THE FLASH SYSTEM: A FIELD STUDY OF HELPING BEHAVIOR by Carole J. Achenbach A thesis submitted in partial fulfillment of the requirements for the degree of Master of Arts in the Department of Psychology in The University of South Florida August, 1971 Thesis supervisor: Dr. Louis A. Penner
Certificate of Approval -Master's Thesis Graduate Council University of South Florida Tampa, Florida CERTIFICATE OF APPROVAL MASTER 1 S THESIS This is to certify that the Master's Thesis of Carole J. Achenbach name of student da e Thesis committee: __ _____ )Sember J s:
ACKNOWLEDGMENTS The author wishes to acknowledge the help of the Florida Highway Patrol without whose cooperation this study would have been impossible. Thanks also are due my committee: Dr. Louis Penner, Dr. Max C. Dertke, and Dr. Donald Stein. I wish also to thank my husband, Dr. Karl Achenbach, whose help and forbearance were essential to the task. ii
I. II. III. IV. v. VI. TABLE OF CONTENTS LIST OF TABLES . . LIST OF FIGURES ABSTRACT . . . INTRODUCTION . A. Victim Characteristics and Helping B. Situational Determinants of Helping METH.OD RESULTS DISCUSSION REFERENCES APPENDIX . . . . . . . . . . . iii iv v vi 1 2 4 8 11 29 33 34
LIST OF TABLES Table 1. Comparisons of Proportions of Helping for Ten Victim Conditions 12 Table 2. Comparisons of Proportions of Flashing for Ten Victim Conditions 13 Table 3. Comparisons of Proportions of Stopping for Ten Victim Conditions 14 Table 4. Comparisons of the Proportions of Helping for Victim Race Categories 18 Table 5. Comparisons of the Proportions of Flashing for Victim Race Categories 19 Table 6. Comparisons of the Proportions of Stopping for Victim Race Categories 20 Table 7. Comparisons of the Proportions of Helping for Victim Sex Categories 22 Table 8. Comparisons of the Proportions of Flashing for Victim Sex Categories 23 Table 9. Comparisons of the Proportions of Stopping for Victim Sex Categories 25 Table 10. Comparison of Representation in General Driver Population with Representa-tion in Helping Driver Population 26 Table 11. Proportions of Helping Behavior for Alone vs Not Alone Vehicles and for Alone vs Not Alone Drivers 28 iv
LIST OF FIGURES Figure 1. Diagram of the experimental setting . Sa v
ABSTRACT To determine the influence of race, sex, diffusion of responsibility and cost on helping behavior, an apparently disabled vehicle was located on the shoulder of an interstate highway. An experimental system for reporting disabled vehicles was recently installed on this highway, such that, passing drivers need only flash their high beams to notify the highway patrol of a disabled vehicle. The race and sex of the single and paired victims, who were seated on the lowered tail-gate of the car, was varied. Data collectors recorded the race and sex of the passing drivers. In addition, they recorded whether the driver was alone or not alone in his car and whether the vehicle was alone on the highway or in close proximity to at least one other car. While significantly more drivers flashed than stopped, the proportion of drivers who flashed was only .04. Victim race and sex were less influential in eliciting helping than whether the victims were alone or paired. Paired victims elicited more help than single victims. White drivers helped significantly more than black drivers and males helped more than females. The data failed to support the concept of diffusion of responsibility, however, there was an effect vi
for the cost of looking foolish. That is, drivers who were alone in their vehicles helped significantly more than drivers who were not alone. vii
INTRODUCTION There has recently been a renewal of interest in some of the more positive forms of social behavior. These prosocial behaviors are apparently unselfish acts of helping and have been labeled "altruistic" or helping behaviors. These behaviors are of particular interest to social investigators because they seem to occur without regard to the law of effect, that is, altruistic acts seem to occur in the absence of reward. This apparent violation of reinforcement theory may be entirely a function of the definition of altruism. The purpose of this paper, however, is not to argue this theoretical point. The reader is referred to Krebs (1969) and MacCauley and Berkowitz (1970) for a theoretical discussion of altruistic behavior and a review of the literature. The purpose of this study was to determine the influence of specific "victim," helper, and situational characteristics on helping behavior. The method of studying helping behavior used in this research was initiated by Darley and Latane and the method is characterized by field or field-like settings in which the effects of systematically varying specific situational characteristics are determined. 1
In this study, the race and sex of the victim(s) were systematically varied and the effects of helper race and sex on the rate of helping were determined. In addition, the influence of two situational determinants, diffusion of responsibility (Darley and Latane, 1968) and the costs of helping, was investigated. Victim Characteristics and Helping Novak and Lerner (1968) found that if a potential helper perceived a victim to be similar to himself, an interaction would seem more attractive and, consequently, be more likely to occur than if the potential helper perceived the victim to be dissimilar. A similar finding was reported by Feldman (1968) who found that a helper would be more likely to give help to a victim he perceived to be similar to himself than he would a victim he perceived as dissimilar. These findings suggest that a victim is most likely to receive help from someone of the same race and sex as himself. A study by Piliavin, Rodin, and Piliavin (1969) supported this hypothesis. They had black male and white male students pretend to be either drunk or ill on a New York subway. Males helped these male victims significantly more than did females. White males helped white victims significantly more than they helped black victims and black males tended to help black victims more than white victims. Additional research on the influence 2
of race and sex in helping situations, however, fails to support the increased probability of interaction due to perceived similarity. In a study of Wispe and Freshley (1971), black female and white female victims had a full 3 bag of groceries apparently break open as they were leaving a grocery store. They found that males helped these female victims significantly more than did females regardless of the race of the victim. Females by and large ignored the victims regardless of race. In another study, Darley and Latane (1970) investigated helping behavior as a function of the sex of a person requesting help as well as the sex of the potential helper. They found that females were more likely to be helped than males, particularly if the helper was male. In addition they found that a pair of female requesters increased the likelihood that help would be given, while a pair of males were slightly less likely to be helped than a single male. The fact that female victims were helped significantly more by males than females is not a surprising finding. It is also not surprising, given the physical and social differences between males and females, to find that males are more likely to help a male victim than are females. That fact that males of either race helped females equally, regardless of race, however, is not such an easily predictable finding.
4 Situational Determinants of Helping and Darley (1970) suggest that before a person will help in an emergency he must first identify the situation as an emergency. If a lone bystander observes an emergency he must decide for himself the seriousness of the situation. If, however, there is more than one person present, each individual bystander may look to the other bystanders for some confirmation of his evaluation of the situation. If each bystander fails to react to the emergency, it is possible for a state of "pluralistic ignorance" (Darley and 1970, p. 15) to develop. Each bystander will refrain from acting because he doesn't see anyone else responding as if the situation is an emergency. In addition, if one bystander in a group acts to help he faces the possibility that the situation is not in fact an emergency and that he may look foolish for over-reacting. Support for this premise is found in the following studies. Latane and Darley (1968) had Ss who were either alone, with two non-reacting "stooges" or with two other Ss in a room that gradually became filled with smoke. Their results showed that Ss who were alone in the room were more likely to report the smoke than Ss who were with either the two "stooges" or the two other Ss. The Ss who were not alone may have refrained from reporting the smoke because they feared that the other two people in the room would think that they were scared. Alternatively, the may have felt
5 safer because there were other people present. The following experiment was designed to rule out such explanations by creating an experimental situation where the Ss were not in potential danger. In this study, and Rodin (1969) asked male undergraduate Ss to wait in a room alone, with a friend, or with a stranger. While in the room they overheard a tape recording of a woman falling and crying out in pain. They found that alone reacted faster than Ss who were with one other person. In this situation the socially acceptable behavior was to aid the "fallen woman." There was no reason for the Ss to feel that their behavior would be evaluated negatively, however, Ss who were not alone still failed to act as fast as Ss who were alone. In addition to identifying a situation as an emergency, Darley and Latane suggest that bystanders may fail to act because of what they call a diffusion of responsibility. If a lone bystander observes a situation where help may be needed, then the responsibility to act and the guilt caused by inaction rest entirely on his shoulders. If, however, there is a group of bystanders, the responsibility to help and the guilt associated with inaction are reduced for all members of the observing group by diffusion. Darley and Latane reason that a single bystander would be most likely to help because he carries all the responsibility and must accept all the blame if he fails to act. Darley and Latane reason further that bystanders who are unable to
6 observe or communicate with other bystanders known to be present will be less likely to help than bystanders who know that they are alone. To test this hypothesis, Darley and Latane (1968) told their Ss that they were either alone or in the experiment with either one or four other unseen subjects. The Ss overheard someone having an epileptic seizure. Their results showed that alone reported the emergency faster than who thought that they were not alone. The subway study cited by Piliavin, Rodin, and Piliavin (1969), however, failed to find support for the concept of diffusion of responsibility in that the number of bystanders on the subway did not effect the rate of helping. Darley and Latane also suggest that the costs that accompany helping behavior tend to inhibit helping. The first cost identified was that of physical exertion and possible physical injury. The second cost, that of looking foolish, has been referred to earlier. It is simply the embarrassment one suffers if one responds to what he perceives to be an emergency and none exists. This cost probably increases with an increase in the number of people observing the potential helper. With respect to the victim and helper characteristics, the following hypotheses were made: 1. Males will be more likely to help than females.
7 2. Females are more likely to be helped than males. 3. Male helpers will tend to help male victims of their own race more than male victims of a different race. 4. Male helpers of either race will help female victims regardless of race. With respect to the situational variable of diffusion of responsibility, it was hypothesized that the driver of a vehicle alone on the highway would be more likely to help than the driver of a vehicle that is not alone on the highway. With respect to Darley and Latane's cost of looking foolish, it was hypothesized that a driver who is alone in his vehicle would be more likely to help than a driver who has at least one passenger in the vehicle with him.
METHOD This study was conducted on Interstate Highway 4 just east of Lakeland, Florida. There is an experimental system in operation on this highway for helping disabled motorists. This system consists of signs erected along the highway that instruct the passing motorists to report any disabled vehicle by flashing their high beams three times at an oval sign with the word FLASH on it. When a passing motorist flashes his lights a photo-electric cell, located just past each flash sign, transmits a signal to Florida Highway Patrol Headquarters in Lakeland. Upon receipt of three such signals a Highway Patrol trooper is dispatched to assist the stranded motorist. This system provides an excellent field setting for research on helping behavior because the system itself is an unobtrusive measure and the physical costs of helping are minimal. In addition the physical cost of helping is not confounded with the cost of looking foolish. Procedure. A tan 1965 station wagon was parked on the shoulder of the highway halfway between an instruction sign and a flash sign (see Figure 1). This location insured that every driver passing the disabled vehicle had 8
INSTRUCTION SIGN "VICTIMS" FLASH SIGN e w I I I I I I I I I I I I I I I I I I I I I DATA COLLECTORS FIGURE 1 DIAGRAM OF EXPERIMENTAL SETTING I I I (INTERSTATE HIGHWAY 4 EAST OF LAKELAND, FLORIDA Ba
9 had at least one opportunity to learn that he could help by flashing his lights. The hood of the station wagon was raised to suggest to the passing motorists that the vehicle was disabled. While the experiment was in progress, the victim or victims seated themselves on the lowered tailgate of the station wagon and were, therefore, in clear view of the on-coming motorists. The four victims were college students. There were ten victim conditions. The single victim conditions included a white female, a black female, a white male, and a black male. The remaining six victim conditions consisted of victim pairs. They were: a white male and female, a black male and female, a white female and a black female, a white male and a black male, a black male with a white female, and a white male with a black female. Each victim condition was terminated after thirty minutes and the order of presentation was determined by chance. Since it was felt that drivers going to and from work and lunch might respond differently from other drivers, the data collection periods were limited to avoid both commuter and lunch hour traffic. All the data was collected on the same week day--a Thursday. The four data collectors were located in the trees on the median strip approximately 100 feet past the flash sign. This location provided a clear view of the highway while allowing the data collectors to be hidden from the view of the passing motorists. The data collectors recorded the incidence and type of helping behavior, that
10 is whether a passing motorist flashed his lights or actually stopped to assist. In addition, the data collectors recorded the race and sex of each driver, whether the driver was alone in his vehicle or had at least one passenger, and whether a vehicle was alone on the highway or in close proximity to one or more other vehicles when passing the victims.
RESULTS The data indicate that while the FLASH system, because it involves negligible cost to the driver, may appear to be an altruist's dream, it fails to elicit more than a minimal rate of helping. While more motorists flashed than stopped, the drivers who flashed represented only 4% of the total population of drivers who passed. This may be due to the fact that the drivers who didn't flash did not understand the instruction signs or that they may have been too busy attending to driving to flash. The overall incidence of helping was low (i.e., 81 out of 1490 drivers flashed or stopped. See Appendix for complete record of data). This extremely low percentage of helping makes the drawing of any definite conclusions tenuous, but the data do suggest certain interesting findings. The statistic used in all analyses was a test for the significant difference between proportions. Tables 1, 2, and 3 give an overview of the relative amount of helping (Table 1), flashing (Table 2), and stopping (Table 3), elicited by each of the ten victim conditions. These tables list the victim conditions across the top. The order in which the conditions are listed varies from table to table because it is determined by the magnitude o f the proportion associated with each 11
Table 1. Comparisons of Proportions of Helping for Ten Victim Conditions1 BM (. 013) N BM NS WF BF WFBF WMBM -WMBF BMBF -WM BMWF -WMWF p <. 05 ** p <. 01 *** p< .001 WF BF (. 032) (. 046) N N NS 2.35 --NS --Victim Conditions WFBF WMBM WMBF (. 056) (. 059) (. 062) NlO N N ** ** ** 3.07 3.06 3.06 NS NS NS NS NS NS --NS NS --NS --BMBF WM BMWF (. 062) (. 064) (. 067) N N N ** ** *** 3.26 3.19 3.37 NS NS 2.19 NS NS NS NS NS NS NS NS NS NS NS NS --NS NS --NS --WMWF (. 088) Nl2 *** 4.16 ** 3.11 2.10 NS NS NS NS NS NS lcell entries are Z scores based on comparisons of the proportion of people helping. The proportion is the number of people who helped divided by the number of people who passed. '"""" N
WF BM BF BMWF WFBF WM BMBF WMBM WMBF WMWF Table 2. Comparisons of Proportions of Flashing for Ten Victim Conditions1 Victim Conditions WF BM (. 005) (. 013) N=l --* p <. 05 ** p < .01 *** p < .001 N NS --BF (. 033) NS ** 2.80 NS --BMWF WFBF (. 033) (. 033) NS N ** ** 2.80 2.80 NS NS NS NS --NS WM BMBF WMBM WMBF (. 056) (. 056) (.059) (. 062) N N N N *** *** *** *** 3.92 4.25 4.15 4.07 ** ** ** ** 2.86 3.07 3.06 3.06 NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS WMWF (. 066) N *** 4.36 *** 3.31 NS NS NS NS NS NS NS lcell entries are Z scores based on comparisons of the proportion of people f lashing. The proportion is the number of people who flashed divided by the number of people who passed. ,........ w
BM WMBM WMBF BMBF WM BF WMWF WFBF WF BMWF p ** p *** p Table 3. Comparisons of Proportions of Stopping for Ten Victim Conditions1 BM (. 000) NO --.05 .01 .001 WMBM (. 000) NO NS --WMBF (. 000) NO NS NS --Victim Conditions BMBF WM BF (. 006) (. 008) (. 013) Nl Nl N NS NS 2.17 NS NS 2.17 NS NS 2.17 --NS NS --NS --WMWF WFBF WF (. 022) (. 022) (. 026) N N N ** ** 2.44 2.75 3.25 ** ** 2.44 2.75 2.88 ** ** 2.44 2.75 2.88 NS NS 2.22 NS NS NS NS NS NS --NS NS --NS --BMWF (. 033) N *** 3.30 *** 3 30 *** 3.30 2.45 2.08 NS NS NS NS lcell entries are Z scores based on comparisons of the proportion of people stopping. The proportion is the number of people who stopped divided by the number of people who passed. t--'
15 condition. The numerical value of the proportion for any victim condition is noted in parentheses just below that condition. The condition listed at the top left has the smallest proportion and the proportions increase in magnitude so that the condition at the top right has the largest proportion. The conditions listed down the side of the table are ordered such that the condition at the top has the smallest proportion and the condition at the bottom has the largest proportion. The entries in the body of the tables represent the Z score obtained when pairs of proportions are compared. Each Z score represents the comparison between the proportion listed directly above and directly to the left of the entry. Where the Z scores are significant, as indicated by the asterisks, the difference is always in the direction of the proportion appearing at the top of the column. Table 1 summarizes the differences between the proportions of help given in each victim condition. Help, as used in this study, includes both flashing and stopping behavior. Its numerical value equals the sum of the frequencies of both flashing and stopping. The components of the proportions in Table 1, are the number of drivers who helped divided by the total number of drivers who passed. The white male-white female victim pair received the largest proportion of help which was significantly more than the help received by the single black female, the single white
female and the single black male. The black male-white female victim pair received a significantly greater proportion of help than either the black male or the white female alone. All victim conditions with the exception of the single white female received significantly more help than the single black male. 16 Table 2 summarizes the differences in the proportions of flashing behavior between victim conditions. In other words, it is Table 1 with stopping behavior taken out. The components of the proportions used in Table 2 are the number of drivers who flashed in any victim condition divided by the total number of drivers who passed in that condition. The white male-white female victim pair received the largest proportion of flashes while the white female victim received the smallest proportion of flashes. All victim conditions with the exceptions of the single black male received a significantly greater proportion of flashes than the white female victim. The single black male received a significantly smaller proportion of flashes than the single white male. The single black male also received a significantly smaller proportion of flashes when he was alone than when he was paired with the black female or the white male. Table 3 summarizes the differences in the proportions of stopping behavior between victim conditions. The components of the proportions used in Table 3 are the number of drivers who stopped to help in any victim condition
17 divided by the total number of drivers who passed in that condition. The black male-white female victim pair had the largest proportion of passing drivers who stopped. Since this study was conducted in the South, one might define this as "rescuing" behavior rather than helping behavior. None of the passing drivers stopped for the single black male, the white male-black female pair, or the white male-black male pair. Tables 4, 5, and 6 show the effect of victim race on helping behavior. Table 4 delineates the effect of victim race on overall helping behavior. The components of the proportions used in Table 4, are the number of drivers who flashed or stopped for any race category divided by the total number of drivers who passed. A white victim pair received a significantly greater proportion of help than any other victim race category except a black victim pair. All victim pairs regardless of racial composition received significantly more help than single black victims. Table 5 shows the influence of victim race on flashing behavior. The components of the proportions used in Table 5 are the number of drivers who flashed for any victim category divided by the total number of drivers who passed. All victim pairs regardless of racial composition received significantly more flashes than single victims of either race. Table 6 demonstrates the effect of victim race on stopping behavior. The components of the proportions used
B w WB BB ww Table 4. Comparisons of the Proportions of Helping for B (. 030) N -* p < .05 ** p (.01 *** p < .001 Victim Race Categories Victim Race w WB BB (. 044) (. 061) (. 062) N=l4 N=36 NlO ** NS 3.10 2.46 --NS NS --NS --ww (. 088) N=l2 *** 3.86 ** 2.75 2.07 NS --18
B w WB BB ww Table 5. Comparisons of the Proportions of Flashing for Victim Race Categories B (. 023) N=7 --* p < .05 ** p< .01 *** p( .001 Victim Race w WB (. 025) (. 045) N=8 N ... 27 ** NS 2.75 ** --2.50 --BB ww (. 056) (. 066) N=9 N ** *** 2.75 3.31 ** ** 2.58 3.15 NS NS -NS -19
B BB WB w ww Table 6. Comparisons of the Proportions of Stopping for Victim Race Categories B (. 006) N=-2 -* p < 05 ** p < .01 *** p < .001 BB (. 006) N=-1 NS --Victim Race WB (. 015) N 2.25 NS --w ww (. 019) (. 022) N N 2.16 2.28 NS NS NS NS --NS --20
in Table 6 are the number of drivers who stopped in any race category divided by the total number of drivers who passed. Single white victims and victim pairs with at least one white victim received significantly more stops than a single black victim. There was no difference in the proportion of stops between single and paired victims of the same race (i.e., WW-W N.S., BB-B N.S.). 21 Tables 7, 8, and 9 show the influence of victim sex on helping behavior. The components of the proportions used in these tables are the number of drivers who helped, flashed, and stopped (respectively) for any sex category divided by the total number of drivers who passed. Table 7 demonstrates the effect of victim sex on overall helping. Male-female victim pairs received significantly more help than either male or female victims alone. Paired female victims received no more help than single female victims; this was also true for male victim pairs versus single male victims. Table 8 shows the proportions of flashing elicited by each victim sex category. Two male victims received more flashes than a single female victim. Male-female victim pairs received significantly more flashes than the single victims of either sex, as well as a female victim pair. There was no difference in the proportions of flashes given to single versus paired victims of the same sex for either males or females.
M F FF MM MF Table 7. Comparisons of the Proportions of Helping for Victim Sex Categories M (. 038) N=-10 -* p <. 05 ** p<..Ol *** p < .001 Victim Sex F FF (. 040) (. 056) N=-13 N=-10 NS NS --NS --MM MF (. 062) (. 075) N=-8 N NS *** 3.36 NS ** 3.18 NS NS --NS --22
F M FF MF MM Table 8. Comparisons of the Proportions of Flashing for Victim Sex Categories F (.017) N=6 --* p < .05 ** p< .01 *** p < .001 Victim Sex M FF (. 033) (. 033) N=9 N=6 NS NS -NS --MF MM (. 054) (. 059) N=31 N=8 *** *** 4.63 3.50 2.10 NS 2.10 NS --NS --23
24 Table 9 shows the influence of victim sex on stopping behavior. Single female victims and victim pairs including at least one female received significantly more stops than either single or paired males. In order to determine if driver characteristics are related to helping behavior, the proportions of drivers in the general population (i.e., drivers who passed the disabled vehicle) were compared with the proportion of drivers in the helping population. These comparisons are presented in Table 10. As can be seen from Table 10, white male drivers were significantly over-represented in the helping population while white female drivers were significantly underrepresented in the helping population. There was no difference between the black male's representation in the general population and his representation in the helping population. There were too few black female drivers in the general population to make any meaningful comparisons. White male drivers are also over-represented in the flashing and stopping populations. Black male drivers are significantly under-represented in the flashing population and white female drivers are significantly under-represented in the stopping population. Diffusion of responsibility has been treated as a dichotomous variable; either a vehicle was alone on the highway when passing the victims or the vehicle was in close proximity to at least one other vehicle when passing
MM M MF F FF Table 9. Comparisons of the Proportions of Stopping for Victim Sex Categories MM (. 000) N=O --* p < 05 ** p < 01 *** p ( .001 Victim Sex M MF (. 003) (. 0 15) N=l N=9 NS ** 3.00 --2.40 --F FF (. 020) ( 022) N=7 N=4 *** ** 3.33 2.75 ** ** 2.83 2.71 NS NS --NS --25
Table 10. Comparison of Representation in General Driver Population with Representation in Helping Driver Population Driver Characteristics White Males White Females Black Males White Males White Females Black Males White Males White Females Black Males p < .05 ** p < .01 *** p < .001 Helping (N=81) General Population Helper Population 82% N=l220 92% N 12.5% N=l87 5% N 5.3% N=79 3% N=2 Flashes (N=60) 82% N=l220 93% N=56 12.5% N=l87 7% N 5.3% N-=79 0% N=O Stops (N) 82% Nl220 90% Nl9 12.5% N=-187 0% N=O 5.3% N=-79 10% N 26 Z score 6.66*** 2.67** NS 7.33*** NS 2 .12* 5.00*** 5.20*** NS
27 the victims. The cost of looking foolish was also treated dichotomously; either the driver was alone in his vehicle or he had at least one other person in the vehicle with him. Table 11 summarizes the analyses of helping behavior for both diffusion of responsibility and the cost of looking foolish. The entries in the table represent the proportion of drivers who helped relative to the number of drivers who passed. The top row of the table illustrates the proportion of help given by drivers who were alone in their vehicles when their vehicles were alone on the highway (column 1), when their vehicles were not alone on the highway (column 2), and their proportion of help when collapsed across vehicle conditions (column 3). The second row of the table shows the proportion of help given by drivers who were not alone in their vehicles across each vehicle condition and their proportion of help collapsed across both vehicle conditions. The bottom row of the table illustrates the proportion of help given in each vehicle condition collapsed across driver conditions. There was no difference between the proportion of help given by drivers of solitary vehicles and drivers of vehicles that were not alone on the highway. Column 3 shows a significant difference between the helping behavior of drivers who were alone in their vehicles and drivers who had at least one passenger. Alone drivers helped significantly more than drivers who were not alone.
28 Table 11. Proportions of Helping Behavior for Alone vs Not Alone Vehicles and for Alone vs Not Alone Drivers1 Vehicle Alone Not Alone Alone .062d .068ac .06Sb Driver 1 Not Alone .043c .040ad .042b .054 .oss Entries sharing the same subscript are significantly dif-ferent from each other. The proportions equal the number of drivers who helped divided by the number who passed. a, c, d p < OS b p < 01
DISCUSSION Despite the small amount of helping, some conclusions can be drawn regarding the effects of victim and driver characteristics on helping behavior. For those drivers who helped by flashing their lights, whether the victim was alone or not seemed to have more influence than the victim's race. Victim pairs, regardless of racial composition, received more flashes than single victims of either race. On the other hand, drivers stopped more often to help single white victims or victim pairs which included at least one white victim than either single or paired black victims. The fact that drivers flashed more for paired rather than single victims is a surprising finding. If this effect were due to the paired female victims, it would not be as interesting since Darley and Latane (1970) found that pairs of females were more likely to be helped than single females. The data on the influence of victim sex, however, shows that male-female and male-male pairs contributed most to this effect. Bryan and Test (1969) have suggested that potential helpers may want to avoid interpersonal contact with members 29
30 of a minority group. This is at least a possible explanation for the fact that white single or paired victims tended to have more drivers stop to help than either single or paired black victims. The fact that the mixed race victim pairs tended to elicit more stopping than either the single or paired black victims is completely due to the BMWF victim condition. As stated previously, this effect may be due to "rescuing" rather than helping motives. The fact that a pair of male victims received a larger proportion of flashes than any other victim sex category is, to say the least, surprising. The author can offer no explanation for this finding. It does, however, strongly suggest the necessity for replication of the study. Drivers who stopped to help did so according to Darley and (1970) findings on the influence of requestor sex. Paired females tended to have more drivers stop to help than single females, while paired males tended to have fewer drivers stop to help than single males. The results of this study suggest that the influence of the race and sex of the victim are different when the helper must come into physical or verbal contact with the victim than when the helper can act through a mechanical intermediary. Flashing behavior was found to be much less amenable to explanation, in terms of the findings of previous research where some form of physical or verbal contact was necessary to help, than stopping behavior.
31 The driver's race had a significant effect on helping behavior. White drivers regardless of sex were more likely to help than black drivers. There was also a clear effect of driver sex on helping. Males were much more likely to help than females. Analysis of the interaction of driver race and sex, indicate that white male drivers were more likely to help than any other drivers. In fact, the results mainly show the behavior of white male drivers. The data presented here suggest that at least in this situation Darley and concept of diffusion of responsibility was not operative. That is, no differences were found between the percentage of helping behavior displayed when the passing cars were alone or with others. The cost of looking foolish, as we have defined it, however, did seemingly affect helping behavior. When another person was present in the car (someone who could possibly chastize the driver for responding), helping behavior was reduced significantly. Besides providing a field investigation of some of the variables that might affect helping behavior, this study provided a test of the FLASH system. The major conclusion that can be drawn is that the FLASH system, although it provides a situation in which help can be given with minimal cost, does not seem to foster helping behavior. An independent "replication" conducted by a reporter for a local
32 newspaper produced a helping rate which was slightly lower than the percentages reported in this study.
REFERENCES Bryan, J. H., & Test, M. J. Models and helping: Naturalistic studies in aiding behavior. Journal of Personality and Social Psychology, 1967, 6, 400-4U7. Darley, J. M., & B. Bystander intervention in emergencies: Diffusion of responsibility. Journal and Social Psychology, 1968, 8, 377-Darley, J. M., & B. Norms and normative behavior: Field studies of social interdependence. In J. Macauley & L. Berkowitz (Eds.), Altruism and helping. New York: Academic Press, 1970. ---Feldman, R. E. Response to compatriot and foreigner who seek assistance. Journal of Personality and Social Psychology, 1968, 10(3), 2TI!-214. ---Latane, B., & Darley, J. M. Group inhibition of bystander intervention in emergencies. Journal of Personality and Social Psychology, 1968, 10, 215-27!. B., & Rodin, J. A lady in distress: Inhibiting effects of friends and strangers on bystander intervention. Journal of Experimental Social Psychology, 1969, 5, 189-202. Novak, D. W., & Lerner, M. J. Rejection as a consequence of perceived similarity. Journal of Personality and Social Psychology, 1968, 9(2, Pt. 1), 147-152. ---Piliavin, I. M., Rodin, J., & Piliavin, J. Good samaritanism: An underground phenomenon? Journal of Personality and Social Psychology, 1969, 13, W / 1spe, L. G., & Freshley, H. B. Race, sex, and sympathetic helping behavior: The broken bag caper. Journal of Personality and Social Psychology, 1971, 17(1), 65. 33
WFBF WF BF BMWF =N 148 83 67. 154 82 3 % 125 83.37. 118 7 9 .1% =FLASH 6 4% 1 67. 5 47. 4 3.3% WM 4 2. 7% 5 3.2% 1 8 % 5 4 .2"/., a HELP 10 6.7% 6 3.8% 6 4 .8% 9 7 6 % 24 13.5% 23 1 2.2% 14 9.37. 22 1 4.7% =FLASH 0 0 % 0 0% 0 0 % 1 4.57. WF =STOP 0 0 % 0 0% 0 0 % 0 0 % =HELP 0 0'7. 0 0% 0 0% 1 4.5'7. tl) "' =N 3 1.6% 1 0 5.3 % 1 0 6.6% 8 5 3'/, ..... =FLASH 0 0 % 0 0% 0 0% 0 0 % BM a STOP 0 0% 0 0 % 1 10% 0 0'7. HELP 0 0 % 0 0% 1 10% 0 0 % 2 1.1% 0 0 % 1 6 % 1 .6% -FLASH 0 0 % 0 0'7. 0 0 % 0 0'/, BF =STOP G 0'7, 0 0 % 0 0 % 0 0'7. =HELP--C 0'7. 0 0% 0 0 % 0 0'/, =N 177 :J9. 8 % 187 99.8 % 1 5 0 99.8% 149 99.7% =FLASH 6 3 .3% 1 .57. 5 3.3 % 5 3.3 % TOTAL =STOP 4 2.2% 5 2 6 % 2 1.3 % 5 3.3% HELP 1 0 5.6% 6 3.2 % 7 4.6 % 1 0 6.7% VICTIMS WM WMBM BMBF WMBF 100 80% 113 83 7 % 128 80 57. 118 92 9 % 7 n 8 n 8 6 2 % 7 5.9% 1 1 % 0 0 % 1 .7% 0 0% 8 8 % 8 7 % 9 7 % 7 5 9'/, 18 14.4% 1 6 11. 8 % 17 10.6% 8 6.2% 0 0'/, 0 0 % 1 5.8% 1 12.5% 0 0% 0 0 % 0 0 % 0 0'/, 0 0'/, 0 0'/, 1 5.8% 1 12.5% 7 5.6% 6 4.4% 14 8 8 % 1 .7% 0 0'/, 0 0 % 0 ()% 0 0 % 0 0% 0 0 % 0 0 % 0 0 % 0 0'7. 0 0 % 0 0 % 0 0% 0 0 % 0 0 % 0 0 % 0 0'7. 0 0% 0 0 % 0 0 % 0 0% 0 0 % 0 0'7. 0 0% 0 0 % 0 0% 0 0 % 0 0'/, 0 0'7. 125 100'/, 1 3 5 99.9% 1 59 99.9% 127 99.8 % 7 5 6'/, 8 5 9'/o 9 5 6 % 8 6.2% 1 .8% 0 0 % 1 .6% 0 0% 8 6 .4% 8 5.9 % 10 6 .2% 8 6.2% BM WMWF 117 80 1 % 99 73.37. 2 1.7% 8 8% 0 0% 2 2 % 2 1. 7 % 10 10.17. 19 13% 26 1 9.2% 0 0 % 1 3.8% 0 0 % 0 0% 0 0 % 1 3 8 % 1 0 6.8% 10 7 4 % 0 0 % 0 0% 0 0% 1 10% 0 0% 1 10% 0 0'7. 0 0% 0 0 % 0 0 % 0 0'7. 0 O'lo 0 0 '/, 0 0 % 1 46 99.9% 135 9 9.9% 2 1 3'7. 9 6 6 % 0 0 % 3 2.2% 2 1.3% 12 8.8% TOTAL 1220 56 19 75 187 4 0 4 79 0 2 2 4 0 0 0 1490 60 2 1 8 1 81.8% 4 5 % 1.5% 6.1% 12.5% 2.1% 07. 2.1% 5 3 % 0 % 2.5% 2.5% .2% 0 % 0% 0% 99.8% 4 % 1.4% 5.4% w VI