eta
cho
(1967) 353-372; @J North- lkand Publishing Co., Ams%erdam
Not to be reproduced by phoaoprint or microfilm without written permission from the publisher
cho
ical Labomtory,
University
of Ut
cht, The Netherlands
*
(1953; 1954), Cxm~~s and
have established that i.3 a ga
RDS
er low probability of large gains,
ability, or high probability of s_rnallgains. This is
terlmed ‘probability preference’. Probability preference is indicative of
someone’s utility for risk. Those who prefer Ilow probabilities to gain
great sums are considered; t e gamblers or speculators; those who
prefer intermediate probabilit s the calculated risk-takers; those who
high probabilities for small gains the cautious.,
INSON (1957) stimulated much r
ch on the relation between
need for achievement and probability
rence, especially in situations in which subjects be!ieved they could influence the results by
their own efforts. Atkinson’s model for predicting risk-taking behavior
comprises six variables: th.e subjective probability of success (p,r,
the subjective probability of failure (I+), the incent;ve value of success (Is),
ve value of failure (If), the need for achieve.ment (MS), and
the motive to avoid failure (I&). The variables are combined multiplicatively in the following equaiion : resulting rnotiva tion =
= Me x Ps x 1, -+ Mf x Pf x .Tf.
This function has a mtiximum when PS ==0.5 if M8 > Mf and a minimum
when Ps = 3.5 if MI : M8.
The motivation is assumed to be strongest when uncertainty regardthe outcome is greatest, that is if 4) = 0.5. Those for whom
f will prefer tasks of intermediate difficulty, with a moderate
e need for achievement is lowest in very high or
risk of failure.
very low probabilities. Those for whom Mf > M, will set either a very
high or vc:ry low level of aspiration ar -“cthus minimize their fear of
failure. reference for low and preference for high probabilities are
* Now at the University of Groaingen.
353
.
o have, moreover, a marked.y low
e siudy reported here was ca ‘ed out in 1961,!‘62 by us to test this
study of motivation to avoid ilure was, on account of the object of
ettable.
our study ‘risk-taking behavior’, cot carried out. This is a
356
failure.
failure, but who have a ve
skill orientation -
ypotkeses to be tested are:
eople considered to be “gamblers’ or ‘speculators
s will have in general a low need for achievement.
;is high need for achievement wi:L in situations which do
to their need for achieveme,?t., i.e. in gambling situations, avoid extr
risks and prefer small or moderate ones.
n a task where Ss are through instruction skill-directed,
higl\ need for achievement will reject extremely li
aspiration ~1stoo easy and etitremely tigh ones as too unrealistic and
fer intermediate ones which are realistic and give the satisfaction
‘e do no;: expect any positive correlation
under skill orientation in individuals
negative correlation will
results under chance and skill orientation.
(3
lere will be a
sitive correlation between risk
er &ill orientation for peo le ~<th a low need fG
tation thalj under c
obabilities will be
eferred under skill
ce orientation, exce
nnce orientation, for whom the
ust try
to
win as
me between
This
two altern
and a chance of 0.9 of losing 0.50 @.
and a chance of 0.5 of losing 02% G.
ng cards: 1 red and
card will be drawn.
hkk
cards. After
G, IX it is Iqlack you have lost
ut a circle
round
play. You must try to
he extent
ot’ your
the alternative
with
cIb :xr~
you
nc
refer.
NIW you must choose a
the two you want to
ible ‘.x to lose 9~ little as possible.
ain or loss wiil be determine
for those from the
by you will be played.
on the basis of these
ts for detmmini~g the probability bweference; expecte
G.in
amu& (Ci)
prob.
0.1
0.2
1.50
0.4
0.61
0.41
a.25
0.17
a.6
088
0.9
l4
0.1
0.2
0.4
8.6
aA8
0.9
1:.23
0.82
0.50
0.33
0.1
0.2
0.4
0.6
0.8
8.9
6.
4.00
2.45
1.63
1.
0.67
3.
2.00
prob.
0.9
0.8
0.6
.4
.2
0.1
1SO
0.9
0.8
0.6
0.4
0.2
0.1
0.33
0.50
0.82
1.23
2.
3.
OS
0.67
i.
1.63
2.45
4
1.33
2.
3.
4.
8.
12.
16
0.8
0.6
0.4
0.2
0.1
0.1
0.2
0.4
0.6
0.8
0.9
12.
8.
3.
2.
1.33
0.9
0.8
0.6
0.4
0.2
0.1
0.1
0.2
24.
16.
9.80
6.53
0.9
0.8
0.6
0.4
Q.8
0.9
4.
2.67
6.53
9.
16.
24.
ariance
1.2
1.8
0.9
Of
24.
5.76
2.1
0.09
points and prob. 0.8
2.0
2.0
150
6.
2.0
Of
ts an
a.
rob. 0.9
of
.
2.0
17 phts
and prob. 0.6
2.0
150
reached the criterion he
ade.
is t
is is a words;lake a sentence fro these words. Every word-maze consists of 5
columns of 12 words. ou have to begin by connecting a word
tEe first line with another word to form a sentence ending with a
lines
from the last line. ou have to coxrect the words by
wor
wtth a pencil. A sentence c:in only be ormed by connecting
from a certain line diagonally with a word on a higher or lo
in one of the neighboring columns. ‘hat is to say you ma
(stay on the sa e iine) and not vertically (stay in the
n), and you may not ski
rve COmake a sente
ou *begin with a wor
words with a pencil line.
e sentence will always end with a wor
first maze than.
zes were for prxtice.
ch,
_-. ___.
_._‘- __ _-
30
20
$3
.60
ecled value
in all
cases, but in ite
less than in the others.
in a new situation
a result of cauticm
ABLE
b ->
ean nu
rou
points under chance and s
A!
30
20
10
kance
65.55
85.80
orientation
kill
.EO
76.85
69.90
for the three groups.
P
0.001
0.028
< 0,025
TABLE 5
Sties for the t
Percentages of the chosen low, medium and hi
groups under chance and skilI orientation.
Chosen prob.
low
medkm
5igh
Group M
Skill
Chance
.125
.37l5
. ml
.3Oo
.52O
.I80
roup
Chance
ill1
.
ill
.775
.
.
A50
.
.
.
.Q7O
.
.
le difkrexzes within th
bilities under chance and 5
Thus all the hifkrences are in the predicted direction: under skill
~~~tatiun the level of the
e case: Ss with a
is now classified
e for the inter-
istribution of tbie Ss according to n
of
points in
experimentI
umber
I
I
<40...
41-68 s * .
>m
* _ *
_Total . . . .
L
<2
2-5 s>S
I---
--_
00
1 0
i712
1
!
8
I2
value of 0.76 and higher, 7 Ss
need for achievement.
need for achievement.
into groups of the same size
eference under skill
aMe 8 gives a survey of the res s. A low number of
level of aspiration.
TABLE 8
~~~~~~~~~onof Sk acxzolrdlingto n
ch and level of aspiration (ex
,,
’
ed for achievemen
allis test m2
not signif-
368
W. Cc
VAN DER MEER
DIscussIoN
Under chance orientation where the outcome d nds enfiifely on
&antx, s’s prefer the bets with the
value whatever
the outcome for the S i
their probability preference may be.
pendent
on his skill, differences in expected value play no part
soever.
It is ditlicult to decide accurately to what ext
variance inkenced the Ss’s choices. According to
the influence of difference in variance is negligible, when the probabilities or the expected values of the bets are not equal. The only way
to discover anything about the Muence of the variance is a co
of the choices in IV and VI (cf. table 2). The Ss from groups
choose in VI significantly more lower probabilities than in
there is a terLency for Ss with a preference for high probabilities to
prefer smalkr variances, WlGle Ss with a preference for intermediate
probabihties also prefer moderate variances (VAX DER EER, 1963).
If this tendency also operated under the conditions of experiment I,
it would result in a stronger preference for hi@ probabilities for cup
II and for intermediate probabilities for group
in IV but not in _VI.
is no relation ai all
However, this explanation is not correct, as t
between the choice of lower probabilities
than in IV and the
variance Freference of the Ss. The latter was determined along with
the §s’s ‘probability preference in the basis experiment, but will not be
discussed here.
Neither were there any differences in need for achievement between
the & who preferred lower probabilities in VI than in IV and those
for whom this was not the case.
ler possible explanation is that Ss in VI were scared off by the
possibility of a big loss and therefore chose a lower probability of gain.
Under skill orientation in group
there is the same tendency, but n&ot
is could indicate a tendency to avoid r’sk or failure.
But this explanation is ot entirely satisfactory, as the bets in the
basis ex iment were o
erhaps the effec
re and not iu t
cause of a contrast
effect with the first four items in which nothing could be lost.
LITTIG (1962) predkted that under skill orientation lower probabilities would be preferred than under chance orientation.
a non-significant difference in the predicted direction. Our findings
t of our study was
critical stand agains
s with low need for
5ks_d-,s(ri
Ss with
a high
ievei 9f aspiration.
elation between xtility for risk and need for achievment
ank correlation was found fcr I:e total group
Ice under ch ante orientation)
tween ut3lity for risk (probability pref
erence for IOWprobabilities
for achievement. Ss with a
) had a low need for achie
nt. Ss with ailow need for
referred low probabilities. Ss with a.high
nt predominant1
need for achievement a%vo high risks and preferrzd high or intermediate probabili;ies. Of the latter group of 17 Ss the-;e were 7 S!; with
an average probability preference > 0.76., that is wilthin the scclpz of
this study with a preference ft)r extremely high probabilities (tablL: 6).
elation between level (of aspiraGon an$ need for achievement
Between level of aspiration and need for achievement there is a
significant non-linear relation. Ss with a low need for achievement
prefer a high level of aspiration, Of the 10 Ss with a preference for
an extremely high level of aspiration, equal to a mean probability
<f 0.20, 8 Ss are in thus group, Ss with a moderate need for achievement prefer a low level of aspiration. Of the Ss uith prefereree for
an extremely low level of aspiration equal to a ean probability )s 0.76,
6 fall into the group with a moderat; level of aspiration. Ss with a
high need for achievement prefer a moderate level of aspiratiorl.
or thl= three risk-groups from the basis experiment separate rank
370
H.
6.
VAN DER MEER
cor&ations were calculated between level of aspiration and need for
t, Sign&ant negative correlations were foun
ation for group
There was no signiificant c
efer some risk, one can say
the S’s who under chance orientati
that the lower their need for achievement the higher their level of
aspiratiun.
Relation between utility for risk and level of aspiration
The above already indicates a relation between utility for risk an
the level of aspiration preferred. For the groups with a low and a
moderate need for achievement sign&ant positive rank correlations
were found, for the group with a high need for achievement there was
a sign.iGcant negative rank correlation.
Our Endings are clearly in conflict with Atkinson’s madel. They also
c&%c;:t with the findings of Atkinson a;7d his associates. But the latter
is probhably not really so. Bf we disregard the division into risk-groups
snd consider the total group a homogenous one, and if we had, in
additilon, distributed the Ss according to their need for achievement
into t-wo instead of three groups, then our results would have agreed
with those of Atkinson. A more thorough differentiation than the one
we applied might possibly lead to more interesting aspects. However
our number of Ss was too small for this.
Without investigati(on into the motivation to avoid failure it is not
possible to say, whether our results agree with those of Heckhausen.
For the time being we can only say they do not contradict them.
A clritical stand was taken to Atkinson”s model. A study was set-up as to
tb-: relation between need for achievement, probability preference under chance
orientation (utility for risk) and probability preference under skill orientation
(level of aspiration). The following relations were found.
(1) A significant negative rank correlation exists between need for achiev+
ment and utility for risk.
Ss with preference for low probabilities have a low need for achievement.
Ss wit” a low need for achievement prefer predominantly low probabilities. Ss
with :i high need for achievement avoid low probabilities.
(2) A. significant non-linear relation exists between need for achievement and
ip:vel of aspiration. Ss Mth 410~ need for achievement prefer a high, in some
cases an extremely high level of aspiration;, Ss Edith a modcri?te need for
achievement prefer a low, in some cases an extremely low level of aspiration;
Ss with a high need for achievement prefer a moderate level of aspiration.
%tive
rank correlation was found betwee
aspiration in the groups which have under chance orientatitin
obabilities. For the group preferring
ve correlation,
oderate need for achifivement there was
for the
oup with high nee:d for achievement
ei’el of aspiration.
chance orientation
er skill orientation
prefer high and intermediate probasignificantly more lower probabilities;
low probabilities prefer under skill
cantly rnt?re higher p,robabilities.
the highest expected value are prethe probabilj ty preference.
Under !Ml orientation
REF!ZRENCES
ATKINSON,J. W., 1957. Motivationzll determinants
of
risk-l,aking
behavior,
tasy, action, and society. Princeton.
analysis of human motivation in terms
9 1958. Towards experimental
of motives,
ctancies, and in
. In: J. W. Atkinson (Ed.):
Motives in
sy, action, and
---------,
1964. An introduction to motiv
Lrawx~, 1’960. The achieve) J. R. BASMAN, R. W. EARL and
ment motive, goal-setting and probability pre -:rc rices.. J. abnorm.
sot. Psychol. 60, 27-36.
LITWIN, 1960. Achievement motive andi test anxiety tollmotive to approach success and motive to avoid failure.
J. abnorm. sot. Psychol. 60, 52-63.
ARK, K. A., R. TEEVAN and H. RIICCIUTI, 1956. Hope of success and fe;ar of
failure as aspects of need for achievement. J. abncrm. sot. Psychol.
53, 182-t56.
Coo~ss, C. H. and D. G. PR~JIT’T,
1960. Components of risk iri dec
Probabilitv and variance preferences. J. exp. Psyc: hoi.
EDWARDS,W., 1953. PY*obability-preferences in gambling. Aml<:r. J.
349-364.
among bets with IJiflering expected
9 1954. Probability-preferences
values. Amer. 9. Psychoi.
, 1954. The reliability of probability
preferences.
Bmer. J. Ysychol.
7, 441-452.
954. Variance preferences
in gambling. Amerl J. Psychoi.
68-95.
HANCOCK, J. G. and R. C. TEEEVM, 1964. Fear of failure and risk-taking
behavior. J. Pers. 32, 200-209.
9
mp
372
H
C.
VAN DE
EER
urcht ini der Leistungsmotivation.
H., 1963. Hoffnung und
Meisenheim am Glan.
d chance orientations on probability
L.rEiG, L. W., 1.962,. E&c~s of skill
preferences. Psychol. Rep.
-9 1963. EfFects of motivation on probability preferences. J. Pess. 3
417-427.
MAHOKX,C. H., 1960. Fear of failure and unreaiistic vocational aspiration.
J. abnorm. sot. Psychol. 60, 253-261.
M~.IL~-R, G. and 3. B. SARASON,1952. A study of anxiety and learnin
J. abnorm. sot. Psychol. 4
with high and low need for
MCCLELLAND,
D. C., 1958. Risk takin
Motives in fantasy, action,
achievement. In J. W. A
and society. Princeton.
‘--,
J. W. ARSON,
R. A. CLARKand E. L. LOWELL,1953. The achievement motille. New York.
VAND&R MEER, H. C., 1963. Decision-making: The influence of probability
preference, variance preference and expected value on strategy in
gambling. Acta Psychol. 21, 231-259.
MEYER, H. H., W. B, WALKERand G. H. LITV~IN,1961. Motive patterns and
risk preferences associated with entr=epreneurship. J. abnorm. sot.
Psychol. 63, 570-57
PEIFXX,C. S., 1958. Values in universe of chance. New York.
e personality correlates of
SCO~EL,A., PH. RATOOSW
aqd J. S. MINAS,1959.
decision making under conditions of
Behav. Sci. 4, 19-28.
:~ECKXAUSEN,