EFFECTS ON THE HOSTILITY OF SPECTATORS
OF VIEWING AGGRESSIVE SPORTS

Robert L Arms, Gordon W. Russell, and Mark L Sandilands


The effects on spectator hostility of viewing aggressive athletic contests were investigated using three 
diverse measures of hostility in a replication of the widely cited Goldstein and Arms (1971) Army-Navy 
football study. Male and female subjects were exposed to either stylized aggression (professional 
wrestling), realistic aggression (ice hockey), or a competitive but nonaggressive control event (swimming) 
in a before-after design While the three measures of hostility yielded somewhat different results for the 
three events, general support was found for the earlier finding of increased spectator hostility as a result of 
observing aggression. Whereas hostility was shown to increase at wrestling and hockey, such increases 
did not occur at the swimming competition. Other aspects of mood change among spectators were also 
investigated. There was a blunting of the quality of interpersonal relations at the three events.

 Hostile outbursts occurring at a number of spectator sports have contributed to a general international 
concern escalating levels of violence. While that concern has primarily dealt with illegal aggression on the 
field of play, attention has also been focused on those relatively rare but newsworthy occasions in which 
violence has erupted among fans viewing sports contests. For example, soccer has been plagued by rampant 
hooliganism both in Britain and on the continent (e.g., Scottish Education Department, 1977). Elsewhere, 
soccer fans supporting their team in a losing cause in Guatemala recently attacked fans of the winning side 
with machetes, hacking five people to death (San Francisco Chronicle, 1977). Smith's (1978) archival study 
of collective violence among ice hockey fans found that approximately 74% of hostile outbursts were 
preceded by extraordinary displays of violence among the players.

At the level of public beliefs, the notion of a "safety valve" or catharsis enjoys widespread currency. This 
view as it applies to spectators viewing aggression on the field of play, predicts that such displays serve to 
reduce hostility and physiological arousal in the viewer. This cathartic viewpoint is bolstered by the 
influential writing of Freud, Lorenz, Tinbergen and their populizers. In general the position has received 
little encouragement from the experimental literature (e.g., Geen and Quanty, 1977).  Compelling as the 
notion of catharsis may be, vicarious participation in aggressive displays serves with few exceptions (e.g., 
Doob and Wood, 1972) to enhance, rather than diminish, hostility in the onlooker.

Generalizations from the experimental laboratory to real-life settings inevitably contain an element of risk. 
However, when the results of investigations conducted in the more complex naturalistic setting agree with 
lab findings, then confidence in their applicability is greatly enhanced. Unfortunately, complementary 
investigations of spectatorship reported in the literature have produced conflicting results. Kingsmore (1970) 
found professional wrestling fans showed less extrapunitive and intrapunitive aggression on the Thematic 
Apperception Test (TAT) and less self-reported aggression following the matches. Also, basketball fans 
showed less extrapunitive aggression after a game. Turner (1970) reported college males showed increases 
in TAT aggression at basketball and football, though not at amateur wrestling.

	The annual Army-Navy football game provided the setting for a field study by Goldstein and Arms (1971) 
whose design pitted the cathartic against the enhancement position. A sample of males leaving the stadium 
after the game score significantly higher on the Buss-Durkee hostility scale (Buss, 1961 ) than an equivalent 
sample entering the stadium before the game. Pre and postevent measures for males at an equally 
competitive, but non aggressive control event (gymnastics) did not differ. The major finding of an overall 
increase in hostility of both Army (winners) and Navy (losers) fans strongly supported a general 
disinhibition position (Bandura, 1973) which predicts that the observation of aggression leads to a lessening 
in the strength of inhibitions against expressing hostility.

Inevitably, a number of plausible rival interpretations remained to explain the heightened postevent 
hostility. The authors suggested a selection bias whereby football fans might be presumed to be more 
volatile than those attracted to gymnastics competition. Other potential sources of bias include individual 
versus team competition and different norms governing expressive behavior at the events. Also cited 
(Goldstein and Arms, 1971) were differences in the density, numbers, and activity levels between the 
football and gymnastics fans. Goldstein (1976) has suggested that the student interviewers may have gained 
in confidence in the pregame stage, and unwittingly, approached and been successful in interviewing more 
aggressive-looking fans after the game. Furthermore, the differential consumption of alcohol at the two 
events cannot be discounted as a contributing factor, particularly if the football fans felt in any way 
"threatened" (Taylor et. al., 1976). Finally, Mann (1974) suggested the overall increase in hostility may 
have arisen from all fans experiencing a dull, lopsided game (Army 27, Navy 0).

	The present investigation was designed as a systematic replication (Sidman, 1960) of the Goldstein and 
Arms (1971) study that would also test the merits of rival explanations advanced to account for their 
results. Departures from the original design were the inclusion of female spectators, the substitution of ice 
hockey for American football, and a provincial team swimming competition for gymnastics.

A special case for cathartic effects due to exposure to stylized aggression has been persuasively argued 
by Noble (1975). Cartoons, roller derby, or professional wrestling-displays in which interpersonal 
mayhem may be seen as fictional or a spoof-would qualify as stylized aggression. Thus, professional 
wrestling was chosen to represent stylized aggression, and hockey, realistic aggression.

Procedure

Students (N = 127 females; N = 87 males) were recruited from an introductory psychology subject pool 
at the University of Lethbridge. Subjects received a 4% research participation bonus, free admission, and 
an offer of return transportation to their sporting event. The study was presented as an investigation of 
spectator's impressions of sports events. Assignment to the events and to the pre and postevent conditions 
was random within the female and male categories. Departures from the random assignment procedure to 
events proved necessary in 5% of the cases (exam conflicts). 

Preevent subjects arrived 30 minutes before the event, were given their tickets, and were then escorted 
to a spare dressing room and were simply asked to complete a set of measures indicating their feelings at 
that time. Before taking their seats (dispersed) these subjects were asked not to discuss the measures with 
others. Postevent subjects arrived 15 minutes before game time, were given their tickets, and were asked to 
remain momentarily seated following the event. At the conclusion of the event, they were escorted to the 
same room to complete the hostility scales.


	To ensure a more complete mapping of the multidimensional domain of hostility, the following 
measures  of the dependent variable were administered : (1) the Buss-Durkee (Buss, 1961) subscales of 
indirect hostility, resentment, and irritability summed by Goldstein and Arms (1971) to provide an overall 
index; the aggression scale of the Nowlis (1965) Mood Adjective Check List (MACL); and a punitive 
measure (Goldstein et al., 1975) based on the prison sentences subjects would assign to individuals 
convicted of serious crimes.

	The complete MACL, including the ad hoc adjectives "sexy" and "aroused:" were administered to provide 
the general pattern of mood changes at each event and as background against which any resulting changes 
in hostility could be better understood. Earlier comment (Stone, 1973) and discussions with pilot subjects 
suggested that a sexual component may be inherent in contests involving either heavily padded athletes or 
contestants in swimsuits. An index of involvement was calculated for each subject by summing self-report 
ratings of frequency of TV viewing, annual attendance, and general interest in their assigned sport. Finally, 
subjects assigned to the wrestling match were asked: "What percentage of professional wrestling action do 
you think is 'faked' or just acting?"


Results

Differences in pre- and postevent levels of spectator hostility were tested by t-tests (two-tailed). As 
presented in  Table 1, significant increases occurred at wrestling and hockey though no changes were 
observed at the control event. At wrestling, the MACL aggression measure proved sensitive to stylized 
aggression (p < 0.01), the combined pre to  postevent increase originating principally with females. The 
overall increase in hostility at the hockey was significant (p < .05) using the punishment index and  
marginally so (p < .06) with the Buss-Durkee scale. The distributions of involvement scores, and 
authenticity ratings by subjects assigned to wrestling, were too strongly skewed to permit meaningful 
analyses of these variable.

Changes in other dimensions of mood assessed by the MACL were tested by t-test While anxiety and 
fatigue were unaffected by experiences at the events, all events  produced a significant decline in feelings of 
Social Affection. Surgency declined significantly at hockey and swimming while feelings of sexiness 
dropped at the swim meet whereas arousal evidenced no change.

Discussion

The coordination and movement of subjects at the venues was accomplished smoothly and without 
incident, mainly the result of improved procedures developed during the pilot stage. The local Western 
Canada Hockey League club (Tier 1, Junior A) badly outclassed the visiting Flin Flon Bombers, winning 
11-1 in a lackluster contest. In addition to being a rout, the game was devoid of major fights, with only the 
occasional flash of playmaking skill in evidence. The good guys also won the main tag-team event on the 
professional wrestling card. Whereas students in the earlier pilot study  remained aloof from the ring action, 
the present subjects entered wholeheartedly into the spirit of the evening, booing, cheering, and trading 
insults with the villains. The swim meet was fiercely competitive with spectators easily as vociferous as 
those at the wrestling event.

Posttest hostility means were, almost without exception, greater than the mean pretest scores. Although 
aspects of the cathartic and enhancement viewpoints were conjoined in the present design, support was 
forthcoming for only the latter position. Subjects exposed to stylized  aggression (wrestling) showed a 
significant increase on the MACL aggression scale, a combined effect arising principally from the female 
data. Furthermore, the choice of professional wrestling to represent stylized aggression was borne out 
insofar as subjects overwhelmingly saw it as a sham. Subjects at the hockey game showed increased 
hostility on the Buss-Durkee and punishment scales, the males in particular contributing to the latter 
increase. Spectators at the control event showed no significant changes in hostility on any of the three 
measures. The present results are consistent with the earlier Goldstein and Arms (1971) findings and 
provide further support for a general disinhibition position (e.g., Bandura, 1973).

Table 1

Mean Pre and Posttest Hostility Scores by Sex at Three Events
_________________________________________________________

		MACL (Aggression)		Buss-Durkee	Punishment	
_________________________________________________________
Event	N	Pre	N	Post	t	Pre	Post	t	Pre	Post	t

Wrestling
Females	32	1.72	22	5.59	3.61	13.39	14.55	.81	22.94	23.71	 1.07
Males		16	3.06	20	3.80	0.60	12.56	12.98	.24	18.72	34.85	1.72
Combined	48	2.17	42	4.74	3.18	13.11	13.80	.64	26.20	29.01	.55

Hockey
Females	24	2.08	23	3.52	1.02	12.23	14.50	1.60	18.52	22.74	.97
Males		17	2.29	16	2.50	.22	14.06	15.81	1.06	12.41	24.19	3.03
Combined	41	2.17	39	3.10	1.02	12.99	15.04	1.90	15.99	23.33	2.45

Swimming (Control)
Females	12	3.08	14	1.36	1.24	13.00	14.86	1.05	25.17	33.82	1.05
Males		8	3.25	10	3.40	.13	11.38	13.40	.85	26.35	15.60	1.77
Combined	20	3.15	24	2.21	.98	12.35	14.25	1.34	25.60	26.23	.11


Although hostility increased through exposure to displays of both stylized and realistic aggression, such 
increases were not registered consistently on all three measures. This makes it apparent that reliance on a 
single measure of aggression could be misleading. Nevertheless, each measure was shown to be sensitive 
on at least one occasion leading to a conclusion consistent with, but less decisive than that reached by 
Goldstein and Arms (1971). That is to say, the observation of aggression on the field of play leads to an 
increase in hostility on the part of spectators though displays of stylized aggression may increase one type 
of hostility in the viewer and realistic aggression, another.

The present results support Turner (1970) and, in extending the generality of the earlier Goldstein and Arms 
(1971) results, largely negate the rival explanations advanced to date. Increases in spectator hostility have 
been demonstrated to occur in a different culture with two additional aggressive sports, one a stylized 
display. Furthermore, the increases in hostility found among avid Army and  Navy fans seem also to be true 
of females and student spectators without strong sporting interests or team loyalty. The likelihood of a 
selection factor operating to attract less volatile spectators to their control event (gymnastics) or the 
unwitting selection of less hostile preevent males was negated by the random assignment of subject to 
events and to pre- and posttest conditions. With the number and density of spectators at wrestling and the 
control event equal in the present study and alcohol not present at the events, it seems unlikely that 
differences could have arisen from these factors. The hockey game was easily as lackluster as Army's 
runaway victory over Navy. However, the final wrestling bout was in doubt until the dying minutes of the 
match when the forces of good overcame the forces of evil. Mann's (1974) suggestion that increased 
hostility on the part of the Goldstein and Arms (1971) subjects arose from the dull, one-sided nature of the 
contest thereby seem less plausible. Lastly, while the football fans may have been thwarted in their efforts 
to beat the other 100,000 spectators onto a crowded freeway at the conclusion of the game, a similar 
challenge did not confront the Lethbridge (Pop. 50,000) subjects.

In addition to hostility, changes along other dimensions of mood were explored using the remaining 
MACL measures. There appeared to be a general deterioration in the quality of interpersonal relations. The 
most pronounced and consistent decrement occurred in Social Affection, a dimension characterized by the 
adjectives "affectionate," "forgiving", "kindly," and "warmhearted." It is particularly noteworthy that 
diminished social affection occurred at the nonaggressive control event as well as at wrestling and hockey. 
Thus, the reduction in social affection cannot be attributed to aggressive content but rather arises from other 
features of the spectator experience. Furthermore, postevent scores on the Surgency factor ("carefree",  
"playful", and "witty") were lower for males viewing the hockey game and for females attending the 
swimming competition. When changes in mood have occurred, such changes have consistently been toward 
a more negative emotional state. The present results call into question an assumption that sports events are 
necessarily rich social occasions where goodwill and warm interpersonal relations are fostered (Mehrabian, 
1976:284).