1
r

in? M ,.‘
1. H } g
L 1M M
’ ‘1“ 1
‘l I .L

w
I

l \
\

f

109
572

THS THE EFFECTS 05 sex or SUBJECT, ssx 0;:
EXPERiMENTER, AUDIENCE DURlNG WARM
up, AND AUDIENCE DURING mama.
EZATIQN om THE LEARMNG AND
pmsoammcg as VERBAL TASKS

 

MA.
MISCMGAN STAN 'UNWE-RSETY

Ems-1 mm was
5-970

  
     

  

LIFRARY

Michigan State
University

THES‘S

ABSTRACT
THE EFFECTS OF SEX OF SUBJECT, SEX OF EXPERIMENTER,
AUDIENCE DURING WARM UP, AND AUDIENCE DURING
MEMORIZATION ON THE LEARNING AND PERFORMANCE
OF VERBAL TASKS '
BY

Kenneth John Good

In his theory of social facilitation, Zajonc predicts
that the presence of an audience hinders learning and
facilitates the performance of learned tasks. He suggests
that the presence of an audience increases 'drive' which,
in turn, facilitates the performance of well-learned, dom—
inant responses.

The purpose of this research was twofold: first, to
test the drive theory predictions of Zajonc against the
alternate explanation that previous results might have
been due to a contrast effect imposed by the temporal se-
quence of events; second, to examine the effect on audience
phenomena of variables relevant to social role theories.

The major hypothesis (I) predicted that the temporal
order of events would be more of an influence on behavior
than the type of behavior (i.e. learning or performing).
It was suggested that if audience conditions were changed
from the first of two tasks to the second, subjects would
infer the relative importance of the second task by com-
paring audience Conditions during the first task with

those during the second task. It was felt that subjects

2) Kenneth John Good
would when adjust their motivation toward the second task
to a level appropriate to their perception of its relative
importance.

Specifically, it was predicted that subjects would ex-
cel on the second of two tasks if they had been alone dur-
ing the first task and before the audience during the sec-
ond task. This hypothesis was supported but only with the
subjects of male experimenters. However, the experiment

was designed such that the subjects first performed one

 

task and then learned another. Since subjects learned more
before the audience than alone, these results are in direct
conflict with Zajonc's theory of social facilitation. The
results for the subjects of female experimenters did not
support either the present hypothesis of Zajonc's theory.

The second hypothesis (II) predicted that the situa-
tional variable, sex of experimenter, and the personal var-
iable, sex of subject, would influence audience effects.
Audience effects were discussed in terms of the compatibil-
ity of a person's behavior with the behaviors defined as
appropriate for that role. It was suggested that there are
at least four factors affecting this compatibility---type
of behavior, kind of task, personal variables, and situa-
tional cues. Zajonc has only investigated the effects of
two types of behavior, learning and performing.

Hypothesis II(a) predicted that subjects might be
affected differently by male and female experimenter-ob-

servers. This hypothesis was supported by two of the three

Kenneth John Good

response measures. Subjects performed better for female
experimenters. However, audience effects were only found
with the subjects of male experimenters.

Hypothesis II(b) predicted that male and female sub-
jects might be affected differently by the presence of an
audience. This hypothesis was also supported by two of the
three response measures. During the first task males per-
formed better than females if alone and there was a trend
for females to perform better than males in the audience
condition. No difference was found between the audience
effect for males and females during the second task.

HYpothesis II(c) predicted that there might be an
interaction of sex of subject and sex of experimenter-ob-
server. This was found only with one of the three response
measures, intrusions (i.e. incorrect responses). However,
the interaction with this response measure was highly sig-
nificant with subjects making approximately two and one
half times as many intrusions before an observer of the

same sex as subjects before an observer of the opposite sex.

Approved?

James L. Phillips, Chairman
Lawrence A. Messga Member

Gordon‘Wood, Member

THE EFFECTS OF SEX OF SUBJECT, SEx OF EXPERIMENTEB,
AUDIENCE DURING NAEN UP, AND AUDIENCE DUEINC
MEMORIZATION ON THE LEARNING AND PERFORMANCE

OF VERBAL TASKS

By

Kenneth John Good

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF ARTS

Department of Psychology

1970

ACKNOWLEDGMENTS

The author wishes to express his sincere appreciation to:

Dr. James L. Phillips for his encouragement, suggestions, and
criticisms throughout all aspects of the preparation of
this thesis,

Dr. Lawrence A. Messe'for his methodological, theoretical, and
stylistic criticisms and suggestions,

Dr. Gordon Wood for his assistance in the selection of verbal
materials,

Mary Ann Belinski, Donna McKittrick, and Fred Belinsky for
being conscientious experimenters,

Chuck Kenoyer for his assistance with electrical apparatus,

my wife, Linda, for her encouragement and assistance in pre-
paring this thesis, and

my daughter, Shannon, for miscellaneous distractions.

ii

TABLE OF CONTENTS

Page

LIST OF TABLES .... .................. ' ................... . v
INTRODUCTION ... ................................. ....... 1
METHOD .... ...... ................ ............ .......... . 19
Subjects ........................................,.. l9
Experimenters ...................................... 19
Apparatus ........................,................. 19
Manipulations ...................................... 20
Procedure and Experimental Design .................. 20
RESULTS ....... ..... ..................................... 26
Results Relevant to Hypothesis I ,,,,,,,.,...,..,.., 26
Results Relevant to Hypothesis II ,,,,,,,,,,,,,,,,,, 30
....,..... 32
Results Relevant to Hypothesis II(b) .,,,,,,,,, 38

El

Results Relevant to Hypothesis II(a)
Results Relevant_to Hypothesis II(c) ,,,,,,,,,,
DISCUSSION .... ......,.. .......... .. ..... ..,.. .. .. 45
OVERVIEW ................................................ 53
LIST OF REFERENCES ...................................... 55
APPENDIX A- EXPERIMENTAL MATERIALS ....... ..... . ..... .... 57
Transcript of Recorded Instructions ................ 57
Preliminary Instructions ...................... 57

warm-up TaSk InStruCtionS 000000000000000000000 57

iii

Page

memorization Task Instructions................ 58
Recall Instructions............ ........ ....... 59
English Nouns for Memory Task...................... 61
Postexperimental Questionnaire..................... 62

APPENDIX B- TABLES OF CELL MEANS,..,.,......,.,,,.,,,,,, 63

iv

LIST OF TABLES

Human audience and co-action studies mentioned
either in Zajonc (1965) or in Zajonc (1966).........

Analysis of variance for English noun recall........

Analysis of variance for the simple effects of
audience conditions in the interaction of sex of
experimenter and audience conditions during warm-up
(A x C) for the response measure, English noun

recallOOOOOOOOOOOOOO0.0.0.0.000.........OOOOOOOOOOCO

Cell means for the interaction of sex of experi-
menter and audience conditions during warm-up
(A x C) for the response measure, English noun

recall-......OOOOOOO......OOOOOOOO......OOOOOOOOOOOOO

Analysis of variance for the simple effects of
audience conditions in the interaction of sex of
experimenter and audience conditions during
learning (A x D) for the response measure, English
noun recall.........................................

Cell means for the interaction of sex of experi-
menter and audience conditions during learning
(A x D) for the response measure, English noun

recall...‘......COOOOOOOOO......OOOOOOOOOOO ..... 0...

Analysis of variance for the simple effects of

sex of experimenter in the interaction of sex

of experimenter_and audience conditions during
warm-up (A x C) for the response measure, English
noun recall.........................................

Analysis of variance for the simple effects of

sex of experimenter in the interaction of sex of
experimenter and audience conditions during
learning (A x D) for the response measure, English
noun recall.........................................

Analysis of variance for B-word formation.. ..... ....i

V

Page

1M
28

29

29

31

31

33

33
35

Table Page

10. Analysis of variance for the simple effects of
sex of experimenter in the interaction of sex
of experimenter and audience conditions during
warm-up (A x C) for the response measure, B-word
formation......... ..... ....................... ...... 36

11. Cell means for the interaction of sex of experi—
menter and audience conditions during warm-up
(A x C) for the response measure, Beword formation.. 36

12. Analysis of variance for intrusions....... ....... ... 37

13. Analysis of variance for the simple effects of sex
of subject in the interaction of sex of subject
and audience conditions during warm-up (B x C)
for the response measure, B-word formation.......... 39

14. Cell means for the interaction of sex of subject
and audience conditions during warm-up (B x C)
for the response measure, B-word formation.......... 39

15. Analysis of variance for the simple effects of sex
of subject in the interaction of sex of_subject
and audience conditions during warm-up (B x C)
for the response measure, intrusions................ 40

16. Cell means for the interaction of sex of subject
and audience conditions during warm-up (B x C)
for the response measure, intrusions................ 40

17. Analysis of variance for the simple effects of sex
of experimenter in the interaction of sex of ex-_
perimenter and sex of subject (A x B) for the re-
sponse measure, intrusions.......................... 43

18. Analysis of variance for-the simple effects of sex,

of subject in the interaction of_sex of experimenter

and sex of subject (A x B) for the response measure,

intruSiOnSOOO......0.00.00.00.00...OOOOOOOOOOO0.0.... 1+3
19. Cell means for the interaction of sex of experi-

menter and sex of subject (A x B) for the re-

sponse measure, intrusions................ ........ .. an
20. Cell means for the response measure, B-words........ 63

21. Cell means for the response meaSure, English noun
recall...........O..........O......r............999’0 64

22. Cell means for the response measure, intrusions..... 65

vi

INTRODUCTION

Social psychology is based on the axiom that people
can be studied not only as units but as social beings who
interact and influence each otheris behavior. A part of
the discipline of social psychology is devoted to studying
how people behave in aggregates or groups in which personal
or situational variables are manipulated or are known to
exist.

Within this general area, much of the emphasis of re-
search has been directed toward the understanding of how
the behavior of other persons affects the behavior of a sub-
ject. However, this paper is concerned with what has been
called the I'audience effect”. That is, how does just being
observed by other persons influence the subject's behavior?
Before considering the goals of the present study, a brief.
summary of research relevant to the study of audience ef-
fects is presented below.

According to Jones and Gerard (1967), the first syste—
matic attempts to study the effect of the presence of others
on individual performance were carried out by educators be-
tween 1898 and l9lh. The interest of these educators was
in empirically testing an applied questioni could school
work be better performed by a student working alone or in
the presence of others? The first psychologist to attempt

l

to answer this question was Floyd Allport (1924). Fol-
lowing a series of experiments from 1916 through 1919, he
concluded that the presence of others increases the quan-
tity and vigor of responses at the expense of their intel-
lectual quality.

In 1930 a student of Allport's renewed interest in
audience effects. Dashiell (1930) demonstrated facilita-
tion of simple multiplication and serial word association
tasks when the subjects were observed by an audience. How-
ever, Pessin and Husband (1933) conducted an experiment
which demonstrated that an audience hindered the learning
of finger mazes. It is not obvious how one could predict
these results from Allport's explanation.

Pessin (1933) gave 60 subjects three lists of seven
nonsense syllables and found that during the learning phase
(i.e. until each subject mastered the list) the subjects
who learned before an audience took longer to learn the
list and made more mistakes than a group that learned the
list alone. After training the subjects to a criterion, he
excused them for a few days. After five familiarization
trials, he found that the group performing before the audi-
ence relearned faster, and thus had a higher “saving score".

Zajonc (1965,1966) has theorized that an audience ”in;
hibits learningI and I'facilitates performance'. He formu-
lated an explanation of the phenomenon which draws upon two
earlier findings. First, Mason and Brady (1964) have demon-

strated that the presence of others is closely related to

increased adreno-cortical activity which in turn is be-
lieved to be associated with "drive". Secondly, Spence
(1956) theorized that increased 'drive' facilitated the
performance of dominant responses.

Thus Zajonc deduced that the presence of an audience
increases drive and as such facilitates the performance of
learned responses. The audience would, however, inhibit
learning of new responses since irrelevant responses which
were dominant before the training would be facilitated. In
this way, Zajonc has reconciled many paradoxical results of
earlier experiments. In support of this position Cottrell,
Hack, Sekerak, and Rittle (1968) have recently shown social
facilitation of verbal responses which were made relatively
dominant by varying the number of learning trials for each
item.

In concluding his review of experiments dealing with
audience and co-action1 effects, Zajonc (1965) advised:

'If one were to draw one practical suggestion from
the review of the social-facilitation effects which are
summarized in this article he would advise the student
to study all alone, preferably in an isolated cubical,
and to arrange to take his examinations in the company
of many other students, on stage and in the presence
of a large audience. The results of his examination

would be beyond his wildest expectations, provided, of
course .he had learned his material quite thoroughly.“

(p. 274)

 

1Co-action studies do not use an audience per se but
involve two or more subjects working on a task in the pre-
sence of each other.

The existence of a relationship between learning, per-
forming, and audience effects is not questioned in this
study. However, what is at issue is whether this relation-
ship accounts for a sufficient amount of the variance in
the subject's behavior to warrant the general advice that
Zajonc (1965) has offered. It is suggested here that if
variables relevant to other theories in social psychology
are manipulated, the learning-performing distinction might
prove to be relatively unimportant. In this study manipu-
lations were made to examine relationships which follow from
two traditions in psychology---contrast or comparison
theories and role theories.

In psychology contrast or comparison theories have been
concerned with the effects of a stimulus situation in rela-
tion to preceding or contemporaneous stimuli. One compari-
son theory in social psychology is that of Thibaut and
Kelley (1959). They have theorized that individuals set
comparison levels, based on previous outcomes, which they
use to evaluate their present situation. For example, in
an absolute sense a child would be rewarded if he were
given a dollar per week as an allowance. However, if he
had been accustomed to receiving five dollars per week,
the comparison might serve as a cue for resentment rather
than gratitude.

In the present context, the comparison effect of
changing audience conditions will be considered in a hypo-

thetical two part experiment. If one group of subjects

remained before the audience during both parts of the ex-
periment and another group remained alone throughout the'
experiment, differences in performance on the second part
could be explained as being due to audience effects.

However, if still another group were before the audi-
ence during the first task and were left alone during the
second task, it could be expected that performance on the
second task would be affected by the change per se. Thus
the audience would provide a comparison level and the change
to the alone condition might provide a cue for a change in
behavior. In the present situation, the subject might in-
fer that what he was doing in the second task is not as
important to the experimenter. This is because people pre-
sumably attend to that which they consider important. As
such, subjects might lower their motivation toward the task
to a level which they consider appropriate for the perceived
importance of the task.

If a final group were added in which the subjects per-
formed the first part alone and the second part before the
audience, performance on the second task might again be in-
fluenced by the change. Here relative to the alone condi-
tion, the audience might serve as a cue that the second
task was more important. Thus, subjects might increase
their motivation to a level appropriate to their perceived
importance of the task.

To summarize the contrast effect of changing audience

cues, one would expect the best performance on the second

task if the subjects performed the first task alone and the
second before the audience since it is the only condition
that would result in an increment of perceived relative im-
portance of the second task. Likewise, one would expect
the worst performance on the second task if the subjects
performed the first task before the audience and the second
task alone since it is the only condition that would result
in a decrement in the perceived importance of the second
task.

In his advice quoted above, Zajonc (1965) suggests
that a person will perform best if he first learns alone
and then performs before an audience. However, this asser—
tion confounds the behavioral sequence (i.e. learning then
performing) with the temporal sequence (i.e. the first task
alone and the second task before the audience). Therefore,
a survey was made of the studies mentioned either in Zajonc
(1965) or Zajonc (1966) to determine the extent to which
these experiments studied or controlled for this confounding.

First, it was found that in many of the studies there
was only one part. That is, the subjects learned or per-
formed either alone or before the audience. Thus there was
no opportunity for a comparison effect. However, in the re—
maining studies, the behavioral sequence was always con-
founded with the temporal sequence. That is, learning was
always the first task and performing was always the second
task. Thus, if the effect of the temporal sequence is as

discussed above, this would tend to bias results in the

direction of supporting the learning-performing distinction.
In the present experiment the behavioral sequence was

reversed. As such, the subjects first performed one task

 

either alone or before the audience and then learned an-
other task either alone or before the audience. Zajonc's
theory predicts that the audience facilitates performance
but hinders learning.

The present comparison interpretation predicts better
performance on the second of two tasks if the subjects were
alone during the first task and before the audience during
the second. Since in this experiment the performing task
preceded the learning task, contrary predictions are made
by Zajonc's theory and this comparison hypothesis.

Turning from contrast and comparison notions, the
relevance of role theories was considered. It was noted
that Zajonc's theory makes general predictions about the
effects of audiences on individuals. However, such gener-
ality ignores the possibility that differences in individu-
als and audiences may greatly influence audience phenomena.

Goffman (1959) has developed a dramaturgical conceptu-
alization of interpersonal behavior which relates the main-
tenance of social roles with audience phenomena. His basic
assumption is that individuals give performances in the pre-
sence of others as a means of presenting themselves in a
certain way for a certain class of observers. He uses the
term "performance" to refer to "...all the activity of an

individual which occurs during a period marked by his

continuous presence before a particular set of observers
and which has some influence on the observers." (p. 22)
He also notes, "The pre-established pattern of action
which is unfolded during a performance and which may be
presented or played through on other occasions may be
called a 'part' or 'routine.'" (p. 16) He goes on to de-
fine a social role in terms of the parts or routines "...
presented by the performer on a series of occasions to the
same kind of audience or to an audience of the same persons.”
(p. 16) Thus Goffman (1959) suggests that an individual's
behavior is influenced by the social role that he adopts
in the presence of an audience and that the role, in turn,
is influenced by characteristics of the audience.

Although Goffman (1959) did not elaborate on the mech-
anics of role maintenance, Jones and Gerard (1967) have
made the following observations:

"Roles are shared norms concerning the behavior of
certain persons in certain settings. Men confront
different social expectations-are constrained by a
different set of norms-than women. Adults differ
from children, leaders from followers, waiters from

customers, in the norms relevant to their roles.“

(p. 177)

Thus, Jones and Gerard suggest that when a person adopts a
particular role, he must limit his behavior to a subset of
behaviors which conform to the social expectations for that
role.

It is suggested here that audience effects be consid-
ered as a special case of role conflict. That is, audience

effects should occur when the person's behavior is in

conflict with the behaviors which are defined as appropri—
ate for the role that he assumes before the audience. For
instance, it might be expected that a person who curses in
private and among close friends would inhibit cursing while
giving a speech if he considered cursing to be an inappro-
priate behavior for the role of speaker. Moreover, it
would be expected that although the individual might swear
when he is among peers, he might inhibit this behavior
around his parents. That is, the behavior could be appro-
priate for the role of friend but inappropriate for the role
of son. Thus the compatibility of specified behaviors with
the behaviors that the person defines as appropriate for
the role which he assumes before the audience must be exa-
mined. As such, there seems to be at least four factors
that would influence this compatibility---the type of be-
havior, the nature of the task, characteristics of the in-
dividual subjects, and the kind of cues provided by the
stimulus situation.

Zajonc has stated that the audience hinders learning
and facilitates performance. This relationship is compati-
ble with the present Formulation. This could be interpreted
to mean that one type of behavior, performing learned tasks,
is compatible with the role of performer while another be-
havior, learning new tasks, is not. On face value, this
seems to be a plausible relationship since the performer
role involves the emission of overt behavior which can be

observed and evaluated. The performance of a learned task

10

also involves overt behavior while learning is essentially
a covert behavior. As mentioned above, the existence of a
relationship between learning, performing, and audience ef-
fects is not questioned in this study. However, it was
noted that Zajonc's predictions consider only one variable,
type of behavior. It is, therefore, suggested that if task,
personal, and situational variables are also manipulated,
the learning-performing distinction might prove to be rela-
tively unimportant in predicting the subjects actual behav-
ior.

Looking first at the nature of the task, it can be
seen that Zajonc might have overgeneralized his findings.
Many well-learned behaviors become dominant but are inhib-
ited in the presence of an audience. People may become
sexually aroused at a party, dance, or theater, but most
postpone sexual behavior until there are conditions of pri-
vacy. Likewise, burglars notoriously look for houses in
which the occupants are away rather than breaking into the
nearest house. Thus dominant responses which are contrary
to the norms, customs, or laws of a society tend to be in-
hibited in public.

Personal variables may be important since they may in-
fluence the type of roles that the person assumes. For in-
stance, Travis (1928) has shown that "normals" are facili-
tated and "stutterers" are hindered by the presence of an
audience during a written free association task. Since the

stutterers had a speech difficulty, they had probably

11

learned to assume nonverbal roles in the presence of others
in order to avoid exhibiting their handicap. If so,.per-
forming the verbal task before the audience would be a task
which was incompatible with the nonverbal role we would ex-
pect them to assume.

For the present study, the subjects' sex was selected
as a personal variable. First, it was felt that, since
Zajonc intended that his theory be quite general, it should
apply to both males and females. Secondly, Jones and Gerard
(1967) point out, "Men confront different social expecta-
tions-are constrained by a different set of norms-than women.“
As such, males are limited to a different subset of behav-
iors than females. Therefore, it is suggested that this
might influence the compatibility of task-required behav-
iors with role-expected behaviors. Thirdly, males and fe-
males differ in their abilities with various tasks as de-
scribed in Anastasi and Foley (1949). Thus, it might be
expected that males and females would be affected differ-
ently by the presence of an audience since they might as-
sume different roles which prescribe different behavioral
repertoires.

Situational variables might also have an influence on
the audience effect. Zajonc (1965) briefly alludes to the
cue function of the audience.3 He points out that in co-ac-
tion experiments the subjects provide cues and they might,
therefore, imitate each other. Although imitation cues

might be important in some experiments, the concern here

12

is with situational cues which provide information about

the type of role that the subject should adopt. For exam-
ple, one might want to see whether audience effects are
different when children serve as observers since people
adopt different roles vis-a-vis children than they do vis-a-
vis adults.

In this study the sex of the experimenter, and there-
fore the audience, was selected as the situational variable.
First, if the generality that Zajonc attributes to his
theory is justified, it is necessary that similar audience
effects occur with male and female observers. Secondly, if
the role interpretation has merit, it might be expected
that subjects would adopt different roles before male ob-
servers than before female observers. As discussed above,
these role differences might lead to differences in the au-
dience effect.

Since the present study focused upon the possible ef-
fects of personal and situational variables on audience ef-
fects, it was necessary to determine the extent to which
previous research studied or controlled for these variables.
Therefore a survey was made of the studies mentioned either
in Zajonc (1965) or Zajonc (1966) which dealt with audience

or coqaotion effects with humans.2 An attempt was made to

 

2Zajonc intended that his theory be relevant to all
organisms. Since the present formulation uses the concept
'role', it limits itself to explaining human behavior. How-
ever, this restriction is justified since studies of animal
behavior have failed to show that "audience effects" exist
separate from "competition effects".

l3

ascertain the sex of the subjects and the sex of the obser-
vers in each study.

Table l is a classification of these studies along the
dimensions of sex of subject and sex of observer. Four
categories were designated for each of these dimensions}
female, mixed (i.e. both male and female), male, and unspe-
cified (i.e. the sex of the individuals was not mentioned
in the article). The letters in parentheses specify the
type of experiment: audience (A), or co-action (C).

Note that in none of the experiments was the audience
composed solely of females. Likewise, none of the studies
reported audience effects for female subjects separately
from male subjects. In contrast, eight of the thirteen ex-
periments employed a strictly male audience (or co-actor)
and seven used only male subjects (or co-actors). Of these,
six used both male subjects and observers or male co-actors.
Upon examining the remaining studies in Table 1, it was
found that none reported comparisons of male and female
subjects or male and female observers.

To summarize, none of the studies which were reviewed
have investigated the effects of a female audience separate
from the effects of a male audience. In addition, none of
these studies considered audience effects for female sub-
jects separate from male subjects.

In an unpublished study by Good (1968), hypotheses
relevant to the present formulation were investigated.

Twenty male and twenty female introductory psychology

l4

 

_IIDBEMWVIHHmemmmwl

gov
Ammaav masons
gov

Ir

donHommmsb

 

Aav Annmav .
unspnsm a cannon

1.1

on Asmmav Datamata
ADV Annmavm

.Hm .po smecaom

Aev Aeomav unease:
on Aommav tsetse
Aevfl

Amwmav snag a EMwwom
Ammmwe aspom e nose

A<v
Ammoflv esteem

Eat:

 

A so>pommo

 

Ase
Ammeav naswne

gov
Azmmfiv phMQWH<
. o
Aeneas agendas

HO. unflm.

sows:

 

 

 

goameom

 

ouaoaooenn=,.

 

GHQ:

 

dexaz.

 

pooflpsm no son

 

030%.

AoomHv osonsN

:H so AmmmHv ononsN :H secede eozofipsoe moHUSpm onpomnoo use monofiesm nmesm 1H canoe

15

students served as subjects. Half of the males and females
first performed a warm up task before an extended audience
which consisted of a television camera and a microphone.
The warm up task was similar to the game Scrabble but the
subjects were asked to form cross—words which contained the
letter B. They were given five minutes for this task. The
other subjects did not warm up at all. This design, of
course, confounded familiarity with the audience with warm
up per se. Therefore, the effect of the temporal sequence
could not be assessed.

After a two-minute delay, all subjects were given five
minutes to learn as many of 50 one-syllable English nouns as
they could. This learning took place either alone or before
the audience. After another two-minute delay, all subjects
were asked to recall as many words as they could. All sub-
jects performed before the audience during the recall phase.

The first result relevant to the present discussion is
that a significant interaction of the sex of the subject and
audience conditions was found (pl .05). Upon further analy-
sis, it was found that the audience effect was only shown
for male subjects. This result is important both because
it suggests that males and females may be affected differ-
ently by the audience and because it demonstrates audience
effects with the extended audience.

Secondly, an overall sex difference was found in which
females performed significantly better than males (pl .005).

However, since only a male experimenter was used, it is

l6

impossible to separate the effects due to the sex of the
subjects and the effects due to the interaction of the sex
of the subject with the sex of the experimenter.

It was suggested that the interaction of the sex of the
subject and the audience conditions might have been due to
the fact that the experimenter, and presumably the audience,
was male. Thus, females may have defined their role with
the male experimenter differently than the male subjects.
More importantly, they might have defined their role differ-
ently if the experimenter had also been female. Therefore,
the present study is basically a replication of this study
with two important changes. First, all subjects performed
the warm up task but only half did so before the audience.
Since the first task was a performing task and the second
was a learning task, the order of the behavioral events was
reversed by using two separate tasks. Thus, the effect of
the temporal sequence of events could be evaluated. Secondly,
half of the subjects participated in an experiment for a
female experimenter and were presumably observed by a female
audience while the other subjects participated for a male
experimenter and audience. Thus, the effect of the situa-
tional variable, sex of experimenter, was investigated in
this study; 5

Based upon the arguments presented above, the following
two hypotheses were generated:

,HXPCTHESIS I: It was predicted that the temporal order of

events would be more of an influence on behavior than the

17

type of behavior in which the subjects were engaged (i.e.
learning or performing). As such, it was predicted that

the subjects would excel during the second of two tasks if
they had been left alone during the first and were before

an audience during the second._ However, since in the pre-
sent design, the first task involved mainly performance

and the second task involved mainly learning, it was pre-
dicted_that subjects would learn more if they had performed
alone and then learned before an audience. Since it was
predicted that subjects would learn more_if they were be-
fore the audience, this hypothesis predicts results directly
contrary to the predictions of Zajonc's theory of social
facilitation.

HYPOTHESIS II: It was predicted that the personal variable,
sex of subject, and_the situational variable, sex of audi-
ence, would influence audience effects. Since, as documented
above, there is a lack of information surrounding the exis-
tence of audience effects with females, this prediction is
made without specifying the directionality of the effect of
these variables. However, any effects found due to the situ-
ational or personal variables would be contrary to the in-
tended generality of Zajonc's theory.

There are three possible first-order interactions of
these variables---sex of experimenter with audience effects,
sex of subject with audience effects, and sex of experimen-
ter with sex of subject. These three relationships will be

considered separately.

18

(a) Subjects might have been affected differently by
male and female audiences. For example, subjects might
have perceived the female observer's role as neutral on-
looker, but perceived the male observer's role as evalua-
tor. On the other hand, subjects might have interpreted
observation by a male as a cue to his being conscientious
while interpreting the same behavior by a female observer
as a sign of her being curious.

(b) Male and female subjects might have been affected_
differently by the presence of an audience. For instance,
the audience might have increased the male subjects' need
for achievement while increasing the female subjects' fear
of failure. 0n the contrary, since females tend to be more
competent on verbal tasks, they might experience less con-
flict between Verbal behavior and the behaviors prescribed
by their role before the audience.

(0) There might have been an interaction of the sex of
the subject and the sex of the experimenter. For example,
subjects might have been more apprehensive about being ob-
served and evaluated by a same-sex observer since he might
have been perceived as more of a peer. 0n the other hand,
subjects might consider failure before a member of the op-

posite sex to be more threatening.

METHOD

gubjects
Forty-nine male and 51 female volunteers enrolled in
introductory psychology at Michigan State University served

as subjects for experimental credit.

Experimenters

Two male and two female students at Michigan State Uni-
versity served as experimenters., The author was one of the
male experimenters. In addition, the author obtained the
names of students in a juniorelevel social psychology class
who were interested in working on an ongoing research pro-
ject for course credits in psychology. From these names
the author randomly selected a male and two females. Thus
experimenter selection biases were minimized since the ex-
perimenters were chosen before they were met personally by

the author.

Apparatus

 

A television camera, a microphone,_a tape recorder, and
an electric light-push button signaling system were used in

this experiment.

19

20

Manipulations

 

As in the pilot study mentioned above, the extended aud-
ience was used in this experiment. In conditions which re-
quired an audience manipulation, the television camera situ-
ated across the table from the subject was turned on. The
camera was directed toward the subject and a red indicator
light on the front of the camera was functioning. In addi-
tion a large microphone was situated directly before the sub-
ject. Cables from the camera and microphone were plugged.
into a labeled wall receptacle in the view of the subject.

In the alone conditions the subjects remained in the
same room. However, the camera was turned off and pointed
toward the wall. The microphone was moved to a shelf to the
left of the subjects. The cables from the television camera
and the microphone were disconnected and wound around the ap-

paratus.

Procedure andExperimental Design

A four-way factorial design with one nested.factor was
used in this study. The independent variableswere sex of
the subject, performing the scrabble task (warmup) either
alone or before the audience, learning the English nouns (mem—
orization task) either alone or before the audience,and in-
dividual experimenters nested within sex of experimenter.
All subjects were asked to recall the learned material be-
fore the audience.

The primary dependent variable was the number of English

9

21

nouns recalled following the learning phase of the experi-
ment. An additional response measure was the number of words
formed during the warm up or scrabble task.3 Also, the num-
ber of intrusions were analysed. An intrusion was defined as
any response which was not counted as a correct response dur-
ing the recall of the English words.' These included incor-
rect responses, misspelled responses, and repetitions of cor-
rect responses.
When each subject arrived for the experiment, he was

taken to a small laboratory room. The experimenter asked the
subject to be seated, to put on the earphones provided, and

to wait for tape recorded instructions.“

The experimenter
then left the room. A separate set of instructions had been
recorded by each experimenter. The text of each experimen-
teris recorded instructions was identical except that male
experimenters said that video tapes of the subjects' behavior
would be reviewed "...by the other men in our research team.”
Female experimenters substituted the word “women“ for "men“
in their instructions.

The experimenters_were not in the same room as the sub-

jects during the instructions or tasks. Thus, a_communica—

tion system had to be devised by which the experimenter could

 

3Since the scrabble task preceded the memorization task,
the analysis of the scrabble task did not include a factor
for audience conditions during learning. _

“A transcript of recorded instructions is presented in
Appendix A.

22

ascertain the subjects' understanding of the instructions
while not violating the privacy of the alone conditions.
Therefore, an electric light-push button signaling system
was used with a control box in the room with the subject.
However, it was felt that subjects in audience conditions
might question the necessity for the control box in view of
the microphone and_television_camera, Therefore, the fol-
lowing recorded_instructions were given:

'All inStructions for this experiment are tape
recorded. ... After each time that instructions are
given, the amber light on the control box to your
left will come on automatically, At that time you
should press the appropriate button to indicate
whether you understand or do not understand the in-
.structions.' , '

Thus, it was subtly suggested to the subject that apparatus
connected to the tape recorder turned on the light at the
end of a set of instructions and that pressing the button
advanced the recorder.

In the first recorded message subjects were_told that
the experiment consisted of three separate parts. Subjects
were also instructed on the use of the control box.

The second recorded message then explained what will be
called the warm-up or scrabble task. Before each subject was
a deck of fifty 2%" x 3” cards with a letter of the alphabet
stenciled on each card using 1' Roman letters. In the deck
were three of each vowel, ten B's, and twenty-two other conso-
nants. Each subject was told that he would have seven minutes

to play a crossword game which was similar to the game Scrab-

ble. He was to lay the cards on the table and spell as many

23

words as possible which contained the letter B. However,
after making the first word, each new B-word had to share a
letter with one of the preceding words. An example which
illustrated the technique with X-words was provided. In ad-
dition, each subject was asked to say each word out loud as
he formed it.5 When the subject indicated that he understood
the instructions, he was told to begin.

When the subject was instructed to stop, the experimenter
reentered the room and counted the number of words spelled by
each subject. He then made any appropriate changes in the
room to create the desired audience manipulation and he pro-
vided materials for the learning task. Then the experimenter
again left the room.

Lying on the table before each subject was a deck of fif-
ty 5" x 8" note cards with a one-syllable English noun on
each card.6 Each subject was told that he would be allowed
seven minutes to memorize as many words as he could in any
order that he desired. The cards were, however, presented
in the same random order for all subjects.

Also, each subject was instructed to say the words out
loud as he was memorizing. In addition, he was told that

he would be_asked to recall the words in the final part of

 

5Cottrell et.al. (1968) have shown that audience ef-
fects are obtained only if the others present can conceivab-
ly observe and evaluate the subject's behavior. This re-
quirement insured that the subject was emitting behavior that
could conceivably be observed by an audience.

6A list of these words is presented in Appendix A.

24

the experiment. When the subjects indicated that they under-
stood the instructions, they were told to begin.

When each subject was told to stop, the experimenter a-
gain entered the room. He made any appropriate changes in
the room to create the audience condition for all subjects,
and he substituted materials appropriate for the recall phase
of the experiment. The experimenter then left the room.

Each subject was then instructed to recall as many words
as he could by printing one response on each sheet of paper,
and then turning that sheet face down. In addition, each
subject was asked to say each word out loud as he was writing
his responses. The subjects were also instructed to leave
the room when they had recalled as many words as they could.

Upon leaving the room each subject was asked to fill out
a six-question post-experimental questionnaire.7 Three ques-
tions were concerned with the subject's perception of the ex-
periment, suggestions for improving the experiment, and the
subject's perception of any deceptions during the experiment.
These questions served as criteria for eliminating subjects
from the analyses. Three other questions concerned with the
subject's personal feelings during the experiment were in-
cluded as filler items.

In order to insure that subjects did not communicate the
purpose of the experiment to subsequent subjects, they were

not debriefed after each experimental session. However,

 

7The post-experimental questionnaire is reproduced in
Appendix A.

25

interested subjects were mailed a brief summary of the pur-

pose of the study after all subjects had participated.

RESULTS

Of the 51 female and 49 male subjects who participated
in this experiment, four subjects were excluded from the an-
alyses. One female indicated on the post-experimental ques-
tionnaire that she believed that the purpose of the experi-
ment was to study audience effects. Another female stated
that she believed that she had been observed during an alone
condition. One male and one female did not follow directions
during recall. Thus, 48 males and 48 females remained in all
analyses reported below.

Separate analyses of variance were calculated for each
dependent variable and analyses of variance for simple effects
were calculated for each interaction which was significant
or marginally significant.8
Results Relevant to Hypothesis I

In Hypothesis I, it was predicted that the temporal se-

quence_oflevents would be more of an influence on behavior

 

8Both mar inally significant (.10 / p‘l .05) and signif-
icant (p L .05 results were indicated In tables and discus-
sed in the text. First, since in Hypothesis I Zajonc's
theory and the present formulation predict results in oppo-
site tails of the distribution, an alpha of .10 may be con-
sidered as a one-tail test of each hypothesis at the .05
level. Secondly, since Hypothesis II is essentially explora-
tory, results at the .10 level were included in order to
suggest hypotheses for further research.

26

27

than the type of behavior (i.e. learning or performing).
The results support this hypothesis but only with the sub-
jects of male experimenters.

The response measure used to test this hypothesis was
the number of English nouns recalled. The analysis of vari-
ance for English noun recall is presented in Table 2. It was
predicted that subjects would recall more nouns if they had
performed the warm-up task alone and then learned the nouns
before the audience. Although the means were in the predicted
direction9, neither the main effect for audience conditions
during warm-up (C) nor the effect for audience conditions dur-
ing learning (D) was significant in Table 2. However, two
interactions of audience conditions and the sex of the exper-
imenter were noted.

First, the interaction of the sex of the experimenter
and audience conditions during warm-up (A x C) was signifi-
cant (p,Z .01). An analysis of variance for the simple ef-
fects of the audience conditions was calculated and is pre-
sented in Table 3. An effect for audience conditions during
warm-up was found only for the subjects of male experimenters
(p Z .05). The cell means for this interaction are presented
in Table 4. It can be seen that subjects of male experimen-
ters recalled more English nouns when warm-up had been alone

than when warm-up had been before the audience.

 

9Tables of cell means, for each level of each indepen-
dent variable used in the analysis of each response measure,
are presented in Appendix B.

28

Table 29 Analysis of variance for English noun recall

 

 

 

 

Source of Variability df MS F p
Sex of Experimenter (A) 1 192.67 5.30 .05
Sex of Subject (B) 1 287.04 7.90 .01
Audience during Warm-up (C) 1 13.50 0.37 n.s.
Audience during Learning (D) l 28.17 0.78 n.s.
Individual Experimenters (E/A) 2 82.60 2.27 n.s.
A x B 1 8.17 0.22 n.s.
A x C 1 273.38 7.52 .01
A x D 1 108.38 2.98 .10
B x C 1 24.00 0.66 n.s.
B x D 1 0.17 0.00 n.s.
B x E/A 2 54.19 1.49 n.s.
C x D l 30.38 0.84 n.s.
C x E/A 2 192.02 5.28 . 1
D x E/A 2 5.44 0.15 n.s.
A x B x C l 18.38 0.51 n.s.
A x B x D 1 84.38 2.32 n.s.
A x C x D l 2.67 0.07 n.s.
B x C x D l 70.04 1.93 n.s.
B X C x E/A 2 44.27 1.22 n.s.
B x D x E/A 2 31.60 0.87 n.s.
C x D x E/A 2 11.35 0.31 n.s.
A x B x C x D l 10.67 0.29 n.s.
B x C x D x E/A 2 29.52 0.81 n.s.
Error 64 36.34

Total 95

29

Table 3- Analysis of variance for the simple effects of audi-
ence conditions in the interaction of sex of experi-
menter and audience conditions during warm-up (A x C)
for the response measure, English noun recall

 

 

Source of Variability df MS F p

 

Audience conditions for
Male Experimenters 1 204.19 5.62” .05

Audience conditions for
Female Experimenters

82069 2028 n0S0
Error 6

36.34

¢+4

Table 4- Cell means for the interaction of sex of experimenter
and audience conditions during warm-up (A x C) for
the response measure, English noun recall

Warm-up Conditions

Alone Audience

 

Female 19.25 21.88

Sex of Experimenter

 

Male 19.79 15.67

 

 

 

 

30

Secondly, there was an interaction of the sex of the ex-
perimenter and audience conditions during learning (p‘L .10).
An analysis of variance for the simple effects of audience
conditions was calculated and is presented in Table 5. Audi-
ence conditions during learning were shown to have an effect
only with the subjects of male experimenters (P.Z .10).10
The cell means for this interaction are shown in Table 6.
Subjects of male experimenters recalled more English nouns
when learning had been before the audience than when learning
had been alone.

To summarize, it was found that the subjects of male ex-
perimenters learned more English nouns if they had warmed up
alone and if they had learned the nouns before the audience.
No audience effects were found with the subjects of female
experimenters. Thus, none of the results showed learning to

be hindered by the presence of the audience.

Results Relevant to Hypothesis II

In Hypothesis II it was predicted that the personal vari-
able, sex of subject, and the situational variable, sex of
experimenter, would influence audience effects. This hypo—
thesis received support from the measures of B-word formation,
English noun recall, and intrusions. As in the statement of
the hypothesis, the results for each of the three first-order

interactions will be presented separately.

 

10p Z .05 for the directional hypothesis

31

Table 5- Analysis of variance for the simple effects of audi-
ence conditions in the interaction of sex of experi-
menter and audience conditions during learning (A x D)
for the response measure, English noun recall

 

 

 

Source of Variability df MS F p
Audience conditions for

Male Experimenters 1 123.52 3.40 .10
Audience conditions for

Female Experimenters l 13.02 0.36 n.s.
Error 64 36.34

Table 6- Cell means for the interaction of sex of experimenter
and audience conditions during learning (A x D) for
the response measure, English noun recall

Learning Conditions

Alone Audience

Female 21.08 20.04

 

Sex 6f Experimenter

 

Male 16.13 19.33

 

 

 

 

32

Results Relevant to HypotheSis II(a) It was predicted
that subjects might be affected differently by male and fe-
male audiences. Interactions were found with both the audi-
ence manipulation during warm-up (A x C) for English noun
recall and B word formation and with the audience manipula-
tion during learning (A x D) for English noun recall.

An interaction of sex of experimenter and audience con-
ditions during warm-up (A x C), for the measure of English
noun recall, is found in Table 2. This is one of the inter-
actions that was presented in the previous section. However,
here an analysis of variance for the simple effects of sex
of experimenter was calculated and is presented in Table 7.
It can be seen that the sex of the experimenter had an ef-
fect during the audience condition (p,[ .001) but not during
the alone condition. The cell means in Table 4 show that
subjects recalled more English nouns for a female experimen-
ter than for a male experimenter in the audience condition.

An interaction of sex of experimenter and audience con-
ditions during learning (A x D) was also found in Table 2.
This is the other interaction presented in the previous sec-
tion. An analysis of variance for the simple effects of sex
of experimenter was calculated and is presented in Table 8.
An effect was found for the sex of the experimenter in the
alone condition (p,Z .01) but not in the audience condition.
The cell means in Table 6 show that subjects in the alone
condition recalled more nouns for female experimenters than

for male experimenters.

33

Table 7- Analysis of variance for the simple effects of sex
of experimenter in the interaction of sex of experi-
menter and audience conditions during warm-up (A x C)
for the response measure, English noun recall

 

Source of Variability df MS F p

 

Sex of Experimenter for

Audience Condition 1 462.52 12.73 .001
Sex of Experimenter for

Alone Condition 1 3.52 0.10 n.s.
Total Error 64 36.34

Table 8- Analysis of variance for the simple effects of sex
of experimenter in the interaction of sex of experi-
menter and audience conditions during learning
(A x D) for the response measure, English noun re-

 

 

 

call
Source of Variability df . . MS F p.
Sex of Experimenter for
Audience Condition 1 6.02 0.17 n.s.
Sex of Experimenter for
Alone Condition 1 295.02. 8.12 .01

Total Error 64

34

The analysis of variance for B-word formation is pre-
sented in Table 9. The interaction of sex of experimenter
and audience conditions during warm-up (A x C)11 was mar-
ginally significant (p 1 .10). An analysis of variance for
the simple effects of sex of experimenter was calculated and
is presented in Table 10. An effect for the sex of the ex-
perimenter was found in the audience condition (P.Z .05) but
not in the alone condition. The cell means for this inter-
action are presented in Table 11. In the audience condition,
subjects formed more B-words for a female experimenter than
for a male experimenter.

The results of the analysis of variance for the response
measure, intrusions, is shown in Table 12. Neither the inter-
action of the sex of the experimenter and audience conditions
during warm-up (A x C) nor the interaction of sex of experi-
menter and audience conditions during learning (A x D) was
significant.

To summarize, subjects formed more B-words and recalled
more English nouns when warm-up had been before a female ex-
perimenter-audience. Subjects also recalled more English
nouns when learning had been alone for a female experimenter.
There were no interactions of the sex of the experimenter
and audience conditions either during warm-up or during

learning with the measure of intrusions.

 

11As noted above, since the scrabble (warm-up) task pre-
ceded the memorization (learning) task, a factor for audi-
ence conditions during learning was not included in the an-
alysis of B-word formation.

Table 9- Analysis of variance for B-word formation

 

 

 

 

Source of Variability df MS F p
Sex of Experimenter (A) l 26.04 2.95 .10
Sex of Subject (B) l 2.67 0.30 n.s.
Audience during Warm-up (C) l 2.67 0.30 n.s.
Individual Experimenters (D/A) 2 26.21 2.97 . 0
A x B l 6.00 0.68 n.s.
A x C l 24.00 2.72 .10
B x C 1 57.04 6.46 .01
B x D/A 2 1.08 0.12 n.s.
C x D/A 2 18.42 2.09 n.s.
A x B x C l 3.38 0.38 n.s.
B x C x D/A 2 10.21 1.16 n.s.
Error 80. 8.83

Total 95

36

Table 10— Analysis of variance for the simple effects of sex
of experimenter in the interaction of sex of exper-
imenter and audience conditions during warm-up
(A x C) for the response measure, B-word formation

 

 

Source of Variability

Sex of Experimenter for
Audience Condition
Sex of Experimenter for

Alone Condition
Total Error

.df MS F P

1 50.02 5.66 .05

1 .02 0.00 n.s.
80 8.83

Table 11- Cell means for the interaction of sex of experi-
menter and audience conditions during warm—up
(A x C) for the response measure, B-word formation

Warm-up Conditions

 

Female

Sex of Experimenter

 

Male

 

Alone Audience
10.83 11.50
10.79 9.45

 

 

 

37

Table 12- Analysis of variance for intrusions

 

 

 

 

Source of Variability df MS F p
Sex of Experimenter (A) l 0.04 0.03 n.s.
Sex of Subject (B) l 0.00 0.00 n.s.
Audience during Warm-up (C) l 0.67 0.48 n.s.
Audience during Learning (D) l 0.38 0.27 n. 5.
Individual Experimenters (E/A) 2 1.69 l. 22 n. s.
A x B 1 15.04 10. 94 .002
A x C 1 1.04 0.76 n. s.
A x D l 0.67 0.48 n.s.
B x C 1 6.00 4.36 .05
B x D l 0.04 0.03 n.s.
B x E/A 2 1.10 0.80 n.s.
C x D l 1.04 0.76 n.s.
C x E/A 2 0.52 0. 38 n.s.
D x E/A 2 0.94 0. 68 n.s.
A x B x c 1 2.04 1.48 n.s.
A x B x D l 1.50 1.09 n.s.
A x C x D l 0.17 0.12 n.s.
B x C x D l 1.04 0.76 n.s.
B x C x E/A 2 2.60 1.89 n.s.
B x D x E/A 2 0.27 0.20 n.s.
C x D x E/A 2 1.35 0.98 n.s.
A x B x C x D l 0.00 0.00 n.s.
B x C x D x E/A 2 0.35 0.26 n.s.
Error 64 1.38

Total 95

38

Results Relevant to Hypothesis II(b) It was predicted
that male and female subjects might be affected differently
by the presence of an audience. Interactions were found with
the audience manipulation during warm-up (B x C) but not with
the audience manipulation during learning (B x D).

In Table 9 the interaction of sex of subject and audience
conditions during warm-up (B x C) was significant (p‘L .05)
for the response measure, B-word formation. An analysis of
variance for the simple effects of sex of subject was cal-
culated and is presented in Table 13. A significant effect
for the sex of the subject was found only in the alone con-
dition (p11 .05). The cell means for this interaction are
presented in Table 14. In the alone condition, male subjects
formed more words than female subjects. Note, however, that
there was a nonsignificant (P A .20) trend in the opposite
direction in the audience condition.

In Table 12, a significant interaction of sex of_subject
and audience conditions duringwarm-up (B x C) for the re—

An analysis of varie

sponse measure, intrusions (p,[ .05). ,
anCe for the simple effects of sex of subject was calculated
and is presented in Table 15. .The effect of the sex of_the
subject was nonsignificant (p Z .20) in both the alone and_
audience conditions. .Upon examining the cell means in Table
16, it is noted that the trend in the alone condition is in
the opposite direction as the trend in the audience condition.

No interactions of sex of subject and audience conditions

were found in Table 2 with the response measure, English noun

recall.

39

Table l3- Analysis of variance for the simple effects of sex
of subject in the interaction of sex of subject and
audience conditions during warm-up (B x C) for the
response measure, B-word formation

 

 

Source of Variability df MS F p

 

Sex of subject for

Audience Condition 1 17.52 1.98 n.s. (.20)
Sex of subject for

Alone Condition 1 42.19 4.78 .05
Total Error 80 8.83

Table 14- Cell means for the interaction of sex of subject
and audience conditions during warm-up (B x C) for
the response measure, B-word formation.

Warmeup Conditions

Alone Audience

 

Female 9.88 ( 11.08
Sex of Subject '

 

 

Male 11.50 9.88

 

 

 

 

40

Table 15- Analysis of variance for the simple effects of sex
of subject in the interaction of sex of subject and
audience conditions during warm—up (B x C) for the
response measure, intrusions

 

 

Source of Variability df MS F p

 

Sex of Subject for

Audience Condition 1 3.00 2.17 n.s. (.20)
Sex of Subject for

Alone Condition 1 3.00 2.17 n.s. (.20)
Total Error 80 1.38

Table 16- Cell means for the interaction of sex of subject
and audience conditions during warm-up (B x C) for
the response measure, intrusions

warm-ap'c.nditions

Alone Audience

 

Female 0.58 1.25
Sex of Subject

 

Male 1.08 0.75

 

 

 

 

41

To summarize, interactions of the sex of the subject
and audience conditions during the warm-up task were found
with the response measures, Baword formation and intrusions.
No interaCtions of sex of subject and audience conditions
during learning were found. No interactions of the sex of
the subject and audience conditions were found with the
response measure, English noun recall.

Results Relevant to Hypothesis II(c) It was predicted
that there might be an interaction of the sex of the experi-
menter and the sex of the subject. This hypothesis was sup-
ported by the measure of intrusions but not by the response
measures, B-word formation or English noun recall.

In Table 12 a significant interaction was found with
the measure of intrusions (p Z .002). Analyses of variance
for the simple effects of both sex of subject and sex of
experimenter were calculated and are presented in Tables 17
and 18. In Table 17, the effect for sex of experimenter
is found with both male and female subjects (p _/_ .05)}
Table 18 shows that the effect for sex of subject was found
for both male and female experimenters (p Z .05). The
cell means for this interaction are shown in Table 19. It
is noted that the most intrusions were made when the sub-
jects were of the same sex as the experimenters and that
the fewest intrusions were made when the subjects were of
the opposite sex as the experimenters.

In Tables 2 and 9, the interaction of sex of experimen-

ter and sex of subject was found to be nonsignificant for

42

the response measures, B-word formation and English noun
recall.

To summarize, the interaction of sex of subject and
sex of experimenter was found only with the dependent

variable, intrusions.

43

Table 17- Analysis of variance for the simple effects of sex
of experimenter in the interaction of sex of experi-

menter and sex of subject (A x B) for the response
measure, intrusions

 

 

Source of Variability df MS F p

 

Sex of Experimenter for

Female Subjects 1 8.34 6.04 .05
Sex of Experimenter for

Male Subjects 1 6.75 4.89 .05
Total Error 80 1.38

Table 18— Analysis of variance for the simple effects of sex
of subject in the interaction of sex of experimenter

and sex of subject (A x B) for the response measure,
intrusions

 

 

Source of Variability df MS F p

 

Sex of Subject for

Female Experimenters l 7.52 5.45 .05
Sex of Subject for
Male Experimenters l 7.52 5.45 .05

Total Error 80 1.38

44

Table 19- Cell means for the interaction of sex of experi-
menter and sex of subject (A x B) for the response
measure, intrusions

Sex of Experimenter

 

Female , Male
Female 1.33 0.50
Sex of Subject

 

 

Hale ‘ 005“ 1029

 

 

 

 

DISCUSSION

In Hypothesis I it was predicted that the temporal se-
quence of events would be more of an influence on behavior
than the type of behavior in which the subject was engaged
(i.e. learning or performing). This hypothesis was supported
for the subjects of male experimenters.

In examining the results, no audience effects were found
for the subjects of female experimenter-observers. This re-
sult, in itself, leads one to question the generality of
Zajonc's theory. However, as discussed earlier, none of the
human-subject experiments mentioned by Zajonc (1965, 1966)
actually investigated the effect of a female audience. There-
fore, this finding could represent the absence of audience
effects with female observers or the conflict of audience
effects with temporal order effects. Further research with
female observers is needed to clarify this issue.

With male experimenter-observers, subjects learned more
English nouns when they had warmed up by performing the scrab-
ble task alone rather than before the audience. They also
learned more nouns if learning had been before the audience.
Thus, subjects of male experimenters excelled on the second
task when they had accomplished the first task alone and the

second before an audience. The results support Hypothesis I

45

46

and the contrast model.

These results are also contrary to the relationship
posited by Zajonc since subjects learned more nouns when
they were before the audience than when they were alone.

That is, by merely reversing the temporal sequence of behav-
ioral events, results were obtained in the direction opposite
that predicted by Zajonc's theory of social facilitation
(1965).

Since this is the first experiment to test for the ef-
fects of the temporal sequence, since results were obtained
only at the .10 level (.05 directional test), and since audi-
ence effects were obtained only with male experimenters, this
finding must be accepted with a degree of caution. It is
also suggested that further research might investigate the
effects of the temporal sequence independent of the behavioral
sequence. For example, one might include four conditions
such that the first task is either learning or performing and
the second is either learning or performing. Another possi-
bility is having longer chains of tasks such that data can be
obtained within individual subjects.

In Hypothesis II it was predicted that the personal vari—
able, sex of subject, and the situational variable, sex of
experimenter, would influence the audience effect. This hypo-
thesis received support from the measures of B-word formation,
English noun recall, and intrusions.

It was predicted in Hypothesis II(a) that the sex of the

experimenter might influence audience effects. The results

47

revealed that subjects formed more B-words and recalled more
English nouns if warm-up had been before a female experimenter-
observer than before a male experimenter-observer. In the
alone condition, no effect was found for the sex of the exper-
imenter.

In the audience condition the subjects had a salient re-
lationship with both the task and the audience. That is, the
subjects had to exhibit certain behaviors to accomplish the
task and had to limit themselves to ya domain of behaviors de-
fined by their roles vis-a-vis the audience.

However, in the alone condition the relationship with
the task is salient but the subjects' roles vis-a-vis the
experimenter probably were not. If the subjects were not
maintaining this role, it is not expected that characteris-
tics of the experimenter would influence their behavior any
more than other roles not salient in the situation (e.g. son,
customer, lover, etc.).

With regard to the earlier discussion which considered
audience effects in terms of the compatibility of roles with
ongoing behavior, it is noted that the question of compati-
bility is only relevant when the subject had assumed a role
vis-a-vis another person (i.e. was in the audience condition).
Thus the differences in the audience condition might have been
due to one or both of the following factors. First, there
might have been a facilitation effect before the female audi-
ence if roles assumed vis-a-vis a female observer tended to

exclude behaviors, irrelevant to the task, in which the

48

subject might engage if he were left alone (e.g. daydreaming,
visual exploration of the room, etc.). Secondly, roles adopted
vis-a-vis a male observer might have had the effect of sub-
ordinated task-oriented behaviors to other behaviors (e.g.
attempting to impress the experimenter, trying to discover

the purpose of the experiment, demonstrating a unique system

of memorization, etc.).

In the interaction of sex of experimenter and audience
conditions during learnin , it was found that subjects re-
called more nouns in the alone condition if the experimenter
was female rather than male. However, in the audience con-
dition there was no difference between the subjects of male
and female experimenter-observers. Thus, subjects continued
to perform better before a female experimenter but did so in
the alone condition rather than the audience condition. This
result is contrary to the preceding discussion. However, it
is in line with the discussion of Hypothesis I (i.e. no audi-
ence effects were found with subjects of females but subjects
learned less when working alone than before the audience for
a male experimenter-observer). Thus, in this particular in-
teraction the comparison model and the role interpretation
lead to predictions of contrary effects. Since the results
support the comparison model, it can be tentatively accepted
that the comparison effects had a stronger influence on behav-
ior . That is, the changing of audience conditions may have
had a greater influence on the audience effect than the sex

of the experimenter.

49

It was predicted in Hypothesis II(b) that male and fe-
male subjects might have been affected differently by the
presence of an audience. Interactions were found with the
audience manipulation during warm-up but not with the mani-
pulation during learning.

Significant interactions of sex of subject and audience
conditions during warm-up were found with the response measures,
B-word formation and intrusions. Males formed more B-words
than females in the alone condition. This is contrary to
the predictions of sex differences by Anastasi and Foley (1949).
Since the subjects were alone, differences can not be ex-
plained by role differences. Thus, no explanation is offered
for this finding. In the audience condition, there was a non-
significant trend in the opposite direction. Likewise, there
was a trend for males to subsequently make more intrusions
when warm-up had been alone and a trend for females to make
more intrusions when warm-up had been before the audience.

Thus, with regard to the audience manipulation during
warm-up, the interaction of subject's sex and audience ef-
fects was given some support. However, the analyses of sim-
ple effects show mostly nonsignificant trends. Thus, al-
though the sex of subject may influence audience effects,
this factor may have only a relatively weak influence on ac-
tual behavior.

There was no interaction of sex of subject and audience

conditions during learning with either the measure of English

50

noun recall on the measure of intrusions. This finding is
contrary to the results found in the earlier study (Good,
1968). There are at least two possible reasons why this
might be.

First, it may be that the temporal sequence or contrast
effect influenced both male and female subjects, but that
only males were affected by the audience per se. Contrast
effects were not investigated in the pilot study. Secondly,
since in the present study both male and female experimenters
were used, some of the variability found in the pilot study
might be accounted for by a higher-order interaction in the
present study.12

In Hypothesis I(c) it was predicted that there might be
an interaction of the sex of the subject and the sex of the
experimenter-observer. This hypothesis received support but
only with the measure of intrusions. The interaction was
highly significant with subjects making approximately two
and one half times as many intrusions before an experimenter-
observer of the same sex as one of the opposite sex. There
are at least two possible explanations for this interaction.
First, it might have been that the subjects were more care-
ful before experimenter-observers of the opposite sex, and,
therefore, they may have made fewer mistakes. 0n the other
hand, subjects might have tried harder for experimenter-ob-

servers of the same sex, and, therefore, they may have simply

 

12Note that in Table 2 there was a nonsignificant (p A .20)
second—order interaction of sex of subject, sex of experimen-

ter, and audience conditions during learning (A x B x D).

51

made more guesses.

Upon reconsidering the experimental situation, it may
have been inappropriate to expect an interaction with the
response measures, B-word formation and English noun recall.
In the first place, audience conditions were counterbalanced
such that only half of the subjects would be expected to have
assumed a role vis-a-vis the experimenter during the perfor-
mance of the scrabble task and half during the learning of
the English nouns. Thus, with the response measures for
these tasks, half of the subjects are expected not to have
been maintaining a role. However, since intrusions could
only be made during recall before the audience, it is ex-
pected that all subjects had assumed a role vis-a-vis the
audience.

Secondly, the response measures, B—word formation and
English noun recall, are undoubtedly influenced by the per-
sons' abilities at the task. Thus if this factor had a rela-
tively stronger influence on behavior, it would tend to mask
the interaction of the sex of the subject-performer and the
sex of the experimenter-observer. However, the number of
intrusions is the number of incorrect responses that the sub-
ject made independent of the number of correct responses.
That is, the number of intrusions is the number of responses
that the person attempted beyond his ability to recall.
Therefore, the measure of intrusions allowed a more direct
test of the relationship predicted in the hypothesis than

the other measures.

52

To summarize, the results show that both hypotheses in
this study received support. Thus, both the comparison model
and the social role model are supported as viable approaches
to the study of audience phenomena. In addition, the results
provided evidence contrary to the predictions of Zajonc's

(1965) theory of social facilitation.

OVERVIEW

The area of audience effects has shared a problem with
the rest of the field of group dynamics and much of psycho-
logy in general. This is the assumption that behavior can be
explained in terms of main effects or simple interactions.

As Messé (1969) has suggested, it is doubtful whether any
main effects exist in the "real world", and there are probab-
ly few, if any, simple interactions.

Many earlier studies of groups have failed to obtain
consistent findings. For instance, Bird (1940) made an ex-
tensive examination of research relevant to leadership traits.
He found that only about five percent of the discovered
traits were common to more than three studies. The problem
seems to stem from the fact that the investigators were in-
terested in finding common attributes applicable to "the
group" where "the group' might be anything from a street gang
to a presidential cabinet.

A more rewarding approach, however, might be trying to
ascertain what kinds of behavior are exhibited by what kinds
of groups working on what kinds of tasks in what kinds of
situations. Here a "kind” is merely defined as a subgroup
in which the relationship in question is believed to hold.

The process of research then advances by further subdividing

53

54

these subgroups until no higher order interactions can be
found.

In the area of audience phenomena the question has been:
What is the effect of "the audience"? It is proposed that a
more rewarding question might be: What is the effect of a
certain kind of audience on a certain kind of person engaged
in a certain kind of behavior to complete a certain kind of
task in a certain kind of situation? Such a question is awk-
ward to state and more difficult to answer. However, the re-
sults of this study indicate that subgroupings, at least at
this rudimentary level, must be made before we can begin to
understand the subtleties and complexities of audience pheno-

mena .

LIST OF REFERENCES

LIST OF REFERENCES

Ader, R., and Tatum, R. Free operant avoidance conditioning
in individual and paired human subjects. .Journal of the
Experimental Analysis of Behavior, 1963, 6{I357-359.

 

Allport, F.H. The influence of the group upon association
and thought. Journal of Experimental Psychology, 1920,
3, 159-182.

Allport, F.H. Social Psychology. Boston: Houghton Mifflin,
1924, (11), 260-291.

 

Anastasi, Anne, and Foley, J.P.Jr. Differential Psychology.
New York: MacMillan Company, 1949,7I19), 646-688.

 

Bergum, 3.0., and Lehr, D.J. Effects of authoritarianism“
on vigilance performance. Journal of Applied Psychology,

1963. 31. 75-77.

Bird, C. Social Psycholpgy. New York: Appleton-Century Com-
pany, 1940.

 

Cottrell, N.B., Wack, D.L., Sekerak, G.J., and Rittle, R.H.
Social facilitation of dominant responses by the pre-
sence of an audience and the mere presence of others.
Journal of Personality and Social Psychology, 1968, 2(3),
245-250.

 

Dashiell, J.F.. An experimental analysis of some group ef-
fects. Journal of Abnormal and Social Psychology, 1930,
pg, 190-199. _

Goffman, E. The Presentation of seir‘in Everyday Life.
Garden City, New YOrk: IDoubleday andICOmpany, Inc.,
1959.

Good, K.J. The effects of audience, prehabituation, and sub-
ject's sex on the learning and recall of English nouns.
Unpublished experiment, 1968.

 

 

Gurnee, H. The effect of collective.learning upon.the.in-
dividual participants. Journal of Abnormal and Social

Psychology, 1939, 4, 529-532.

 

55

r1

)1

4"

1“

 

 

 

56

Jones, E.E., and Gerard, H.B. Foundations of Social PsyCh-
olo . New York: John Wiley and SOns, Inc., I967, (15)
600.

 

Mason, J.W., and Brady, J.V. The sensitivity of psychoendo-
crine systems to social and physical environment. In
P. Leiderman and D. Shapiro (Eds.), Psychobiological
Approaches to Social Behavior. Stanford: StanfOrd‘Uni-
versityIPress, 1964.

 

Meumann, E. Haus-und Schularbiet: Experimente an kinder der
volksschule, Die Deutsche Schule, 1904, (8), 278-303,
337-359: 416-431.

Messe, L. Personal Communication, 1969.
Pessin,J., and Husband, R.W. Effects of social stimulation,

on human maze learning. Journal of Abnormal and Social
Psychology, 1933, afi, 148:I54.

 

 

Pessin, J. The comparative effects_of.social_and.mechanical
stimulation on memorizing. American Journal of Psych-
olo , 1933, 45, 263-270.

 

Spence, K.W. Behavior Theory and Conditioning. New Haven,
Conn.: Yale University Press, 1956.

 

Seidman, D., Bensen, S.B., Miller, L., and Meeland, T.
Influence of a partner_on tolerance for selfqadministered
electric shock. Journal of Abnormal and Social Psych—

ology, 1957, 4, 210-212.

Thibaut, J.W. and Kelley, H.H. The SoCial Psychology of
Groups. New York: John Wiley and Sons, Inc., 1959.

Travis, L.E. The effect of a small audience upon eye-hand
coordination, Journal of Abnormal and Social Psychology,
1925, 20, 142-146.

Travis, L.E. The influence of the group upon the stutterer's
speed.in.free association- Journal of Abnormal and
Social Psychology, 1928, 2 , 45-51.

Triplett, N. The dynamogenic-factors.in pacemaking and com-
petition. American Journal of Psychology, 1897. 2.

 

507-533-
Zajonc, H.B. Social facilitation. Science, 1965, 132, 269-
274.

Zajonc, H.B. Social PsyChology: An Experimental Approach.
Belmont, Calif}: IWadsworth Publishing Company, Inc.,
1966, (1-3), 1-28.

APPENDICES

APPENDIX A

EXPERIMENTAL MATERIALS

EXPERIMENTAL MATERIALS

Transcript of Recorded Instructions
Preliminary InstruCtions:

"All instructions for this experiment are tape recorded.
Therefore, you are to remain seated and continue to wear the
earphones until the entire experiment is completed. This ex-
periment is divided into three separate parts with separate
instructions for each part. After each time that instruc-
tions are given, the amber light on the control box to your
left will come on automatically. At that time you should
press the appropriate button to indicate whether you under-
stand or do not understand the instructions. Incidentally,

I will monitor part of this experiment by means of the closed
circuit audio-visual system that you see in this room. In
addition, video tapes will be made for review by the other
men (women) in our research team. Before we begin, please
indicate whether or not you understand these preliminary in-
structions.”

(5 second pause)

WarmIUp TaSk InStructions:

 

nBefore you is a deck of 50 cards with one letter of
the alphabet on each card. When I tell you to begin, you
are to arrange the letters into words. The words are to read

either down or across such that after making the first word,

57

58

at least one letter of each new word is shared with one of
the preceding words as in the example provided.
(5 second pause)
In this example all of the words contain the letter 'X'.
However, your task is to construct as many words as you can
which contain the letter 'B' as in 'baker'. In addition, I
want you to say out loud each word that you form. You will
be given seven minutes to perform this task. You will be
told when to begin and when to stop. Before we begin, please
indicate whether or not you understand these instructions.
(5 second pause)

"Remember, you are to spell-out as many words as you can
which contain the letter 'B', and also remember to say each
word out loud as you are working...Ready...Begin.

(7 minute delay)

"Stop! Remain seated and continue to wear the earphones.
There will be a short delay while I prepare for the second
part of the experiment.‘

(Delay to allow 2 minutes from “Stop" above to "Begin" below)

Memorization Task Instructions:

 

"Before you is a deck of fifty five-by-eight cards. A
one-syllable English noun is printed on each card. When I
tell you to begin, you are to remove the cover sheet and be-
gin memorizing. You will be given seven minutes to memorize
as many words as you can, in any order that you desire. In"
addition, I want you to say the words out loud as you are

looking at them and while you are thinking about them. You

59

are not expected to memorize all of the words. However, you
should spend the entire seven minutes trying to memorize as
many of the words as possible. I will tell you when to begin.
When I tell you to stop, you are to turn the cards face down.
Before we begin, please indicate whether or not you under-
stand the instructions.

(5 second pause)

"Remember, you are to spend the entire time memorizing
and also remember to keep saying the words out loud as you
are memorizing...Ready...Begin.

(7 minute delay)

"Stop! Remain seated and continue to wear the earphones.
There will be a short delay while I prepaFe for the third part
of the experiment.“

(Delay to allow 2 minutes from "Stop" above to "Begin" below)
Recall Instructions:

”You are now asked to recall as many words as you can
in any order that you desire. When you are told to begin,
say the first word you can recall out loud and then print
your response on the paper provided using the 'magic marker'.
Print only one word on each sheet of paper and turn each
sheet face down after you have made a response. Continuing
this procedure, take as much time as you need in order to
recall as many words as you can. When you are finished, you
may remove the earphones and leave the room. Before we be-
gin, please indicate whether or not you understand the in-

structions.

60

(5 second pause)
”Remember, you are to respond both by saying the words
out loud and by printing the words on the sheets of paper

provided...Ready...Begin.'

61

English Nouns fOr Memory Task*

1. FAITH l9. WEALTH 37. FUR
2. OAK 20. BAR 38. BOWL
3. KEY 21. AUNT 39. GUN
4. TRUCE 22. TAIL 40. NOOSE
5. MAID 23. CAP 41. FLIGHT
6. TASK 24. LEAGUE 42. SERF
7. CAKE 25. HORN 43. PRIDE
8. SILK 26. STUB 44. SALE
9. DAWN 27. RANK 45. TROOP

10. HEEL 28. ROUTE 46. BUSH

11. PACK 29. PAN 47. CROP

12- AID 30. NOISE 48. SWORD

13. GAS 31. KNIFE 49. LAMP

14- RAIL 32. NUT 50. BAND

15. NEST 33. LOG

16. BOOK 34. GIFT

17. DUST 35. FAME

l8. DOOR 36. DEBT

 

*The words are presented in the same random order that
the cards were arranged for each subject.

62
Bostexperimental questionnaire
1. What do you think was the purpose of this experiment?

2. Based on your answer to question.#l, what would you
change about this experiment?

3. During this experiment, did you feel that you were
deceived in any way? If yes, explain.

4. How pleasant or unpleasant was this experiment for you?

/0.00.00/0.0...O/OOOOOOO/OOOOOOO/OOOOOOO/OOOOOOO/OOOOOOO/

Quite Rather Somewhat Somewhat Rather Quite
Pleasant Unpleasant

5. How interesting or uninteresting was this experiment

for you?
/OOOOOOO/OOOOOOO/OOOO.../0.00000/OOOOOOO/OOOOOOO/OOOOOOO/
Quite Rather Somewhat Somewhat Rather Quite
Uninteresting Interesting

6. To what extent did it bother you to be observed during
this experiment?

/OOOOOOO/OOOOOOO/OOOO.../OOOOOOO/'OOOOOOO/OOOOOOO/OOOOOOO/
Not at Slight- Some- Moder- Consid- Very Extreme-
Kfll 1y what ately erably Much ly
Bothered Bothered

...8

Q“!

‘ \

\‘ \\ \‘ .

o o o o o o c c ‘ 9 I a o o o o ‘. v- o I a n o o o
.
\ 7‘
a a o O a o c o a O O h 0 1 O n "I a o a 1
\
o o 9 I o o u o o o o o n o o n w 1 o 2 o a a o
0

0‘0.

\

0......

I C.

H

APPENDIX B

TABLES OF CELL MEANS

Ject

 

Sex of Sub

63

 

 

 

 

 

 

 

 

 

 

.m n. eI -... - . .
m m a 32% 2:3
n M”
e
.m
.1
r e ... , f
e 1 . ,
m. m 3.5 ,_ 223
E e
F
r.
o I
u e ........ _
e S l om.mH mn.w.mv
l a
a M
m
e
F.
l
m mm.ma mw.ua
e . .
F.
00S®HU§¢ 0QOH4
maleamz

monozum .OLSmmoe omsommOL 63p Lou msmme Hamo tom mance

64

 

 

 

 

 

 

 

 

 

 

 

e, t : i ......
m ooém omém 8.: om.mm
6.
1 P
. a e
4% M“ 1m m” ... , ,. .1 . ... . . ., ,
up m m SJN $.S $.Nm 8.8
..D .1 e
”u .r R.
.qo .e
.o “E e
_x f a 3.8 8.3 8:3 8.3
e “O M” T
S 3..
1 x
a .e
m S
e e . .. . .
F 1 o o o o
m on ma S 9 mm 3 mm 3
.e
F
mocofics< ocoa< mommaos< onoa<
mnasammq
moccaos< onH<
mauspw3

Hamomp :50: Smfiawsm .OLSmmms omnommoa exp pom mamOe HHOU lam manme.

 

65

 

 

 

 

 

 

 

 

 

 

 

.e; -
m om; 26. $6 $6
8
.I .r
a ,e
a A .... m
...... m m R6 R6 $6 $6
.6 .1 e
u r F
S _e
P
x f m. 86 $6 $6 86
e 0 _w“
5 e
l x r
a .e
m S
e We
F a 84 mm; mm; on;
m
00:6H054 onoa< cocoaus< ozoa¢
mcacnmoq
mosoausa _ Ocoa<
manage:

meamSppcd .OpSmmos Oncogmoa 02p how mamas Hamo INN OHQME

 

F59192989

HIGAN STATE UNIVERSITY LI||BRAR

JII II

III IIII Willa