Information processing, situation specificity, and the generality of risk-taking behavior

Abstract

Constructed 2 structurally similar risk-taking tasks to evaluate intertask consistency of individual differences. Only the mode of response differed between tasks. A total of 92 college students served as Ss. In 1 task, Ss chose their preferred bet within each of a number of pairs of bets. In the other, Ss set selling prices for these same bets. A measure of S's preference for "long-shot" gambles was obtained from each response. Reliable individual differences were found for each measure. However, the intermeasure correlation was relatively low considering the high degree of similarity between tasks. It is argued that the 2 response modes triggered different methods of processing information about probabilities and payoffs in a way that perturbed individual differences and reduced intertask consistency. Information-processing considerations may be one important component of the situation specificity prevalent in risk-taking behavior. Results suggest that high correlations are unlikely between risk-taking measures in structurally different settings or between risk-taking and other behaviors. (24 ref.) (PsycINFO Database Record (c) 2012 APA, all rights reserved)

Full text

Journal
of
Personality
and
Social Psychology
1972,
Vol.
22, No. 1,
128-134
INFORMATION
PROCESSING,
SITUATION
SPECIFICITY,
AND
THE
GENERALITY
OF
RISK-TAKING
BEHAVIOR1
PAUL SLOVIC
2
Oregon
Research
Institute
An
effort
was
made
to
construct
two
structurally similar risk-taking tasks
in
order
to
evaluate
intertask consistency
of
individual differences. Only
the
mode
of
response
differed
between tasks.
In one
task, subjects chose their preferred
bet
within each
of a
number
of
pairs
of
bets.
In the
other, they
set
selling
prices
for
these same bets.
A
measure
of the
subject's preference
for
"long-shot"
gambles
was
obtained
from
each
response. Reliable individual
differences
were
found
for
each measure. However,
the
intermeasure correlation
was
relatively
low
considering
the
high degree
of
similarity
between
tasks.
It is
argued
that
the two
response modes triggered
different
methods
of
processing information
about probabilities
and
payoffs
in a way
that perturbed individual
differences
and
reduced
intertask
consistency. Information-processing considerations
may
be one
important component
of the
situation
specificity
prevalent
in risk-taking
behavior.
These results imply that high correlations
are
unlikely between risk-
taking measures
in
structurally
different
settings
or
between risk taking
and
other
behaviors.
Kogan
and
Wallach
(1967)
posed
a
ques-
tion
fundamental
to our
understanding
of
risk-taking behavior: "How important
are
individual
consistencies relative
to
situational
constraints when
it
comes
to
predicting some-
one's
inclination
for
risk
or
conservatism?
[p.
208]."
Slovic
(1962)
found
little correlation
among several
different
kinds
of
risk-taking
measures administered
to the
same subjects.
This
result
implied
that
most
of the
variation
among individuals
was
situationally deter-
mined,
and it
casts
doubt upon
the
existence
of
a
general
trait
of
risk-taking propensity.
Since then, this failure
to find
transsituational
generality
for
individual differences
in
risk
taking
has
been replicated
in
studies
by
Bfichacek
(1968), Flanders (1970), Good-
1
This research
was
supported
by the
Personnel
and
Training Research Programs, Psychological Sci-
ences
Division,
Office
of
Naval Research, under con-
tract N00014-68-C-0431, Contract Authority
NR-
153-311,
and by
Grants
MH
12972
and MH
1S414
from
the
National
Institute
of
Mental Health.
The
author
is
indebted
to
Douglas
MacPhillamy
and
Barbara
Madigan
for
assistance
in
conducting
this
experiment
and
analyzing
the
results
and to
Robyn Dawes, Lewis Goldberg, Sarah Lichtenstein,
Douglas
MacPhillamy, Leonard Rorer,
and
Bernard
Weiner
for
their
comments
on the
manuscript.
2
Requests
for
reprints should
be
sent
to
Paul
Slovic,
Oregon Research Institute, P.O.
Box
3196,
Eugene, Oregon 97403.
man
(1970),
Greene (1963, 1964), Higbee
(1971),
Johnson (1963), Weinstein (1969),
and
Weinstein
and
Martin
(1969).
Some
investigators, such
as
Cohen
and
Christianson (1970), have reacted
to
this
negative experimental evidence with disbe-
lief,
arguing
that
it flies in the
face
of
per-
sonal experience that
attests
to the
coherence
of
individual behavior. Others have criticized
these studies
in
more
specific
ways
in an at-
tempt
to
explain
why
they failed
to find the
expected levels
of
generality.
One
criticism
is
that
they have failed
to
include
a
proper
sampling
of
risk-taking behaviors. However,
the
studies cited above have employed
a re-
markably thorough
assortment
of
risk-taking
tasks
in
problem-solving, athletic, social, voca-
tional,
and
pure gambling situations. Only
those
tasks
highly similar
in
structure
and
involving
the
same sorts
of
payoffs
(e.g.,
all
financial, all
social,
etc.)
have
shown
any
generality (e.g., Slakter, 1969), and,
as
simi-
larity decreases, these intertask correlations
rapidly decrease (Kogan
&
Wallach,
1964;
Weinstein, 1969).
Researchers
such
as
Flanders
(1970),
Ko-
gan and
Wallach (1964),
and
Slovic (1964)
have hypothesized
that
more generality would
exist
among
tasks
which provided real, rather
than hypothetical,
payoffs.
Higbee
(1971)
has
proposed
just
the
opposite—that
imagined
or
128