EXAMINING A NORMATIVE MODEL OF
STRESS USING CHOICE UNDER RISK
ZARMINA KHAN; BRYAN GRANT, MSC; JIM NEUFELD, PHD
Final sample: 26 participants, including 10 males and 16 females (Age M =
18.85, S.D. = 2.11)
PHASE 1: MEASURES
Desirability of Control
Intolerance of Uncertainty
Endler Multidimensional Anxiety
PHASE 2: LEARNING
Individuals learned the probabilities of each of 10 letters being paired with a
stressor (an 85 db, one second burst of white noise)
Letter stimuli
J
M
B
L
Z
P
V
A
G
D
Prob. of stressor 11% 22% 33% 40% 42% 44% 64% 67% 69% 71%
BACKGROUND
Purpose: test a formal model of coping with stress
called decisional control (DC)
DC: a process whereby individuals who find
themselves in multifaceted stressing situations make
probabilistic judgments to reduce the occurrence of
an adverse event (Lees & Neufeld, 1999)
Need for Cognition
Uncertainty Tolerance
General Decision-Making Style
Research question: assuming that DM’s have correctly
learned the probabilities of adverse event occurrence,
we are investigating whether environmental
experiences lead to noticeable model drift in decisionmaking over time
Two models of possible fit were explored:
1. Individualized: each individual’s subjective probability ratings
2. Conditional: objective probability ratings
ANOVA revealed no significant differences:
G2 values: Conditional (M = 103.79, SD = 63.57) and Individualized (M =
106.234, SD = 71.67)
Pearson 2 values: Conditional (M = 208.95, SD = 264.87) and Individualized (M
= 190.967, SD = 254.97)
Marginal Means of 2 Values for the Models
Marginal Means of G2 Values for the Models
108.00
215.00
106.00
106.23
104.00
103.79
208.95
205.00
195.00
190.967
102.00
185.00
100.00
175.00
PHASE 3: PRACTICE
•
•
•
•
Participants were presented with the rules of the DC framework
Within each architecture, there are two levels of choice: the top (bin)
level and the bottom (element) level
With the various combinations of choice available at each of the two
levels, a total of nine different architectures were presented
A green box indicates full choice, a red box indicates no choice, and a
grey box indicates an uncertain choice
Previous research: tested a formal, mathematically
specified model under uncertainty (Grant, 2016),
which involves situations in which a decision maker
(DM) has prior experienced knowledge pertaining to a
stressor’s occurrence
Present research: explore responding for choices
under risk, in which a DM is provided prior knowledge
regarding stressor occurrence instead of experienced
knowledge
RESULTS
Estimated Marginal Means
Novel research in the domain of decisional control has
recently tested a mathematically-specified, normative
model of coping with stress using choice under
uncertainty to understand the cognitive underpinnings
of choice and linked threat. The present research
tested the same normative model under risk using the
same methodology and a similar sample of
undergraduate students. An ANOVA revealed that the
three models, Individualized, Group, and Conditional,
performed similarly. A correlation analysis also
revealed individuals do not appear to differ on DC
amenability based on individual differences.
METHODS
Estimated Marginal Means
ABSTRACT
Conditional
Individualized
Conditional
Individualized
Model
Model
A correlation analysis revealed relationships among various psychometric
variables but not between model fit indexes and these variables
DOC
EMAS-PD
EMAS-SE
EMAS-NS
EMAS-DR
IUSTot
IUSF1
IUSF2
NFC
UTS
GDMS-D
GDMS-A
GDMS-S
GDMS-I
GDMS-R
GInd
PInd
GCon
DOC
EMAS-PD
-.44
EMAS-SE
-.28
.33
EMAS-AM
-.55
.51
.47
EMAS-DR
-.19
.21
-.05
.38
IUSTot
-.03
.35
.35
.50
-.10
IUSF1
.14
.20
.27
.25
-.14
.91
IUSF2
-.24
.40
.38
.65
-.10
.86
.60
NFC
-.09
.09
-.05
.26
.07
.48
.39
.45
UTS
-.14
.29
.48
.45
-.11
.62
.67
.47
.17
GDMS-D
-.49
.25
.36
.53
.04
.34
.09
.53
.11
.26
GDMS-A
.13
-.02
.18
.00
-.44
.28
.29
.15
.07
.27
.08
GDMS-S
.14
.05
.21
.23
-.23
.46
.31
.51
.43
.22
.09
.50
GDMS-I
-.34
.14
-.05
.30
.20
-.05
-.23
.07
.21
-.15
.61
-.07
-.13
GDMS-R
-.25
.24
.01
.36
.22
.07
-.10
.14
.25
-.03
.45
-.19
.01
.80
GInd
-.23
-.11
-.16
.27
.18
.10
-.10
.17
.04
-.10
-.12
.02
.24
.12
.12
PInd
-.15
-.14
-.08
.26
.16
.06
-.10
.15
-.09
-.05
-.15
-.03
.21
.06
.11
.94
GCon
-.15
-.05
-.12
.27
.22
.00
-.08
.11
.08
-.09
-.14
-.04
.19
.14
.10
.87
.12
PCon
-.03
-.08
-.09
.22
.18
.09
.06
.09
.07
-.01
-.19
.04
.19
.12
.13
.77
.02
.95
CONCLUSIONS
Figure 1. In a NU condition, participants have only information at the element level and neither information or control at the bin level.
They can select any letter within the group indicated, but the letter assigned to them is random within that group.
PHASE 4: TESTING
Participants completed three blocks of trials, with all nine architectures
presented 12 times within each block. Letter selection was left to
participants discretion
•Individuals perform well when probabilities of stressor occurrence are
known (i.e. under risk), and risky choices place less importance on
subjective expected utilities than uncertain choices do
•Individuals do not appear to differ on DC aptitude as a function of the
dispositional characteristics examined
If you have additional comments, questions, or inquiries regarding the present research, please
contact Zarmina at [email protected]