Note on the Drawing Power of Crowds of Different Size

Abstract

Reports on the relationship between the size of a stimulus crowd, standing on a busy city street looking up at a building, and the response of passersby. As the size of the stimulus crowd was increased a greater proportion of passersby adopted the behavior of the crowd. Ss were 1424 pedestrians. The results suggest a modification of the J. S. Coleman and J. James (see 36:1) model of the size of free-forming groups to include a contagion assumption. (PsycINFO Database Record (c) 2012 APA, all rights reserved)

Full text

Journal
of
Personality
and
Social Psychology
1969,
Vol.
13, No.
2,
79-82
NOTE
ON THE
DRAWING
POWER
OF
CROWDS
OF
DIFFERENT
SIZE1
STANLEY
MILGRAM,3
LEONARD
BICKMAN,3
AND
LAWRENCE
BERKOWITZ
The
City University
of New
York
This
study
reports
on the
relationship
between
the
size
of a
stimulus
crowd,
standing
on a
busy city street looking
up at a
building,
and the
response
of
passersby.
As the
size
of the
stimulus
crowd
was
increased
a
greater
propor-
tion
of
passersby adopted
the
behavior
of the
crowd.
The
results
of
this
study
suggest
a
modification
of the
Coleman
and
James model
of the
size
of
free-
forming
groups
to
include
a
contagion assumption.
Ill a
typical
urban setting, when
a
group
of
people
engage
in an
action
simultaneously,
they have
the
capacity
to
draw others into
the
crowd.
The
actions
of the
initial group
may
serve
as a
stimulus
for
others
to
imitate
this action.
A
careful
analysis
of the
details
of
crowd
formation
is of
obvious interest
to a
society
in
which collective action plays
an
increasingly
important
part
in
social
life.
One
theoretical
formulation
that bears
on
this
problem
is
that
of
Coleman
and
James
(1961).
Coleman
and
James assumed
that
there
is a
"natural
process"
by
which
free-forming
groups acquire
and
lose members
and
thus
reach specific maximum sizes.
They
have
de-
veloped
a
model that generates
a
size distri-
bution
that
closely
approximates
the
actually
observed
size distribution
of
many thousands
of
groups.
The
central assumption
of
their
model
of
acquisition
and
loss
are "a
constant
tendency
of a
group member
to
break away,
independent
of the
group, thus
producing
a
his
study
arose
out of a
graduate seminar
in
social
psychology
conducted
by the first
author
at
The
City University
of New
York, Among those
who
took
part
in the
present study were Stuart
Baum,
Sheryl
Bruder,
Fay
Crayne, Victor
Ernoult,
Susan
Flinn,
Bert
Flugman,
Henry
Glickman,
Mi-
chael
Hoffman,
Marcia Kay,
Jo
Lang, Elaine Lieber-
man,
Nicholas
Papouchis,
Arthur
Shulman,
Henry
Solomon,
Sheila Sperber,
and
Mark
Silverman.
The
study
was
supported
by The
City University
of New
York
and by a
small
grant
from
the
National
Insti-
tute
of
Mental Health, Number 16284-01.
2
Requests
for
reprints
should
be
sent
to
Stanley
Milgram,
Graduate Center,
The
City University
of
New
York,
33
West
42nd
Street,
New
York,
New
York 10036.
3
Now at
Smith College, Northampton, Massa-
chusetts.
loss rate
for the
group proportional
to
size;
and an
acquisition rate
for
each group pro-
portional
to the
number
of
single individuals
available
to be
'picked
up [p.
44].'
"
Thus
the
growth
of a
group
is
independent
of the
size
of
the
group
and
dependent only upon
the
number
of
persons
who are
available
to
join
the
group. However, Coleman
and
James
pointed
out
that
"a
contagion
assumption—
that
is, an
assumption
that
a
person
is
more
likely
to
join
a
large group than
a
small
one
[p.
44],"
might
be
needed
in
their model.
(Their
use of the
term "contagion"
is not
entirely accurate, since this term does
not
signify
in any
direct
way
that
a
large group
is
more
effective
in
attracting
new
persons than
a
small one.
It is
preferable,
in
this connec-
tion,
to use the
phrase
"assumption
of
initial
group
size.")
This
paper reports
on the
effects
which
crowds
of
different
sizes
had on
passersby,
following
the
quantitative approach
to the
study
of
crowd behavior
outlined
by
Milgram
and
Toch
(1969).
A
few of the
basic concepts used
in
this
study need
to be
clarified.
First
there
is the
stimulus crowd.
This
was
provided
by the in-
vestigators
and
varied
in
number
from
1 to
IS.
If the
crowd
is to
draw onlookers, then
it
must
be
exposed
to an
available population.
The
population
may be
finite,
and
thus
ex-
haustable,
or it may be
continually replen-
ished
as in the
present study.
The
population
may
also
be in
various states
oj
activity,
that
is,
sitting around
(as at a
beach)
or
moving
along paths.
The
available population
in the
case
of the
present study consisted
of the
stream
of
pedestrians moving along
a
major
city
throughfare.
Finally,
the
crowd must
79

80
S.
MlLGRAM,
L.
BlCKMAN,
AND L.
BERKOWITZ
exhibit
some
sort
of
observable action
that
the
population
can
imitate
or in
some manner
respond
to. In the
present
study
the
stimulus
crowd
stood
on the
pavement
and
looked
up.
at the
window
of a
nearby building.
This
action,
or
parts
of it,
could
be
adopted
by the
passersby.
The
passerby could simply look
up
at the
building where
the
crowd
was
star-
ing
without breaking
stride,
or he
could make
a
more
complete
imitative
action
by
stopping
and
standing alongside
the
crowd.
Analyses
were
undertaken
for
both
types
of
responses.
In
sum,
the
investigators wanted
to see in
what degree crowds, varying
in
size
from
1 to
15
persons,
and all
performing
the
same
ob-
servable action, would draw persons into their
activities.
METHOD
Subjects
The
subjects were 1,424 pedestrians
on a
busy
New
York City street
who
passed along
a
50-foot
length
of
sidewalk during
thirty
1-minute
trials.
The
study
was
conducted
on two
winter
afternoons
in
1968.
Procedure
A
50-foot
length
of
sidewalk
was
designated
as
the
area
of
observation.
At a
signal,
flashed
from
the
sixth-floor window
of an
office
building across
the
street
from
this area
of
sidewalk,
a
group
of
confederates
(stimulus crowd) entered
the
middle
of
the
observation
area, stopped,
and
looked
up at
the
sixth-floor window.
Thij
gaze
was
maintained
for
60
seconds.
At the end of
this period
the
group
was
signaled
to
disperse.
After
the
area
was
cleared
of
the
gathered crowd
the
procedure
was
repeated
using
a
different
size
stimulus crowd. Five randomly
ordered
trials were conducted
for
each
of the six
different
size stimulus crowds.
The
stimulus crowds
were
composed
of 1, 2, 3, S, 10, and 15
persons.
Motion
pictures were taken
of the
observation
area
for
the 60
seconds during which
the
stimulus crowd
maintained
its
gaze
at the
window.
Data Analysis
The
motion pictures were analyzed
to
determine
the
total number
of
persons
who
passed through
the
observation
area
and
their behavior. Pairs
of
judges
counted
the
number
of
persons entering
the field;
within
this group,
the
number
of
persons
who
looked
up; and finally the
number
of
persons
who
stopped.
RESULTS
The first
question
is
whether
the
number
of
persons
who
stop
alongside
the
crowd
in-
creases
as the
size
of the
stimulus
crowd
in-
creases.
The
data
are
provided
in
Figure
1
(broken
line). While
4% of the
passersby
stopped alongside
a
single individual looking
up,
40%
of the
passersby stopped alongside
a
stimulus crowd
of
IS.
An
analysis
of
vari-
ance
was
performed
on the
mean percentage
of
persons
who
stopped alongside
the
crowd
100i
90.'
80
70
60
50
40
30
20
10
who
look
up
who
stop
.---•*'
>----•""
235
10
SIZE
OF
STIMULUS
CROWD
15
FIG.
1.
Mean percentage
of
passersby
who
look
up and who
stop,
as a
function
of
the
size
of the
stimulus crowd.

DRAWING
POWER
OF
CROWDS
81
(Table
1).
This
analysis indicates
that
the
size
of the
stimulus crowd significantly
affects
the
proportion
of
passersby
who
stand
along-
side
it.
But the
influence
of the
stimulus crowd
is
not
limited
to
those
who
stop
and
stand
alongside
it. For a
larger number
of
passersby
partially
adopt
the
behavior
of the
crowd
by
looking
up in the
direction
of the
crowd's
gaze, while not,
however,
breaking stride
and
standing alongside
it.
Here
again
the
influence
of
the
stimulus crowd increases along with
its
size.
While
one
person
induced
42
% of
the
passersby
to
look
up
(whether
or not
they
also stopped),
the
stimulus crowd
of
IS,
all
looking
in the
same direction, caused
86%
of
the
passersby
to
orient themselves
in the
same
direction (Figure
1,
solid
line).
An
analysis
of
variance again
confirms
the
difference
in
means
(Table
2).
A
trend
analysis
for
unequal intervals
was
performed
on the
data
(Gaito,
196S).
There
is
a
significant linear trend
(F =
101.7,
p <
.01)
and a
nonsignificant quadratic trend
(F
=
.42)
for the
passersby
who
stopped. How-
ever,
for the
passersby
who
looked
up,
there
are
both significant linear
(F =
57.2,
p<
.01)
and
quadratic
(F =
11.6,
p <
.01)
com-
ponents.
This
bears
on a
recent discussion
of
Gerard,
Wilhelmy,
and
Conolley
(1969).
In
their
study,
conformity
increased
in
linear
fashion
as a
function
of
group size,
in
con-
trast
to
Asch (1951),
who
found
a
curvi-
linear relationship.
The
present study shows
that
a
single
set of
group-size manipulations
can
generate
both
types
of
functions, depend-
ing
on the
specific dependent variable selected
for
analysis.
A
comparison
of
those
who
stop
and
those
who
look
up
shows
that
while both behaviors
increase with
the
size
of the
stimulus crowd,
TABLE
1
ANALYSIS
OF
VARIANCE
or
THE
PROPORTION
OK
PASSERSBY
WHO
STOP
AS A
FUNCTION
OF
THE
SIZE
OF THE
STIMULUS
CROWD
TABLE
2
ANALYSIS
OF
VARIANCE
OF THE
PROPORTION
OF
PASSERSBY
WHO
LOOK
UP AS A
FUNCTION
OF
THE
SlZE
OF THE
STIMULUS
CROWD
Source
Between
Within
Total
5S
.423
.099
.522
A!
S
24
29
MS
.085
.004
F
20.63*
Source
Between
Within
Total
55
.628
.187
.815
<y
5
24
29
MS
.125
.008
F
16.28*
>p
<
.001.
* p <
.001.
the
percentage
of
those
who
only look
up is
always
higher
than
those
who
stop,
regard-
less
of the
size
of the
stimulus crowd.
It
appears
that
the
more demanding,
in
time
or
effort,
the
behavior
the
less likely
it is
that
the
passerby
will
join
it.
Two
additional
points need
to be
made.
First,
it is
clear that while
the
effects
of a
precipitating
group
of a
given
size
for the
subsequent growth
of the
crowd were studied,
the
size
of the
stimulus crowd increased
as
soon
as
persons joined
it.
Thus,
the
effect
of
a
stimulus crowd
of
constant size
was not
studied.
In
order
to do
this
it
would
be
nec-
essary
to
withdraw
a
member
of the
stimulus
crowd
as
soon
as a
passerby
joined
it.
Second,
the
maximum size which
the
crowd
attains
is
dependent
not
only
on the
initial
size
of the
crowd,
but
also
on the
nature
of
the
stimulus
to
which
the
passerby
is
directed.
In the
present study, passersby were oriented
by
the
gaze
of the
crowd
to a
scene
that
had
no
special
holding
power.
(Pedestrians
looked
up to the
sixth
floor
of an
office
building
where some dimly perceived
figures
were peer-
ing
back
from
inside.
It was not a
scene
of
compelling interest.)
If,
instead,
an
acrobat
were
performing
on the
building ledge,
the
interest
of the
scene would likely hold crowd
members
for a
longer period
of
time,
and the
crowd
would grow
to a
larger maximum size
within
a
1-minute
interval
(the size
of the
crowd
at any
given moment being equal
to
the
initial stimulus crowd plus additions minus
withdrawals.)
There
is
some logical basis
for
joining
larger crowds:
all
other things being
equal,
the
larger
the
crowd
the
more likely

82
S.
MlLGRAM,
L.
BlCKMAN,
AND L.
BERKOWITZ
its
members
are
attending
to a
matter
of
interest.
The
results
of
this study show
that
the
number
of
persons
who
will
react
to, and
join
in,
the
observable
behavior
of a
stimulus
crowd
is
related
to the
size
of the
stimulus
crowd.
These
findings contradict
the
acquisi-
tion assumption
of the
Coleman
and
James
model.
The
acquisition
rate
is
not,
as
they
assume, dependent only upon
the
number
of
persons available
to
join
the
group. (For
the
present
study,
the
mean number
of
such indi-
viduals
was not
significantly
different
for the
different
size stimulus crowds.)
An
assump-
tion
of
initial group size
is
indeed necessary.
REFERENCES
ASCH,
S. E.
Effects
of
group pressure upon
the
modi-
fication
and
distortion
of
judgment.
In H.
Guetz-
kow
(Ed.),
Groups, leadership,
and
men.
Pitts-
burgh: Carnegie Press,
1951.
COLEMAN,
J.
S.,
&
JAMES,
J. The
equilibrium size
distribution
of
freely-forming
groups.
Sociometry,
1961,
24,
36-45.
GAITO,
J.
Unequal intervals
and
unequal
N
in
trend
analysis.
Psychological
Bulletin,
1965,
63,
125-127.
GERARD,
H. B.,
WILHELMY,
R.
A.,
&
CONOIXY,
E. S.
Conformity
and
group
size.
Journal
oj
Personality
and
Social
Psychology,
1968,
8,
79-82.
MILGEAM,
S.,
&
TOCH,
H.
Collective behavior:
Crowds
and
social movements.
In G.
Lindzey
&
E.
Aronson
(Eds.),
The
handbook
oj
social psy-
chology.
Vol.
4.
(2nd ed.) Reading,
Mass.:
Addi-
son-Wesley,
1969.
(Received February
6,
1969)