaged in their pre-arranged sequence of a slowly developing argument which led, in turn, to a fight. The total sequence, from start to finish, lasted approximately five minutes. That is, in each case a seemingly spontaneous fight _ emerged out of the situation being acted out. The fight con- tinued until it was stopped by S or until a period of 45 seconds had elapsed. The transitional sequence from the beginning of the psychodrama to the start of the fight was as follows: The argument developed slowly from the initial disagreement inherent in the assigned roles and situation. The intensity of the argument gradually increased. Then, one of the actors, seem- ingly from the intensity of the psychodrama, appeared to lose his temper, and he slapped the other actor on the face. The actor being slapped responded as follows: "Look this is only acting. This is just supposed to be an experiment....(5 sec. -10- pause) You can't get away with hitting me like that." That is, there was a clean break from the psychodrama setting into the situation of a "real" fight. At this point the actors began fighting in a violent and extremely convincing manner. The trained coders rated and scored S;S_verbal and nonverbal be- haviors through the one-way observation mirrors (See section on measures of S behavior). The four experimental conditions, exactly as they were read aloud to the actors, along with the situations to be acted out, are listed below:** 1) Male—Male- "You (pointing to the actor previously agreed upon) will play the father. You (pointing to the other actor) will be his son. The situation is that the son has let his hair grow long, has a beard, and has taken to wearing 'hippy' clothes. The son is not at all concerned with dressing in a fashion which the father considers 'appropriate for a young man.‘ The father voices his displeasure. The son, however, is little impressed by the father's statements. The father does not want the son to 'become a hippy.'" 2) Female-Female- "You (pointing to the actress previously agreed upon) will play the mother. You (pointing to the other actress) will be her daughter. The situation is that the daughter has taken to wearing 'hippy' clothes, does not appear to be concerned with 'becoming an educated woman', nor does she care about acting in a manner which the mother considers 'appropriate for a young ** The actor playing the "parent" in.each condition is the one who slapped the other actor. -11- woman'. The mother voices her displeasure. The daughter, however, is little impressed by the mother's statements. The mother does not want the daughter to 'become a hippy'." 3) Male-Female (male injuring female)- "You (pointing to the actor) will play the father. You (pointing to the actress) will play the daughter. The situation is that the daughter has taken to wearing 'hippy' clothes, does not appear to be concerned with 'becoming an educated woman', nor does she care about acting in a manner which the father considers 'appropriate for a young woman'. The father voices his displeasure. The daughter, however, is little impressed by the father's statements. The father does not want the daughter to 'become a hippy'." 4) Female-Male (female injuring male)- "You (pointing to the actress) will play the mother. You (pointing to the actor) will play the son. The situation is that the son has let his hair grow long, has a beard, and has taken to wearing 'hippy' clothes. The son is not at all concerned with dressing in a fashion which the mother considers 'appropriate for a young man'. The mother voices her displeasure. The son, however, is little impressed by the mother's statements. The mother does not want the son to 'become a hippy'." BEHAVIOR OF CONFEDERATES DURING EACH EXPERIMENTAL TRIAL The three confederates (undergraduate assistants) were used So as to confront the actual S with a group situation--that is, one which more realistically approximates the phenomena in -12- such "bystander" situations. The use of the confederates represents an attempt to use a design representative of actual conditions. The three confederates performed a stan- dardized sequence of verbal and nonverbal behaviors. These behaviors called attention to what was going on (provided S with more information than merely his own perceptions of the fight), but in no way implied that S should take any type of action. For example, the confederates would ask, with puzzled and worried expressions, "What is going on?", another would ask, "Are they really fighting?" Another, with a worried ex- pression would ask, "What happened?" Similarly, the confed- erates would rub their hands as if anxious, shift their feet, and in general appear nervous. As soon as the fight was stOpped by S, or when the 45 seconds of fighting had elapsed, and the actors had stopped fighting, while still maintaining the atmosphere of anger at each other, S returned to the room. CRITERION FOR PLAUSIBILITY OF EXPERIMENTAL MANIPULATIONS Upon returning to the room, S appeared surprised that the actors were already finished (except in those cases where S had already acted to stOp the fight). He asked what had happened. Neither the actors nor the confederates responded to this query. After a brief silence, S indicated that he would like someone to explain what had happened, as people were obviously upset about something. S, in a tone of irritation -13- then stated that since nobody would respond, he wanted to speak to each person individually in the hall outside the room. If S then began to relate what had happened, he was asked to come out in the hall and explain further. Other- wise S's name was read from the "list of S's" and upon acknowledging his name, he was asked to come out to the hall in order to explain what had happened during st absence. This rather "baroque" procedure was used in order to spare S from avoidable embarrassment, upon learning the true nature of the experiment. Once in the hall, S was asked what had happened during E's absence. S' 5 response was the criterion for including or excluding S from the da ta analysis. In all ___.__.-_, . a... .— ,_.—-—‘...‘V‘ -- __,.. m - ‘h—a—r’ but seven cases, S began to relate how the actors had gotten "carried away" and had actually started fighting. When S_ was satisfied that S had believed the fight to be authentic, the true nature of the experiment was explained, and debrief- ing followed. If during his explanation S smiled, laughed, or suggested that something "funny" seemed to be going on, then that S was dropped from the data analysis. That is, if there was any indication that S doubted the authencity of the fight before S explained the tnue nature of the experiment, then he was not included in the data analysis. The seven cases where this occurred, were distributed as follows over the four conditions: male-male: 2; female-female: 2; male-female: 2; female—male: l. DEBRIEFING PROCEDURE All SS, upon completion of the trial and the test for -14— plausibility, were debriefed. The debriefing procedure con- sisted of three parts. 1) The true nature of the experiment was explained. The deception involved was elucidated, along with the reasons for having deceived S, It was explained that the research dealt with a social issue of considerable relevance and that was the reason for having resorted to deception. 2) S was encouraged to discuss his feelings with S, Invariably there was a sense of confusion being experienced by S. Sufficient time was spent on this phase, so as to actually encourage S, in great detail, to express the resentment and anger at having been deceived. 3) S was then invited to come behind the one-way ob- servation mirrors and to observe a number of subsequent trials. S was told that he might remain as long as he desired. Not only was S able to rexperience what had just happened to him, but he was able to further discuss his feelings with the trained coders who sat with him. During this phase of the debriefing, S, was asked to complete a short self-report questionnaire (appendix D). Finally, S was told that the nature of the experiment depended upon secrecy, and upon future Ss not knowing what was involved. S was asked not to discuss the details of the experiment, even with friends, for a period of one week. MEASURES OF SUBJECT'S BEHAVIOR A system of behavior (verbal and nonverbal) categories was specially developed for the present study (Appendix C). -15- There are three main measures of behavior: 1) a measure of socially-responsible behavior; 2) a measure of interaction with others in the room; and 3) whether the behavior was ver- bal or nonberbal. In addition, S;§_visible level of arousal was measured on a five-point scale. Behavior Categories: 1) Measure of social responsiblity- a) Positive Behaviors-— those designed, in some manner, to take positive action to stop the fight, on some level of effectiveness. b) Negative Behaviors- those designed, in some manner, to actively move away from or to avoid interfering with the fight. This category differs from the next one in that Negative behaviors mean that S showed some change in behavior in reaction to the fight, but that this behavior was such as to not involve S in stopping the fight. c) Neutral Behaviors- either no action at all, or bizarre behaviors which are unrelated to the fight situation with which S is being confronted. 2) Measure of interaction with others- a) Intrapersonal Behaviors- those which in— volve only S himself, and which are apparently not directed toward other persons in the room. b) Interpersonal (peer) Behaviors- those which are directed toward S's "peers" - the confederates. That is, behaviors directed toward those in the room, other than the actors. -l6- c) Interpersonal (actor) Behaviors- those which are directed toward the actors- the fighters. 3) Measure of verbal vs. nonverbal behavior- a) Verbal Behaviors- S uses words or voice. b) Non-Verbal Behaviors- all other behaviors other than verbal. The organization of categories is such that there are 18 theoretically possible ratings. In practice not all were used, since no Ss showed Neutral behaviors. The complete range of possible scorings, along with the symbols used in scoring are found in Appendix B. An example of the scoring procedure is provided, along with some scoring procedures followed in this study. Level of Arousal Measure: V/// The term "level of arousal", as used in this study, represents an attempt to codify a series of observable be- haviors along the continuum from "negligible arousal" or relaxed and calm, through "extreme arousal" or being dis- traught and on the verge of being uncontrollable. As can be quickly seen, this measure includes the many diverse behaviors typically subsumed under the rubic, "indicators of anxiety". The intent, in using the term arousal, is to avoid the pit- falls and personalized meaning which inhere to the term anxiety. In the present study, a five—point scale was used, with the above-described extremes as the anchoring points. Negligible arousal was coded as "1", while extreme arousal -17- was coded as "5" (Appendix C). Each coder rated S's level of arousal during the fight, as part of the overall ratings, which were filled out once the trial was over. Score Sheet for Coders: A three-page score sheet was completed by each coder, for each S, The coder provided both a descriptive and cod- ified portrait of S's behavior during the fight. The score sheet was so designed, that S was able to code elements of the descriptive portrait, without knowing the sex of SIor the experimental condition. The coder obtained the following information for each S, 1) Body: posture, movements, gestures. 2) Verbalizations: tone of voice, and exact words 3) Facial Expressions: 4) Other Relevant Information: A facsimile of the score sheet is found in Appendix C. RELIABILITY‘OF THE RATINGS A group of five raters were trained in the use of the behavioral category system, and the score sheet. A minimum of two raters were present at each trial, and in 80% of the trials, there were three trained raters present. Interjudge reliabilities were as follows: 1) Positive-Negative Dimension: 97.6% 2) Intrapersonal-Interpersonal (peer)-Interper- sonal (actor) Dimension: 91.5% -18- 3) Verbal—Nonverbal Dimension: 100% The agreement among raters, therefore, was very high for these major categories of behavior. The arousal rating for each S was taken as the mean of the ratings for that trial. HYPOTHESES The following hypotheses were formed: 1) Over all experimental conditions, female SS will show more Negative behaviors (non-involvement), while male S§_will show more Positive behaviors (attempts to stop the fight). 2) Over all experimental conditions, female SS will show more Intrapersonal behaviors (not interacting with other people), while male §§ will show more Interpersonal (actor) behaviors (interacting directly with the actors). Both sexes will use Interpersonal (peer) behaviors (inter- acting with other "peers" in the room) with approximately equal frequency. 3) Over all experimental conditions, female SS will express more Verbal behaviors, while male S§_will show more Non-verbal behaviors. 4) Over all experimental conditions, female SS will show higher levels of arousal than male SS. 5) Within each sex, SS_will react differentially, according to the experimental condition. It is generally -19- predicted that with higher arousal, all SS_will show a tendency to utilize Negative and Intrapersonal behaviors, while with low levels or arousal, SS_will show more of a ten- dency to utilize Positive and Interpersonal (actor) behaviors. 5a) Condition Male-Male (man beating up a man). In this condition, female S§_will show a moderate level of arousal, while male SS_will show the lowest level of arousal for any of the four conditions. Male §§.Will utilize the greatest number of Positive and Interpersonal (actor) behaviors, while female SS_will utilize some inter- mediate number of these behaviors. 5b) Condition Female-Female (woman beat- ing up a woman). In this condition, female S§_will show the lowest level of arousal for any of the four conditions, while male S§_will show a moderate level of arousal. In this condition female SS_will utilize the greatest number of Positive and Interpersonal (actor) behaviors, while male §§ will utilize some intermediate number of these behaviors. 5c) Condition Male-Female (man beating up a woman) and Condition Female—Male (woman beating up a man). In these conditions, both male and female S§_wi11 show the highest levels of arousal, for any of the four conditions. Both Isle and female §§ will utilize the least number of Positive and Interpersonal (actor) behaviors, and will corres- pondingly utilize the maximum number of Negative and Intra- personal behaviors. Male SS, however, will show fewer Nega- tive and Intrapersonal behaviors than will female SS, -20- RESULTS GENERAL FINDINGS Table 1 lists the number of Se in each condition, by sex. This table only lists those Ss who were included in the final data analysis. The data from seven Ss was not in- cluded in the final analysis, because these SS failed to pass the criterion for plausibility of the experimental manipula- tions. Table 1. Distribution of Ss in each condition, by sex (N = 35). Condition Male- Female- Male- Female- Total Male Female Female Male S Male 4 5 5 . 3 ‘ 17 E a; x Female 4 5 3 6 18 Total 8 10 8 9 35 In the remainder of this section, the hypotheses will be restated and the pertinent data presented. Hypothesis 1. Over all experimental conditions, female Ss will show more Negative (non—involvement) behaviors, while male Ss will show more Positive (attempts to stOp fight) behaviors. -21- Table 2. Distribution of Positive and Negative be- haviors, by sex of S (all conditions com- bined) (N = 35). "' Behavior Male Ss Female Ss Total N = 17 N = 18 Positive 10 2 12 Negative 7 16 23 x2 = 4.11* d.f. = 1 *p<: .05 for a non-directional hypothesis As seen from Table 2, IS male Ss showed positive be- haviors, as compared to only S female Ss. Correspondingly, only 1 male Ss showed negative behaviors, as compared to SS female Ss. The hypothesis is supported by the data at a statistically significant level. Hypothesis 2. Over all experimental female Ss will show more Intrapersonal (not other people) behaviors, while male Ss will personal (actor) (interacting directly with haviors. Both sexes will use Interpersonal conditions, interacting with show more Iggggf the actors) be- (peer) (inter- acting with other "peers" in the room) behaviors with approx- imately equal frequency. -22- Table 3. Distribution of Intrapersonal, Interpersonal (peer), and Interpersonal (actor) behaviors, by sex of S (all conditions combined) (N= 35). Behavior Male Ss Female Ss Total N = 17 N = 18 Intrapersonal 3 9 12 Interpersonal (peer) 4 7 11 Interpersonal (actor) 10 2 12 x2: 8.99* d.f. = 2 *p<: .025 for a non directional hypothesis As seen from Table 3, S_male Ss showed Intrapersonal behaviors as compared to S_female Ss. Correspondingly, IS. male Ss showed Interpersonal (actor) behaviors as compared to only S female Ss. This hypothesis is supported by the data at a statistically significant level. Hypothesis 3. Over all experimental conditions, fe- male Ss will show more Verbal behaviors, while male Ss will show more Non-verbal behaviors. -23- Table 4. Distribution of Verbal and Non-Verbal behaviors, by sex of S (all conditions combined) (N = 35). Behavior Male Ss Female Ss A Total N = 17 N = 18 Verbal 6 3 9 Non-verbal ll 15 26 x2 = 0.76 d.f. 1 Not Significant As seen from Table 4, S male Ss expressed Verbal be- haviors as compared to S_female Ss. Male Ss expressed II Non-verbal behaviors as compared to IS_female Ss. This hypothesis is not supported by the data. In fact, the results indicate a statistically non-significant result in the reverse direction. That is, male Ss used more Verbal behaviors. Hyppthesis 4. Over all experimental conditions, fe- male Ss will show higher levels of arousal than male Ss. -24- Table 5. Level of arousal by sex of S (all con- ditions combined) (N = 35). Male 85 Female 85 N = 17 N = 18 _ MEAN 2.89 2.59 S.D. 0.84 0.79 t = 1.06 df = 32 (Winer, 1962, p. 37) Not Significant As seen from Table 5, a t-test between mean male arousal and mean female arousal (all conditions combined) produced a statistically non-significant result. In fact, the mean arousal score for male Ss over all conditions (X'= 2.89, S.D. = 0.84) was higher than that for female Ss (X'= 2.59, S.D. = 0.79). This is the reverse of what had been hypothesized. This is of importance, in that the subsequent hypo- theses were based on the assumption that high arousal would act as an inhibitor of Positive and Interpersonal (actor) behaviors. However, the expected simple linear relationship was not observed. In spite of this, the following hypotheses are listed as originally put forth. -25- Hypothesis 5. Within each sex, SS_will react differentially, according to the experimental condition. It is generally predicted that with higher arousal, all SS_will show a tendency to utilize Negative and Intra- personal behaviors, while with low levels of arousal, SS 'will show more of a tendency to utilize Positive and Inter- personal (actor) behaviors. Hypothesis 5a. Condition Male-Male (man beating up a man). In this condition, female SS_will show a moderate level of arousal, while male §§ will show the lowest level of arousal for any of the four conditions. Male SS_will utilize the greatest number of Positive and Interpersonal (actor) behaviors, while female SS_will utilize some inter- mediate number of these behaviors. "Hypothesis 5b. Condition Female-Female (woman beating up a woman). In this condition female SS_will show the lowest level of arousal for any of the four conditions, while male SS_will show a moderate level of arousal. In this condition female S§_will utilize the greatest number of Positive and Interpersonal (actor) behaviors, while male SS will utilize some intermediate number of these behaviors. 'Hypothesis 5c. Condition Male-Female (man beating up a woman) and Condition Female-Male (woman beating up a man). _26_. In these conditions, both male and female SS_will show the highest levels of arousal, for any of the conditions. Both male and female §§.Wi11 utilize the least number of Positive and Interpersonal Xactor) behaviors, and will correspond- ingly utilize the maximum number of Negative and Intraper- sonal behaviors. Male SS, however, will show fewer Negative and Intrapersonal behaviors than will female SS, Results bearing on the above hypothesis are pre- sented on the following pages. As seen from Table 6, performing a 2 x 2 x 2 analysis of variance upon the data, for level of arousal, revealed several effects which approached statistical significance. Of the main effects, only sex of "victim" approached statis- tical significance (p'<:.lO). Main effects for sex of "aggressor" and sex of S were not statistically significant. The interaction effect for sex of "victim" and sex of "agg- ressor" approached statistical significance (p< .10) . A test for simple effects with male "victims" revealed a statistically significant effect (p‘<:.05). That is, SS were more aroused in the male-male condition than in the male-female condition. A test for simple effects with male "aggressors", also revealed a statistically significant effect (p'<:.025). That is, SS_were more aroused in the male-male condition than in the male-female condition. No other inter- action effects approached a statistically significant level. -27- mmo.uv_d.aa mo. v ma; OH. V as me.em Hmuoe No.0 am eo.oH fineness aHauHs ea.H mo.H H mo.H I smsum ”coHpomHmuaH ae.o mm.o H mm.o m mo xmm a =H0mmmuma<= l mo xmw "coauomnmucH mH.o mo.o H mo.o m we saw was =eHHoH>= m0 xmm “cofluomnwusH moo.o moo.o H moo.o mHmsmmumHmEmm w mamzlmamfiom «eamo.h hm.e H hh.v onEomlmHmz a mHmsumHmz ne.o mm.o H mm.o onEmmlonEmm w onEomlonz «Imm.e am.~ H ma.~ mHmZImHmemm w OHMZIOHMZ *moov mmoN H mmom =E._uu.0._u>: mo xmm w =H0mmonmm<= m0 xom "cofluomumucH am.H SH.H H SH.H m_mo xwm mH.H Hs.o H Hs.o =H0mmmumma= Ho xmm 4mm.m om.m H om.m =sHuoH>= Ho xwm m w: we mm oousom . HmmmH .HmcHsv .Amm i zv Homsoum mo Ho>oH com: muoommm CAME mm ~m MD Now one .:H0mmoummm= mo xmm .=Eflpofl>= mo xom :uH3 .oHnmu mosseum> mo mammascd .m manna -28- Table 7. Mean arousal scores and standard deviations for each condition, by sex of S_(N = 35). Condition Male- Female- Male- Female- Male Female Female Male =8 N=10 N=8 N=9 S B Male x = 3.95 2.75 2.52 2.47 X 0': 0.07 0 80 0.61 0 61 N = 4 5 5 3 O F Female §'= 2.93 2.42 2.33 2.63 O'= 0 79 1.07 0 55 0 47 S N = 4 5 3 6 As seen from Table 7, male arousal is highest in the Male-Male condition and lowest in the Male-Female condition. This is the reverse of what had been hypothesized for male SS, Female arousal is highest in the Male—Male condition, and lowest in the Male-Female Condition. This is the reverse of what had been hypothesized for female SS. Level of arousal appears to have a differential rela- tionship to behavior, depending upon §L§ sex. Over the four experimental conditions, for male SS, mean level of arousal is positively correlated with the percentage of male S§_who showed positive behaviors in that condition. The Spearman Rank Correlation Coefficient (Siegel, 1956, p. 202) is +£199° This is significant at p = .05. On the other hand, for female 85, over the four conditions, mean level of arousal is 3333: - tively correlated with the percentage of female SS, in that -29- condition, who showed positive behavior. The Spearman Rank Correlation Coefficient is -.95. This is not quite statis- tically significant (due to the small size, n=4), but is very close to p = .05. Table 8. Distribution of Positive and Negative be- haviors, for male SS, by condition (N = 17). Condition Male— Female- Male- Female- Total Male Female Female Male =4 N=5 N=5 N=3 B E Positive 4 4 0 2 10 H A V Negative 0 l 5 l 7 I O R Fisher's Exact Test, p<: .05 for a non-directional hypo- thesis. In the above table, the expected frequencies are less than 5. Accordingly, it is necessary to combine categories, so as to increase the expected frequencies of each cell (Siegel, 1956) p. 178). Due to the nature of the hypotheses, the most reasonable procedure is to combine the Male-Male condition with Female-Female, and to combine the Male-Female condition with Female-Male. The data were tested for signifi— cance by means of Fisher's Exact Test (Siegel, 1956, p. 96). As seen from Table 8, male SS_showed S_positive behaviors in the combined Male-Male and Female-Female conditions, -30.. as compared to only S Positive behaviors in the combined Male-Female and Female-Male conditions. The directional hypothesis that male §§.Will show more positive behaviors in the combined conditions Male-Male and Female-Female, than in the combined conditions Male-Female and Female-Male, is supported by the data at a statistically significant level. Table 9. Distribution of Positive and Negative behaviors, for male 85 in the male-male and male-female condIEions (N = 9) Male- Male- Total Condition Male Female =4 N=5 B E Positive 4 0 4 H A V Negative 0 5 5 I O 'R' Fisher's Exact Test, p.<:.025 for a non-directional hypothesis. As is seen in Table 9, male SS_showed 5 positive be- haviors (100%) in the condition male-male, as compared to S positive behaviors (0%) in the condition male-female. The hypothesis that male 83 will show different amounts of posi- tive behaviors in these two conditions is supported by the data at a statistically significant level. -31- Table 10. Distribution of Positive and Negative be- haviors for male SS, in the female-female and male-female conditions (N = 10). Condition Female- Male— Total Female Female N = 5 N = 5 B E Positive 4 0 4 H A v Negative 1 5 6 I O R Fisher's Exact Test, p< .05 for a non-directional hypothesis. As seen in Table 10, male SS showed S_positive behaviors (80%) in the condition female-female, as compared to S positive behaviors (0%) in the condition male-female. The hypothesis that male SS_will show different amounts of positive behaviors in these two conditions is supported by the data at a statistically significant level. -32- Table 11. Distributuion of Positive and Negative behaviors, for female SS_ by condition (N = 18) . Male- Female- Male- Female- Total Condition Male Female Female Male B E Positive 0 l l 0 2 H A V Negative 4 4 2 6 16 I O R Fisher's Exact Test, Not Significant. For the reasons given above, it was necessary to com- bine conditions in the above-described manner. As seen from Table 11, female SS_showed only 92§_positive behavior in the combined conditions Male-Male and Female-Female, and only 222 positive behavior in the combined conditions Male-Female and Female—Male. The hypothesis that female SS_will show more positive behaviors in the combined conditions Male-Male and Female-Female, than in the combined conditions Male-Female and Female-Male, is not supported by the data. -33- Table 12. Distribution of Intrapersonal, Interpersonal (peer), and Interpersonal (actor) behaviors, for male SS, by condition (N = 17). Male- Female- Male- Female— Total Condition Male Female Female Male N-4 N=5 N=5 N=3 B E Intrapersonal 0 0 2 l 3 H A V Interpersonal 0 l 3 0 4 I (peer) O R Interpersonal 4 4 0 2 10 (actor) X2 = 6.07* d.f. = 2 (for combined categories) ..... * p<:.05 for a non-directional hypothesis For the reasons given above, it was necessary to combine categories in the above-described manner. As seen from Table 12, male S§_showed no Intrapersonal behaviors in the combined conditions Male-Male and Female-Female, as compared to showing S Intrapersonal behaviors in the com- bined conditions Male-Female and Female—Male. Correspond- ingly, male SS showed S_Interpersonal (actor) behaviors in the combined conditions Male-Male and Female-Female, as com- pared to only S Interpersonal (actor) behaviors in the com- bined conditions Male-Female and Female-Male. The hypothesis that male S§_will show more Interpersonal (actor) behaviors in the combined conditions Male-Male and Female-Female, than in the combined conditions Male-Female and Female-Male, is -34- supported by the data at a statistically significant level. A chi-square technique reported by Karon (1968, p.106) was used to correct for continuity. Table 13. Distribution of Intrapersonal, Interpersonal (peer), and Interpersonal (actor) behaviors, for female SS, by condition (n = 18). Male- Female- Male- Female- Total Condition Male Female Female Male N=4 N=5 N=3 N=6 B E Intrapersonal 2 0 2 5 9 H A Interpersonal 2 4 0 l 7 V (peer) I O Interpersonal 0 l l 0 2 R (actor) X = 4.32 d.f. - 2 (for combined categories) Not Significant For the reasons given above, it was necessary to com- bine categories in the above-described manner. As seen from Table 13, female SS showed only S Intrapersonal behaviors in the combined conditions Male-Male and Female-Female, while showing Z_Intrapersonal behaviors in the combined conditions Male-Female and Female-Male. On the other hand, female SS used S Interpersonal (peer) behaviors in the combined con- -35- ditions Male-Male and Female-Female, while showing only one Interpersonal (peer) behavior in the combined condi- tions Male-Female and Female-Male. The hypothesis that female SS_will show fewer Intrapersonal behaviors in the combined conditions Male-Male and Female-Female, than in the combined conditions Male-Female and Female—Male, is not supported by the data, at a statistically significant level. Table 14. Distribution of Verbal and Non-Verbal behaviors, for male SS, by condition (N = 17). Male- Female- Male- Female- Total Condition Male Female Female Male =4 N=5 N=5 N=3 B E H Verbal l 3 0 2 6 A V I Non-verbal 3 2 5 1 ll 0 R FiSher's Exact Test, Not Significant. For the reasons given above, it was necessary to com- bine categories in the above-described manner. As seen from Table 14, male SS showed S_Verbal behaviors in the combined conditions Male-Male and Female—Female, as compared to S Verbal behaviors for the combined conditions Male-Female and Female-Male. The non-directional hypothesis that male S§_will show different amounts of verbal behavior, over the four con- -36‘ ditions, is not supported by the data at a statistically significant level. However, in those conditions where a woman was the aggressor (Female-Female and Female-Male), male SS showed S Verbal behaviors as compared to only S_Verbal behavior for those conditions in which a man was the aggressor (Male- Male and Male—Female). This is not statistically significant, by Fisher's Exact Test (p<:.10). Table 15. Distribution of Verbal and Non-verbal behaviors, for female SS, by condition (N = 18). Condition Male- Female- Male- Female- Total Male Female Female Male =4 N=5 N=3 N=6 B E H Verbal 0 l l l 3 A V I Non—verbal 4 4 2 5 15 O R Fisher's Exact Test, Not Significant. For the reasons given above, it was necessary to com- bine categories in the above-described manner. As seen from Table 15, female SS showed only 932 verbal behavior in the combined conditions Male-Male and Female—Female, as compared to S_verbal behaviors in the combined conditions Male-Female and Female-Male. The non-directional hypothesis that female 'S§_will show different amounts of Verbal behavior, over the -37- four conditions, is not supported by the data at a statistically significant level. COMPARISON OF INTERFERERS ’AND NON-INTERFERERS Table 16. Characteristic facial expressions of interferers and non-interferers (N = 35). Facial Interferers Non-Interferers Significance _ _ Level (Fisher's N_12 N—23 Exact Test- Non-directional hypothesis Worried* 4 12 Not Significant Nervous Smile** 1 12 ;)<:.025 Smile Changed To Worry*** 9 6 5><:.025 * Worried refers to facial expressions where S appeared worried, tense, concerned, frightened, shocked, etc. ** Nervous smile refers to facial expressions where S laughed or smiled nervously throughout the fight. *** Smile changed to worry refers to facial expressions where ‘S_went from a nervous smile or laugh to an expression of worry and concern, as the argument turned into a fight. The facial expressions listed in Table 16, were de- rived from ratings of SS' observed behavior during the ex- perimental manipulations. As seen from Table 16, non-interferers laughed or smiled nervously throughout the fight more frequently than did interferers. Only one of 12 interferers smiled nervously -38— throughout the fight, while 12 of the 23 non-interferers did so. By Fisher's Exact Test, this is statistically significant at p < .025. It is also seen from Table 16, that 9 of the 12 interferers changed their facial expressions from an in- itially nervous smile or laugh to one of worry and concern, as the argument became a fight, while only 6 of the 23 non- interferers did so. By Fisher's Exact Test, this is stat- istically significant at p‘<:.025. As is also seen from Table 16, 12 of the 23 inter- ferers appeared worried and frightened, while only 4 of the 12 interferers did so. However, this was not statistically significant. 'COMPARISONS AMONG 'NON‘INTERFERERS Table 17. Characteristic facial expressions among male and female non-interferers (N = 23). Facial Male Female Significance Level Expression N = 7 N = 16 (Fisher's Exact Test for non- directional hypo- thesis) Worried 1 11 p < . 05 Nervous Smile* 3 9 Not Significant Smile Changed To Worry* 3 3 Not Significant ............ IThese are the same facial expressions described in Table 12, and were derived in an identical manner. -39- As seen from Table 17, ll of the 16 female non- interferers appeared scared or worried, while only one of the 17 male non—interferers did so. By Fisher's Exact Test, this is statistically significant at p< .05. As seen from Table 17, 9 of the 16 female non-inter- ferers laughed or smiled nervously throughout the fight, while only 3 of the 7 male non—interferers did so. Three of the seven male non-interferers changed their expression from nervous smiling or laughter, to worry or concern as the fight progressed, while only 3 of the 16 male non-interferers did so. This was not statistically significant, however. Table 18. Most frequently used rationalizations for non-interference, among male and female non-interferers ( N = 23). Rationalization Male Female Significance Level N = 7 N = 16 (Fisher's Exact Test for non-direc- tional hypothesis) Doubted Authenticity* 2 8 Not Significant Interference Was Imminent** 5 6 Not Significant * refers to rationalizations for non-interference such as, "I didn't know whether they were really fighting", "I didn't think it was real", etc. i.e., doubting the authenticity of the fight. ** refers to rationalizations for non-interference such as, "If it had continued any further, I would have interfered- I was just on the verge of doing this", "One more blow and that would have been enough", "I had just about come to the conclusion I wouldkhave to StOp it, when it ended", etc. i.e., implying interference by S was imminent, had the fight continued. -40- The rationalizations listed in Table 18, were de- rived from a self-report questionnaire which §§ were asked to complete after they had been debriefed, and the true nature of the experiment had been explained to them. The various individual responses were grouped into a number of categories, based upon similarity of content. The two cat- egories listed in Tabie 18, were the most frequently used rationalizations for non-interference. Typically SS chose one of the other of these ration- alizations, with only S_SS_utilizing both of these categories. By Fisher's Exact Test with Tocher's Modification (Siegel, 1956, p. 101), this is statistically significant at p = .05. ‘ SS_who did SgE_appear worried or frightened used these two categories of rationalizations more frequently than other non-interfering SS, By Fisher's Exact Test with Tocher's Modification, this is statistically significant at p = .05. Conversely, those SS_who appeared worried or frightened, rarely used these types of rationalizations. As is seen from Table 18, female §§ more frequently rationalized their non-interference by doubting the authen— ticity of the fight (8 out of 16), while male SS more fre- quently rationalized their non-interference by implying that their interference was imminent, had the fight continued (5 out of 7). These differences are not statistically signifi- cant, however. Two other observations are noteworthy: 1) Only S_S§_indicated that their failure to in- terfere was based on the expectation that someone else in the -41_ room would assume responsibility for stopping the fight. In each of these four cases, S also gave one of the two most frequently-used rationalizations for non-interference. 2) S92£_female SS did not interfere because they felt that being a woman, they might have been injured. SSESS of these SS also gave one of the two most frequently-used rationalization to explain their non-interference. SUMMARY OF RESULTS The major results may be summarized as follows: 1. Combining all conditions: a. Female SS showed significantly fewer positive behaviors (attempts to interfere with the fight), than did male Ea (x2: 4.11, df = 1, p<.05). b. Female SS showed significantly more in- trapersonal behaviors (not interacting with others), while male SS showed more interpersonal (actor) behaviors (inter- acting directly with the actors) (x2 = 8.99, df = 2, p< .025). Both sexes showed interpersonal (peer) behaviors (interacting with the confederates) with approximately equal frequency. c. Although not statistically significant, male 83 used more verbal behaviors than did female SS. d. Although not statistically significant, male SS showed a higher overall level of arousal than did female'SS, -42- 2. Comparing behaviors in the different con- ditions: a. Level of arousal was related to sex of the "victim" (Analysis of Variance, df - l, 27; F = 3.55; E><:-10)- Interaction effect between sex of "victim" and sex of "aggressor" also approached significance (df = l, 27; F = 4.08; pw<:.10). A.test for simple effects with male "victims" was significant (df = l, 27; F = 4.82; p‘<:.05). A test for simple effects with male "aggressors" was also significant (df - 1,27; F = 7.69; p‘<:.025). Main effects for sex of "aggressor" and sex of S were not statistically significant, nor were any of the other interaction effects. b. For male SS, mean level of arousal is 'positively correlated with the percentage of positive be- haviors shown in a given condition (Spearman r = +1.00, p = .05). For fewer SS, mean level of arousal is negatively correlated with the percentage of positive behaviors in a given condition (Spearman r = -.95, p not significant due to small n). c. Male SS showed significantly more posi- tive behaviors in the combined conditions male-male and fe- male-female, than in the combined conditions male-female and female-male (Fisher's Exact Test, p <:.05). While 100% of male‘SS_showed positive behaviors in the condition male-male, 0% of the male SS_ showed positive behaviors in the male- female condition. This was statistically significant (Fisher's Exact Test, p‘<:.025). Female SS did not show statistically -43- significant differences in the number of positive be- haviors shown over the four conditions. d. Male SS_showed significantly more inter- personal (actor) behaviors in the combined conditions male- male and female-female, while showing more intrapersonal be— haviovs in the combined conditions male-female and female- male (X2 = 6.07, df = 2, pw<:.05). Female SS_showed more intrapersonal behaviors in the combined conditions male- female and female-male than in the conditions male-male and female-female (X2 = 4.32, df = 2). This was not statistically significant. e. There were no statistically significant differences in the amounts of verbal behavior used by male and female SS over the four conditions. However, in those conditions where a woman was the aggressor (female-female and female-male), male SS showed more verbal behaviors than did male SS in the conditions where a man was the aggressor (Fisher's Exact Test, p<:.10). 3. Comparing interferers and non-interferers: a. Significantly more non—interferers either laughed or smiled nervously throughout the fight than did in- terferers (Fisher's Exact Test, p <:.025). b. Although not statistically significant, non-interferers also appeared more worried and frightened than did interferers. -44- c. Interferers changed facial expressions from ,/ /. an initial nervous smile or laugh, to one of worry and con- \/ cern, as the argument became a fight, significantly more frequently than did non-interferers (Fisher's Exact Test, p <:.025). 4. Comparing male and female non-interferers: a. Significantly more female non-interferers appeared scared, worried, or frightened than did male non- interferers (Fisher's Exact Test, p <:.0§). b. Although not statistically significant, more female non-interferers laughed or smiled nervously throughout the fight, while more male non-interferers changed their ex- pressions from nervous smiling or laughter, to worry or con- cern, as the fight progressed. c. Although not statistically significant, more females rationalized their non-interference by doubting the authenticity of the fight, while more male non-interferers ra- tionalized their non-interference by indicating that their interference was imminent, had the fight progressed any further. -45_ DISCUSSION The results indicate that when male college under- graduates are confronted with a person being injured by direct physical aggression, they react differentially, accord- ing to their sex, and the sex of the people fighting. It must be kept in mind that discussion of results can only be generalized to the population from which the sample is drawn, namely male college undergraduates. Experimentation with samples from the appropriate populations should pre- cede a wider generalization of the results. Two variables studied by Darley and Latane' (1968 a & b), but which were not manipulated in the present study, re- quire comment: 1) Diffusion of Responsibility: This term refers to the fact that as the size of the bystander group increases, the probability of any one person interfering decreases. In the present study, the group size was held constant, and therefore the influence of this variable may be presumed con- stant over all conditions. 2) The Intervention Sequence (Noticing-Deciding- Taking Responsibility): This is a postulated three-step sequence, leading ultimately to intervention. It is postu- lated that before a person will interfere, he must notice what is happening, he must decide that there is an emergency, -46- and he must take personal responsibility for intervening. a) Noticing: The experimental procedure was such that it was virtually impossible for S not to notice that one person was injuring another. The fight took place within ten feet of S, and directly in front of him. b) Deciding There is an Emergency: There could be little doubt that the fight was a situation of dan- ger. The person being injured screamed, yelled, and was thrown against the wall of the room-complete with loud noise. He was clearly being injured. The apparent anger of the attacker was intense. Similarly, the confederates went through a series of behaviors which called attention to the fight. They had concerned expressions. However, they did not imply, in any manner, that a specific course of action should be taken. Although 10 SS (8 female and 2 male) subsequently questioned the authenticity of the fight (self-report questionnaire), all'SS_inc1uded in the data analysis (N = 35) verbally in- dicated to S, immediately after the fight, that they thought a fight had just taken place. That is, all §§ included in the data analysis passed the criterion of plausibility referred to in the Method section. c) Taking Responsibility: Before S left the room, he indicated that the group members should pay close attention to the acting, because they were to rate each other's performance, at the end of the session. This suggested that ' S should pay attention to what was going on, since he was re- sponsible for evaluating what transpired. More important, -47- however, is that SIS non-interfering SS implied (responses to self-report after fight) that they felt a personal re- sponsibility for interfering when someone was injured. Apparently this stated sense of responsibility was subsequently mediated by rationalization or distortion of events, in the case of the non—interferers. The results of the present study appear to be free from confounding effects attributable to the above-discussed variables. Diffusion of responsibility may be viewed as having an invariant influence, while the three aspects of the intervention sequence appear to have been present in each trial. These results thus add a new dimension to the original and highly creative research of Darley and Latane': the factor of sex differences-differences in the sex of the ob- server, and differences in the sex of the people fighting. Darley and Latane' (1968a) distinguish between "direct" and "reportorial" intervention. Direct intervention (breaking up a fight, extinguishing a fire, swimming out to save a drowner) often requires skill, knowledge, or physical power. It may involve danger... A second way of dealing with an emergency is to report it to someone qualified to han- dle it, such as the police." (p. 382). They further note that the cultural expectation is for men to take more responsibility than women in situations requiring direct intervention. On the other hand, according to Darley and Latane', there are apparently no such cultural expecta- tions for men to take more responsibility than women, where reportorial intervention is possible—both should be expected -48- to interfere. The preponderance of male interferers and the ver- itable absence of female interferers (over all conditions combined) is in agreement with other findings. Berkowitz, Klanderman, and Harris (1964) studied whether SS took initiative to help another person, who was dependent upon them, in a task-learning situation. They found that men tend to take more responsibility and initiative for helping de- pendent people, than do women. As with the present study, this situation required direct intervention. Darley and Latane' (1968a) studied intervention be- havior in a situation where SS overheard a purported epilep— tic seizure. They found that male and female SS_both reported the emergency to S with equal speed. This is an apparent contradiction to the findings in the present study. However, the disagreement is only apparent. The only dependent varia- ble in the Darley and Latane' study was reportorial inter- vention, since direct intervention was impossible. The pre- sent study allowed both reportorial and direct intervention. Herein lies the resolution. As already indicated, cultural expectations for men and women differ for direct intervention, but should be similar in situations allowing for reportorial intervention. In the present study, frequency of direct intervention (over all conditions combined) was significantly related to the sex of S. This is interpreted as being in accord with cultural expectations for direct intervention. -49- One of the most important findings was that the frequency with which male SS intervened, depended signifi- cantly upon the sex of the people fighting. It had been hypothesized that the failure of men to interfere, when a man was injuring a woman, would be due to the inhibiting effects of a high arousal level. However, for male §§ there was a correlation of + 1.00 between degree of arousal and the percent of interventions in a given condition. That is, male SS_appeared least aroused when a man was injuring a woman, and most aroused when one man was injuring another. When two men were fighting, 100% of the male SS intervened, and when two women were fighting, 80% of the male SS inter— vened verbally. In these two conditions, male SS appeared to be considerably more aroused than in the other two con- ditions (male-female and female-male). Much public attention has recently been focused upon what Frank (1968, p. 74) has called the "bystander" phenom- ¢/// enon. In particular, there has been considerable puzzlement and distress regarding situations in which a woman is being injured by a man. It is repeatedly asked, "Why didn't anyone come to her aid?" Aliendation, indifference, and degenera- tion of moral values have all been proposed as explanations for this unhappy situation. It appears, however, that such explanations do not accurately account for what is involved - at least in the case of a college man, watching another college man, injure a college woman. -50— Why didn't the male SS intervene when a man was in— juring a woman? For male S§_there were considerably fewer signs of visible arousal in this condition. In light of this, it is difficult to postulate that male S§_failed to act because they were experiencing anxiety, due to arousal, There was less arousal. In fact, it may be an error to con- ceive of noninterference by these male SS as resulting from some type of inhibition. It may be that these male SS_were obtaining some type of vicarious gratification, by watching a “pw‘ man injure a woman. There is some support for this reasoning. This was the only condition in which a male S (one S) laughed or smiled continuously throughout the entire fight. In addition, more male SS (three SS) showed expressions which changed from a smile or laugh to one of worry-—as the fight became more violent--in this condition, than in any other condition. Other, indirect, evidence also suggests this possibility. All but one of the non-interferers claimed that he would have interfered within the next few moments, had the fight contin- ued. However, the fact remains that he did not, even though by that time in other conditions, male SS_had long since in- tervened. The common usage of this rationalization for non- interference suggests that these SS felt ambivalence. On the one land they felt that it was expected that they should in- tervene, and so they rationalized their failure to act in a manner consistent with cultural expectations. On the other hand, they failed to interfere in spite of this felt pressure to do so. _51- The belief that a man should be more likely to in- terfere when a woman is being injured (as contrasted with one man injuring another) may partially be descended from the overly romantic and chivalrous notion of women that characterized Victorian life. Such unrealiStic, perhaps dereistic, notions impose cultural expectations in such a manner that the obverse aspect of male-female relationships was one of deepseated anger and resentment. In our society, men are allowed to Openly express anger and aggression against other men, with little or no social reprisal. On the other hand, regardless of the anger or rage felt toward a woman (justified or not!!) it is extremely unacceptable for a man to physically injure a woman. It may be that the observed results are restricted to the sampled population--male undergraduate students. Such a conclusion requires further research. We are, however, free to speculate on why S§_were getting vicarious gratifi— cation. It is reasonable to suggest that a generalized hos- tility toward women is more prevalent among male college students than among older men. College males, in general, have had less satisfactory relationships with women. More frequently than not they have seen women in the role of frustrator, rather than as a person who can provide affec- tion, intimacy, satisfaction, and comfort as well as sexual gratification. This coupled with the natural anger and aggression that a person Sggl§_when frustrated by another per- son, would lead college males, in particular, but perhaps also -52- men in general, to find vicarious gratification in watching another man injure or assault a young woman. The desire to physically retaliate is vicariously consummated. Again, it must be stressed that the specificity or generality of the observed behaviors, requires elucidation through further research with different population groups. In discussion with colleagues, a number of other explanations have been forwarded. It is noteworthy that there was unequivocal rejection of the above-discussed explanation. That is, these colleagues all doubted that the male S_failed to interfere because he was receiving vicarious gratification. The psychoanalytic defense mechanism of denial is perhaps appropriate to mention in this context. At any rate, some of these alternative explanations will be briefly discussed. One suggestion is that SS failed to intervene because they were in an unfamiliar situation. That is, there was a lack of familiarity with the surroundings and the situation. This might have been so for the surroundings, but it should have been equally so for all SS, regardless of the experimen- tal condition. If, however, one considers that the situation of a man attacking a woman is rarely observed, then we must further consider this explanation of unfamiliarity. It is, in fact, true that we rarely see a man injuring a woman. It is also true that men rarely, if ever, witness one woman beating up another woman. The results showed, however, that significantly more men interfered in the latter condition, than in the former. That is, it seems that in one unfamiliar -53- situation (female—female) the male S intervened with considerable regularity (80%), while in another unfamiliar situation (male-female), Ell.°f the male SS failed to in- terfere. In other words, this exPlanation may account for some of what was observed,but it does not seem to account for the differential behavior observed in the various con- ditions. Another explanation suggests that before intervening, ' S, evaluates the possible "costs" of such an action, to himself. The validity of such explanations, in general, is not a topic for discussion at present, however. Such ex- planations do not seem to account for the differential be- haviors observed in this study. Such explanations claim that SS_did not interfere because they would have to pay the "cost" of having their personal freedom restricted by virtue of the helping act. Such explanations further suggest that by becoming involved, S may have to pay the "cost" of restrictions on his freedom even subsequent to the actual intervention. That is, S may have to seek help for one or both of the fighters, or perhaps S will have to take them shoumd be asked: "Why did home, etc. Again the question only the'SS in the male-female interfering?" Also, it may be in terms of physical danger to When two men are fighting, and when two men were fighting! unanimously refrain from added that the "cost" to S himself is objectively greater yet all male S§_interfered The above explanations deal with unitary approaches to the observed phenomena. As with many social phenomena, -54- the problem may be considerably more complex. The situation used in the present research may arouse conflicting and com- plex motives in S, As a result such unitary explanations would perhaps be oversimplifications. A definitive ex- planation for the above findings must await further research. The relationship between arousal and behavior was also dependent upon the sex of S. For male SS there is a positive correlation between level of arousal and the percen- tage of intervening behaviors in a given condition. For fe- male SS, the relationship is the reverse. The higher the arousal, the fewer the positive behaviors. This is impor- tant in View of the findings of Darley and Latane' (1968a) who reported that non-interfering SS appeared to be more aroused than interfering SS, The relationship appears to be more complex. In the male-female condition, male §§ were less aroused than in any other condition, and yet they did not intervene. When male S§_were most aroused (male- male condition), they interfered with the greatest frequency. Darley and Latane's (1968a) conclusion that the non- interfering bystander is not an apathetic or alienated indi- vidual is supported by the present data. However, their interpretation of the reason for non-intervention appears to be in conflict with the additional findings of the present study. They explain their SS'non-intervention in the follow- ing manner. S is in a state of unresolved conflict as to whether or not he should intervene. The greater the emotional -55- arousal observed in their non-interfering SS is considered a manifestation of this conflict. In the present study, an alternative explanation is required, perhaps due to the differences in the two experiments. In the present study, a situation allowing direct intervention, non—interfering male §§ were ISSS aroused than were interfering male SS, It is unlikely that apathy or alienation are major factors, for given the intensity of the fights, and the close proxi- mity of the fights to S, it is difficult to imagine any person remaining apathetic. On the contrary, it was suggested that S was obtaining considerable vicarious gratification, in watching a man beat up a woman. S was obtaining sufficient gratification to overcome any proclivity he might have had toward intervention. Unquestionably there was a conflict. The conflict, however, could be as follows: Should S maximally enjoy the situation or should he interfere? Interference, the socially acceptable behavior was renounced, with few, if any, visible signs of guilt or anxiety. It is true that more male SS tended to look at their peers in this condition, than in all the other conditions combined. It is also true that when asked why they had not interfered, 71% of these male §§. replied that their intervention was imminent. These are the signs that suggest that the above--indicated conflict was in- deed taking place. However, in spite of this experienced pressure, to respond in a socially acceptable fashion, the gratification derived from watching a man beat up a woman was -56- sufficient to overcome any such tendency toward inter- vention. This behavior is in contrast to that shown by male SS in the male-male condition. In the male-male condition, all male SS intervened before half the time allotted for each fight had elapsed. The reasons given by SS for their non-interference re- veal an interesting sex difference. When male SS_were asked why they hadn't stopped the fight, most indicated that such intervention was imminent. Female SS_more frequently re- sponded to this question by doubting the authenticity of the fight. In the language of Darley and Latane' (1968b), these female SS_failed to interpret the situation as an emer- gency. Given that interfering as well as most non—interfering male SS_conSidered the situation to be an emergency, why didn't the female SS do likewise. It is highly questionable that the female SS lack the perceptual and cognitive abilities of the male SS, That is, they both may be assumed to be equally astute in assessing the objective realities of a situation. But yet the female S§_typically doubted the authen- ticity of the fight, in contrast to their male counterparts. That is, female SS tended to rationalize their non-interference by distortion or denial of the fight. Male SS_on the other hand tended to rationalize their non-intervention by stating that their intervention was imminent. SS_who rationalized their non-interference by doubting the authenticity of the fight, did not appear tense, worried -57- or frightened. Apparently they were able to reduce their fear and anxiety by denying or distorting what had happened. When debriefed, they were able to answer that they had not interfered because they doubted the authenticity of the fight. The confusion that these §§ experienced was readily apparent. Immediately after the fight, these SS told S that a fight had taken place during his absence. That is, they passed the criterion for plausibility of the experiment. However, once the true nature of the experiment was explained to them, these same SS_were then able, with apparent equanimity, to state that they had doubted the authenticity of the fight and had therefore not intervened. The speed with which such reversals took place, and the apparent dis- regard for conflicting elements of reality, affords an in- sight into the potency of such defensive processes. The speed with which this change took place, and the seemingly callous disregard for the conflicting elements of reality was impressive. On the other hand, S§_who did not deny or distort the emergency confronting them, were placed in the position of having to deal with their reasons for non—interference in an alternative fashion. This suggests that there are several characteristic modes of responding, when a person is confronted with physical aggression, in which someone is being injured. Some peOple responded by denying or distorting what is happening. -58- This apparently enables them to reduce their perceived anxiety-anxiety which attends the conflict of deciding whether or not to interfere. Others respond by watching the fight intently, apparently receiving vicarious gratifi- cation. Such people appear to rationalize their non-in- terference by insisting that their intent to intervene had been present from the outset, and that their actual intervention had been imminent. Others are frightened by what they ob- serve and indicate that their own fright, the inaction of the bystanders, or the fear of being injured prevents them from interfering. Other people, and unfortunately they represent a minority, tend to respond to such situations by interven- ing—-by attempting to stop the fight. SIRECTIONS’FOR 'FUTURE‘RESEARCH The tendency is to deny the validity of the conclu- sions and speculations found in the previous pages. They are offensive to us, and reminds us of certain socially unacceptable desires within ourselves. Rather than denying the findings presented here, it is more fitting to explore this particular experimental sit— uation in greater detail. This will serve to clarify the relevant variables influencing this rather unpleasant state of affairs. For example, the present study used a sample drawn from a pOpulation of college undergraduates. Replications should be carried out with a variety of -59- other samples, each representing different subpopulations within American society. Similarly, the study should be replicated with persons in varying age and racial groups, as it may be that the above results are a function of the bystander's socio-economic-educational level as well as his age. Finally, all of these replications should ideally be carried out in a more naturalistic setting. ETHICAL CONS IDERATIONS This research utilized procedures which might stim- ulate certain questions of ethics. Several considerations are relevant to this issue: 1) The research proposal was discussed and approved by three members of the psychology department faculty. This enabled S to survey other opinions, regard- ing the ethical implications of the research. All three faculty members concurred that the present study did not appear to violate contemporary ethical principles. Spec- ifically, the research was preceded by a careful weighing of the potential dangers to S, against the foreseeable benefit to others. After considerable assessment, S felt that (a) the potential psychological danger to S was minimal, and (b) that the debriefing procedures were sufficiently comprehensive to protect S from psychological harm. The research was also evaluated in terms of its potential for providing information of scientific value. It was felt that the research had potential for providing valuable information on a relevant and pressing social problem. -60- 2) S reviewed his responsibility to act as the protector of S§L_health and privacy. In this regard it was decided that any time the procedure appeared to be producing even the slightest hint of psychological harm, no further trials would be conducted until appropriate changes could be implemented. No such changes were found to be necessary. 3) There were no explicit restraints, which kept §§ from leaving the experiment, had they desired to do so. 4) Two of the three faculty members observed one or more experimental trials, and thereby provided mean- ingful supervision of the research. In view of these considerations, it is felt that there was a careful assessment of the potential dangers. More than sufficient care was taken to avoid situations which would psychologically harm S, S was firmly convinced that S was protected from any potential psychological harm. If he had not been so convinced, the research would not have been con- ducted. -61- SUMMARY AND CONCLUSIONS The purpose of this study was to observe the re- actions of college men and women who were confronted with a situation in which one person was injuring another by direct physical aggression. The experiment purported to evaluate the ability of people to extemporaneously act out roles and situations assigned to them--a type of psycho— drama. Out of this situation emerged a "spontaneous" fight between two apparent SS. Six participants were present in each trial--all were ostensibly SS, However, there was only one S in each trial. In addition to the two fighters, who were trained actors, there were three undergraduate assist- ants, who served to make the group situation more plausible. S was not present in the room during the fight sequence. A 2 x 4 design was used with sex and experimental conditions as the independent variables. There were four experimental conditions, each representing a fight between two male and/or female actors. In each case one actor was clearly beating up the other. The conditions were: 1) male- male; 2) female-female; 3) male-female (male injuring female); and 4) female-male (female injuring male). Of the forty SS originally obtained for the experiment, five male and five female S§_were randomly assigned to each condition. Failure of SE to appear for the experiment, plus the failure of seven SS to satisfactorily meet the criterion for plausibility ~62- of the experimental manipulations, necessitated obtaining additional SS, Difficulty in obtaining additional SS limited the total number of SS used in the data analysis to §§.(17 male and 18 female). The verbal and nonverbal behaviors of each S were scored and rated by trained coders through one-way observa- tion mirrors. The major results may be summarized as follows: 1. Combining all conditions: a. Female S§_showed significantly fewer positive behaviors (attempts to interfere with the fight), than did male S_s (x2 = 4.11, df = 1, p< .05). b. Female SS showed significantly more in- trapersonal behaviors (not interacting with others), while male §§ showed more interpersonal (actor) behaviors (inter- acting directly with the actors) (X2 = 8.99, df = 2, pw<:.025). c. Although not statistically significant, male SS_used more verbal behaviors than did female SS. d. Although not statistically significant, male'Ss showed a higher overall level of arousal than did fe- male'Ss. 2. Comparing behaviors in the different condi- tions: a. Level of arousal was related to sex of the "victim" (Analysis of Variance, df - l, 27; F = 3.55; -63- p <:.10). Interaction effect between sex of "victim" and sex of"aggressor" also approached significance (df = l, 27; F = 4.08; p< .10) . A test for simple effects with male "victims" was significant (df = 1,27; F = 4.82; p}<:.05). A test for simple effects with male "aggressors" was also significant (df - 1,27; F = 7.69; p <:.025). Main effects for sex of "aggressor" and sex of S were not statistically significant, nor were any of the other interaction effects. b. For male §§ mean level of arousal is positively correlated with the percentage of positive be- haviors shown in a given condition (Spearman r = +1.00, p = .05). For female SS, mean level of arousal is negatively correlated with the percentage of positive behaviors in a given condition (Spearman r = -.95, p not significant to due small n). c. Male SS showed significantly more posi- tive behaviors in the combined conditions male-male and fe- male-female, than in the combined conditions male-female and female-male (Fisher's Exact Test, p <:.05). While 100% of male SS_showed positive behaviors in the conditions male- male, 0% of the male SS_showed positive behaviors in the male-female condition. This was statistically significant (Fisher's Exact Test, p‘<:.025). Female S§_did not show statistically significant differences in the number of posi- tive behaviors shown over the four conditions. d. Male S§_showed significantly more in- terpersonal (actor) behaviors in the combined conditions male- male and female-female, while showing more intrapersonal be- -64- haviors in the combined conditions male-female and female- male (x2 = 6.07, df = 2, p< .05). Female S_s_ showed more intrapersonal behaviors in the combined conditions male- female and female-male than in the conditions male-male and female-female (X2 = 4.32, df = 2). This was not statistically significant. e. There were no statistically significant differences in the amounts of verbal behavior used by male and female SS over the four conditions. However, in those conditions where a woman was the aggressor (female-female and female-male), male S§_showed more verbal behaviors than did male SS in the conditions where a man was the aggressor (Fisher's Exact Test, p‘<:.10). 3. Comparing interferers and non-interferers: a. Significantly more non-interferers either laughed or smiled nervously throughout the fight than did interferers (Fisher's Exact Test, p < .025) . b. Although not statistically significant, non-interferers appeared more worried and frightened than did interferers. c. Interferers changed facial expressions from an initial nervous smile or laugh, to one of worry and concern, as the argument became a fight, significantly more frequently than did non-interferers (Fisher's Exact Test, p<.025). -65- 4. Comparing male and female non-interferers: a. Significantly more female non-interferers appeared scared, worried, or frightened than did male non- interferers (Fisher's Exact Test, p'<:.05). b. Although not statistically significant, more female non-interferers laughed or smiled nervously throughout the fight, while more male non-interferers changed their expressions from nervous smiling or laughter, to worry or concern, as the fight progressed. c. Although not statistically significant, more females rationalized their non-interference by doubting the authenticity of the fight, while more male non-interferers rationalized their non-interference by indicating that their interference was imminent, had the fight progressed any further. The differences in the behavior of male and female SS, with all conditions combined, were attributed to the influ- ence of cultural expectations for sex role behavior. However, since male SS failed to interfere in the condition male- female it was speculated that they were deriving vicarious sexual and/or hostile gratification from watching a man in- jure a woman. These results, if extended to actual life situations, suggest some disturbing conclusions. First of all, the re- sults suggest that women are unlikely to directly interfere in a situation where one person is injuring another--regard- -66- less of the sex of the participants. The results also suggest that while college men are likely to directly in- terfere when two men are fighting, they are unlikely to in— terfere if a college man is beating up a college woman. Much of the recent public attention in this area has focused upon situations in which a woman was being injured by a man. Popular reaction has been one of puzzlement and distress. It is repeatedly asked, "Why didn't anyone help her?" Ex- planations of alienation, indifference, or degeneration of moral values have all been proposed as explanations for this regrettable situation. It appears that such explanations do not totally account for what is happening, at least in the case of a man injuring a woman. College men interfere, at least in fight situations, in a differential fashion: They interfere, or fail to do so, based on the sex of the people involved in the fight. This leads to the speculation that men will not interfere, when women are being injured by another man, because they are possibly obtaining sexual or hostile vicarious gratification which conflicts with socially respon- sible thoughts, feelings, and actions. If this is the case, and the present results tend to suggest such a disturbing conclusion, then situations similar to the Kitty Genovese incident will probably continue to occur. It appears that a woman being assaulted in the city streets may be functionally alone, even if (or perhaps, as Darley] and Latane' note,'e3pecially if) she is surrounded by many bystanders. 10. 11. 12. 13. -67- REFERENCES Berkowitz, L., Klanderman, S., and Harris, R. Effects of experimenter awareness and sex of subject on reactions to dependency relationships. Sociometry, 1964, S1, 327-329. Darley, J. and Latane', B. Bystander intervention in emergencies: diffusion of responsibility. S; ’Person. &'SOC.'PsYchol., 1968, S (4), 377—383. (a) Darley, J. and Latane', B. When will people help in a crisis. Psychology Today, 1968, 2&7), 54. (b) Darley, J. and Latane', B. Personal communication. 1969. Frank, J. Sanity & Survival. New York: Vintage, 1968. Hays, W. L. Statistics for Psychologists. New York: Holt, Rinehart, & Winston, 1963. Karon, B. P. Problems of validities. In A. I. Rabin (Ed.): ’Projective Techniques in Personality Assessment. New York: Springer, 1968. Kaufman, H. The unconcerned bystander. Paper presented at the meeting of the American Psychological Association, San Francisco, 1968. Latane', B. & Rodin, J. A lady in distress: inhibiting effects of friends and strangers on bystander in- tervention. J. Exp. Soc. Psychol., 1969, in press. Milgram, S. Behavioral study of obedience. J. abnorm. 'Soc. Psychol., 1963, S1 (4), 371-378. Milgram, S. Group pressure and action against a person. J. abnorm. soc. Psychol., 1964, SS (2), 137-143. Milgram, S. Liberating effects of group pressure. S; 'Person.'& SOC. Psychol., 1965, S(2), 127-134. E) Milgram, S. Some conditions of obedience and dis- obedience to authority. Human Relations, 1965, ’ SS(R), 57-76. (b) 14. 15. 16. -68- Siegel, S. Nonparametric Statistics. New York: McGraw-Hill, 1956. Tilker, H. A., Socially responsible behavior as a function of observer responsibility and victim behavior. Unpublished doctoral dissertation, Michigan State University, 1967. Winer, B. J., Statistical Principles in Experimental 'Design. New York: McGraw-Hill, 1962. APPENDICES -70- APPENDIX A INSTRUCTIONS TO SUBJECTS This experiment is a pilot study to test whether it is feasible to implement a program of psychodrama or "role playing" for college students. Our purpose here will simply be to see how well psychologically healthy students are able to utilize this technique. In actual practice, psychodrama is used as a means of supplementing more conventional forms of psychotherapy. The actors are patients, who are assigned roles by the thera- pist. The therapist typically suggests a situation to be acted out, and this is given to the patients (actors), just prior to the psychodrama session. The goal in psychodrama is to help the patient achieve an emotional catharsis ("release"), and to study the effect ppon the patient of acting out roles related to his past traumatic experiences and present problems. Learning to express oneself easily and spontaneously and to meet new situations effectively as they arise, in the course of psychodrama, frees the individual from emotional blocks, develops insight into and appreciation of the feelings and behavior of others, and promotes adequacy and flexibility in social skills. In the course of our meeting here today, I will pair you in groups of two. Each of you will be given a role to play. These roles will all be members of a typical family. For example, a male-female pair, might be asked to act out the role of father and mother, respectively. I will give each pair a brief description of the setting in which they are to act, before they take their turn. Those of you who are not acting at any given time, should pay attention to those who are, as this may help you to improve your own performance, and also it will be of help to me to have your comments and suggestions. Remember, you will NOT be graded. There are no scores at all. This is only a pIIbt study to see how well psycho- drama can be utilized with university students. However, I would like you to try your best to express your emotions as spontaneously as possible, giving the best possible portrayal of the role and situation that you can. I know that none of you are professional actors, but neither are the peOple with whom psychodrama is used. I only want you to try the best you can and to be as free as possible in expressing yourself and your emotions. Do you have any questions? -71- APPENDIX B BASIC CATEGORIES FOR RATING SUBJECT'S BEHAVIOR POSITIVE MEASURES (+) Intrapersonal (+I) Verbal (+IV) Behav1oral (+13) Behavioral = Non-verbal behavior Interpersonal (peer) (+Ip) Verbal (+IpV) Behavioral (+IpB) Interpersonal (actor) (+Ia) Verbal (+IaV) Behavioral (+IaB) NEUTRAL MEASURES: (N) Intrapersonal (NI) Verbal (NIV) Behavioral (NIB) Interpersonal (peer) (NIp) Verbal (NIpV) Behavioral (NIpB) Interpersonal (actor) (NIa) Verbal (NIaV) Behavioral (NIaB) NEGATIVE MEASURES: (-) Intrapersonal (-I) Verbal (-IV) Behavioral (-IB) -72- Interpersonal (peer) (-Ip) Verbal (-IpV) Behavioral (-IpB) Interpersonal (actor) (-Ia) Verbal (-IaV) Behavioral (-IaB) APPENDIX C SCORE SHEET FOR RATERS Page 1 RATER SUBJECT NAME & STUDENT NUMBER SUBJECT NUMBER SEX CONDITION (1-6) M or F MM, FF, MF, or FM (7) (8) -74- Page 2 SUBJECT NUMBER CATEGORY (9-11) DESCRIPTIVE PORTRAIT: Give SPECIFIC & EXACT Details of S's BeEavior, with regard to SELF, PEERS, and/or ACTORS. BODY (Posture, Movements, & Gestures): (12-14) """ ' ‘ VERBALIZATIONS (Tone of voice & Exact words): (15-17) FACIAL EXPRESSIONS: (18-20) OTHER INFORMATION (regarding S's Behavior): (21-23) SUBJECT NUMBER -75_ Page 3 AROUSAL RATING: (24) (25-27) (34+36) (43-45) (Circle the Appropriate Number) 1 = Neutral or neglibible arousal 3 Moderate arousal, but not uncon— trolled 5 = Extreme arousal, distraught--on the verge of being uncontrollable. (28-30) , (31-33) (37-39) (40-42) (46-48) (49-51) -76- APPENDIX D SELF REPORT QUESTIONNAIRE COMPLETED BY S AFTER DEBRIEFING Please answer these few questions as best you can. Use the back side if necessary. 1) Describe how you felt during the fight. What were your feelings? (52-53) 2) What thoughts were going on in your mind dur- ing the fight? What were you thinking? (54-55) 3) Did you have any ideas or thoughts about . . 9 (56-57) stopping the fight. 4) Why do you think that you didn't stop the fight? (58-59) 5) Did the fight seem to be realistic? Did you (60-61) think that they were actually fighting? 6) What are your general comments and reactions ' 9 (62-63) to What has Just happened to you. IIHIIIIIII|||Il||IlllllIIIIlllllllllIllllIlHllIIIHHI 312930110