Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Choice biases in making decisions for oneself vs. others
(USC Thesis Other)
Choice biases in making decisions for oneself vs. others
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
Choice Biases in Making Decisions for Oneself vs. Others
by
Zhiqin Chen
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PSYCHOLOGY)
December 2018
ii
To my parents and three brothers
iii
Acknowledgements
I want to express my deepest gratitude to my advisor, Dr. Richard John, from whom I
learned about the subject of decision making. I became fascinated by all topics about decision
making and decided to pursue a career in this area. I also feel very lucky to have him as my
mentor in my PhD studies. Over the years, Richard has offered me great freedom, support,
encouragement, and guidance in exploring my research interests. Without his support and
guidance, I would not be where I am now.
I also want to express special thanks to my committee members, Dr. John Monterosso,
Dr. Antoine Bechara, Dr. Detlof von Winterfeldt, and Dr. Morteza Dehghani, for their valuable
time, thoughtful comments, and helpful suggestions. I appreciate all of my lab-mates and friends
in the department, Dr. Jinshu Cui, Dr. Kenneth Nguyen, Sarah Kusumastuti, Mengtian Zhao,
Matthew Baucum, Xiaobei Zhang, and others for their help and valuable comments in piloting
my studies.
iv
Table of Contents
Acknowledgements ................................................................................................................... iii
List of Tables............................................................................................................................. vi
List of Figures ......................................................................................................................... viii
Abstract ...................................................................................................................................... x
Chapter 1 Introduction ................................................................................................................ 1
1.1 Self-Other Decision Differences .................................................................................................... 1
1.2 Theoretical Determinants............................................................................................................... 2
1.3 Affect and Self-Other Decisions .................................................................................................... 4
1.4 Overview ...................................................................................................................................... 6
Chapter 2 Honoring Sunk Cost ................................................................................................... 8
2.1 Introduction ................................................................................................................................... 8
2.2 Study 1 General Others................................................................................................................ 15
2.2.1 Method ............................................................................................................................................... 15
2.2.2 Results and Discussion ........................................................................................................................ 17
2.3 Study 2 Specific Others ............................................................................................................... 21
2.3.1 Method ............................................................................................................................................... 21
2.3.2 Results and Discussion ........................................................................................................................ 24
2.4 General Discussion ...................................................................................................................... 28
Chapter 3 Preferences toward Risk, Loss, and Ambiguity ......................................................... 31
3.1 Introduction ................................................................................................................................. 31
3.2 Method ........................................................................................................................................ 36
3.3 Results ........................................................................................................................................ 49
3.3.1 Decision Dilemma 1 Risk Aversion toward Gains ................................................................................ 50
3.3.2 Decision Dilemma 2 Risk Seeking toward Losses ................................................................................ 53
3.3.3 Decision Dilemma 3 Loss Aversion ..................................................................................................... 56
3.3.4 Decision Dilemma 4 Ambiguity Aversion toward Gains ...................................................................... 59
3.4 Discussion ................................................................................................................................... 61
3.5 General Discussion ...................................................................................................................... 67
Chapter 4 Delay Discounting .................................................................................................... 70
4.1 Introduction ................................................................................................................................. 70
4.2 Method ........................................................................................................................................ 74
4.3 Results ........................................................................................................................................ 81
4.4 Discussion ................................................................................................................................... 87
Chapter 5 General Discussion ................................................................................................... 91
References ................................................................................................................................ 97
v
Appendix A: A 20-item Scale to Measure the Magnitude of Honoring Sunk Cost in Making
Decisions for Oneself in Chapter 2 ......................................................................................... 113
Appendix B: Decision Scenarios and Sample Choices Used to Study Delay Discounting in
Making Decisions for Oneself in Chapter 4 ............................................................................ 125
vi
List of Tables
Table 2.1: Summary of Means and Standard Deviations for the Number of Choices
that Honored Sunk Cost in Each Type of Decision in Study 1
18
Table 2.2: Matrix of Decision Target by Decision Response in Study 1
18
Table 2.3: Summary of Means and Standard Deviations for the Number of Choices
that Honored Sunk Cost in Each Type of Decision in Study 2
24
Table 2.4: Matrix of Decision Target by Decision Response in Study 2
24
Table 3.1: Option Pairs for All Choices in the Decision Dilemma of Risk Aversion
toward Gains
41
Table 3.2: Option Pairs for All Choices in the Decision Dilemma of Loss Aversion
41
Table 3.3: Summary of Each Range of the Certainty Equivalent (CE) and Its Related
Risk Aversion Index and Implied Preference in the Decision Dilemma of
Risk Aversion toward Gains
45
Table 3.4: Summary of Each Range of the Certainty Equivalent (CE) and Its Related
Risk Seeking Index and Implied Preference in the Decision Dilemma of
Risk Seeking toward Losses
46
Table 3.5: Summary of Each Range of the Acceptable Losing Amount (ALA) and Its
Related Loss Aversion Index and Implied Preference in the Decision
Dilemma of Loss Aversion
47
Table 3.6: Summary of Each Range of the Certainty Equivalent (CE) and Its Related
Ambiguity Aversion Index and Implied Preference in the Decision Dilemma
of Ambiguity Aversion toward Gains
48
Table 3.7: Summary of Means and Standard Deviations for the Index of Each
Preference in Each Type of Decision for All Four Decision Dilemmas
50
Table 3.8: Matrix of Decision Target by Decision Response
50
Table 4.1: Each Pair of the Sooner and Later Gains in Each Choice in the Small-
Distant and Small-Near Conditions
76
Table 4.2: Each Pair of the Sooner and Later Gains in Each Choice in the Large-
Distant and Large-Near Conditions
77
Table 4.3: Summary of Each Range of the Sooner Gain (SG) and Its Related Delay
Discounting Index and Implied Discount Rate in the Small-Near Condition
80
vii
Table 4.4: Summary of Means and Standard Deviations for the Delay Discounting
Index in Each Type of Decision for All Four Conditions
81
Table 4.5: Matrix of Decision Target by Decision Response
81
viii
List of Figures
Figure 2.1: Mean number of choices that honored sunk cost in the decisions made on
behalf of a close person, suggestions to a close person, and predictions of a
close person’s decisions in Study 1
20
Figure 2.2: Mean number of choices that honored sunk cost in friend’s and stranger’s
decisions with respect to suggestions to others and predictions of others’
decisions in Study 2
26
Figure 3.1: Flow of the titrated elicitation method presenting different for sure gains to
participants with a fixed risky choice (flip the coin: win $5,000 if heads or
$0 if tails) in the decision dilemma of risk aversion toward gains
43
Figure 3.2: Percentage of participants who chose the for sure gain (accept the offer:
receive $2,500 for sure) and percentage of participants who chose the risky
choice (flip the coin: win $5,000 if heads or $0 if tails) in each type of
decision in the decision dilemma of risk aversion toward gains
51
Figure 3.3: Mean risk aversion index for friend and stranger in the two decision
responses (suggestion and prediction)
52
Figure 3.4: Percentage of participants who chose the for sure loss (settle the case: pay
$2,500 for sure) and percentage of participants who chose the risky choice
(go forward with the appeal process: 50% chance of paying $5,000; 50%
chance of paying $0) in each type of decision in the decision dilemma of
risk seeking toward losses
54
Figure 3.5: Mean risk seeking index for friend and stranger in the two decision
responses (suggestion and prediction)
55
Figure 3.6: Percentage of participants who chose the status quo (not invest – $0 for
sure) and percentage of participants who chose the risky choice (invest:
50% chance of gaining $5,000; 50% chance of losing $5,000) in each type
of decision in the decision dilemma of loss aversion
56
Figure 3.7: Mean loss aversion index for friend and stranger in the two decision
responses (suggestion and prediction)
58
Figure 3.8: Percentage of participants who chose the for sure gain (accept the offer and
settle the case: receive $2,500 for sure) and percentage of participants who
chose the uncertain choice (wait for the investigation: receive full coverage
– $5,000 or nothing – $0, chances are unknown) in each type of decision in
the decision dilemma of ambiguity aversion toward gains
59
ix
Figure 3.9: Mean ambiguity aversion index for friend and stranger in the two decision
responses (suggestion and prediction)
61
Figure 4.1: Flow of the titrated elicitation method presenting different small sooner
(SS) gain options to participants with a fixed large later (LL) gain option of
receiving $100 one year from now in the small-near condition
79
Figure 4.2: Mean delay discounting indices for friend and stranger in the two decision
responses (suggestion and prediction) in the large-distant condition
84
Figure 4.3: Mean delay discounting indices for friend and stranger in the two decision
responses (suggestion and prediction) in the large-near condition
85
x
Abstract
This research includes exploratory studies examining self-other decisions related to six choice
biases: honoring sunk cost, risk aversion toward gains, risk seeking toward losses, loss aversion,
ambiguity aversion toward gains, and delay discounting. For each choice bias, self-other decision
differences were examined both between decisions made for oneself and suggestions offered to
others, and between decisions made for oneself and predictions of others’ decisions. By focusing
on the decisions made for others, this research examined the effects of the decision target (friend
vs. stranger) and the decision response (suggestion vs. prediction) in each choice bias. From a
representative US sample (N = 1063), this research observed diverse results across the above six
choice biases, which have not been adequately described in previous findings. This research
proposes a hypothetical affect-based framework to account for these results.
1
Chapter 1 Introduction
1.1 Self-Other Decision Differences
Differences in self-other decisions have been found widely in previous research. Doctors
have been found to make more conservative medical treatment decisions on behalf of their
patients compared to those they make for themselves (Garcia-Retamero & Galesic, 2012).
Parents were more risk averse in making health and safety decisions on behalf of their adolescent
children compared to their personal decisions (Dore, Stone, & Buchanan, 2014). Compared to
decisions people made for themselves, they were found to be less loss averse when they made
decisions on behalf of an unknown person, including lottery choices (Andersson, Holm, Tyran,
& Wengström, 2016), risk free choices, and social life decisions (Polman, 2012b). However,
Stone, Yates, and Caruthers (2002) identified no differential risk preference with respect to gains
in monetary decisions made for oneself and decisions made on behalf of a specific person (both
unknown and a close friend).
Compared to making decisions on behalf of others, offering suggestions to others occurs
more commonly. Differences between personal decisions and suggestions offered to others has
been discovered in diverse decisions and choices, including financial (e.g., Roszkowski &
Snelbecker, 1990), medical treatment (Atanasov, Anderson, Cain, Schulkin, & Dana, 2015;
Zikmund-Fisher, Sarr, Fagerlin, & Ubel, 2006), and relationship (Beisswanger, Stone, Hupp, &
Allgaier, 2003) decisions, and choices of jobs, college majors, college courses (Kray &
Gonzalez, 1999; Kray, 2000), gifts (Jonas, Schulz-Hardt, & Frey, 2005), and trips (Jonas & Frey,
2003).
With respect to predicting others’ decisions, previous research largely has studied risk
preferences. Compared to risk preferences in personal monetary decisions, Hsee and Weber
2
(1997) found that people predicted that others would be less risk averse. Predictions of others’
decisions were found to be more regressive (less risk averse toward gains and less risk seeking
toward losses) than were both personal monetary (Faro & Rottenstreich, 2006) and waiting time
(Krishnamurthy & Kumar, 2002) decisions. However, Stone and Allgaier (2008) found no
significant differences between decisions made for oneself and predictions of others’ decisions in
both relationship and monetary decisions. Many theories had been proposed to account for self-
other decision differences.
1.2 Theoretical Determinants
With respect to how people think about others’ decisions, theory of mind (Premack &
Woodruff, 1978) addresses the ability of understanding others in different situations and the
simulation heuristics proposed by Kahneman and Tversky (1982) accounts for how people
predict others’ behavior in different situations. However, these related theories or findings focus
only on understanding others’ feelings or behavior and do not directly account for differences
between personal decisions and decisions made for others in the same decision problems.
From the perspective of judging for self vs. others, people have been found to believe that
others are more susceptible to various cognitive biases (e.g., Armor, 1998; Ehrlinger, Gilovich,
& Ross, 2005; Pronin, 2007; Van Boven, Dunning, & Loewenstein, 2000; Wilson, Houston,
Etling, & Brekke, 1996), which is referred to as the phenomenon of the bias blind spot (Pronin,
Lin, & Ross, 2002). Pronin (2008) discussed further that people judge themselves based more on
their internal feelings and thoughts, but judge others based more on their external behavior or
appearance. However, evaluation for others differs from making decisions for others. According
to construal level theory (Trope & Liberman, 2010), evaluation for others involves higher-level
construals compared to making decisions for others, as the former implies a broader view and
3
more general assessment of others, including personality, behavior in the past and future etc.,
while the latter concerns only specific events.
With respect to making decisions for others, people are found to process information
differently in personal decisions and others’ decisions. Compared to personal decisions, people
have been found to put more weight on only a few important attributes (Kray & Gonzalez, 1999)
or perceived social values (social values theory; Beisswanger et al., 2003; Stone & Allgaier,
2008; Stone, Choi, de Bruin, & Mandel, 2013), be concerned more about positive/optimistic
outcomes (Beisswanger et al., 2003; Zikmund-Fisher et al., 2006), depend more on general
public preferences (Kray, 2000), and engage in more balanced information searches (Jonas &
Frey, 2003; Jonas, Schulz-Hardt, & Frey, 2005) when offering suggestions to others. People also
have been shown to exhibit more pre-decisional distortion, in that they favor information that
supports their choices, and less post-decisional distortion, in that they search for information that
matches their preferences (Polman, 2010), focus on promotion rather than prevention (Polman,
2012a), and generally search more information in terms of both alternatives and attributes of
each alternative (Liu, Polman, Liu, & Jiao, 2018) when making decisions on behalf of others in
contrast to personal decisions. People also were less likely to be affected by their original options
(status quo bias; Samuelson & Zeckhauser, 1988) in both decisions made on behalf of others
(Polman, 2012b) and suggestions to others (Lu & Xie, 2014), in contrast to personal decisions.
These studies have spanned a broad range of choice contexts.
In addition, with respect to the relations between psychological distance and construal
level theory discussed by Trope and Liberman (2010), a more socially distant other elicits high-
level mental construals characterized by more general and abstract views. However, this theory
was only used to explain the findings that people exhibited more creative performance, such as
4
more creative ideas and more insights (Polman & Emich, 2011), and paid more attention to
desirability than feasibility (Lu, Xie, & Xu, 2013) in decisions they made on behalf of others in
contrast to their personal decisions. The risk-as-feelings hypothesis (Loewenstein, Weber, Hsee,
& Welch, 2001) was used to account for self-other decision differences in risk preferences in
various decisions, including monetary (Hsee & Weber, 1997; Faro & Rottenstreich, 2006),
medical treatment (Garcia-Retamero & Galesic, 2012), and waiting time (Krishnamurthy &
Kumar, 2002) decisions. In summary, all of the above determinants in self-other decisions seem
to indicate context-dependent and domain-specific self-other decision differences.
1.3 Affect and Self-Other Decisions
Affect, which refers commonly to subjective emotional experiences (Lazarus, 1991), has
been found to play a key role in decision making (for review, see Lerner, Li, Valdesolo, &
Kassam, 2015). Many previous decision theories have considered affect an important
determinant in decision making, such as regret theory (Loomes & Sugden, 1982), theory of
disappointment aversion (Gul, 1991), SP/A theory (security-potential/aspiration; Lopes & Oden,
1999), and decision affect theory (Mellers, Schwartz, & Ho, 1997). Similarly, the more general
dual-process model (Kahneman & Frederick, 2002) proposes that emotional influences are as
important as cognitive influences in decision making. In addition, the neural theory – somatic
marker hypothesis (Bechara, 2004; Bechara & Damasio, 2005) confirms further the significant
role of emotion in decision making.
Two types of emotions are considered to influence decision making, including “expected
emotions”, which refer to emotions that people expect to experience in the future as a
consequence of decision outcomes, and “immediate emotions”, which refer to the current
emotions that people experience at the moment they make a decision (Loewenstein & Lerner,
5
2003). According to Lerner et al.’s (2015, Figure 2, p.815) emotion-imbued choice (EIC) model,
current emotions are caused largely by “expected outcomes”, “incidental influences”,
“characteristics of decision maker”, and “characteristics of options”. Therefore, for any decision
context, a certain level of affective intensity would be expected to be evoked at the moment of
decision making based on the outcomes expected and the characteristics of options. The intensity
of the emotions experienced is believed to influence decisions and behavior directly
(Loewenstein, 1996). When reaching an adequate level of intensity, the emotion is considered to
overcome cognitive and deliberative processing to influence decisions and behavior directly
(Loewenstein & Lerner, 2003).
With respect to self-other decisions, Albrecht, Volz, Sutter, Laibson, and von Cramon
(2011) showed that people displayed less emotional involvement when they made intertemporal
choices on behalf of an unknown person compared to when they made choices for themselves.
Polman and Vohs (2016) found that people reported they experienced more enjoyment while
making decisions on behalf of others than while making decisions for themselves. Kray (2000)
found that advisors experienced less regret than did personal decision makers. Garcia-Retamero,
Okan, and Maldonado (2015) found that depressed individuals and non-depressed individuals
behaved differently when they predicted others’ risky decisions. Wray and Stone (2005) found
that both self-esteem and anxiety served as moderators that enhance self-other decision
differences in risk preference in relationship decisions. All of these results suggest that self-other
decision difference might be based on affect.
6
1.4 Overview
This research includes exploratory studies about decisions made for oneself vs. others,
and attempts to account for self-other decision differences in a wide range of domains.
Compared to previous self-other decision research, this research differs in two major ways.
First, this research examines self-other decisions with respect to six different choice
biases. Chapter 2 includes two studies regarding the choice bias of honoring sunk cost in various
everyday decisions. People’s choices or decisions are recognized to honor sunk cost if they tend
to continue certain actions simply due to prior investment of money, time, or energy (Arkes &
Blumer, 1985). Chapter 3 covers four common preferences toward risk, loss, and ambiguity: 1)
risk aversion toward gains, 2) risk seeking toward losses, 3) loss aversion, and 4) ambiguity
aversion toward gains, respectively. When people face a choice between an option of receiving
$50 for sure and a risky choice with a 50% chance of winning $100 and a 50% chance of
winning nothing ($0), and they prefer the sure option, their choice will be considered to
demonstrate risk aversion toward gains. When people face a choice between an option of losing
$50 for sure and a risky choice with a 50% chance of losing $100 and a 50% chance of losing
nothing ($0), and they prefer the risky choice, their choice will be considered to demonstrate risk
seeking toward losses. When people face a choice between the status quo ($0) and a risky choice
with a 50% chance of winning $100 and a 50% chance of losing $100, and they prefer
maintaining the status quo, their choice will be considered to demonstrate loss aversion. When
people face a choice between an option of receiving $50 for sure and an uncertain choice with an
unknown chance of winning $100 or $0, and they prefer the sure option, their choice will be
considered to demonstrate ambiguity aversion toward gains. Chapter 4 studies delay discounting
with respect to expediting a future gain. The study of delay discounting examines intertemporal
7
choices between a smaller sooner reward and a larger later reward, such as receiving $50 now vs.
receiving $100 one year from now.
Second, this research studies making decisions for others more specifically. With respect
to the type of decision response, this research examines and compares decisions made on behalf
of others, suggestions offered to others, and predictions of others’ decisions. With respect to the
decision target, this research defines the other specifically. Based on the vividness of the mental
image, other is categorized first as a general or a specific other. A general other usually lacks
detailed descriptions in the scenarios, such as someone in the US, someone you don’t know, etc.
In contrast, a specific other in the scenarios can be a specific friend or a well-defined person.
Moreover, this research differentiates the general other further as either a close person or an
unknown person, and differentiates the specific other further as either a close friend or a stranger
whom people met for the first time. These distinct definitions of the decision response and the
decision target enable comparisons between decisions made for oneself and suggestions to
others, and between decisions made for oneself and predications of others’ decisions in general,
as well as investigations of the effects of both the decision target and the decision response in
decisions made for others in each choice bias. More details are included in each chapter.
8
Chapter 2 Honoring Sunk Cost
2.1 Introduction
Imagine that you and one of your close friends go out to dinner at a restaurant. After
dinner, you both order your favorite dessert, each of which costs approximately $10. Your
friend’s dessert is tasty, but very rich, and after 2 or 3 bites, he feels very full. However, he does
not want to take it home. Do you think that your friend should eat a few more bites or finish the
dessert before leaving? If he asks your suggestion, would you suggest that he eat more? What if
the desserts are free? If it was you, would you eat more?
People’s decisions are recognized to honor sunk cost if they tend to continue certain
actions simply because of prior investment of money, time, or/and energy (Arkes & Blumer,
1985). Take the initial example: if your friend chooses to eat a few more bites simply because he
paid for the dessert, that choice is considered to honor sunk cost. Various personal decisions had
been found to honor sunk costs, including business and investment (e.g., Arkes & Blumer, 1985;
Garland, 1990), medical treatment (e.g., Braverman & Blumenthal-Barby, 2012; Coleman, 2010)
and everyday (e.g., Frisch, 1993; Strough, Schlosnagle, Karns, Lemaster, & Pichayayothin,
2014; Olivola, 2018) decisions. However, honoring sunk cost potentially can result in terrible
consequences. For example, project managers may invest more in an unsuccessful ongoing
project rather than invest in a new and more promising project because of prior monetary
investment. Medical professionals may prefer to maintain an initial ineffective treatment rather
than search for a new and better one as a result of the efforts they invested in it. People may stay
in the theater and continue watching a boring movie rather than do something else for fun simply
because of the cost of the ticket, etc.
9
The phenomenon of honoring sunk cost also is referred to as the sunk cost effect or
fallacy, which is associated commonly with the idea of escalation of commitment (e.g., Staw,
1976; Staw & Fox; 1977). The two are correlated highly but differ from each other. Escalation of
commitment refers to “…the tendency for decision makers to persist with failing courses of
action” (Brockner, 1992), which emphasizes the (potentially) unsuccessful aspect of previous
actions and refers usually to studies of business decisions. Honoring sunk cost is only one causal
factor associated with escalation of commitment (e.g., Garland, 1990). Reasons other than sunk
cost, such as personal responsibility for failures (Staw, 1976; Conlon & Parks, 1987),
psychological connectedness (Gunia, Sivanathan, & Galinsky, 2009), and the percentage of
project completion (Garland & Conlon, 1998), etc. can stimulate escalation of commitment.
The sunk cost effect usually does not hold the constraints of (potential) failure of
previous actions or the specific decision context. It concerns essentially the effect of sunk costs.
For example, people are considered to demonstrate the sunk cost effect if they prefer a less
enjoyable, but more expensive Michigan ski trip to a more enjoyable, but less costly Wisconsin
ski trip simply because the former costs them more (see Experiment 1 in Arkes & Blumer, 1985).
Usually, research on honoring sunk cost studies personal decisions of continuing one’s initial
actions
1
either because of one’s own sunk costs (intra-individual sunk costs; e.g., Arkes &
Blumer, 1985; Thaler, 1980) or those of others (inter-individual sunk costs; e.g., Bornstein,
Emler, & Chapman, 1999; Braverman & Blumenthal-Barby, 2012; Olivola, 2018). The studies
1
Honoring sunk cost in personal decisions about continuing others’ initial actions usually have been studied in
business (e.g., Gunia, Sivanathan, & Galinsky, 2009) and medical treatment (e.g. Bornstein, Emler, & Chapman,
1999) decisions. Decisions in these studies always stress failures of the original decision maker’s actions, which
fundamentally are associated with the research on escalation of commitment as the continuation behavior could be
attributable to other issues rather than sunk cost, such as the relationship between the original and subsequent
decision maker.
10
reported in this chapter examine the sunk cost effect in everyday decisions with either one’s or
others’ intra-individual sunk costs in the context of making self-other decisions.
Differences in making self-other decisions with respect to risk preference have been
observed widely in various decisions, including monetary (e.g., Andersson, Holm, Tyran, &
Wengström, 2016; Hsee & Weber, 1997; Faro & Rottenstreich, 2006; Polman, 2012b;
Roszkowski & Snelbecker, 1990; Stone, Yates, & Caruthers, 2002), medical treatment (e.g.,
Garcia-Retamero & Galesic, 2012; Zikmund-Fisher, Sarr, Fagerlin, & Ubel, 2006), relationship
(e.g., Beisswanger, Stone, Hupp, & Allgaier, 2003; Stone & Allgaier, 2008), and physical safety
(e.g., Stone, Choi, de Bruin, & Mandel, 2013) decisions. People also have been found to
demonstrate different information processing, such as weighing attributes of a choice differently
(Kray, 2000; Kray & Gonzalez, 1999) and searching for different information (Buehler, Griffin,
& Ross, 1994; Jonas & Frey, 2003; Jonas, Schulz-Hardt, & Frey, 2005; Polman, 2010; Lu, Xie,
& Xu, 2013) in making decisions for others in contrast to making decisions for themselves.
However, research on the role of honoring sunk cost in making decisions for others, to say
nothing of differences in making self-other decisions, is extremely limited.
Bornstein et al. (1999) found that medical professionals did not demonstrate the sunk cost
effect when evaluating others’ medical treatment decisions and did demonstrate the effect when
evaluating others’ everyday decisions. However, evaluating others’ decisions differs from
making decisions for others studied in this chapter. Braverman and Blumenthal-Barby (2012)
found that medical professionals did not honor the sunk cost of their time, patients’ money (inter-
individual cost), or both on initial treatment when they offered their patients treatment
recommendations. However, their study did not investigate differential self-other decision-
11
making. The studies reported in this chapter were designed to examine differences in making
self-other decisions in the magnitude of honoring sunk cost in everyday decisions.
Affect and Honoring Sunk Cost
The primary underlying reasons proposed to account for honoring sunk cost include
waste aversion (e.g., Arkes, 1996; Arkes & Blumer, 1985; Bornstein & Chapman, 1995; Frisch,
1993), loss aversion (e.g., Kahneman & Tversky, 1979; Thaler, 1980), regret aversion (e.g.,
Gilovich & Medvec, 1994), and cognitive dissonance or self-justification (e.g., Bornstein &
Chapman, 1995; Brockner, 1992), etc. This rationale suggests that affect is associated strongly
with honoring sunk cost. Anger (Coleman, 2010) and mindfulness (Hafenbrack, Kinias, &
Barsade, 2014) also have been found to influence the magnitude of honoring sunk cost. In
addition, studies have found that honoring sunk cost is associated primarily with prior investment
of money rather than with time (e.g., Heath, 1995; Soman, 2001; Soster, Monga, & Bearden,
2010) or effort (e.g., Cunha & Caldieraro, 2009). Compared to time and effort, money is
identified more clearly, less indefinite, or easier to measure. As Loewenstein, Weber, Hsee, and
Welch (2001) discussed, compared to vague outcomes, vivid outcomes help evoke more
anticipatory emotions associated with the outcomes expected. Therefore, I would argue that past
monetary investment will induce stronger probable negative emotions associated with potential
loss than will past investment of time or effort, and honoring sunk cost is influenced greatly by
the probable negative emotions induced. Arkes and Blumer (1985) found that people were less
likely to demonstrate the sunk cost effect when evaluating an ongoing investment decision for a
company in which they had no personal involvement than was the company’s president who
made the same decision for the company. Thus, it is possible to infer that no personal
involvement in a company would induce almost no negative emotions involved about decisions
12
for the company, which results subsequently in a lower likelihood to honor sunk cost. Hence,
differences in the magnitude of honoring sunk cost would be expected to find in making self-
other decisions.
Overview and Hypotheses
The studies reported in this chapter address everyday decisions with intra-individual sunk
cost
2
—prior monetary investment—in the context of making self-other decisions to determine
whether there is a difference in the magnitude of honoring sunk cost between decisions made for
oneself and others. Hsee and Weber (1997) found that self-other decision differences in risk
preference existed only when people had an abstract or vague image of others (“other students in
the U.S.” or “other students on this campus”) and disappeared when participants had a vivid or
specific image of others (“look around and see who sits closest to you”). I define the former as
general others and the latter as specific others based on the vividness of mental images of others
and studied both cases (Study 1–general others, Study 2–specific others). The main comparisons
for each study are listed in the following. Based on Hsee and Weber’s (1997) findings about
general and specific others, I hypothesize first that the self-other decision differences in each of
the following related comparisons will be attenuated for specific others’ decisions in contrast to
general others’ decisions.
Decisions made for oneself vs. decisions made on behalf of others. In everyday situations,
adults may make decisions on behalf of their partner, young children, or aged parents, but
seldom on behalf of their friends or strangers. Therefore, I tested only those situations that
involve making decisions on behalf of a close person in Study 1. Stone and his colleagues
2
In the situation of making decisions for others, intra-individual sunk cost implies that the prior money was invested
by the same person who initiated the actions and sequentially continues or stops the actions. Participants in the
current research are asked to make decisions or offer suggestions to this person, or predict his/her decisions.
13
detected differential risk preferences between making decisions for oneself and on behalf of a
specific friend in relationship (Beisswanger et al., 2003; Stone & Allgaier, 2008) and physical
safety (Stone et al., 2013) decisions and proposed that social values theory (Stone & Allgaier,
2008) explains the differential decisions made on behalf of others that rely primarily on
perceived social value in a specific situation. Taking waste aversion as a social value, according
to social values theory, I hypothesize that decisions made on behalf of a close person will be
more likely to honor sunk cost compared to those made for oneself.
Decisions made for oneself vs. suggestions offered to others. Although Beisswanger et al.
(2003) pointed out that advisors focus more on positive outcomes of their friends’ relationship
decisions than their personal decisions, Dana and Cain (2015) indicated that advisors in general
would weigh adverse perspectives of outcomes heavily and offer conservative advice. However,
both arguments are still consistent with each other based on Kray and Gonzalez’s (1999) findings
that advisors place more emphasis on a few important attributes than do those making personal
decisions. Further, people were found to depend more on most people’s preferences (Kray, 2000)
and engage in more balanced information searches (Jonas & Frey, 2003; Jonas et al., 2005) when
offering suggestions to others compared to when making decisions for themselves. Taking either
waste aversion or loss aversion triggered by prior monetary investment as the most important
issue in making decisions with sunk cost, I hypothesize that suggestions offered to others would
be influenced more by sunk cost than those made for oneself.
Decisions made for oneself vs. predictions of others’ decisions. Hsee and Weber (1997)
found self-other decision differences in risk preference when participants were asked to predict a
general other’s decisions rather than those of a specific other and used the risk-as-feelings
hypothesis (Hsee & Weber, 1997; Loewenstein et al., 2001) to explain the differences.
14
Considering the relation between affect and honoring sunk cost discussed above, I hypothesize
that the magnitudes of honoring sunk cost between decisions made for oneself and predictions of
general others’ decisions are different in Study 1, but not in Study 2 about specific others’
decisions.
Friends’ decisions vs. strangers’ decisions. To include possible effects from
social/interpersonal distance (Trope & Liberman, 2010) in self-other decisions, friends’ and
strangers’ decisions were considered separately in this research. Typically, people feel more
socially connected to friends than strangers, which would make them treat friends’ decisions
differently from stranger’s decisions. Consequently, I hypothesize that the decision target (friend
vs. stranger) will have an effect on the magnitude of honoring sunk cost in everyday decisions.
Decisions made on behalf of others vs. suggestions offered to others vs. predictions of
others’ decisions. To reveal the possible effects of decision response, making decisions for
others was divided into three different decision responses, making decisions on behalf of others,
offering suggestions to others, and predicting others’ decisions. When making decisions on
behalf of others, decision makers usually assume certain responsibilities that involve a superior-
subordinate relationship, such as professional-client (e.g., Garcia-Retamero & Galesic, 2012;
Zikmund-Fisher et al., 2006) or parent-child (e.g., Dore, Stone, & Buchanan, 2014), etc. In
offering suggestions to others, the advisor always pays great attention to the probable outcomes
of the advice that recipients would face (e.g., Beisswanger et al., 2003; Dana & Cain, 2015).
Compared to decisions made on behalf of others and suggestions offered to others, predictions of
others’ decisions require people to take almost no responsibility, have any personal interactions
with the others, or pay extra attention to the probable outcomes the others may encounter.
15
Therefore, I hypothesize that the decision response will have an effect on the magnitude of
honoring honor sunk cost in everyday decisions.
Decisions made for oneself vs. predictions of one self’s decisions. People have been
found to perceive their future selves differently from their present selves (see Hershfield, 2011).
Although self-perceptions differ from decisions made for oneself, it would be interesting to
determine whether our predictions of our own decisions reflect or mis predict decisions we make
for ourselves. According to construal level theory (Trope & Liberman, 2010), prediction of
future experience is distant and at a high level of construal, and the actual experience is
situational and at a low level of construal. Therefore, I hypothesize that a differential magnitude
of honoring sunk cost exists between decisions made for oneself and predictions of one self’s
decisions.
2.2 Study 1 General Others
2.2.1 Method
Participants
A total of 323 undergraduate students were recruited from a university subject pool.
Among all participants recruited, 26 either failed to pass the attention check question or finished
the survey in an unexpectedly short time. These were excluded in the subsequent analysis. Thus,
the study had 297 participants (Mage = 20 years, age range: 18–34 years) with 198 females and 96
males
3
. Participants were assigned randomly to one of six decision groups
4
: making decisions for
oneself, predicting one self’s decisions, making decisions on behalf of someone with whom
3
Three subjects did not report their gender.
4
Decision group was treated as a between-subject factor for all chapters in this research. There are two main
reasons. First, each chapter in this research examined either five or six decision groups rather than two groups.
Second, between-subject design could help avoid carry over effects from within-subject design.
16
participants have a close relationship (a close person), offering suggestions to a close person,
predicting a close person’s decisions, and predicting an unknown person’s decisions
5
.
Materials and Procedure
Each participant completed a questionnaire online. Each questionnaire included eight
main pairs of analogous vignettes about everyday personal decisions, in which one had a high
prior monetary investment and one had no or a low prior monetary investment, which were
adapted from Strough et al.’s (2014; Study 1: Q1, Q2, Q3, Q5, Q8, Q9, Q12, Q15)
6
study to
measure the magnitude of honoring sunk cost. The analogous vignettes in each pair were
presented separately. Four pairs were assigned randomly to display no/low-cost vignettes first,
while the other four showed high-cost vignettes first. Following the first eight different vignettes
(one from each pair), 18 items about the need for cognition scale (Cacioppo, Petty, & Kao,
1984), 11 items about the regulatory focus questionnaire (Higgins, Friedman, Harlow, Idson,
Ayduk, & Taylor, 2001), and five items about the maximizing scale (Lai, 2010) were inserted as
fillers. Then, the other eight complementary vignettes were presented in the questionnaire. All
vignettes were changed slightly to correspond to each different situation of making decisions for
others.
This study examined the way people address general others’ decisions. The participants
were not given any specific descriptions about the other people in this condition. In the condition
of making a decision for a close person, each vignette began with “Think about someone with
whom you have a close relationship.” Then participants were asked to decide what that person
5
Making everyday decisions on behalf of an unknown person and offering suggestions to an unknown person were
not included here. In this study, I tested a “general other” without a vivid image of this person. Compared to the six
situations chosen here, these two situations appeared to be more unrealistic and lacked sufficient motivation for
participants to become involved in the situations.
6
The selection of eight items depends heavily on the adequacy of the fit between the vignette and the situation of
making decisions for others.
17
should do, recommend what that person should do, or predict what that person would decide to
do. In making a decision for an unknown person, each vignette began with “Someone you don’t
know…” and asked participants to predict what that person would decide to do.
Measurement
Each vignette ended with an action choice question with the five same action options
with different levels of continuation (see Strough et al., 2014) across all six decision groups. As
measured by Strough et al. (2014), within a pair of analogous vignettes in each decision group, if
a participant chose a higher level of continuation in the high-cost situation than in the no/low-
cost situation, this choice was considered to honor sunk cost. Therefore, each participant’s
magnitude of honoring sunk cost was measured by the number of choices that honored sunk cost
ranging from 0 to 8. This is the dependent variable in this study.
2.2.2 Results and Discussion
Table 2.1 summarizes the mean number of choices that honored sunk cost in decisions
for oneself (DFS), predictions of one self’s decisions (POS), decisions made on behalf of a close
person (DFC), suggestions to a close person (STC), predictions of a close person’s decisions
(POC), and predictions of an unknown person’s decisions (POU). In general, the numbers of
choices that honored sunk cost in all six types of decisions are relatively small. On average,
participants exhibited sunk cost effect in fewer than half of the items. Across all types of
decisions, DFC has the fewest number of choices that honored sunk cost and POU has the
greatest number of choices that honored sunk cost. To help better capture the main comparisons
and analyses in this study, a matrix of decision type was constructed based on the decision target
and the decision response, as shown in Table 2.2.
18
Table 2.1
Summary of Means and Standard Deviations for the Number of Choices that Honored Sunk Cost
in Each Type of Decision in Study 1
Decision Type N M SD
DFS 51 1.86 1.61
POS 50 2.18 1.48
DFC 49 1.49 1.37
STC 48 1.69 1.55
POC 49 2.20 1.66
POU 50 3.34 1.99
Note. The number of choices that honored sunk cost ranged from 0 to 8, and the higher number
indicated a higher magnitude of honoring sunk cost.
Table 2.2
Matrix of Decision Target by Decision Response in Study 1
Decision Response
Decision Target Decision Suggestion Prediction
Self DFS POS
A close person DFC STC POC
An unknown person POU
With respect to self-other decision differences in Table 2.2, three planned contrast
analyses were performed between DFS and DFC, DFS and STC, and DFS and the predictions of
general others’ decisions (combination of POC and POU; M = 2.78, SD = 1.91). The only
significant result was that between DFS and predictions of general others’ decisions, t(291) = -
3.25, p = .001, which implies that people are more likely to honor sunk cost in predictions of
general others’ decisions in contrast to personal decisions. This result can be explained by the
bias blind spot (Pronin, Lin, & Ross, 2002) that people believe that others are more susceptible
to various cognitive biases. The difference in the magnitude of honoring honor sunk cost
between DFS and predictions of general others’ decisions was consistent with my hypothesis, as
19
well as Hsee and Weber’s (1997) findings about risk preference in predicting general others’
decisions.
No significant results were found either between DFS and DFC, t(291) = 1.15, p = .25, or
between DFS and STC, t(291) = 0.54, p = .59, which implies that the magnitude of honoring
sunk cost in personal decisions did not differ from decisions made on behalf of a close person or
suggestions they offered to a close person. This result conflicts with my hypothesis of a
differential magnitude of honoring sunk cost in making decisions for oneself and on behalf of
others based on social values theory. Stone and colleagues proposed social values theory based
largely on studies of relationship (Beisswanger et al., 2003; Stone & Allgaier, 2008) and physical
safety decisions (Stone et al., 2013), while this study concerns everyday personal decisions. It is
possible to argue that social values theory is context-dependent and thus, does not work well for
the everyday decisions in this study. Another possible explanation would be that a vague image
of a close person may blur the line between oneself and that close person easily and have little
effect in self-other decision differences.
With respect to making decisions for others in Table 2.2, the results of a planned contrast
between POC and POU showed that the decision target had a significant effect on the number of
choices that honored sunk cost in the predictions of others’ decisions, t(291) = -3.49, p = .001,
which implies that a close person is predicted to be less likely to honor sunk cost than is an
unknown person. This is consistent with my hypothesis that social distance is important in
studying making decisions for others.
For all decisions related to a close person in Table 2.2, the results of a one-way analysis
of variance (ANOVA) among DFC, STC, and POC showed that the decision response did not
have a significant effect on the number of choices that honored sunk cost in decisions for a close
20
person, F(2, 143) = 2.85, p = .06, 𝜂
"
#
= .038, which implies that people display similar
magnitude of honoring sunk cost for all three types of decision response in making decisions for
a close person (see Figure 2.1). This result contrasted with my hypothesis that decision response
is important in making decisions for others. As discussed above, the differences among the three
responses (making decisions on behalf of others, offering suggestions to others, predicting
others’ decisions) are attributable primarily to different levels of responsibility, personal
interaction, and attribute weighting. However, a vague image of a close person might weaken a
decision maker’s responsibilities and feelings of connectedness. Therefore, it is possible that
there are no differences among the three decision responses in decisions for a close person.
Figure 2.1. Mean number of choices that honored sunk cost in the decisions made on behalf of a
close person, suggestions to a close person, and predictions of a close person’s decisions in
Study 1. The number of choices that honored sunk cost ranged from 0 to 8, and the higher
number indicated a higher magnitude of honoring sunk cost. The height of the bar represents the
mean. Standard errors are represented by the error bars attached to each column.
For personal decisions in Table 2.2, the results of a planned contrast between DFS and
POS did not show any significant difference in the number of choices that honored sunk cost,
21
t(291) = -0.98, p = .33. This result contradicts my hypothesis about the difference between
making decisions for oneself and predicting one’s own decisions. In this study, I changed the
choice question from “you decide to…” in the DFS only to “you think you would decide to…” in
the POS. This might not be sufficiently salient to allow participants to detect the difference
between personal decisions and predictions. Therefore, in Study 2, I modified the scenarios to
inform participants that each event/decision would take place in the near future to further
examine the difference between DFS and POS in the magnitude of honoring sunk cost.
In addition, the number of choices that honored sunk cost was not correlated significantly
with the need for cognition, regulatory focus, or maximizing measures in any decision group in
this study. Sex differences were found only in the POS, t(48) = 2.65, p = .011, which showed
that females (M = 2.47, SD = 1.39) were more likely to predict that they would honor sunk cost
than were males (M = 1.25, SD = 1.42).
2.3 Study 2 Specific Others
2.3.1 Method
Participants
A total of 263 participants were recruited from Amazon’s Mechanical Turk (AMT).
Among all participants recruited, 15 either did not pass the attention check question or finished
the survey in an unexpectedly short time and thus, were excluded in the subsequent analysis. The
final pool included 248 participants (107 males, 141 females). Participants’ ages ranged from 25
to over 65 and the annual household income ranged from less than $25,000 to more than
$150,000. All participants were assigned randomly to one of six decision groups, including
making decisions for oneself, predicting one self’s decisions, offering suggestions to a close
22
friend, predicting a close friend’s decisions, offering suggestions to a stranger, and predicting a
stranger’s decisions.
Materials
This study also used pairs of analogous vignettes about everyday personal decisions, in
which each pair included a no-cost and a high-cost prior monetary investment to measure the
magnitude of honoring sunk cost. Rather than 8 pairs of vignettes, there was a total of 20 pairs
(see Appendix A) in this study, including 10 adapted from Strough et al. (2014; Study 1: Q1, Q3,
Q5, Q6, Q7, Q9, Q11, Q12, Q13, Q15), and 10 created exclusively for this study. Largely, for
the 10 new items, I added life decisions related to modern technology, such as taking online
courses, playing video game, and replacing cell phones, etc. To motivate participants to become
involved in each decision group, especially in offering suggestions to others, a decision dilemma
was created across all decision groups for each vignette in this study regardless of the prior cost,
with respect to a personal struggle or conflicted issue, such as boredom vs. self-benefit, etc.,
when people choose to continue initial actions. The prior monetary investment remained the only
difference in each pair of vignettes. An action choice question followed each vignette. Each
question had three available action options
7
: (a) stop doing; (b) continue doing to some extent,
and (c) complete the action. The prior monetary investment varied from $15 to $500 across all
vignettes.
With the decision target of a close friend, participants were asked to think about a close
friend who has beliefs, attitudes, values, and interests similar to theirs and then write down the
7
Strough et al. (2014) used five action options in their study. Take their first question in Study 1 about eating dessert
as an example. They offered five action options: (a) stop eating the dessert entirely; (b) eat 1/4 of the dessert; (c)
eat 1/2 of the dessert; (d) eat 3/4 of the dessert, and (e) eat all of the dessert. It might be difficult for participants to
visualize 1/4, 1/2, or 3/4 of the dessert when making decisions, especially for others. Additionally, all actions in
(b), (c), and (d) indicate continuing initial actions to some extent. Therefore, in this study, I combined action
options (b), (c), and (d) into one action option—continue doing to some extent.
23
first name and the gender of that friend at the beginning of the questionnaire. Subsequently, both
the first name and gender information were inserted in each vignette and its action choice
question for the participant to read. This particular close friend was assumed to face all 20
everyday personal decision dilemmas and participants were asked either to offer suggestions to
this friend or predict his/her decisions. For the decision target of a stranger, each vignette
presented a unique explicit situation informing participants that they are meeting a stranger for
the first time and having a conversation with that person about an everyday personal decision
dilemma. 20 different strangers were assumed to face 20 different everyday personal decision
dilemmas separately. Half of the strangers were described as female and half as male.
Procedure
In this study, rather than finishing all 20 pairs of vignettes at one time, participants
worked on the no-cost and high-cost versions of the 20 pairs separately at an approximately one-
week interval. In each questionnaire, half of the vignettes were under the no-cost situation and
the other half were under the high-cost situation. Participants either were assigned randomly to
work on half of the specific vignettes with no prior monetary investment (no-cost situation) first
or work on the same half with a high prior monetary investment (high-cost situation) first. The
order effects between high-cost and no-cost versions were counterbalanced for all decision
groups.
Measurement
I coded each action option first as an action score: 1 indicated option (a)–stop doing; 2
indicated option (b)–continue doing to some extent, and 3 indicated option (c)–complete the
action. Then, for each pair of vignettes, if a participant demonstrated a higher action score in the
high-cost situation compared to the no-cost situation, the choice was considered to honor sunk
24
cost. The total number of choices that honored sunk cost (range: 0–20) was used to measure the
magnitude of honoring sunk cost for each participant in each decision group.
2.3.2 Results and Discussion
Table 2.3 summarizes the mean number of choices that honored sunk cost in decisions
for oneself (DFS), predictions of one self’s decisions (POS), suggestions to a friend (STF),
predictions of a friend’s decisions (POF), suggestions to a stranger (STST), and predictions of a
stranger’s decisions (POST). In general, the numbers of choices that honored sunk cost in all six
types of decisions are relatively small. On average, participants demonstrated sunk cost effect in
only about a quarter of the 20 items. Across all types of decisions, STF has the lowest number of
choices that honored sunk cost while POST has the highest number of choices that honored sunk
cost. To help better capture the main comparisons and analyses in this study, a matrix of decision
type was constructed based on the decision target and the decision response, as shown in Table
2.4.
Table 2.3
Summary of Means and Standard Deviations for the Number of Choices that Honored Sunk Cost
in Each Type of Decision in Study 2
Decision Type n M SD
DFS 46 5.57 2.96
POS 38 5.74 2.93
STF 36 4.08 2.42
POF 41 5.76 2.71
STST 37 5.49 2.36
POST 50 6.26 2.75
Note. The number of choices that honored sunk cost ranged from 0 to 20, and the higher number
indicated a higher magnitude of honoring sunk cost.
Table 2.4
Matrix of Decision Target by Decision Response in Study 2
25
Decision Response
Decision Target Decision Suggestion Prediction
Self DFS POS
A close friend STF POF
A stranger STST POST
With respect to self-other decision differences in Table 2.4, two planned contrast analyses
were conducted between DFS and suggestions to others (combination of STF and STST; M =
4.79, SD = 2.48), and between DFS and predictions of others’ decisions (combination of POF
and POST; M = 6.03, SD = 2.73). No significant results were found either between DFS and
suggestions to others, t(242) = 1.53, p = .13, or between DFS and predictions of others’
decisions, t(242) = -0.36, p = .72. The result of no difference in the magnitude of honoring sunk
cost found between DFS and predictions of others’ decisions is the same as Hsee and Weber’s
(1997) findings in their risk preference studies about specific others’ decisions. However, no
difference in the magnitude of honoring honor sunk cost between DFS and suggestions to others
differed from my hypothesis that there would be a difference between them. Typically, the cause
of the difference between DFS and suggestions to others in previous studies is attributed to
different information processing between them (e.g., Beisswanger et al., 2003; Dana & Cain,
2015). In each vignette in Study 2, I added a conflict that presented both the benefits and
drawbacks of continuing initial actions regardless of prior monetary investment. My original
hypothesis about a differential magnitude of honoring sunk cost between DFS and suggestions to
others was based on the assumption that prior monetary investment is the primary concern in
making decisions about sunk cost. The contradictory results suggest that, in addition to prior
monetary investment, both benefits and drawbacks of continuing initial actions might attract
personal decision makers’ and advisors’ attentions, and the idiosyncratic concern of attribute
results in no difference in honoring sunk cost.
26
With respect to making decisions for others in Table 2.4, the results of a 2 (decision
target: friend vs. stranger) X 2 (decision response: suggestion vs. prediction) between-subjects
ANOVA revealed a significant main effect of the decision target, F(1, 160) = 5.49, p = .02, 𝜂
"
#
= .033, showing that participants’ choices were more likely to honor sunk cost for a stranger’s
decisions (M = 5.93, SD = 2.60) than for a friend’s decisions (M = 4.97, SD = 2.69), and a
significant main effect of the decision response, F(1, 160) = 9.03, p = .003, 𝜂
"
#
= .053, showing
that participants’ choices were more likely to honor sunk cost in the predictions of others’
decisions than in the suggestions to others. There was no interaction between the decision target
and the decision response, F(1, 160) = 1.22, p = .27, 𝜂
"
#
= .008. These results (see Figure 2.2)
are consistent with my hypotheses about the effects of decision target and decision response in
making decisions for others.
Figure 2.2. Mean number of choices that honored sunk cost in friend’s and stranger’s decisions
with respect to suggestions to others and predictions of others’ decisions in Study 2. The number
of choices that honored sunk cost ranged from 0 to 20, and the higher number indicated a higher
magnitude of honoring sunk cost. The height of the bar represents the mean. Standard errors are
represented by the error bars attached to each column.
27
For personal decisions in Table 2.4, the results of a planned contrast between DFS and
POS did not show any significant difference in the number of choices that honored sunk cost,
which is the same result as that found in Study 1. Compared to Study 1, the vignettes about
predictions in this study were modified to inform participants that decisions take place in the
future (sometime after they see the question) and they need to make predictions. My hypothesis
about a difference in the magnitude of honoring sunk cost between DFS and POS was based on
the implications of construal level theory, that predictions of future experiences and actual
experiences differ. Previous studies have found that people mis-predicted their future behavior,
such as future craving for drugs (Loewenstein & Schkade, 1999), smoking (Slovic, 2000), etc.
Craving for drugs and smoking differ from the everyday personal decisions reported in this
chapter, as they involve addiction and impatience in the long term and fundamentally,
participants were asked to predict their future feelings and behavior related to habits. The other
possible reason for the absence of a difference between DFS and POS is that this study asked
people to predict events in the near future. People may not have different perceptions about their
present and future selves or different feelings between current events and events in the near
future.
In addition, the number of choices that honored sunk cost was not correlated significantly
with age, gender, income, or employment status in any decision group. However, education was
found to have a strong, negative correlation with the number of choices that honored sunk cost,
r(34) = -.56, p < .001, only in the STF. This implies that when offering suggestions to a friend,
people with a higher education are less likely to honor sunk cost.
28
2.4 General Discussion
In general, the numbers of choices that honored sunk cost in all types of decisions are
relatively low, with only about 25% of the time demonstrating sunk cost effect in both studies,
which implies no sunk cost effect in everyday decisions on average. However, the studies
reported in this chapter still revealed self-other decision differences in the magnitude of honoring
sunk cost in everyday decisions. Additionally, I developed a 20-item scale to measure the
magnitude of honoring sunk cost in everyday decisions that can be used to test both making
decisions for oneself and others. For everyday decisions in the near future, the difference in the
magnitude of honoring sunk cost between predictions of one self’s decisions and decisions made
for oneself was negligible.
With respect to self-other decision differences in the magnitude of honoring sunk cost in
everyday decisions, across both studies, the difference was found only between decisions made
for oneself and predictions of a general other’s decisions, particularly when the general other was
unknown. No self-other decision differences were found when people had a vivid image of
others. Hsee and Weber (1997) applied the risk-as-feelings hypothesis to account for their
findings about self-other decision difference in risk preference between a specific and a general
other. As discussed earlier, there is a strong relation between affect and honoring sunk cost. This
can be applied to partially account for the results reported in this chapter that people tend to pass
their feelings about sunk cost to others when they have a vivid image of others. With respect to a
general other’s decisions in the first study, in addition to Hsee and Weber’s (1997) findings of
risk preference toward a general unknown other’s decisions, the results suggest that the self-other
decision difference in the magnitude of honoring sunk cost might be mediated by the
disconnected feelings for general others from oneself. The results can infer further that self-other
29
decision difference in choice bias is mediated generally by the greater social distance when
people have a vague image of others. Future studies are expected to examine this mediation
effect.
Moreover, compared to the magnitude of honoring sunk cost in personal decisions,
people were found to predict general others to be more likely to honor sunk cost in the same
decision problems. This is consistent with the bias blind spot (Pronin, Lin, & Ross, 2002) that
people believe others are more likely to demonstrate cognitive biases. As no self-other decision
differences in the magnitude of honoring sunk cost were found between decisions made for
oneself and predictions of specific others’ decisions, the results suggest that the bias blind spot
might only fit in the situations about general others.
In making decisions for others, both studies found that the decision target (friend vs.
stranger) affected the magnitude of honoring sunk cost in everyday decisions regardless of
participants’ mental images of others. Compared to decisions for strangers, participants were less
likely to honor sunk cost in the decisions for someone with whom they have a close relationship,
either for suggestions offered to that person or for predictions of that person’s decisions. The
difference in the magnitude of honoring sunk cost here is consistent with our common sense that
generally, we treat friends’ decisions differently from those of strangers. Considering the
different level of social closeness between friends and strangers and the nonmonetary issues in
the decisions about sunk cost, the results reported in this chapter suggest that when a decision
includes attributes other than money, such as self-improvement, self-benefit, discomfort, or
boredom, etc., that involve motivations or emotions, people may weigh the attribute of money
lightly in their friends’ decisions.
30
The studies reported in this chapter also reveal that in making decisions for others, the
decision response (suggestion vs. prediction) affects only the magnitude of honoring honor sunk
cost in everyday decisions when people have a vivid or specific image of others. Compared to
suggestions offered to others, predictions of others’ decisions usually include minimal personal
involvement in the decisions or social interaction with others. However, the results reported in
this chapter suggest that when people have a vague image of others, the differential decision
response toward others’ decisions vanishes.
31
Chapter 3 Preferences toward Risk, Loss, and Ambiguity
3.1 Introduction
When our parking meter approaches its time limit, we may choose to stay in our office
for another five or 10 minutes and risk receiving a $60 ticket rather than leave a few minutes
earlier. We buy insurance to cover various possible losses, although sometimes the likelihood of
such losses is extremely small (e.g., earthquake insurance in California). In the stock market, we
hold onto losses with the hope of gaining and do not realize that we are accepting the risk of
losing more. We also are very likely to avoid an investment opportunity with equal chances of
winning and losing the same amounts. Many of our everyday decisions involve money and
certain degrees of risk or loss. Further, realistically, we seldom obtain precise probability
information about the occurrences of all possible outcomes in our everyday decisions, which are
considered to involve uncertainty or ambiguity.
For these decisions that involve risk, loss, or ambiguity, we seek others’ help and ask for
their suggestions, presumably to make better decisions. These others could be significant others,
relatives, close friends, colleagues, or professionals. This raises several questions. Do the
suggestions that we obtain from others differ from their personal decisions? What would others
think about our personal decisions? Would suggestions that others offer us differ from their
predictions of our personal decisions?
In various monetary decisions, self-other decision differences in risk preference have
been found between making decisions for oneself and making decisions on behalf of others (e.g.,
Polman, 2012b), between making decisions for oneself and offering suggestions to others (e.g.,
Roszkowski & Snelbecker, 1990), and between making decisions for oneself and predicting
others’ decisions (e.g., Hsee & Weber, 1997; Faro & Rottenstreich, 2006). People also were
32
found to be less loss averse when making monetary choices on behalf of others than when
making choices for themselves (Polman, 2012b). Monetary decisions involving risk, loss, or
ambiguity contain similar information that consists of monetary amounts (gain or loss) and
probabilities (known or unknown). Compared to previous studies, this chapter systematically
studies self-other decisions in four common preferences toward risk, loss, and ambiguity: 1) risk
aversion toward gains, 2) risk seeking toward losses, 3) loss aversion (Kahneman & Tversky,
1979), and 4) ambiguity aversion toward gains (Ellsberg, 1961).
Hsee and Weber (1997) found that when people had a vivid image of others, self-other
decision differences in risk preference between decisions made for oneself and predictions of an
unknown other’s decisions disappeared. However, for a specific other’s (physically present)
decisions, Polman (2012b) found self-other decision differences in loss aversion between
decisions made for oneself and decisions made on behalf of others. In our daily lives, we are
commonly involved in making decisions for a specific other rather than for a general other
8
(with
a vague or abstract image). Thus, this study focuses only on specific others’ decisions and
divides these decisions further into those of close friends and strangers. Considering that making
decisions on behalf of a specific other usually occurs between superiors and subordinates, such
as between a professional and a client (e.g., Garcia-Retamero & Galesic, 2012; Zikmund-Fisher,
Sarr, Fagerlin, & Ubel, 2006), or between a parent and a child (e.g., Dore, Stone, & Buchanan,
2014), the decision maker typically is required to assume certain responsibilities. With that true,
this study examines situations that everyone has experienced or engages in daily, including
offering suggestions to a specific other and predicting a specific other’s decisions.
Affect-Based Hypotheses
8
A general other is more likely to refer to a group of people such as someone in the US, someone on this campus,
someone you do not know, etc.
33
Current emotions, referring to as emotions or feelings experienced at the moment of
decision making, have been viewed as an important determinant in decision making (for review,
see Lerner, Li, Valdesolo, & Kassam, 2015; Loewenstein & Lerner, 2003). Loewenstein, Weber,
Hsee, and Welch (2001) proposed the risk-as-feelings hypothesis, arguing that emotional
responses to risky situations largely determine risk preferences. In addition, as Camerer (2012)
stated, loss aversion is considered to be associated highly with, and driven by, negative emotions,
and is “…an exaggerated emotional reaction of fear”. Compared to certain outcomes, van Dijk
and Zeelenberg (2006) found that uncertainty about outcomes reduces emotions in decision
making. All of these studies have suggested that there are relations between affect and
preferences toward risk, loss, and ambiguity.
The risk-as-feelings hypothesis has been applied in previous studies to account for the
absence of self-other decision differences in risk preference toward either gains or losses
between decisions made for oneself and predictions of specific others’ decisions in which people
were considered to transfer their emotions to a specific unknown or close other (e.g., Hsee &
Weber, 1997; Faro & Rottenstreich, 2006). Considering specific others’ decisions examined in
this study, I hypothesize first that there are no self-other decision differences in risk aversion
toward gains and risk seeking toward losses. However, in mixed gambles involving both gains
and losses, self-other decision differences were found in loss aversion between decisions made
for oneself and decisions made on behalf of specific unknown others (e.g., Andersson, Holm,
Tyran, & Wengström, 2016; Polman, 2012b). Hence, I hypothesize that there are self-other
decision differences in loss aversion. With the same magnitudes of outcomes, choices used to
study risk aversion toward gains and choices used to study ambiguity aversion toward gains are
nearly the same, except that the former have specific probability information and the latter do not
34
have. Frisch and Baron (1988) defined ambiguity as “…the subjective experience of missing
information relevant to a prediction”. Based on the hypothesis of no self-other decision
differences in risk preference toward either gains or losses in this study, I hypothesize
accordingly that there are no self-other decision differences in ambiguity aversion toward gains.
According to 3 et al.’s (2015, Figure 2, p.815) emotion-imbued choice (EIC) model,
current emotions that decision makers experience when making decisions are caused primarily
by “expected outcomes”, “incidental influences”, “characteristics of decision maker”, and
“characteristics of options”. Therefore, based on possible outcomes and features of options in
specific choices, it is possible to evaluate and compare the intensities of different current
emotions elicited by different choices. Hsee and Weber (1997) found that self-other decision
differences in risk preference were more salient when the stakes were larger than smaller.
Considering that outcomes with larger stakes
9
are more likely to provoke more intense emotions
at the moment of decision making than are outcomes with smaller stakes, their results suggest
potentially that the intensity of current emotions might influence self-other decision differences.
Based on the relations between affect and preferences toward risk, loss, and ambiguity discussed
above, I thus propose that affect-based hypothesis I accounts for self-other decision differences
in preferences toward risk, loss, and ambiguity, in which the more intense the affect potentially
elicited by choices that decision makers face, the more salient self-other decision differences
become.
People often experience intense emotions (both positive and negative) in close
relationships (for review, see Berscheid & Ammazzalorso, 2003) and close relationship blurs the
9
To compare decision differences across the four preferences toward risk, loss, and ambiguity, this study only
examines choices with large stakes. In addition, people are also more likely to seek help from others when their
decisions involve big amounts in contrast to small amounts.
35
boundaries between oneself and close others easily (e.g., Aron, Aron, Tudor, & Nelson, 1991;
Cialdini, Brown, Lewis, Luce, & Neuberg, 1997). Considering the relations between affect and
preferences toward risk, loss, and ambiguity, I hypothesize that the decision target (friend vs.
stranger) influences preferences toward risk, loss, and ambiguity in making decisions for others.
Advisors are always considered to pay great attentions to probable outcomes that their
advice recipients might face (e.g., Beisswanger, Stone, Hupp, & Allgaier, 2003; Dana & Cain,
2015), as they usually want to be liked by the advisees in order to maintain a good relationship
(Dana & Cain, 2015). Compared to suggestions offered to others, predictions of others’ decisions
involve almost no interpersonal communications and are less likely to be influenced by the affect
evoked by responsibilities or reputation protections. Therefore, I hypothesize that the decision
response (suggestion vs. prediction) influences preferences toward risk, loss, and ambiguity in
making decisions for others. Furthermore, considering the different level of affective intensity
involved in different decision target or different decision response, I propose affect-based
hypothesis II to account for the effects of both the decision target and the decision response in
making decisions for others, in which a higher level of affective intensity potentially elicited by
choices that decision makers face is more likely to result in the effect of either the decision target
or the decision response in preferences toward risk, loss, and ambiguity.
Chapter Overview
This study addresses four common preferences toward risk, loss, and ambiguity: 1) risk
aversion toward gains, 2) risk seeking toward losses, 3) loss aversion, and 4) ambiguity aversion
toward gains. Each preference was studied under a specific decision dilemma with a unique real-
life scenario. An index of each preference was applied to measure the intensity of each
preference for each participant. For each preference, the percentage of participants who chose
36
specific choices, which imply their related preferences toward risk, loss, or ambiguity, was
examined first. Then, self-other decision differences in the preference index were examined
between decisions made for oneself and suggestions offered to others, and between decisions
made for oneself and predictions of others’ decisions. In addition, the effects of both the decision
target and the decision response on the preference index in making decisions for others were
investigated. Furthermore, the correlations between the preference index and empathy scores,
numeracy scores, and demographic factors were studied separately. After presentations of all
results for the four decision dilemmas, discussion of evaluating and comparing the affective
intensities potentially elicited by choices that decision makers face across all four decision
dilemmas and of the way the results reflected all of the hypotheses discussed above were given.
Finally, a general discussion presents the main contributions of this work as well as its
limitations.
3.2 Method
Participants
A total of 272 participants were recruited from Amazon’s Mechanical Turk (AMT) to
finish an online questionnaire. Among all recruited participants, 20 either finished the
questionnaire in an unexpectedly short time or did not pass the attention check question, and thus
were excluded in the subsequent analyses. The final data set included 252 participants (130
males, 122 females). They are diverse in the age from 18 to over 65 and in the annual household
income ranging from below $25,000 to above $150,000. There are 46.82% (118 out of 252) of
participants who reported to be between 25 to 34 years old, and 45.63% (115 out of 252) of
participants who reported to have a Bachelor’s degree. Participants were randomly assigned to
one of five independent decision groups, including making decisions for oneself, offering
37
suggestions to a close friend, predicting a close friend’s decisions, offering suggestions to a
stranger, and predicting a stranger’s decisions.
Materials
Four separate decision dilemmas including scenarios and choices related to 1) risk
aversion toward gains, 2) risk seeking toward losses, 3) loss aversion, and 4) ambiguity aversion
toward gains were put into one questionnaire in a random order. In addition, 28 items regarding
the empathy scale (Davis, 1980), eight items regarding the subjective numeracy scale (Fagerlin
et al., 2007), and several demographic questions were added into each questionnaire.
For each decision dilemma, it contains a unique real-life scenario and a sequence of two-
option choices to measure the intensity of each preference. Scenarios for the decision dilemmas
of risk aversion toward gains and risk seeking toward losses were adapted from Hsee and Weber
(1997). Scenarios for the decision dilemmas of loss aversion and ambiguity aversion toward
gains were exclusively created for this study. All four decision dilemmas with one sample choice
in each with respect to making decisions for oneself are presented in this following.
Risk aversion toward gains. You bought a lottery ticket a week ago and are now informed
of winning a 50-50 chance to collect a $5,000 prize. However, the lottery has an option to forego
the 50-50 coin flip for $5,000, and instead accept a fixed prize amount for sure. You do not know
the exact amount of the fixed offer but are informed of a few possible offers. Your task is to
decide whether to accept the offer or flip the coin for the $5,000 prize. For each of the following
possible offers, please indicate whether you would accept the offer or flip the coin for the $5,000
prize.
For my lottery winnings, I would…
A. accept the offer: receive $2,500 for sure
38
B. flip the coin: win $5,000 if heads or $0 if tails
Risk seeking toward losses. You were recently involved in a lawsuit and lost the case.
You may appeal the court’s decision. Your attorney believes there is a 50% chance of winning
the appeal and paying nothing, and a 50% chance of losing the appeal and paying the full
damages ($5,000). The appeal would be handled for free by your attorney friend and not require
any more of your time in court. Your attorney offers a second option, to settle the case for less
than the damages awarded. You do not know the exact amount required to settle the case but are
informed of a few possible settlement amounts. Your task is to decide whether to take the
settlement or appeal the case. For each of the following possible settlement amounts, please
indicate whether you would settle the case or go forward with the appeal.
For my lost case, I would…
A. settle the case: pay $2,500 for sure
B. go forward with the appeal process:
50% chance of paying $5,000
50% chance of paying $0
Loss aversion. You are considering a financial investment. You consulted a financial
adviser a week ago and are now offered a few possible investment choices. Your task is to decide
whether to invest or not. For each of the following possible choices, please indicate whether you
would invest or not.
I would…
A. invest
50% chance of gaining $5,000
50% chance of losing $5,000
39
B. not invest
Ambiguity aversion toward gains. You recently had a car accident and fortunately there
were no personal injuries, but the estimated cost to repair the damages totaled about $5,000. As
you are unsure about your own degree of responsibility for the accident, the insurance company
offers you two options to cover the expenses related to the damages. One option involves an
investigation to determine which driver was at fault for the accident. If the investigation
determines that you are more than 50% at fault, you will receive nothing from the insurance
company. If the investigation determines that you are less than or equal to 50% at fault, you will
receive full coverage for damages ($5,000). The facts of the accident are in dispute, and there is
no way to be sure how the insurance investigation will come out. The other option is to skip the
investigation and settle the case for less than the total damages. You do not know the exact
amount of the settlement offer but are informed of a few possible offers. Your task is to decide
whether to accept the offer or wait for the results of the investigation. For each of the following
possible offers, please indicate whether you would accept the offer or wait for the investigation.
I would…
A. accept the offer and settle the case: receive $2,500 for sure
B. wait for the investigation: receive full coverage ($5,000) or nothing ($0), chances are
unknown
The scenario in each of the above decision dilemmas was adjusted slightly to fit into
other decision groups. Referring to close friends’ decisions, participants were first asked to think
about a close friend who had beliefs, attitudes, values, and interests similar to his/hers and then
write down the first name of this close friend. Subsequently, their friend’s name was inserted in
the context of scenarios for participants to read. For strangers’ decisions, each scenario contained
40
a unique vignette informing participants that they were meeting someone for the very first time
and starting a conversation with this stranger. Stranger is someone sitting next to them on a
plane, someone from a local hiking group, a new co-worker at a happy hour, or someone at their
friend’s birthday party separately in each scenario. Half of the strangers were described as
female and half as male.
For each sequence of two-option choices following each scenario in each decision
dilemma, the sequence format is the same as the studies of Hsee and Weber (1997), and Shupp,
Loveridge, Skidmore, Lim, and Rogers (2017). It is named as Multiple Price List (as cited in
Shupp et al., 2017). Each choice contains a pair of options including an option of a for sure small
amount and an option of a risky or uncertain choice with a large amount. For each choice in each
decision dilemma, one option was always fixed and option pairs were the same across all
decision groups. For all risky choices in all decision dilemmas in this study, they had two
outcomes and a constant probability of 50% was applied. The selection of equal probability of
two outcomes could potentially help minimize the requirement of numeracy skills from
participants to understand the risky choice, which actually helps bring less noise but similar
predictive power in the risk elicitation study compared to complex probability presentations
(Dave, Eckel, Johnson, & Rojas, 2010).
Table 3.1 shows the details of the option pairs for all choices in the decision dilemma of
risk aversion toward gains. All outcomes are positive. The amounts in each option pair for all
choices in Table 3.1 were also used in the decision dilemmas of risk seeking toward losses and
ambiguity aversion toward gains. However, instead of informing participants of gaining certain
amounts in both options, the decision dilemma of risk seeking toward losses told them that they
were losing these amounts. In other words, all outcomes are negative in the decision dilemma of
41
risk seeking toward losses. In the decision dilemma of ambiguity aversion toward gains, in
contrast to showing a specific chance (50%) of wining $5,000 in the decision dilemma of risk
aversion toward gains, participants were informed that the chance of receiving $5,000 was
unknown. In all of these three decision dilemmas, the risky or uncertain option was fixed and the
for sure amount varied from $500 to $4,500 with increments of $500 between two continuous
ones as shown in Table 3.1.
Table 3.1
Option Pairs for All Choices in the Decision Dilemma of Risk Aversion toward Gains
Option Pair
Choice For Sure Amount Risky Choice (50%)
1 $500 $5,000 or $0
2 $1,000 $5,000 or $0
3 $1,500 $5,000 or $0
4 $2,000 $5,000 or $0
5 $2,500 $5,000 or $0
6 $3,000 $5,000 or $0
7 $3,500 $5,000 or $0
8 $4,000 $5,000 or $0
9 $4,500 $5,000 or $0
For the decision dilemma of loss aversion, the option pairs for all choices are different
from the other three decision dilemmas, which are shown in Table 3.2. In this decision dilemma,
participants always faced a choice between an option of a for sure amount – “not invest” ($0, or
status quo) and a risky option – “invest”. The risky option is a mixed gamble with 50% of
winning $5,000 and 50% of losing some amount. The winning amount of $5,000 was fixed here
and the losing amount varied from $1,000 to $9,000 with increments of $1,000 between two
continuous ones as shown in Table 3.2.
Table 3.2
Option Pairs for All Choices in the Decision Dilemma of Loss Aversion
42
Option Pair
Choice For Sure Amount Risky Choice (50%)
1 $0 $5,000 or -$1,000
2 $0 $5,000 or -$2,000
3 $0 $5,000 or -$3,000
4 $0 $5,000 or -$4,000
5 $0 $5,000 or -$5,000
6 $0 $5,000 or -$6,000
7 $0 $5,000 or -$7,000
8 $0 $5,000 or -$8,000
9 $0 $5,000 or -$9,000
Procedure
After receiving a questionnaire online, each participant started with four sequences of
choices along with specific scenarios as to the four decision dilemmas and then answered
questions regarding the empathy scale, the subjective numeracy scale and several demographic
questions. In each decision dilemma, titrated elicitation method was applied to present the
sequence of choices to participants. Let’s look at the decision dilemma of risk aversion toward
gains as an example specifically.
As the option of the risky choice was always fixed as shown in Table 3.1, Figure 3.1
shows the details of the titrated elicitation method, which leads to 10 possible flow paths of
presentations of the for sure amounts that a participant might go through in the decision dilemma
of risk aversion toward gains. Participants started with the midpoint of the for sure amounts in
Table 3.1 and faced a choice between “accept the offer: receive $2,500 for sure” or “flip the
coin: win $5,000 if heads or $0 if tails”. If participants accepted the for sure offer here, the next
for sure amount they would be offered was the minimum one – $500. If participants rejected the
for sure offer here, the next for sure amount they would be offered was the maximum one –
$4,500. The presentation flow simply depended on participants’ choice in each option pair until
an ending happens. Therefore, though there were nine for sure amounts in total, participants were
43
merely offered either two or four ones as shown in Figure 3.1. Across all four decision
dilemmas, participants always started with the midpoint of the for sure amounts (or the losing
amounts in the decision dilemma of loss aversion) and then followed the presentation flow
guided by the titrated elicitation method.
Figure 3.1. Flow of the titrated elicitation method presenting different for sure gains to
participants with a fixed risky choice (flip the coin: win $5,000 if heads or $0 if tails) in the
decision dilemma of risk aversion toward gains. CE indicates a certainty equivalent of the fixed
risky choice.
44
The procedure of presenting choices to participants from a list of two-option choices in
this study is different from that had been applied in the previous studies with a Multiple Price
List by Hsee and Weber (1997) and Shupp et al. (2017). Hsee and Weber (1997) created a
specific chaotic order for participants to finish all choices in a list. Shupp et al. (2017) asked
participants to work on a list of all choices in an ascending order. However, no matter what order
different studies chose to present all choices in a Multiple Price List to participants, working on
all choices could easily cause unnecessary reverse preference response noises, which were
actually found by Hsee and Weber (1997) and they treated these responses as missing values in
their studies. The titrated elicitation method could help this study eliminate the reverse
preference response noises.
Measurement
For the decision dilemma of risk aversion toward gains, the starting choice (Choice 5 in
Table 3.1) was between a for sure gain of $2,500 and a risky choice of winning $5,000 with a
50% chance (and winning $0 with a 50% chance). The for sure gain here equals the expected
value of the risky choice. If participants prefer the for sure gain over the risky choice, their
choices are considered to demonstrate risk averse toward gains. If participants are indifferent
between the for sure gain and the risky choice, their choices are considered to demonstrate risk
neutral. If participants prefer the risky choice over the for sure gain, their choices are considered
to demonstrate risk seeking toward gains. Furthermore, based on the titrated elicitation method
illustrated in Figure 3.1, each presentation flow hits an end, which provides a range of a certain
equivalent (CE) that a participant treats indifferently from the option of a risky choice. Across all
possible ranges of the CE, each range can be transformed into a risk aversion index from 1 to 10
to measure preference intensity as shown in Table 3.3. One indicates extremely risk averse
45
toward gains and 10 indicates extremely risk seeking toward gains. A risk aversion index of 5 or
6 indicates approximately risk neutral.
Table 3.3
Summary of Each Range of the Certainty Equivalent (CE) and Its Related Risk Aversion Index
and Implied Preference in the Decision Dilemma of Risk Aversion toward Gains
Range of CE Risk Aversion Index Preference
CE £ $500 1 Extremely Risk Averse
$500 < CE £ $1,000 2
$1,000 < CE £ $1,500 3
$1,500 < CE £ $2,000 4 Risk Averse
$2,000 < CE £ $2,500 5
Risk Neutral
$2,500 < CE £ $3,000 6
$3,000 < CE £ $3,500 7 Risk Seeking
$3,500 < CE £ $4,000 8
¯
$4,000 < CE £ $4,500 9
$4,500 < CE 10 Extremely Risk Seeking
For the decision dilemma of risk seeking toward losses, the starting choice was between a
for sure loss of $2,500 and a risky choice of losing $5,000 with a 50% chance (and losing $0
with a 50% chance). The for sure loss here equals the expected value of the risky choice. If
participants prefer the risky choice over the for sure loss, their choices are considered to
demonstrate risk seeking toward losses. If participants are indifferent between the for sure loss
and the risky choice, their choices are considered to demonstrate risk neutral. If participants
prefer the for sure loss over the risky choice, their choices are considered to demonstrate risk
averse toward losses. Similarly, based on the titrated elicitation method, a range of a participant’s
CE indifferent from the fixed risky choice could be obtained. Across all possible ranges of the
CE, each range can be transformed into a risk seeking index from 1 to 10 to measure preference
intensity as shown in Table 3.4. One indicates extremely risk seeking toward losses and 10
46
indicates extremely risk averse toward losses. A risk seeking index of 5 or 6 indicates
approximately risk neutral.
Table 3.4
Summary of Each Range of the Certainty Equivalent (CE) and Its Related Risk Seeking Index
and Implied Preference in the Decision Dilemma of Risk Seeking toward Losses
Range of CE Risk Seeking Index Preference
CE £ -$500 1 Extremely Risk Seeking
-$500 < CE £ -$1,000 2
-$1,000 < CE £ -$1,500 3
-$1,500 < CE £ -$2,000 4 Risk Seeking
-$2,000 < CE £ -$2,500 5
Risk Neutral
-$2,500 < CE £ -$3,000 6
-$3,000 < CE £ -$3,500 7 Risk Averse
-$3,500 < CE £ -$4,000 8
¯
-$4,000 < CE £ -$4,500 9
-$4,500 < CE 10 Extremely Risk Averse
For the decision dilemma of loss aversion, the starting choice (Choice 5 in Table 3.2) was
between a for sure amount of $0 (status quo, nothing to lose) and a risky choice of winning
$5,000 with a 50% chance and losing $5,000 with a 50% chance. The risky choice here is a
mixed gamble with both gains and losses, and thus the preference toward loss in this study can
be viewed as risk preference toward mixed gambles. In the starting choice, the for sure amount
of $0 equals the expected value of the risky choice. If participants prefer the for sure amount
over the risky choice, their choices are considered to demonstrate loss averse. If participants are
indifferent between the for sure amount and the risky choice, their choices are considered to
demonstrate loss neutral. If participants prefer the risky choice over the for sure amount, their
choices are considered to demonstrate loss seeking.
Unlike the other decision dilemmas in this study, the for sure amount was the fixed
option in this decision dilemma and the winning amount in the risky choice was also fixed at
47
$5,000. The losing amount in the risky choice varied from $1,000 to $9,000. Based on the
titrated elicitation method, a range of an acceptable losing amount (ALA) in the risky choice that
makes a participant feel indifferent between the for sure amount and the risky choice can be
obtained for each participant. Across all possible ranges of the ALA, each range can be
transformed into a loss aversion index from 1 to 10 to measure preference intensity as shown in
Table 3.5. One indicates extremely loss averse and 10 indicates extremely loss seeking. A loss
aversion index of 5 or 6 indicates approximately loss neutral.
Table 3.5
Summary of Each Range of the Acceptable Losing Amount (ALA) and Its Related Loss Aversion
Index and Implied Preference in the Decision Dilemma of Loss Aversion
Range of ALA Loss Aversion Index Preference
ALA £ -$1,000 1 Extremely Loss Averse
-$1,000 < ALA £ -$2,000 2
-$2,000 < ALA £ -$3,000 3
-$3,000 < ALA £ -$4,000 4 Loss Averse
-$4,000 < ALA £ -$5,000 5
Loss Neutral
-$5,000 < ALA £ -$6,000 6
-$6,000 < ALA £ -$7,000 7 Loss Seeking
-$7,000 < ALA £ -$8,000 8
¯
-$8,000 < ALA £ -$9,000 9
-$9,000 < ALA 10 Extremely Loss Seeking
For the decision dilemma of ambiguity aversion toward gains, the starting choice was
between a for sure gain of $2,500 and an uncertain choice of winning $5,000 or $0 with an
unknown chance. The for sure gain here equals the expected value of the uncertain choice if the
unknown chance is considered to be 50%. If participants prefer the for sure gain over the
uncertain choice, their choices are considered to demonstrate ambiguity averse toward gains. If
participants are indifferent between the for sure gain and the uncertain choice, their choices are
considered to be ambiguity neutral. If participants prefer the uncertain choice over the for sure
48
gain, their choices are considered to demonstrate ambiguity seeking toward gains. Similarly,
based on the titrated elicitation method, a range of a participant’s CE indifferent from the fixed
uncertain choice could be obtained. Across all possible ranges of the CE, each range can be
transformed into an ambiguity aversion index from 1 to 10 to measure preference intensity as
shown in Table 3.6. One indicates extremely ambiguity averse toward gains and 10 indicates
extremely ambiguity seeking toward gains. An ambiguity aversion index of 5 or 6 indicates
approximately ambiguity neutral.
Table 3.6
Summary of Each Range of the Certainty Equivalent (CE) and Its Related Ambiguity Aversion
Index and Implied Preference in the Decision Dilemma of Ambiguity Aversion toward Gains
Range of CE Ambiguity Aversion Index Preference
CE £ $500 1 Extremely Ambiguity Averse
$500 < CE £ $1,000 2
$1,000 < CE £ $1,500 3
$1,500 < CE £ $2,000 4 Ambiguity Averse
$2,000 < CE £ $2,500 5
Ambiguity Neutral
$2,500 < CE £ $3,000 6
$3,000 < CE £ $3,500 7 Ambiguity Seeking
$3,500 < CE £ $4,000 8
¯
$4,000 < CE £ $4,500 9
$4,500 < CE 10 Extremely Ambiguity Seeking
For other measurements in this study, the empathy scale measures people’s reactions to
others’ observed experiences, which consists of four 7-item subscales: 1) the perspective-taking
scale, 2) the fantasy scale, 3) the empathic concern scale, and 4) the personal distress scale. A
high score implies strong personal reactions to others’ experiences. For the subjective numeracy
scale, this study looks at both the overall numeracy score and the scores of its two subscales:
numeracy ability and numeracy preference. A high score implies high numeracy skills.
49
3.3 Results
The final data set consists of 90
10
participants who made decisions for themselves, 37
participants who offered suggestions to their close friends, 41 participants who predicted their
close friends’ decisions, 45 participants who offered suggestions to strangers, 39 participants
who predicted strangers’ decisions. Table 3.7 summarizes the mean index of each preference in
decisions for oneself (DFS), suggestions to a close friend (STF), predictions of a close friend’s
decisions (POF), suggestions to a stranger (STST), and predictions of a stranger’s decisions
(POST).
Across all five decision types, STST has the lowest mean risk aversion index, indicating
the greatest risk averse toward gains. POST has the lowest mean risk seeking index, indicating
the greatest risk seeking toward losses. POF has the lowest mean loss aversion index, indicating
the greatest loss averse. STST has the lowest mean ambiguity aversion index, indicating the
greatest ambiguity averse toward gains. Across all four decision dilemmas, both means of the
risk seeking index and means of the ambiguity aversion index are close to either five or six for
almost all decision types, which imply risk neutrality toward losses and ambiguity neutrality
toward gains separately in these two decision dilemmas. For each decision dilemma, to help
better capture the main comparisons and analyses regarding self-other decision differences and
making decisions for others in this study, a matrix of decision target by decision response was
constructed, as shown in Table 3.8.
10
For these 90 participants, there were 49 participants who worked on an earlier version of the questionnaire with
respect to making decisions for oneself. They only answered their first question – Question 5 – without further
answering preference intensity questions for all four decision dilemmas. Therefore, I can only include them in the
percentage of participants who chose the option of a for sure amount vs. those who chose the option of a risky or
uncertain choice in Question 5. All other analyses regarding preference intensity only included the other 41
participants’ data for the decision group of making decisions for oneself.
50
Table 3.7
Summary of Means and Standard Deviations for the Index of Each Preference in Each Type of
Decision for All Four Decision Dilemmas
Risk Aversion
Index
Risk Seeking
Index
Loss Aversion
Index
Ambiguity
Aversion Index
Decision
Type
n M SD M SD M SD M SD
DFS 41 3.44 2.40 5.22 2.89 4.46 3.18 4.85 2.60
STF 37 3.49 2.09 6.03 2.71 2.89 2.07 5.11 2.61
POF 41 3.37 2.36 4.95 2.53 2.78 2.23 5.78 2.35
STST 45 2.49 1.47 4.96 2.80 3.31 2.46 4.62 2.43
POST 39 3.90 2.15 4.69 2.57 3.85 2.60 5.21 2.17
Total 203 3.31 2.14 5.15 2.72 3.46 2.59 5.10 2.45
Table 3.8
Matrix of Decision Target by Decision Response
Decision Response
Decision Target Decision Suggestion Prediction
Self DFS
A close friend STF POF
A stranger STST POST
3.3.1 Decision Dilemma 1 Risk Aversion toward Gains
Figure 3.2 summarizes the percentage of participants who chose the for sure gain (accept
the offer: receive $2,500 for sure) and the percentage of participants who chose the risky choice
(flip the coin: win $5,000 if heads or $0 if tails) for each type of decision in the decision dilemma
of risk aversion toward gains. The percentages of participants who preferred the for sure gain
over the risky choice, or, in other words, demonstrated risk averse toward gains were 90.00%,
70.27%, 82.93%, 93.33%, and 76.92% in the DFS, STF, POF, STST, and POST, respectively.
Overall, people were risk averse toward gains heedless of the decision type.
51
Figure 3.2. Percentage of participants who chose the for sure gain (accept the offer: receive
$2,500 for sure) and percentage of participants who chose the risky choice (flip the coin: win
$5,000 if heads or $0 if tails) in each type of decision in the decision dilemma of risk aversion
toward gains.
With respect to self-other decision differences in Table 3.8, two planned contrast analyses
were conducted between suggestions to others (combination of STF and STST; M = 2.94, SD =
1.83) and DFS, and between predictions of others’ decisions (combination of POF and POST; M
= 3.62, SD = 2.26) and DFS. The first contrast revealed that there was no difference in the risk
aversion index between suggestions to others and DFS, t(63.64) = -1.06, p = .29, which suggests
that advisors and personal decision makers share similar risk preferences toward gains. The
second contrast revealed that there was no difference in the risk aversion index between
predictions of others’ decisions and DFS either, t(76.46) = 0.43, p = .67, which suggests that
predictors and personal decision makers share similar risk preferences toward gains.
With respect to making decisions for others in Table 3.8, a 2 (decision target: friend vs.
stranger) ´ 2 (decision response: suggestion vs. prediction) between-subjects analysis of variance
52
(ANOVA) was conducted on the risk aversion index. The results are summarized in Figure 3.3
and revealed a significant main effect of the decision response in the risk aversion index, F(1,
158) = 4.05, p = .046, 𝜂
"
#
= .025, suggesting that advisors were more risk averse toward gains
than predictors were. No significant main effect was found from the decision target in the risk
aversion index, F(1, 158) = 0.53, p = .47, 𝜂
"
#
= .003, implying that participants treated friends’
decisions (M = 3.42, SD = 2.22) and strangers’ decisions (M = 3.14, SD = 1.94) similarly with
respect to the risk preference toward gains in this decision dilemma. However, the interaction
effect was significant, F(1, 158) = 5.71, p = .02, 𝜂
"
#
= .035, showing that participants were less
risk averse toward gains in the STF compared to in the STST. Meanwhile, participants were
more risk averse toward gains in the POF compared to in the POST.
Figure 3.3. Mean risk aversion index for friend and stranger in the two decision responses
(suggestion and prediction). The risk aversion index ranged from 1 to 10, smaller numbers (less
than 5) indicating greater risk averse toward gains. The risk aversion index of either 5 or 6
indicates approximately risk neutral. The height of the bar represents the mean. Standard errors
are represented by the error bars attached to each column.
53
Additionally, in this decision dilemma, no significant correlations were found between
the risk aversion index and any numeracy scores in any types of decisions. For the empathy
scores, the risk aversion index in the STF was found to have a weak, negative correlation with
the fantasy score, r(35) = -.33, p = .044, and have a moderate, negative correlation with the
personal distress score, r(35) = -.57, p < .001. For the demographic factors, the risk aversion
index in the STF had a weak, positive correlation with age, r(35) = .39, p = .02. The risk aversion
index in the POST had a weak, positive correlation with income, r(37) = .34, p = .03. In the STF,
there was a significant difference in the risk aversion index between male (M = 4.57, SD = 1.83)
and female (M = 2.83, SD = 1.99), t(35) = 2.66, p = .01, showing that male participants were less
risk averse toward gains than female participants. In the POF, a significant difference in the risk
aversion index was also found between males (M = 4.14, SD = 2.43) and females (M = 2.47, SD
= 1.98), t(39) = 2.37, p = .02, showing that male participants were less risk averse toward gains
than female participants.
3.3.2 Decision Dilemma 2 Risk Seeking toward Losses
Figure 3.4 summarizes the percentage of participants who chose the for sure loss (settle
the case: pay $2,500 for sure) and the percentage of participants who chose the risky choice (go
forward with the appeal process: 50% chance of paying $5,000; 50% chance of paying $0) for
each type of decision in the decision dilemma of risk seeking toward losses. The percentages of
participants who preferred the risky choice over the for sure loss, or, in other words,
demonstrated risk seeking toward losses were 43.33%, 29.73%, 51.22%, 51.11%, and 53.85% in
the DFS, STF, POF, STST, and POST, respectively. No evidence points out that people were
risk seeking toward losses in any types of decisions in this decision dilemma.
54
Figure 3.4. Percentage of participants who chose the for sure loss (settle the case: pay $2,500 for
sure) and percentage of participants who chose the risky choice (go forward with the appeal
process: 50% chance of paying $5,000; 50% chance of paying $0) in each type of decision in the
decision dilemma of risk seeking toward losses.
With respect to self-other decision differences in Table 3.8, two planned contrast analyses
were conducted between suggestions to others (combination of STF and STST; M = 5.44, SD =
2.79) and DFS, and between predictions of others’ decisions (combination of POF and POST; M
= 4.83, SD = 3.53) and DFS. The first contrast revealed that there was no difference in the risk
seeking index between suggestions to others and DFS, t(198) = 0.52, p = .60, which suggests that
advisors and personal decision makers share similar risk preferences toward losses. The second
contrast revealed that there was no difference in the risk seeking index between predictions of
others’ decisions and DFS either, t(198) = -0.77, p = .44, which suggests that predictors and
personal decision makers share similar risk preferences toward losses.
With respect to making decisions for others in Table 3.8, a 2 (decision target: friend vs.
stranger) ´ 2 (decision response: suggestion vs. prediction) between-subjects ANOVA was
55
conducted on the risk seeking index. The results are summarized in Figure 3.5 and did not reveal
any significant main effects of both the decision target, F(1, 158) = 2.53, p = .11, 𝜂
"
#
= .016, and
the decision response, F(1, 158) = 2.56, p = .11, 𝜂
"
#
= .016, in the risk seeking index. The
interaction effect was not significant either, F(1, 158) = 0.94, p = .33, 𝜂
"
#
= .006. The results
suggest that participants did not demonstrate differential risk preferences toward losses either
between for the friends’ decisions (M = 5.46, SD = 2.66) and for the strangers’ decisions (M =
4.83, SD = 2.68), or between in the suggestions offered to others and in the predictions of others’
decisions in this decision dilemma.
Figure 3.5. Mean risk seeking index for friend and stranger in the two decision responses
(suggestion and prediction). The risk seeking index ranged from 1 to 10, smaller numbers (less
than 5) indicating greater risk seeking toward losses. The risk seeking index of either 5 or 6
indicates approximately risk neutral. The height of the bar represents the mean. Standard errors
are represented by the error bars attached to each column.
Additionally, in this decision dilemma, no significant correlations were detected between
the risk seeking index and any empathy scores in any types of decisions. For the numeracy
scores, the risk seeking index in the STST was found to have a weak, negative correlation with
56
the overall numeracy score, r(43) = -.32, p = .03. For the demographic factors, the risk seeking
index in the POST had a weak, positive correlation with income, r(37) = -.39, p = .02.
3.3.3 Decision Dilemma 3 Loss Aversion
Figure 3.6 summarizes the percentage of participants who chose the status quo (not invest
– $0 for sure) and the percentage of participants who chose the risky choice (invest: 50% chance
of gaining $5,000; 50% chance of losing $5,000) for each type of decision in the decision
dilemma of loss aversion. The percentages of participants who preferred the status quo over the
risky choice, or in other words, demonstrated loss averse were 70.00%, 91.89%, 80.49%,
80.00%, and 69.23% in the DFS, STF, POF, STST, and POST, respectively. Noticeably, people
demonstrated loss averse across all five types of decisions in this decision dilemma.
Figure 3.6. Percentage of participants who chose the status quo (not invest – $0 for sure) and
percentage of participants who chose the risky choice (invest: 50% chance of gaining $5,000;
50% chance of losing $5,000) in each type of decision in the decision dilemma of loss aversion.
With respect to self-other decision differences in Table 3.8, two planned contrast analyses
were conducted between suggestions to others (combination of STF and STST; M = 3.12, SD =
57
2.28) and DFS, and between predictions of others’ decisions (combination of POF and POST; M
= 3.30, SD = 2.46) and DFS. The first contrast revealed that suggestions to others had lower loss
aversion index than DFS had, t(60.86) = -2.45, p = .02, which implies that advisors are more loss
averse than personal decision makers. The second contrast revealed that predictions of others’
decisions had lower loss aversion index than DFS had, t(64.43) = -2.03, p = .046, which implies
that predictors are more loss averse than personal decision makers. Taken together, both results
suggest that people are more loss averse toward others’ decisions (regardless of decision
response) than toward personal decisions.
With respect to making decisions for others in Table 3.8, a 2 (decision target: friend vs.
stranger) ´ 2 (decision response: suggestion vs. prediction) between-subjects ANOVA was
conducted on the loss aversion index. The results are summarized in Figure 3.7 and revealed a
significant main effect of the decision target in the loss aversion index, F(1, 158) = 4.01, p
= .047, 𝜂
"
#
= .025, suggesting that participants were more loss averse toward friends’ decisions
(M = 2.83, SD = 2.14) than toward strangers’ decisions (M = 3.56, SD = 2.52). However, no
significant main effect was found from the decision response in the loss aversion index, F(1,
158) = 0.33, p = .57, 𝜂
"
#
= .002, indicating that participants treated suggestions offered to others
and predictions of others’ decisions similarly with respect to the preference toward loss in this
decision dilemma. The interaction effect was not significant either, F(1, 158) = 0.76, p = .39, 𝜂
"
#
= .005.
58
Figure 3.7. Mean loss aversion index for friend and stranger in the two decision responses
(suggestion and prediction). The loss aversion index ranged from 1 to 10, smaller numbers (less
than 5) indicating greater loss averse. The loss aversion index of either 5 or 6 indicates
approximately loss neutral. The height of the bar represents the mean. Standard errors are
represented by the error bars attached to each column.
Additionally, in this decision dilemma, all of the following significant correlations were
only observed in the DFS. For the empathy scores, the loss aversion index was found to have
moderate, negative correlations both with the fantasy score, r(39) = -.32, p = .04, and with the
empathic concern score, r(39) = -.37, p = .02. For the numeracy scores, the loss aversion index
had a moderate, negative correlation with the numeracy preference score, r(39) = -.47, p = .002.
For the demographic information, the loss aversion index had a moderate, negative correlation
with age, r(39) = -.41, p = .007. There was a significant difference in the loss aversion index
between male (M = 5.48, SD = 3.11) and female (M = 2.88, SD = 2.66), t(39) = 2.76, p = .009,
showing that male participants were less loss averse than female participants regarding personal
decisions.
59
3.3.4 Decision Dilemma 4 Ambiguity Aversion toward Gains
Figure 3.8 summarizes the percentage of participants who chose the for sure gain (accept
the offer and settle the case: receive $2,500 for sure) and the percentage of participants who
chose the uncertain choice (wait for the investigation: receive full coverage – $5,000 or nothing –
$0, chances are unknown) for each type of decision in the decision dilemma of ambiguity
aversion toward gains. The percentages of participants who preferred the for sure gain over the
uncertain choice, or in other words, demonstrated ambiguity averse toward gains were 66.67%,
59.46%, 53.66%, 73.33%, and 64.10% in the DFS, STF, POF, STST, and POST, respectively.
Roughly, the results show that participants demonstrated less ambiguity averse toward their
friends’ decisions in contrast to both personal decisions and strangers’ decisions.
Figure 3.8. Percentage of participants who chose the for sure gain (accept the offer and settle the
case: receive $2,500 for sure) and percentage of participants who chose the uncertain choice
(wait for the investigation: receive full coverage – $5,000 or nothing – $0, chances are unknown)
in each type of decision in the decision dilemma of ambiguity aversion toward gains.
60
With respect to self-other decision differences in Table 3.8, two planned contrast analyses
were conducted between suggestions to others (combination of STF and STST; M = 4.84, SD =
2.51) and DFS, and between predictions of others’ decisions (combination of POF and POST; M
= 5.50, SD = 2.67) and DFS. The first contrast revealed that there was no difference in the
ambiguity aversion index between suggestions to others and DFS, t(198) = 0.03, p = .98, which
suggests that advisors and personal decision makers share similar preferences toward ambiguity.
The second contrast revealed that there was no difference in the ambiguity aversion index
between predictions of others’ decisions and DFS either, t(198) = 1.37, p = .17, which suggests
that predictors and personal decision makers share similar preferences toward ambiguity.
With respect to making decisions for others in Table 3.8, a 2 (decision target: friend vs.
stranger) ´ 2 (decision response: suggestion vs. prediction) between-subjects ANOVA was
conducted on the ambiguity aversion index. The results are summarized in Figure 3.9 and did not
reveal any significant main effects of both the decision target, F(1, 158) = 1.98, p = .16, 𝜂
"
#
= .012, and the decision response, F(1, 158) = 2.77, p = .10, 𝜂
"
#
= .017, in the ambiguity aversion
index. In other words, participants showed no differential preferences toward ambiguity either
between for the friends’ decisions (M = 5.46, SD = 2.48) and for the strangers’ decisions (M =
4.89, SD = 2.32), or between in the suggestions offered to others and in the predictions of others’
decisions. There was no significant interaction effect either, F(1, 158) = 0.014, p = .91, 𝜂
"
#
= .000.
61
Figure 3.9. Mean ambiguity aversion index for friend and stranger in the two decision responses
(suggestion and prediction). The ambiguity aversion index ranged from 1 to 10, smaller numbers
(less than 5) indicating greater ambiguity averse. The ambiguity aversion index of either 5 or 6
indicates approximately ambiguity neutral. The height of the bar represents the mean. Standard
errors are represented by the error bars attached to each column.
Additionally, in this decision dilemma, no significant correlations were found between
the ambiguity aversion index and any numeracy scores in any types of decisions. All of the
following significant correlations were only observed in the STF. For the empathy scores, the
ambiguity aversion index was found to have a strong, negative correlation with the personal
distress score, r(35) = -.55, p < .001. For the demographic factors, the ambiguity aversion index
had a moderate, positive correlation with age, r(35) = .41, p = .01.
3.4 Discussion
Levels of affective intensity potentially elicited by choices
Based on the EIC model, emotions that decision makers experienced at the moment of
decision making could be inferred from expected outcomes and characteristics of options in
specific choices. In the decision dilemma of risk aversion toward gains, participants were asked
62
to provide a CE (<$5,000) that they would gain for sure, which was equivalent to a risky choice
of winning $5,000 or $0 with a 50% chance. In the decision dilemma of loss aversion,
participants were told that they had a 50% chance of winning $5,000; however, they were
required to provide an amount that they could accept losing with the other 50% chance, which
could possibly make them leave the status quo ($0) and accept the risk. Literally, together with
the same 50% chance of winning $5,000, asking participants to consider how much they could
accept losing with the other 50% chance as well as leave the status quo would trigger more
intense emotions than asking them to provide a CE that they would gain for sure. Thus, one can
argue that participants experienced a higher level of affective intensity potentially elicited by
choices in the decision dilemma of loss aversion than in the decision dilemma of risk aversion
toward gains.
In the decision dilemma of risk seeking toward losses, participants were asked to provide
a CE (<$5,000) that they would lose for sure, which was equivalent to a risky choice of losing
$5,000 or $0 with a 50% chance. This decision dilemma has the same magnitudes of expected
outcomes in all choices as the decision dilemma of risk aversion toward gains, except that the
former examined only losses, while the latter examined only gains. Based on the characteristics
of the options in both decision dilemmas, the essential ideas of CEs are winning less for sure in
the decision dilemma of risk aversion toward gains and losing less for sure in the decision
dilemma of risk seeking toward losses. Compared to losing less for sure, winning less for sure
can be treated as a negative incentive. Thus, based on previous findings that negative incentives
evoke stronger emotions than do positive incentives (for review, see Cacioppo & Gardner, 1999),
one can argue that participants experienced a higher level of affective intensity potentially
63
elicited by choices in the decision dilemma of risk aversion toward gains than in the decision
dilemma of risk seeking toward losses in this study.
The decision dilemma of ambiguity aversion toward gains shares the same expected
outcomes with the decision dilemma of risk aversion toward gains but does not include
probability information. According to van Dijk and Zeelenberg’s (2006) findings, uncertain
outcomes reduce emotions involved in decision making. Thus, one can argue that participants
experienced a higher level of affective intensity potentially elicited by choices in the decision
dilemma of risk aversion toward gains than in the decision dilemma of ambiguity aversion
toward losses in this chapter.
In summary, based on the EIC model, participants hypothetically could be considered to
have experienced the highest level of affective intensity potentially elicited by choices in the
decision dilemma of loss aversion across all four decision dilemmas in this study. Participants
also were considered to have experienced higher level of affective intensity in the decision
dilemma of risk aversion toward gains than in both decision dilemmas of risk seeking toward
losses and ambiguity aversion toward gains. More importantly, lower levels of affective intensity
that participants experienced in both decision dilemmas of risk seeking toward losses and
ambiguity aversion toward gains compared to the affective intensity in the other two decision
dilemmas were confirmed by the preference index results shown in Table 3.7. Generally,
participants were found to be evidently risk averse toward gains in the decision dilemma of risk
aversion toward gains and loss averse in the decision dilemma of loss aversion, regardless of the
decision type. However, across all five types of decisions, participants seemed to be risk neutral
in the decision dilemma of risk seeking toward losses and ambiguity neutral in the decision
dilemma of ambiguity aversion toward gains. The risk neutral attitudes found in the decision
64
dilemma of risk seeking toward losses are consistent with Weber and Chapman’s (2005)
findings, that people are less prone to take risk when amounts are large than small in both the
domains of gains and losses. Considering the relations between affect and preferences toward
risk, loss, and ambiguity, one can argue that choices with neutral attitudes involved less emotions
compared to those with evidently risk averse or loss averse attitudes. Therefore, in general,
participants would have experienced lower levels of affective intensity in the decision dilemmas
of risk seeking toward losses and ambiguity aversion toward gains than in the decision dilemmas
of risk aversion toward gains and loss aversion.
Self-other decision differences
Across all four decision dilemmas, self-other decision difference was observed only in
the loss aversion index. These results are consistent with my hypotheses that there would be self-
other decision differences in loss aversion, but not in risk aversion toward gains, risk seeking
toward losses, and ambiguity aversion toward gains. As discussed above, among the four
decision dilemmas, participants were considered to have experienced the highest level of
affective intensity potentially elicited by choices in the decision dilemma of loss aversion.
Finding a difference only in the decision dilemma with the highest level of affective intensity
supports the proposed affect-based hypothesis I, that a higher level of affective intensity
potentially elicited by choices results in more salient self-other decision differences in
preferences toward risk, loss, and ambiguity. Thus, affective intensity potentially elicited by
choices hypothetically serves as a moderator that enhances self-other decision differences in
preferences toward risk, loss, and ambiguity.
In the decision dilemma of loss aversion, compared to making decisions for oneself,
people were found to be more loss averse toward others’ decisions in both the contexts of
65
offering suggestions to others and predicting others’ decisions. However, both Andersson et al.
(2016) and Polman (2012b) found that people were less loss averse when making monetary
choices on behalf of specific unknown others compared to when making choices for themselves,
which differs from the result in this study. This inconsistency may be attributable to different
decision types and different sizes of stakes in choices. Rather than asking people to make
decisions on behalf of others, this study asked people to offer suggestions to others and predict
others’ decisions. It is possible that people have different preferences toward loss in different
types of decisions. The other possible cause is the size of stakes, as the stakes studied in both
Andersson et al. (2016; approximately less than $60) and Polman (2012b; less than $7) were
much smaller than those studied here involving a few thousand dollars. It is possible that
compared to personal decisions, people are more loss averse toward others’ decisions when the
stakes are large but are less loss averse toward others’ decisions when the stakes are small.
Making decisions for others
Across all four decision dilemmas, an influence of the decision target was observed only
in the loss aversion index. In the decision dilemma of loss aversion, people were found to be
more loss averse toward a close friend’s decisions than a stranger’s decisions. As loss aversion
involves strong negative emotions (Camerer, 2012), these results suggest that participants who
made decisions for close friends might have experienced stronger negative emotions during the
process of making decisions than those who made decisions for strangers. This is consistent with
previous findings that people experience intense emotions in close relationships (for review, see
Berscheid & Ammazzalorso, 2003).
Across all four decision dilemmas, an influence of the decision response was observed
only in the risk aversion index. In the decision dilemma of risk aversion toward gains,
66
participants were found to be more risk averse toward gains in the suggestions to others
compared to be in the predictions of others’ decisions. As discussed above, suggestions for
others’ decisions are more affect-oriented than are predictions, which imply that the emotions
potentially evoked by risk would be more likely to influence suggestions. This helps explain the
difference in the risk aversion index between suggestions and predictions.
An interaction effect between the decision target and the decision response also was
found in the risk aversion index. The results showed that participants predicted that their close
friends would be more risk averse that would strangers. Faro and Rottenstreich (2006) found that
people predicted unknown others to be less risk reverse toward gains than themselves, but not a
close friend. According to their findings and the potentially blurred boundaries between oneself
and a close friend, one can infer that people would predict that a stranger would be less risk
averse toward gains than would a close friend. The results in this study are consistent with this
inference. For suggestions offered to others, the results showed that participants who offered
suggestions to close friends were less risk averse toward gains than those who offered
suggestions to strangers, which imply that people are more likely to suggest that friends take a
risk to win a large gain than to suggest that strangers do so. It seems that people are oriented
more to better outcomes in the suggestions they offer to a close friend than in those offered to a
stranger with respect to risk aversion toward gains.
Across all four decision dilemmas, no effects of either the decision target or the decision
response were found in both the risk seeking index and the ambiguity aversion index. With
respect to different levels of affective intensity potentially elicited by choices across the four
decision dilemmas, participants were considered to have experienced lower levels of affective
intensity in the decision dilemmas of risk seeking toward losses and ambiguity aversion toward
67
gains than in the decision dilemmas of risk aversion toward gains and loss aversion. The results
here support the proposed affect-based hypothesis II, that a higher level of affective intensity
potentially elicited by choices would be more likely to trigger the effect of either the decision
target or the decision response in preferences toward risk, loss, and ambiguity.
As well as no self-other decision differences found in the decision dilemmas of risk
seeking toward losses and ambiguity aversion toward gains, it seems that when the level of the
affective intensity elicited by choices is weak, people simply shift their personal preferences to a
specific other (either a close friend or stranger) regardless of the decision response (either
suggestion or prediction). As discussed by Loewenstein and Lerner (2003), when the current
emotion is at lower levels of intensity, people easily can defeat the emotion’s effect on decision
making. Overall, the findings of this study revealed the important role of affective intensity
potentially elicited by choices in self-other decision differences and making decisions for others
with respect to preferences toward risk, loss, and ambiguity. However, the findings of the
correlations between the empathy scores and each preference index suggest that empathy could
not account for any findings with respect to self-other decisions in this study. It is plausible that
people empathize with others, but still treat others’ decisions either differently from or the same
as personal decisions.
3.5 General Discussion
This study involves personal decisions and decisions made for specific others with
respect to four common preferences toward risk, loss, and ambiguity involving large stakes.
Unlike previous studies, this study compared all four preferences, including 1) risk aversion
toward gains, 2) risk seeking toward losses, 3) loss aversion, and 4) ambiguity aversion toward
gains, in self-other decisions. This study examined self-other decision differences in each
68
preference between decisions made for oneself and suggestions offered to others, and between
decisions made for oneself and predictions of others’ decisions. These two comparisons extend
previous comparisons in risk preferences only between decisions made for oneself and
predictions of others’ decisions (e.g., Hsee & Weber, 1997; Faro & Rottenstreich, 2006) and in
loss aversion only between decisions made for oneself and decisions made on behalf of others
(e.g., Andersson, Holm, Tyran, & Wengström, 2016; Polman, 2012b). Additionally, by focusing
only on decisions made for others, this study investigated both the decision target’s and the
decision response’s effects on each preference. The findings underscore the importance of
defining both the decision target and the decision response specifically in studying self-other
decisions.
Moreover, I have proposed two affect-based hypotheses to account for self-other decision
differences and the effects of both the decisions target and the decision response across the four
different preferences toward risk, loss, and ambiguity. The essential idea of the proposed affect-
based hypotheses is that a high level of affective intensity potentially elicited by choices
strengthens the self-other decision differences and the effects of both the decision target and the
decision response in preferences toward risk, loss, and ambiguity. The affect-based hypothesis I
about self-other decision differences proposed in this study can account for a wide range of
various inconsistent results in the literature related to risk preferences. With respect to specific
others’ monetary choices involving only small stakes, it can account for both the presence of
self-other decision differences in loss aversion (e.g., Andersson, Holm, Tyran, & Wengström,
2016; Polman, 2012b), and the absence of the differences in risk preferences toward gains and
losses (e.g., Hsee & Weber, 1997; Stone & Allgaier, 2008; Stone, Yates, & Caruthers, 2002), as
studies related to loss aversion are very likely to have a higher level of affective intensity
69
potentially elicited by choices, as discussed here. It also can account for both the presence of
self-other decision differences in risk preference in relationship and physical safety decisions
(Beisswanger et al., 2003; Stone, Choi, de Bruin, & Mandel, 2013), and the absence of the
differences in risk preference in monetary decisions (e.g., Stone & Allgaier, 2008; Stone, Yates,
& Caruthers, 2002), as relationship and physical safety decisions are very likely to involve more
intense emotions than are monetary decisions.
However, there are two limitations in this study. One is that the level of affective
intensity that decision makers experienced while making decisions was inferred largely from the
expected outcomes and the characteristics of options in choices. Although the rank of affective
intensity across different preferences toward risk, loss, and ambiguity in different decision
dilemmas was confirmed approximately by the results of the preference indices, affective
intensity was not measured directly. Future studies are expected to measure affective intensity
directly from self-reports to confirm the important role of affective intensity in the self-other
decision differences and making decisions for others with respect to preferences toward risk,
loss, and ambiguity. The other limitation is that all of the findings in this study are based on a
single decision context defined by the scenarios. Only one scenario was used in each decision
dilemma. It is very likely that some of these results would vary across different scenarios. Future
studies should extend to various scenarios to study self-other decisions with respect to
preferences toward risk, loss, and ambiguity.
70
Chapter 4 Delay Discounting
4.1 Introduction
There was a recent article on ABC News entitled Credit card debt surpasses $1 trillion in
the US for first time (Pak, 2018). We may feel that this number is too great to associate with our
daily lives. However, have you ever experienced not having sufficient money to pay the full
balance on your credit card? Overspending, or expediting a future expense (e.g., savings for
retirement) is a common problem, especially when it is easy to obtain credit today (e.g., credit
cards). As this problem is so common, we might wonder whether our friends or others can help
prevent us from it in our daily lives. This study addresses whether others expedite a future gain
in the decisions they make for us differently from those they make for themselves.
Different risk preferences have been found between monetary decisions made for oneself
and decisions suggested to others (e.g., Roszkowski & Snelbecker, 1990), and between decisions
made for oneself and predictions of others’ decisions (e.g., Hsee & Weber, 1997; Faro &
Rottenstreich, 2006). However, there are few studies about delay discounting in self-other
decisions. The studies of delay discounting examine intertemporal choices between a smaller
sooner (SS) reward and a larger later (LL) reward. Vlaev et al. (2017) asked students to make
intertemporal choices both for themselves and on behalf of an unknown person in the room.
Their setting was a reciprocal environment as participants’ final payments depended not only on
their own choices but also on others’ choices for them. They found that people discounted more
in making decisions on behalf of others compared to in making personal decisions. Albrecht,
Volz, Sutter, Laibson, and von Cramon (2011) also examined intertemporal choices in both
choices made for oneself and choices made on behalf of a specific unknown other (“a participant
in a subsequent experimental session”). Participants were informed that this unknown person’s
71
payments would depend on their choices. They found that when a SS gain could be received
immediately, people discounted less in the choices they made on behalf of others than in the
choices they made for themselves. Both studies studied intertemporal choices with relatively
small stakes (less than $100).
Compared to previous studies, this study examined self-other decisions in intertemporal
preference with respect to expediting a future gain (with both small and large stakes) within a
fixed duration in specific everyday decision scenarios. In addition, instead of asking people to
make decisions on behalf of others, this study asked people to offer suggestions to others and
predict others’ decisions. Others’ decisions were identified specifically as close friends’ and
strangers’ decisions.
Affect and Delay Discounting
A series of functional magnetic resonance imaging (FMRI) studies has shown that
anticipation of gains involves positive emotions (Knutson & Peterson, 2005). Compared to small
expected gains, large expected gains are anticipated to induce stronger positive emotions.
According to temporal construal theory (Liberman & Trope, 1998; Trope & Liberman, 2000,
2003), people tend to form an abstract mental construal for a temporally distant object and a
concrete mental construal for a temporally proximal object. Based on the effects of vividness
(Loewenstein, Weber, Hsee, & Welch, 2001), an instant gain with a more specific and vivid
mental image would evoke stronger emotions in the process of making decisions than would a
future gain. Ong, Goodman, and Zaki (2018) found that people expected to experience more
positive emotions about events that were occurring now compared to those in the future both for
themselves and others. All of these results suggest that an expected instant gain would induce
stronger positive emotions than an expected future gain. Therefore, when people face a choice to
72
expedite a future gain, it is possible to evaluate and compare the intensity of the positive affect
potentially elicited by different expected gains based on the size of the gain and the time it is
received.
In intertemporal choices, if people prefer a SS gain over a LL gain, they are considered
impatient. The more impatient they are, the more they will discount a future gain compared to a
smaller gain they will accept sooner. To contend with impatience, people need to overcome their
emotions. Thus, intertemporal choices essentially are affect-based choices and discount rates
could, in turn, be used to infer different levels of affective intensity potentially elicited by
choices that decision makers have experienced during the process of making decisions.
Compared to an option to receive a SS gain immediately, Albrecht et al. (2011) found
that there was less affective engagement in the process of making decisions when both SS and
LL gain options occurred in the future. They found further that people are less emotionally
involved and more patient when they make choices on behalf of a specific unknown other
compared to when they make choices for themselves; however, this applied only for the
intertemporal choices associated with receiving a SS gain immediately. Their results potentially
suggest that a higher level of affective intensity experienced during the process of making
decisions was more likely to result in self-other decision difference in intertemporal preference.
Therefore, I propose affect-based hypothesis I to account for self-other decision differences
addressed in this study, in which the more intense the affect potentially elicited by choices that
decision makers face, the more salient self-other decision differences in discount rate become.
Previous research has studied delay discounting between current self vs. future self and
found that when people perceived more connections between their current and future selves, they
discount less for future gains (Bartels & Urminsky, 2011; Ersner-Hershfield, Wimmer, &
73
Knutson, 2009; Hershfield, 2011). According to Aron, Aron, Tudor, and Nelson (1991), people
blur the boundaries between oneself and others easily in a close relationship. Thus, with respect
to making decisions for others in this study, I hypothesize that people will demonstrate
differential discount rates between close friends’ and strangers’ decisions and will show a lower
discount rate for their close friends’ future gains.
When offering suggestions to others, people are concerned about responsibility and prefer
advice that will help maintain good relationships with the recipients of their advice (Dana &
Cain, 2015). Compared to predictions of others’ decisions, suggestions offered to others involve
more affective engagement. Affect has been considered an important determinant in
intertemporal preferences (for review, see Urminsky & Zauberman, 2015), and emotions such as
sadness (Lerner, Li, & Weber, 2013) and gratitude (Desteno, Li, Dickens, & Lerner, 2014) have
been found to influence delay discounting. Therefore, I hypothesize that people demonstrate
differential discount rates between suggestions to others and predictions of others’ decisions. In
addition, people are found to experience intense emotions in close relationships (for review, see
Berscheid & Ammazzalorso, 2003). Considering the different level of affective intensity
involved in different decision target or different decision response, I propose affect-based
hypothesis II to account for the effects of both the decision target and the decision response in
making decisions for others, in which a higher level of affective intensity potentially elicited by
choices that decision makers face is more likely to result in the effect of either the decision target
or the decision response in discount rate.
Study Overview
This study examined delay discounting with respect to expediting a future gain (vs.
deferring an instant gain, see Malkoc & Zauberman, 2006) in self-other decisions. The future
74
gain varied by size (small vs. large) and the time at which it would be received (distant future vs.
near future), which generated four separate conditions to measure discount rate. This design
allowed a comparison of different levels of affective intensity potentially elicited by
intertemporal choices across the four conditions. A delay discounting index was applied to
measure the discount rate in each condition for each participant.
In each condition, self-other decision differences in the delay discounting index were first
examined between decisions made for oneself and suggestions offered to others, and between
decisions made for oneself and predictions of others’ decisions. By focusing only on the
decisions made for others, the effects of both the decision target and the decision response on the
delay discounting index were then investigated. Additionally, the correlations between the delay
discounting index and empathy scores, impulsiveness scores, and demographic factors were
studied separately.
4.2 Method
Participants
A total of 315 participants were recruited from Amazon’s Mechanical Turk (AMT) to
complete an online questionnaire. Among all those recruited, 49 either failed to pass the attention
check question or finished the survey in an unexpectedly short time, and thus were excluded in
the following analyses. The final dataset included 266 participants (131 males, 134 females
11
)
who ranged in age from 18 to over 65 and had an annual household income that ranged from less
than $25,000 to greater than $150,000. There are 64.62% (124 of 266) of participants who
reported to be between 25 and 34 years old. The education information reported indicates that all
participants had obtained at least a high school degree or above and 43.23% (115 of 266) of
11
One participant did not report gender.
75
participants reported that they had earned a Bachelor’s degree. Participants were assigned
randomly to one of five independent decision groups: making decisions for oneself, offering
suggestions to a close friend, predicting a close friend’s decisions, offering suggestions to a
stranger, and predicting a stranger’s decisions.
Materials
The participants mainly answered questions related to expediting a future gain. There are
four future gains examined in this study, including a small distant future gain, a small near future
gain, a large distant future gain, and a large near future gain. For convenience of data collection,
all four conditions were provided in a random order in one questionnaire. Each questionnaire
consisted of main questions regarding all four conditions, as well as 28 items regarding the
empathy scale (Davis, 1980), 30 items regarding the Barratt impulsiveness scale (Patton,
Stanford, & Barratt, 1995), and several demographic questions.
For each condition, the study consisted of a unique real-life scenario and a sequence of
two-option choices. All four real-life scenarios were created exclusively for this study and
included scenarios about cash rewards from a grocery store, a welcome bonus from a bank,
winning a case in a lawsuit, and winning the lottery. (See Appendix B.) To fit each decision
group, the scenario for each condition was altered slightly. With respect to a close friend’s
decisions, participants were asked first to think about a close friend who had beliefs, attitudes,
values, and interests similar to theirs and then to write down that friend’s first name.
Subsequently, their friend’s name was inserted in the context of all scenarios and choices in the
questionnaire. With respect to a stranger’s decisions in each condition, participants read a
distinct vignette informing them that they were meeting a specific stranger and beginning a
conversation with this stranger. Four different vignettes about meeting a stranger included
76
meeting someone at a café, at a social event, at their friend’s birthday party, and at their high
school reunion.
All choices in all four conditions were similar and each regarded an intertemporal choice
between a pair of options: a SS gain vs. a LL gain. In the study of each condition, the LL gain
always was a fixed option and participants were asked to make several choices to expedite this
future gain. In the small-distant condition, participants were asked to expedite $100 two years
from now. In the small-near condition, participants were asked to expedite $100 one year from
now. In the large-distant condition, participants were asked to expedite $10,000 two years from
now. In the large-near condition, participants were asked to expedite $10,000 one year from now.
The duration between the LL gain and the SS gain was fixed across all four conditions at one
year for all choices.
To compare different levels of affective intensity potentially elicited by choices across all
four conditions, the small-distant condition and the small-near condition shared the same pairs of
the sooner and later gains in the choices (Table 4.1), and the large-distant condition and the
large-near condition shared the same pairs of the sooner and later gains in the choices (Table
4.2). For all conditions, the nine different sooner gains were distributed equally from 10% to
90% of the later gain. The temporal distance between a near future event and a distant future
event was one year. The reason for choosing one year as both the expediting duration and the
temporal distance is that in real-life financial decisions, we seldom face a choice of receiving
$100 five years from now and $200 ten years from now, or a choice of receiving $100 two days
from now and $200 ten days from now.
Table 4.1
Each Pair of the Sooner and Later Gains in Each Choice in the Small-Distant and Small-Near
Conditions
77
Gains
Choice Sooner Later
1 $10 $100
2 $20 $100
3 $30 $100
4 $40 $100
5 $50 $100
6 $60 $100
7 $70 $100
8 $80 $100
9 $90 $100
Table 4.2
Each Pair of the Sooner and Later Gains in Each Choice in the Large-Distant and Large-Near
Conditions
Gains
Choice Sooner Later
1 $1,000 $10,000
2 $2,000 $10,000
3 $3,000 $10,000
4 $4,000 $10,000
5 $5,000 $10,000
6 $6,000 $10,000
7 $7,000 $10,000
8 $8,000 $10,000
9 $9,000 $10,000
Procedure
After receiving one questionnaire online, each participant began with four sequences of
choices related to the four conditions. Then they answered questions regarding the empathy
scale, the Barratt impulsiveness scale, and the demographic questions. The titrated elicitation
method (Mazur, 1987; Kirby & Herrnstein, 1995) was applied to present the sequence of choices
to participants for each condition. The following will take the small-near condition as an
example to demonstrate this method in detail.
78
As the LL gain option of receiving $100 one year from now was always fixed in the
small-near condition, Figure 4.1 shows the details of the titrated elicitation method, which leads
to 10 possible flow paths that present the SS gain options with different sooner gains to
participants. Participants started with the midpoint of the sooner gains in table 4.1 and faced a
choice between “the smaller welcome bonus – $50 immediately” or “the welcome bonus – $100
one year from now”. If participants chose the SS gain here, the next sooner gain they would be
offered was the minimum one – $10. If participants chose the LL gain here, the next sooner gain
they would be offered was the maximum one – $90. The presentation flow simply depended on
the participant’s choice in each option pair until an ending happens. Therefore, although there
was a total of nine sooner gains in the study of each condition, participants saw only two or four
of them as shown in Figure 4.1. Across all four conditions, participants always started with the
midpoint of the sooner gains and then followed the presentation flow guided by the titrated
elicitation method.
79
Figure 4.1. Flow of the titrated elicitation method presenting different small sooner (SS) gain
options to participants with a fixed large later (LL) gain option of receiving $100 one year from
now in the small-near condition. SG indicates a sooner gain.
Measurement
Based on the titrated elicitation method illustrated in Figure 4.1, there was a total of 10
possible flow paths for each participant to reach an end in the study of each condition. Because
the LL gain option always was fixed across all four conditions, each end determined a range of a
sooner gain that participants felt indifferent to receiving one year ahead. Across all possible
80
ranges of the sooner gain in each condition, each range can be transformed into a delay
discounting index from 1 to 10 to measure the discount rate. Table 4.3 summarizes each range of
the sooner gain and its related delay discounting index and implied discount rate in the small-
near condition. The ranges of the sooner gain in the other three conditions can be presented in the
similar way. Across all four conditions, a smaller delay discounting index indicates that
participants wish to discount the LL gain more, i.e., they show a higher discount rate.
Table 4.3
Summary of Each Range of the Sooner Gain (SG) and Its Related Delay Discounting Index and
Implied Discount Rate in the Small-Near Condition
Range of SG Delay Discounting Index Discount Rate
SG £ $10 1 Highest (Discount Most)
$10 < SG £ $20 2
$20 < SG £ $30 3
$30 < SG £ $40 4
$40 < SG £ $50 5
$50 < SG £ $60 6
$60 < SG £ $70 7
$70 < SG £ $80 8
$80 < SG £ $90 9
$90 < SG 10 Lowest (Discount Least)
For other measurements in this study, the empathy scale measures people’s reactions to
others’ observed experiences and consists of four 7-item subscales, as Davis (1980) discussed,
including the perspective-taking scale, the fantasy scale, the empathic concern scale, and the
personal distress scale. A high score implies a strong tendency to demonstrate empathy or a
strong feeling of others’ experiences. The Barratt impulsiveness scale measures impulsiveness
and consists of three subscales, including attentional impulsiveness related to the focus of
attention, moto impulsiveness regarding action tendency, and non-planning impulsiveness
concerning forethought (Patton et al, 1995). A high score implies a high level of impulsiveness.
81
4.3 Results
Table 4.4 summarizes the average delay discounting index in each condition in decisions
for oneself (DFS), suggestions to a close friend (STF), predictions of a close friend’s decisions
(POF), suggestions to a stranger (STST), and predictions of a stranger’s decisions (POST).
Across all five types of decisions, STST has the lowest delay discounting index in the small-
distant condition, while POST has the lowest delay discounting index in the small-near
condition, the large-distant condition, and the large-near condition. The lowest delay discounting
index implies a highest discount rate. Across all four conditions, all delay discounting indices in
the conditions of a large future gain were higher than were those in the conditions of a small
gain. To help better capture the main comparisons and analyses with respect to self-other
decision differences and making decisions for others in each condition, a matrix of decision
target by decision response was constructed, as shown in Table 4.5.
Table 4.4
Summary of Means and Standard Deviations for the Delay Discounting Index in Each Type of
Decision for All Four Conditions
Small-Distant
Condition
Small-Near
Condition
Large-Distant
Condition
Large-Near
Condition
Decision
Type
n M SD M SD M SD M SD
DFS 57 6.04 1.97 6.04 2.00 8.02 1.99 7.84 2.15
STF 59 5.95 2.31 6.12 2.40 8.22 2.13 8.12 1.97
POF 67 5.97 2.21 5.78 2.34 7.84 2.55 7.78 2.35
STST 40 5.45 2.36 6.03 2.37 8.08 2.07 7.58 2.25
POST 43 5.72 2.12 5.74 2.17 6.98 2.11 6.95 2.42
Total 266 5.86 2.19 5.94 2.25 7.86 2.22 7.70 2.24
Table 4.5
Matrix of Decision Target by Decision Response
Decision Response
Decision Target Decision Suggestion Prediction
82
Self DFS
A close friend STF POF
A stranger STST POST
Self-other decision differences
Planned contrasts between suggestions to others (combination of STF and STST) and
DFS was conducted on the delay discounting index for each condition separately. In the small-
distant condition, no significant difference was observed between DFS and suggestions to others
(M = 5.75, SD = 2.33), t(261) = -.91, p = .36. In the small-near condition, no significant
difference was observed between DFS and suggestions to others (M = 6.08, SD = 2.38), t(261)
= .10, p = .92. In the large-distant condition, no significant difference was observed between
DFS and suggestions to others (M = 8.16, SD = 2.09), t(261) = .35, p = .72. In the large-near
condition, no significant difference was observed between DFS and suggestions to others (M =
7.90, SD = 2.10), t(261) = .01, p = .99. All of these results imply that advisors have similar
discount rates as personal decision makers regardless of the type of future gain people expect to
expedite.
Planned contrasts between predictions of others’ decisions (combination of POF and
POST) and DFS was performed on the delay discounting index for each condition separately. In
the small-distant condition, no significant difference was observed between DFS and predictions
of others’ decisions (M = 5.87, SD = 2.17), t(261) = -.53, p = .60. In the small-near condition, no
significant difference was observed between DFS and predictions of others’ decisions (M = 5.76,
SD = 2.27), t(261) = -.74, p = .46. In the large-distant condition, no significant difference was
observed between DFS and predictions of others’ decisions (M = 7.50, SD = 2.41), t(261) = -
1.69, p = .09. In the large-near condition, no significant difference was observed between DFS
and predictions of others’ decisions (M = 7.45, SD = 2.40), t(261) = -1.3, p = .19. All of these
83
results suggest that predictors demonstrate similar discount rates as personal decision makers
regardless of the type of future gain people expect to expedite.
Making decisions for others
For each condition, a 2 (decision target: friend vs. stranger) ´ 2 (decision response:
suggestion vs. prediction) between-subjects analysis of variance (ANOVA) was conducted on
the delay discounting index. In the small-distant condition, the ANOVA results revealed no
significant main effects, either for the decision target, F(1, 205) = 1.38, p = .24, 𝜂
#
= .007, or for
the decision response, F(1, 205) = 0.21, p = .65, 𝜂
"
#
= .001. The interaction effect also was not
significant, F(1, 205) = 0.15, p = .70, 𝜂
"
#
= .001. These results suggest that participants
demonstrated no differential discount rates either between friends’ decisions (M = 5.96, SD =
2.25) and strangers’ decisions (M = 5.59, SD = 2.23), or between suggestions offered to others
and predictions of others’ decisions in this condition.
In the small-near condition, the ANOVA results also revealed no significant main effects,
either for the decision target, F(1, 205) = 0.04, p = .85, 𝜂
"
#
= 0, or for the decision response, F(1,
205) = 0.89, p = .35, 𝜂
"
#
= .004. The interaction effect also was not significant, F(1, 205) = 0.01, p
= .93, 𝜂
"
#
= 0. These results suggest that participants demonstrated no differential discount rates
either between friends’ decisions (M = 5.94, SD = 2.37) and strangers’ decisions (M = 5.88, SD =
2.26), or between suggestions offered to others and predictions of others’ decisions in this
condition.
In the large-distant condition, the ANOVA results revealed a significant main effect for
the decision response, F(1, 205) = 5.38, p = .02, 𝜂
#
= .03, showing that advisors demonstrated a
higher delay discounting index than did predictors, which implies that participants displayed a
lower discount rate in their suggestions to others compared to their predictions of others’
84
decisions. No significant main effect was found for the decision target, F(1, 205) = 2.47, p = .12,
𝜂
"
#
= .01, indicating that there was no difference in the discount rate between friends’ decisions
(M = 8.02, SD = 2.36) and strangers’ decisions (M = 7.51, SD = 2.15). There also was no
significant interaction effect, F(1, 205) = 1.25, p = .27, 𝜂
"
#
= .006. The results are summarized in
Figure 4.2.
Figure 4.2. Mean delay discounting indices for friend and stranger in the two decision responses
(suggestion and prediction) in the large-distant condition. The delay discounting index ranged
from 1 to 10, smaller numbers indicating a higher discount rate. The height of the bar represents
the mean. Standard errors are represented by the error bars attached to each column.
In the large-near condition, the ANOVA results revealed a significant main effect of the
decision target, F(1, 205) = 4.61, p = .03, 𝜂
#
= .02, demonstrating that participants showed a
higher delay discounting index for friends’ decisions (M = 7.94, SD = 2.18) than for strangers’
decisions (M = 7.25, SD = 2.35). This implies that participants exhibited a lower discount rate in
friends’ decisions compared to strangers’ decisions. No significant main effect was found
attributable to the decision response, F(1, 205) = 2.30, p = .13, 𝜂
"
#
= .01, indicating that people
85
displayed no differential discount rates when offering suggestions to others and when predicting
others’ decisions. The interaction effect also was not significant, F(1, 205) = 0.19, p = .66, 𝜂
"
#
= .001. The results are summarized in Figure 4.3.
Figure 4.3. Mean delay discounting indices for friend and stranger in the two decision responses
(suggestion and prediction) in the large-near condition. The delay discounting index ranged from
1 to 10, smaller numbers indicating a higher discount rate. The height of the bar represents the
mean. Standard errors are represented by the error bars attached to each column.
Other correlations
For the correlations between the delay discounting index and the empathy scale in each
type of decision, significant correlations were found only in making decisions for strangers. In
the STST, the delay discounting index had a moderate, positive correlation with the perspective-
taking score, r(38) = .32, p = .04 in the small-distant condition, and a moderate, negative
correlation with the personal distress score, r(38) = -.36, p = .02, in the small-near condition. In
the POST, the delay discounting index had a moderate, negative correlation with the perspective-
taking score, r(41) = -.40, p = .01, in the large-near condition.
86
For the correlations between the delay discounting index and the impulsiveness scale in
each type of decision, the correlations were significant in almost all types of decisions except for
the POST. In the DFS, the delay discounting index had a moderate, negative correlation with the
non-planning impulsiveness score, r(55) = -.34, p = .01, in the small-near condition. In the STF,
the delay discounting index was found to have moderate, negative correlations with both the
moto impulsiveness score, r(57) = -.33, p = .01, and the non-planning impulsiveness score, r(57)
= -.31, p = .02, in the large-distant condition. In the POF, the delay discounting index had a
moderate, positive correlation with the attentional impulsiveness score, r(65) = .37, p = .002, in
the small-distant condition. In the STST, the delay discounting index had moderate, negative
correlations with both the attentional impulsiveness score, r(38) = -.40, p = .01, and the moto
impulsiveness score, r(38) = -.36, p = .02, in the small-near condition. The delay discounting
index also was found to have a moderate, negative correlation with the attentional impulsiveness
score, r(38) = -.33, p = .04, in the STST in the large-near condition.
For the demographic factors, neither age nor income were significantly correlated with
the delay discounting index in any types of decisions. Education was found to have only a
moderate, positive correlation with the delay discounting index, r(65) = .32, p = .008, in the POF
in the small-near condition. Sex differences in the delay discounting index were found in the
STST in the small-distant condition, t(37) = 2.07, p = .045, showing that males (M = 6.54, SD =
2.60) tended to demonstrate higher delay discounting indices or smaller discount rates than
females (M = 4.92, SD = 2.13) did when offering suggestions to strangers. Sex difference also
was found in the delay discounting index in the POST in the small-near condition, t(41) = -2.10,
p = .04, indicating that males (M = 5.09, SD = 2.16) tended to demonstrate lower delay
discounting indices or higher discount rates than did females (M = 6.43, SD = 2.01) when
87
predicting strangers’ decisions. The delay discounting indices in either the large-distant condition
or the large-near condition were not significantly correlated with any demographic factors.
4.4 Discussion
Discounting for four different future gain scenarios was studied in this chapter, including
$100 two years from now in the small-distant condition, $100 one year from now in the small-
near condition, $10,000 two years from now in the large-distant condition, and $10,000 one year
from now in the large-near condition. Based on the relations between positive affect and
expected gains, as well as the same expediting duration used for all conditions, one can rank the
intensities of positive affect potentially elicited by the expected gains that participants
experienced at the moment of decision making across the four conditions: large-near condition >
large-distant condition > small-near condition > small-distant condition.
For each type of decision across all four conditions, all delay discounting indices in the
conditions with respect to expediting a large future gain were higher than were those in the
conditions with respect to expediting a small future gain, which indicates that people generally
show a lower discount rate toward a large future gain regardless of the decision type. This is
consistent with previous findings that the discount rate is greater for a smaller reward (e.g.,
Thaler, 1981). As discussed above, a lower discount rate implies high affective engagement,
which suggests that participants should have experienced lower levels of affective intensity in
the conditions with a small future gain than with a large future gain. The delay discounting index
results further support the rank of the affective intensity potentially elicited by the expected gains
across the four conditions.
With respect to self-other decisions, no differences were found in the delay discounting
index between decisions made for oneself and suggestions to others, or between decisions made
88
for oneself and predictions of others’ decisions in all four conditions. Albrecht et al. (2011)
found that people demonstrated different discount rates in personal decisions and decisions made
on behalf of a specific unknown person when the intertemporal choice included the option of
receiving a SS gain immediately. This chapter compared personal decisions with suggestions to
others and with predictions of others’ decisions in a more general way, in that others’ decisions
included both decisions for close friends and for strangers. The results here do not conflict with
the findings of Albrecht et al. (2011), as people might treat close friends’ decisions differently
from strangers’ decisions. Rather than conflicting with the proposed affect-based hypothesis I,
that a higher level of affective intensity potentially elicited by choices results in more salient
self-other decision differences in discount rate, the findings suggest that the positive affect
elicited by the expected gains in this chapter may not have been sufficiently intense to cause self-
other decision differences in discount rate in general.
With respect to making decisions for others, the effect of the decision target in the delay
discounting index was detected only in the large-near condition and the effect of the decision
response in the delay discounting index was found only in the large-distant condition. No effects
of the decision target and decision response were found on the delay discounting index in both
the small-near and the small-distant conditions. As both the large-near and the large-distant
conditions potentially evoke higher levels of affective intensity during the process of making
decisions compared to the two other conditions, the results here support the proposed affect-
based hypothesis II, that a higher level of affective intensity potentially elicited by choices would
be more likely to trigger the effect of either the decision target or the decision response in
discount rate.
89
In the large-near condition, participants discounted less for a future gain in a close
friend’s choices than in a stranger’s choices. This is consistent with my hypothesis and suggests
further that a close relationship helps reduce the discount rate. In the large-distant condition,
people were found to discount less for a future gain in the suggestions offered to others
compared to their predictions of others’ decisions. This is consistent with my hypothesis about
the difference between suggestions and predictions. The results suggest further that people may
assign more weight to good consequences in suggestions to others in contrast to predictions of
others’ decisions.
The findings of the correlations between the delay discounting index and the empathy
scores suggest that the empathy could not explain any findings with respect to making
intertemporal choices for others here. As the empathy scale used in this chapter measures
empathy toward others in general and does not distinguish friends from strangers, it is possible
that the empathy could not account for the effect of the decision target here. It is also plausible
that people who describe themselves as being empathic treat others’ specific decisions in a
different way. In almost all types of decisions except for the POS, moderate to weak correlations
were found between the delay discounting index and the impulsiveness scores, which suggest a
possible connection between impulsive behavior and the discount rate measured in this study.
The delay discounting index in both the large-distant and the large-near conditions were not
correlated with any demographic factors for all types of decisions, which suggest that income,
education, age, and sex did not influence the discount rate regardless of the decision target and
the decision response, when people were asked to expedite a large future gain.
In summary, this study proposed two affect-based hypotheses to account for self-other
decision differences and the effects of the decision target and the decision response in the
90
discount rate across four different conditions. The findings essentially suggest a hypothetical
moderation effect of the affective intensity potentially elicited by the expected gains in the
influences from both the decision target and the decision response in the discount rate with
respect to making intertemporal choices for others. As an initial proposal, future studies should
be conducted to confirm this.
One limitation of this research is that the affective intensity during the process of making
decisions was evaluated and ranked according to the expected gains in the choices based on
previous findings of the connection between affect and intertemporal preferences. Although the
rank of the affective intensity was roughly confirmed by the delay discounting index results here,
future studies are expected to measure the affective intensity directly, such as self-reports etc. to
confirm its important role in intertemporal preferences with respect to self-other decisions. In
addition, this research examined only intertemporal choices about expediting a future gain with a
fixed duration of one year. According to previous research, delaying an instant gain is different
from expediting a future gain, as the former is always treated as an immediate loss (Loewenstein,
1988; Shelley, 1993). Future studies are expected to explore different types of intertemporal
choices.
91
Chapter 5 General Discussion
This research includes exploratory studies examining self-other decisions involving six
choice biases: 1) honoring sunk cost, 2) risk aversion toward gains, 3) risk seeking toward losses,
4) loss aversion, 5) ambiguity aversion toward gains, and 6) delay discounting. The results
revealed that self-other decision differences and the influences from both the decision target
(friend vs. stranger) and the decision response (suggestion vs. prediction) in making decisions for
others were contingent on the type of choice bias. These results have not been adequately
explained in the previous research. This research proposes a hypothetical affect-based framework
to account for the results across different choice biases.
Chapter 2 addressed self-other decision differences in the magnitude of honoring sunk
cost in both general others’ and specific others’ decisions. A difference between decisions made
for oneself and predictions of others’ decisions was found only for general others’ decisions, but
not specific others’ decisions. This result is consistent with previous findings for risk preferences
(e.g., Hsee & Weber, 1997; Faro & Rottenstreich, 2006), which suggest that self-other decision
differences with respect to both general others’ and specific others’ decisions are not restricted to
risk preferences. In addition, unlike previous studies, the studies reported in Chapter 2 examined
decisions made for a general close person and decisions made for a general unknown person and
the findings suggest that the disconnected feelings for general others might serve as a mediator in
self-other decision differences in choice bias for general others’ decisions.
Both Chapters 3 and 4 focused on specific others’ decisions, including suggestions to
close friends and strangers, and predictions of close friends’ and strangers’ decisions. Across five
different preference contexts (risk aversion toward gains, risk seeking toward losses, loss
aversion, ambiguity aversion toward gains, and delay discounting), self-other decision
92
differences were observed only in the loss aversion index both between decisions made for
oneself and suggestions to specific others, and between decisions made for oneself and
predictions of specific others’ decisions. As discussed in Chapter 3, among the four decision
dilemmas with respect to preferences toward risk, loss, and ambiguity, participants in the
decision dilemma of loss aversion were considered to experience the highest level of affective
intensity potentially elicited by choices. In addition, the evoked affect here was negative. The
intertemporal choices studied in Chapter 4 involved gains only, which would evoke positive
affect. As individuals generally experience more intense emotions elicited by negative incentives
than by positive incentives (for review, see Cacioppo & Gardner, 1999), the findings support the
proposed affect-based hypothesis I in this research that when the level of affective intensity
potentially elicited by choices is high enough, there are self-other decision differences in choice
bias for specific others’ decisions.
In addition to studying self-other decision differences, this research examined further the
effects of both the decision target and the decision response in the decisions made for specific
others. In Chapter 2, both the decision target and the decision response were found to have
effects on the magnitude of honoring sunk cost. However, the effect of the decision target was
found only on the loss aversion index in Chapter 3 and the discount rate in the condition of
expediting a near future large gain in Chapter 4. The effect of the decision response was found
only on the risk aversion index in Chapter 3 and the discount rate in the condition of expediting a
distant future large gain in Chapter 4.
According to the rank of affective intensity potentially elicited by choices discussed in
Chapter 3, participants were considered to have experienced higher levels of affective intensity
in the decision dilemmas of loss aversion and risk aversion toward gains than in the decision
93
dilemmas of risk seeking toward losses and ambiguity aversion toward gains. As discussed in
Chapter 4, participants were considered to have experienced higher levels of affective intensity
potentially elicited by the expected gains in the conditions of expediting a large future gain than
a small future gain. The effect of either the decision target or the decision response was not
observed on the risk seeking index or the ambiguity aversion index in Chapter 3, or the discount
rates in both conditions of expediting a small future gain in Chapter 4. As Loewenstein and
Lerner (2003) discussed, when the emotions experienced at the moment of decision making are
less intense, people can defeat the effect of emotions easily. Therefore, the affective intensity
potentially elicited by choices could be considered as an important determinant in making
decisions for specific others. The findings in both Chapters 3 and 4 support the proposed affect-
based hypothesis II that a higher level of affective intensity potentially elicited by choices is
more likely to trigger the effect of either the decision target or the decision response in choice
bias for specific others’ decisions.
Returning to the 20-item scale developed to measure the magnitude of honoring sunk cost
in the second study of Chapter 2, regardless of the prior cost, each item includes a personal
struggle or conflict, such as boredom vs. self-benefit, etc., when one chooses to continue initial
actions. As discussed in Chapter 2, affect is highly associated with honoring sunk cost. The
existing conflict in each item could potentially help induce more intense emotions. Thus, the
results related to the effects of both the decision target and the decision response on the
magnitude of honoring sunk cost further confirm a hypothetically important role of the affective
intensity potentially elicited by choices in making decisions for specific others.
Taken together, the results in this research provide convergent evidence for the proposal
of a hypothetical affect-based framework to account for self-other decision differences and
94
effects of both the decision target and the decision response in decisions made for others across
different choice biases. The framework is summarized in the following rules:
a) Mental image of other first needs to be considered in studying self-other decisions.
b) For self-other decisions between decisions made for oneself and predictions of general
others’ decisions, disconnected feelings for general others might serve as a mediator in
self-other decision differences in choice biases.
c) For decisions made for oneself vs. specific others, affective intensity potentially
elicited by choices might serve as a moderator in self-other decision differences in choice
biases
12
.
d) For decisions made for specific others, higher levels of affective intensity potentially
elicited by choices are more likely to result in differences related to the decision target
(friend vs. stranger) or the decision response (suggestion vs. prediction) in choice biases.
As discussed in Chapter 3, the rule c) in the proposed hypothetical affect-based framework could
account for a wide range of various inconsistent results related to self-other decision differences
in risk preferences in the literature. Future studies should test and confirm this framework.
In addition, the proposed hypothetical affect-based framework is compatible with
theoretical explanations that have been used to account for self-other decision differences in the
literature, as discussed in Chapter 1. For example, with respect to different information
processing, although people put more weight on positive outcomes for their friends’ decisions,
they might still make the same decisions for themselves as for their friends. With respect to
construal level theory, the temporal attribute is useful for evaluating different levels of affective
intensity potentially elicited by intertemporal choices based on the time when the gains will be
12
Here, decisions made for specific others are either the suggestion offered to specific others or the predictions of
specific others’ decisions. Others include both close friends and strangers.
95
received. The risk-as-feelings hypothesis essentially provides an affective framework for this
research with which to evaluate different levels of affective intensity potentially elicited by risky
choices.
For specific findings about self-other decision differences, compared to the magnitude of
honoring sunk cost in personal decisions, participants were found to predict general others to be
more likely to honor sunk cost. Participants were also found to predict specific others to be more
loss averse than themselves. It seems that when the self-other decision differences exist between
decisions for oneself and predictions of others’ decisions for certain choice biases, people tend to
consider others to be more likely to demonstrate the choice bias (consistent with the bias blind
spot; Pronin, Lin, & Ross, 2002) as well as avoid negative consequences.
With respect to the differences between personal decisions and suggestions to others for
specific choice biases, people were found to be more loss averse in the suggestions offered to
specific others than in the decisions made for themselves. One can argue that when the self-other
decision differences exist between decisions for oneself and suggestions to others for certain
choice biases, people might have additional concerns with respect to personal feelings or utilities
related to the interpersonal relationship with whom the suggestion is offered. For example, one
might not want to be get blamed when others face negative consequences after following their
suggestion, or else those who give suggestions to others may ponder over personal benefits from
offering recommendations such as feeling good, being helpful etc.
For the effect of the decision target for specific choice biases, participants were found to
be less likely to honor sunk cost, be more loss averse, and discount less for future gains in their
close friend’s decisions than strangers’ decisions. Given that individuals often experience intense
emotions in close relationship (for review, see Berscheid & Ammazzalorso, 2003), decision
96
makers generally would be more likely to avoid options with negative consequences in their
close friends’ decisions, as opposed to strangers’ decisions.
For the effect of the decision response for specific choice biases, participants were found
to be less likely to honor sunk cost, be more risk averse toward gains, and discount less for future
gains in the suggestions offered to specific others than in the predictions of specific others’
decisions. Given that suggestions involve more emotions than predictions, and advisors usually
want to maintain a good relationship with the advice recipients (Dana & Cain, 2015), decision
makers tend to choose more conservative options or options with positive outcomes in the
suggestions they offer to others in contrast to others’ decisions they predict.
Finally, the studies in this research have two limitations. First, there were no direct
measures for the level of affective intensity potentially elicited by choices in this research.
Possible levels of affective intensities were evaluated solely based on previous research on other
measures of affect. In the future work, the PANAS (Watson, Clark, & Tellegen, 1988) could be
added to measure the level of affective intensity that decision makers experience at the moment
of decision making. This could allow comparison of self-other decision differences and study
effects of both the decision target and the decision response for decisions made for others across
different domains. Second, this research studied only hypothetical choices. Future studies should
extend to field experiments involving relevant, concrete outcomes.
97
References
Albrecht, K., Volz, K. G., Sutter, M., Laibson, D. I., & von Cramon, D. Y. (2011). What is for
me is not for you: Brain correlates of intertemporal choice for self and other. Social
Cognitive and Affective Neuroscience, 6(2), 218–225. http://doi.org/10.1093/scan/nsq046
Andersson, O., Holm, H. J., Tyran, J.-R., & Wengström, E. (2016). Deciding for others reduces
loss aversion. Management Science, 62(1), 29–36. http://doi.org/10.1287/mnsc.2014.2085
Arkes, H. R. (1996). The psychology of waste. Journal of Behavioral Decision Making, 9(3),
213–224. http://doi.org/10.1002/(SICI)1099-0771(199609)9:3<213::AID-
BDM230>3.0.CO;2-1
Arkes, H. R., & Blumer, C. (1985). The psychology of sunk cost. Organizational Behavior and
Human Decision Processes, 35(1), 124–140. http://doi.org/10.1016/0749-5978(85)90049-4
Aron, A., Aron, E. N., Tudor, M., & Nelson, G. (1991). Close relationships as including other in
the self. Journal of Personality and Social Psychology, 60(2), 241–253.
http://doi.org/10.4324/9780203311851
Armor, D. A. (1998). The illusion of objectivity: A bias in the perception of freedom from bias
(Doctoral dissertation). Retrieved from ProQuest Dissertations and Theses database. (UMI
No. 9905589)
Atanasov, P., Anderson, B. L., Cain, J., Schulkin, J., & Dana, J. (2015). Comparing physicians
personal prevention practices and their recommendations to patients. Journal for
Healthcare Quality, 37(3), 189–198. http://doi.org/10.1111/jhq.12042
98
Bartels, D. M., & Urminsky, O. (2011). On intertemporal selfishness: How the perceived
instability of identity underlies impatient consumption. Journal of Consumer Research,
38(1), 182–198. http://doi.org/10.1086/658339
Bechara, A. (2004). The role of emotion in decision-making: Evidence from neurological
patients with orbitofrontal damage. Brain and Cognition, 55(1), 30–40.
http://doi.org/10.1016/j.bandc.2003.04.001
Bechara, A., & Damasio, A. R. (2005). The somatic marker hypothesis: A neural theory of
economic decision. Games and Economic Behavior, 52(2), 336–372.
http://doi.org/10.1016/j.geb.2004.06.010
Beisswanger, A. H., Stone, E. R., Hupp, J. M., & Allgaier, L. (2003). Risk taking in
relationships: Differences in deciding for oneself versus for a friend. Basic and Applied
Social Psychology, 25(2), 121–135. http://doi.org/10.1207/S15324834BASP2502_3
Berscheid, E., & Ammazzalorso, H. (2003). Emotional experience in close relationships. In G. J.
O. Fletcher & M. S. Clark (Eds.), Blackwell handbook of social psychology: Interpersonal
processes (308–330). Malden, MA: Blackwell Publishers.
Bornstein, B. H., & Chapman, G. B. (1995). Learning lessons from sunk costs. Journal of
Experimental Psychology: Applied, 1(4), 251–269.
Bornstein, B. H., Emler, C., & Chapman, G. B. (1999). Rationality in medical treatment
decisions: Is there a sunk-cost effect? Social Science & Medicine, 49(2), 215–222.
http://doi.org/10.1016/S0277-9536(99)00117-3
99
Braverman, J. A., & Blumenthal-Barby, J. S. (2012). Assessment of the sunk-cost effect in
clinical decision-making. Social Science and Medicine, 75(1), 186–192.
http://doi.org/10.1016/j.socscimed.2012.03.006
Brockner, J. (1992). The escalation of commitment to a failing course of action: Toward
theoretical progress. Academy of Management Review, 17(1), 39–61.
http://doi.org/10.5465/AMR.1992.4279568
Buehler, R., Griffin, D., & Ross, M. (1994). Exploring the “planning fallacy”: Why people
underestimate their task completion times. Journal of Personality and Social Psychology,
67(3), 366–381. http://doi.org/10.1037/0022-3514.67.3.366
Cacioppo, J. T., & Gardner, W. L. (1999). Emotion. Annual Review of Psychology, 50, 191–214.
Cacioppo, J. T., Petty, R. E., & Kao, C. F. (1984). The efficient assessment of need for cognition.
Journal of Personality Assessment, 48(3), 306–307.
http://doi.org/10.1207/s15327752jpa4803_13
Camerer, C. (2012). Three cheers-psychological, theoretical, empirical-for loss aversion. Journal
of Marketing, 42(2), 129–133.
Cialdini, R. B., Brown, S. L., Lewis, B. P., Luce, C., & Neuberg, S. L. (1997). Reinterpreting the
empathy-altruism relationship: When one into one equals oneness. Journal of Personality
and Social Psychology, 73(3), 481–494.
Coleman, M. D. (2010). Sunk cost and commitment to medical treatment. Current Psychology,
29(2), 121–134. http://doi.org/10.1007/s12144-010-9077-7
100
Conlon, E. J., & Parks, J. M. (1987). Information requests in the context of escalation. Journal of
Applied Psychology, 72(3), 344–350. http://doi.org/10.1037//0021-9010.72.3.344
Cunha, M., & Caldieraro, F. (2009). Sunk-cost effects on purely behavioral investments.
Cognitive Science, 33(1), 105–113. http://doi.org/10.1111/j.1551-6709.2008.01005.x
Dana, J., & Cain, D. M. (2015). Advice versus choice. Current Opinion in Psychology, 6, 173–
176. http://doi.org/10.1016/j.copsyc.2015.08.019
Dave, C., Eckel, C. C., Johnson, C. A., & Rojas, C. (2010). Eliciting risk preferences: When is
simple better? Journal of Risk and Uncertainty, 41(3), 219–243.
http://doi.org/10.1007/s11166-010-9103-z
Davis, M. H. (1980). A multidimensional approach to individual differences in empathy. JSAS
Catalog of Selected Documents in Psychology, 10, 85.
Desteno, D., Li, Y., Dickens, L., & Lerner, J. S. (2014). Gratitude: A tool for reducing Economic
impatience. Psychological Science, 25(6), 1262–1267.
http://doi.org/10.1177/0956797614529979
Dore, R. A., Stone, E. R., & Buchanan, C. M. (2014). A social values analysis of parental
decision making. The Journal of Psychology, 148(4), 477–504.
http://doi.org/10.1080/00223980.2013.808603
Ehrlinger, J., Gilovich, T., & Ross, L. (2005). Peering into the bias blind spot: People's
assessments of bias in themselves and others. Personality and Social Psychology Bulletin,
31(5), 680-692. http://doi.org/10.1177/0146167204271570
101
Ellsberg, D. (1961). Risk, ambiguity, and the Savage axioms. Quarterly Journal of Economics,
75(4), 643–669.
Ersner-Hershfield, H., Wimmer, G. E., & Knutson, B. (2009). Saving for the future self: Neural
measures of future self-continuity predict temporal discounting. Social Cognitive and
Affective Neuroscience, 4(1), 85–92. http://doi.org/10.1093/scan/nsn042
Fagerlin, A., Zikmund-Fisher, B. J., Ubel, P. A., Jankovic, A., Derry, H. A., & Smith, D. M.
(2007). Measuring numeracy without a math test: Development of the subjective numeracy
scale. Medical Decision Making, 27(5), 672–680.
http://doi.org/10.1177/0272989X07304449
Faro, D., & Rottenstreich, Y. (2006). Affect, empathy, and regressive mispredictions of others’
preferences under risk. Management Science, 52(4), 529–541.
http://doi.org/10.1287/mnsc.1050.0490
Frisch, D. (1993). Reasons for framing effects. Organizational Behavior and Human Decision
Processes, 54(3), 399–429. http://doi.org/10.1006/obhd.1993.1017
Frisch, D., & Baron, J. (1988). Ambiguity and rationality. Journal of Behavioral Decision
Making, 1(3), 149–157. http://doi.org/10.1002/bdm.3960010303
Garcia-Retamero, R., & Galesic, M. (2012). Doc, what would you do if you were me? On self-
other discrepancies in medical decision making. Journal of Experimental Psychology:
Applied, 18(1), 38–51. http://doi.org/10.1037/a0026018
102
Garcia-Retamero, R., Okan, Y., & Maldonado, A. (2015). The impact of depression on self-other
discrepancies in decision making. Journal of Behavioral Decision Making, 28(1), 89–100.
http://doi.org/10.1002/bdm.1833
Garland, H. (1990). Throwing good money after bad: The effect of sunk costs on the decision to
escalate commitment to an ongoing project. Journal of Applied Psychology, 75(6), 728–
731. http://doi.org/10.1037/0021-9010.75.6.728
Garland, H., & Conlon, D. E. (1998). Too close to quit: The role of project completion in
maintaining commitment. Journal of Applied Social Psychology, 28(22), 2025–2048.
http://doi.org/10.1111/j.1559-1816.1998.tb01359.x
Gilovich, T., & Medvec, V. H. (1994). The temporal pattern to the experience of regret. Journal
of Personality and Social Psychology, 67(3), 357–365. http://doi.org/10.1037/0022-
3514.67.3.357
Gul, F. (1991). A theory of disappointment aversion. Econometrica, 59(3), 667–686.
Gunia, B. C., Sivanathan, N., & Galinsky, A. D. (2009). Vicarious entrapment: Your sunk costs,
my escalation of commitment. Journal of Experimental Social Psychology, 45(6), 1238–
1244. http://doi.org/10.1016/j.jesp.2009.07.004
Hafenbrack, A. C., Kinias, Z., & Barsade, S. G. (2014). Debiasing the mind through meditation:
Mindfulness and the sunk-cost bias. Psychological Science, 25(2), 369–376.
http://doi.org/10.1177/0956797613503853
103
Heath, C. (1995). Escalation and de-escalation of commitment in response to sunk costs: The
role of budgeting in mental accounting. Organizational Behavior and Human Decision
Processes, 62(1), 38–54. http://doi.org/10.1006/obhd.1995.1029
Hershfield, H. E. (2011). Future self-continuity: how conceptions of the future self transform
intertemporal choice. Annals of the New York Academy of Sciences, 1235, 30–43.
http://doi.org/10.1111/j.1749-6632.2011.06201.x.Future
Higgins, E. T., Friedman, R. S., Harlow, R. E., Idson, L. C., Ayduk, O. N., & Taylor, A. (2001).
Achievement orientations from subjective histories of success: Promotion pride versus
prevention pride. European Journal of Social Psychology, 31(1), 3–23.
http://doi.org/10.1002/ejsp.27
Hsee, C. K., & Weber, E. U. (1997). A fundamental prediction error: Self–others discrepancies
in risk preference. Journal of Experimental Psychology: General, 126(1), 45–53.
http://doi.org/10.1037/0096-3445.126.1.45
Jonas, E., & Frey, D. (2003). Information search and presentation in advisor–client interactions.
Organizational Behavior and Human Decision Processes, 91(2), 154–168.
http://doi.org/10.1016/S0749-5978(03)00059-1
Jonas, E., Schulz-Hardt, S., & Frey, D. (2005). Giving advice or making decisions in someone
else’s place: The influence of impression, defense, and accuracy motivation on the search
for new information. Personality and Social Psychology Bulletin, 31(7), 977–990.
http://doi.org/10.1177/0146167204274095
104
Kahneman, D., & Frederick, S. (2002). Representativeness revisited. In T. Gilovich, D. Griffin,
& D. Kahneman (Eds.), Heuristics and biases (pp. 49–81). Cambridge, England:
Cambridge University Press.
Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk.
Econometrica, 47(2), 263–292.
Kahneman, D., & Tversky, A. (1982). The simulation heuristic. In D. Kahneman, P. Slovic, & A.
Tversky (Eds.), Judgment under uncertainty: Heuristics and biases (pp. 201–208). New
York, NY: Cambridge University Press.
Kirby, K. N., & Herrnstein, R. J. (1995). Preference reversals due to myopic discounting of
delayed reward. Psychological Science, 6(2), 83–89.
Knutson, B., & Peterson, R. (2005). Neurally reconstructing expected utility. Games and
Economic Behavior, 52(2), 305–315. http://doi.org/10.1016/j.geb.2005.01.002
Kray, L. J. (2000). Contingent weighting in self-other decision making. Organizational Behavior
and Human Decision Processes, 83(1), 82–106. http://doi.org/10.1006/obhd.2000.2903
Kray, L. J., & Gonzalez, R. (1999). Differential weighting in choice versus advice: I’ll do this,
you do that. Journal of Behavioral Decision Making, 12(3), 207–217.
http://doi.org/10.1002/(SICI)1099-0771(199909)12:3<207::AID-BDM322>3.0.CO;2-P
Krishnamurthy, P., & Kumar, P. (2002). Self-other discrepancies in waiting time decisions.
Organizational Behavior and Human Decision Processes, 87(2), 207–226.
http://doi.org/10.1006/obhd.2001.2980
105
Lai, L. (2010). Maximizing without difficulty: A modified maximizing scale and its correlates.
Judgment and Decision Making, 5(3), 164–175.
Lazarus, R. S. (1991). Emotion and adaptation. New York: Oxford University Press.
Lerner, J. S., Li, Y., Valdesolo, P., & Kassam, K. S. (2015). Emotion and decision making.
Annual Review of Psychology, 66, 799–823. http://doi.org/10.1146/annurev-psych-010213-
115043
Lerner, J. S., Li, Y., & Weber, E. U. (2013). The financial costs of sadness. Psychological
Science, 24(1), 72–79. http://doi.org/10.1177/0956797612450302
Liberman, N., & Trope, Y. (1998). The role of feasibility and desirability considerations in near
and distant future decisions: A test of temporal construal theory. Journal of Personality and
Social Psychology, 75(1), 5–18. http://doi.org/10.1037/0022-3514.75.1.5
Liu, Y., Polman, E., Liu, Y., & Jiao, J. (2018). Choosing for others and its relation to information
search. Organizational Behavior and Human Decision Processes, 147(May), 65–75.
http://doi.org/10.1016/j.obhdp.2018.05.005
Loewenstein, G. F. (1988). Frames of mind in intertemporal choice. Management Science, 34(2),
200–214.
Loewenstein, G. F. (1996). Out of control: Visceral influences on behavior. Organizational
Behavior and Human Decision Processes, 65(3), 272–292.
http://doi.org/10.1006/obhd.1996.0028
106
Loewenstein, G., & Lerner, J. S. (2003). The role of affect in decision making. In R. Davidson,
H. Goldsmith, & K. Scherer (Eds.), Handbook of affective science (pp. 619-642). Oxford:
Oxford University Press.
Loewenstein, G., & Schkade, D. (1999). Wouldn’t it be nice? Predicting future feelings. In D.
Kahneman, E. Diener, & N. Schwarz (Eds.), Well-being: The foundations of hedonic
psychology (pp. 85–105). New York: Russell Sage Foundation.
Loewenstein, G. F., Weber, E. U., Hsee, C. K., & Welch, N. (2001). Risk as feelings.
Psychological Bulletin, 127(2), 267–286. http://doi.org/10.1037/0033-2909.127.2.267
Loomes, G., & Sugden, R. (1982). Regret theory: An alternative theory of rational choice under
uncertainty. The Economic Journal, 92(368), 805–824.
Lopes, L. L., & Oden, G. C. (1999). The role of aspiration level in risky choice: A comparison of
cumulative prospect theory and SP/A theory. Journal of Mathematical Psychology, 43(2),
286–313. http://doi.org/10.1006/jmps.1999.1259
Lu, J., & Xie, X. (2014). To change or not to change: A matter of decision maker’s role.
Organizational Behavior and Human Decision Processes, 124(1), 47–55.
http://doi.org/10.1016/j.obhdp.2013.12.001
Lu, J., Xie, X., & Xu, J. (2013). Desirability or feasibility: Self-other decision-making
differences. Personality and Social Psychology Bulletin, 39(2), 144–155.
http://doi.org/10.1177/0146167212470146
107
Malkoc, S. A., & Zauberman, G. (2006). Deferring versus expediting consumption: The effect of
outcome concreteness on sensitivity to time horizon. Journal of Marketing Research, 43(4),
618–627. http://doi.org/10.1509/jmkr.43.4.618
Mazur, J. E. (1987). An adjusting procedure for studying delayed reinforcement. In M. L.
Commons, J. E. Mazur, J. A. Nevin, & H. Rachlin (Eds.), Quantitative analyses of
behavior, Vol. 5. The effect of delay and of intervening events on reinforcement value (pp.
55–73). Hillsdale, NJ, US: Lawrence Erlbaum Associates, Inc.
Mellers, B. A., Schwartz, A., & Ho, K. (1997). Decision affect theory: Emotional reactions to the
outcomes of risky options. Psychological Science, 8(6), 423–429.
Olivola, C. Y. (2018). The interpersonal sunk-cost effect. Psychological Science,
095679761775264. http://doi.org/10.1177/0956797617752641
Ong, D. C., Goodman, N. D., & Zaki, J. (2018). Happier than thou? A self-enhancement bias in
emotion attribution. Emotion, 18(1), 116–126. http://doi.org/10.1037/emo0000309
Pak, N. (2018, March 8). Credit card debt surpasses $1 trillion in the US for first time. ABC
News. Retrieved from https://abcnews.go.com/Business/credit-card-debt-surpasses-trillion-
us-time/story?id=53608548
Patton, J. H., Stanford, M. S., & Barratt, E. S. (1995). Factor structure of the Barratt
impulsiveness scale. Journal of Clinical Psychology, 51(6), 768–774.
Polman, E. (2010). Information distortion in self-other decision making. Journal of Experimental
Social Psychology, 46(2), 432–435. http://doi.org/10.1016/j.jesp.2009.11.003
108
Polman, E. (2012a). Effects of self-other decision making on regulatory focus and choice
overload. Journal of Personality and Social Psychology, 102(5), 980–993.
http://doi.org/10.1037/a0026966
Polman, E. (2012b). Self-other decision making and loss aversion. Organizational Behavior and
Human Decision Processes, 119(2), 141–150. http://doi.org/10.1016/j.obhdp.2012.06.005
Polman, E., & Emich, K. J. (2011). Decisions for others are more creative than decisions for the
self. Personality and Social Psychology Bulletin, 37(4), 492–501.
http://doi.org/10.1177/0146167211398362
Polman, E., & Vohs, K. D. (2016). Decision fatigue, choosing for others, and self-construal.
Social Psychological and Personality Science, 7(5), 471–478.
http://doi.org/10.1177/1948550616639648
Premack, D., & Woodruff, G. (1978). Does the chimpanzee have a theory of mind? Behavioral
and Brain Sciences, 4, 515–526.
Pronin, E. (2007). Perception and misperception of bias in human judgment. Trends in Cognitive
Sciences, 11(1), 37-43. http://doi.org/10.1016/j.tics.2006.11.001
Pronin, E. (2008). How we see ourselves and how we see others. Science, 320(5880), 1177-1180.
http://doi.org/10.1126/science.1154199
Pronin, E., Lin, D. Y., & Ross, L. (2002). The bias blind spot: Perceptions of bias in self versus
others. Personality and Social Psychology Bulletin, 28(3), 369-381.
http://doi.org/10.1177/0146167202286008
109
Roszkowski, M. J., & Snelbecker, G. E. (1990). Effects of “framing” on measures of risk
tolerance: Financial planners are not immune. Journal of Behavioral Economics, 19(3),
237–246. http://doi.org/10.1016/0090-5720(90)90029-7
Samuelson, W., & Zeckhauser, R. (1988). Status quo bias in decision making. Journal of Risk
and Uncertainty, 1, 7–59. http://doi.org/10.1007/BF00055564
Shelley, M. K. (1993). Outcome signs, question frames and discount rates. Management Science,
39(7), 806–815. http://doi.org/10.1287/mnsc.39.7.806
Shupp, R., Loveridge, S., Skidmore, M., Lim, J., & Rogers, C. (2017). Risk, loss, and ambiguity
aversion after a natural disaster. Economics of Disasters and Climate Change, 1, 121–142.
http://doi.org/10.1007/s41885-017-0013-2
Slovic, P. (2000). What does it mean to know a cumulative risk? Adolescents’ perceptions of
short-term and long-term consequences of smoking. Journal of Behavioral Decision
Making, 13(2), 259–266. http://doi.org/10.1002/(sici)1099-0771(200004/06)13:2<259::aid-
bdm336>3.3.co;2-y
Soman, D. (2001). The mental accounting of sunk time costs: Why time is not like money.
Journal of Behavioral Decision Making, 14(3), 169–185. http://doi.org/10.1002/bdm.370
Soster, R. L., Monga, A., & Bearden, W. O. (2010). Tracking costs of time and money: How
accounting periods affect mental accounting. Journal of Consumer Research, 37(4), 712–
721. http://doi.org/10.1086/656388
110
Staw, B. M. (1976). Knee-deep in the big muddy: A study of escalating commitment to a chosen
course of action. Organizational Behavior and Human Performance, 16(1), 27–44.
Staw, B. M., & Fox, F. V. (1977). Escalation: The determinants of commitment to a chosen
course of action. Human Relations, 30(5), 431–450.
http://doi.org/10.1177/001872677703000503
Stone, E. R., & Allgaier, L. (2008). A social values analysis of self-other differences in decision
making involving risk. Basic and Applied Social Psychology, 30(2), 114–129.
http://doi.org/10.1080/01973530802208832
Stone, E. R., Choi, Y., Bruine de Bruin, W., & Mandel, D. R. (2013). I can take the risk, but you
should be safe: Self-other differences in situations involving physical safety. Judgment and
Decision Making, 8(3), 250.
Stone, E. R., Yates, A. J., & Caruthers, A. S. (2002). Risk taking in decision making for others
versus the self. Journal of Applied Social Psychology, 32(9), 1797–1824.
http://doi.org/10.1111/j.1559-1816.2002.tb00260.x
Strough, J., Schlosnagle, L., Karns, T. E., Lemaster, P., & Pichayayothin, N. (2014). No time to
waste: Restricting life-span temporal horizons decreases the sunk-cost fallacy. Journal of
Behavioral Decision Making, 27(1), 78–94. http://doi.org/10.1002/bdm.1781
Thaler, R. H. (1980). Toward a positive theory of consumer choice. Journal of Economic
Behavior & Organization, 1(1), 39–60. http://doi.org/10.1016/0167-2681(80)90051-7
111
Thaler, R. H. (1981). Some empirical evidence on dynamic inconsistency. Economics Letters,
8(3), 201–207. http://doi.org/10.1016/0165-1765(81)90067-7
Trope, Y., & Liberman, N. (2000). Temporal construal and time-dependent changes in
preference. Journal of Personality and Social Psychology, 79(6), 876–889.
http://doi.org/10.1037//0022-3514.79.6.876
Trope, Y., & Liberman, N. (2003). Temporal construal. Psychological Review, 110(3), 403–421.
http://doi.org/10.1037/0033-295X.110.3.403
Trope, Y., & Liberman, N. (2010). Construal-level theory of psychological distance.
Psychological Review, 117(2), 440–463. http://doi.org/10.1037/a0018963
Urminsky, O., & Zauberman, G. (2015). The psychology of intertemporal preferences. In G.
Keren & G. Wu (Eds.), The Wiley Blackwell handbook of judgment and decision making
(pp. 141–181). Chichester, England: John Wiley & Sons.
http://doi.org/10.1002/9781118468333.ch5
Van Boven, L., Dunning, D., & Loewenstein, G. (2000). Egocentric empathy gaps between
owners and buyers: Misperceptions of the endowment effect. Journal of Personality and
Social Psychology, 79(1), 66-76. http://doi.org/10.1037/0022-3514.79.1.66
van Dijk, E., & Zeelenberg, M. (2006). The dampening effect of uncertainty on positive and
negative emotions. Journal of Behavioral Decision Making, 176, 171–176.
http://doi.org/10.1002/bdm.504
112
Vlaev, I., Wallace, B., Wright, N., Nicolle, A., Dolan, P., & Dolan, R. (2017). Other people’s
money: The role of reciprocity and social uncertainty in decisions for others. Journal of
Neuroscience, Psychology, and Economics, 10(2–3), 59–80.
http://doi.org/10.1037/npe0000063
Wilson, T. D., Houston, C. E., Etling, K. M., & Brekke, N. (1996). A new look at anchoring
effects: Basic anchoring and its antecedents. Journal of Experimental Psychology: General,
125(4), 387-402. http://doi.org/10.1037/0096-3445.125.4.387
Weber, B. J., & Chapman, G. B. (2005). Playing for peanuts: Why is risk seeking more common
for low-stakes gambles? Organizational Behavior and Human Decision Processes, 97(1),
31–46. http://doi.org/10.1016/j.obhdp.2005.03.001
Wray, L. D., & Stone, E. R. (2005). The role of self-esteem and anxiety in decision making for
self versus others in relationships. Journal of Behavioral Decision Making, 18(2), 125–144.
http://doi.org/10.1002/bdm.490
Zikmund-Fisher, B. J., Sarr, B., Fagerlin, A., & Ubel, P. A. (2006). A matter of perspective:
Choosing for others differs from choosing for yourself in making treatment decisions.
Journal of General Internal Medicine, 21(6), 618–622. http://doi.org/10.1111/j.1525-
1497.2006.00410.x
113
Appendix A: A 20-item Scale to Measure the Magnitude of Honoring Sunk
Cost in Making Decisions for Oneself in Chapter 2
Instructions: In the following section, you will be asked about your own decisions. Each of the
following 20 items presents an everyday decision problem with 3 different action options. For
each item, please choose the option that best captures your action.
1A. You have been taking a free online business course for self-fulfillment. There is a total of 10
weekly classes. Now, after 2 weeks, you feel very bored and really want to drop the course.
What would you do?
1B. You have been taking an online business course for self-fulfillment, which cost you $400.
The money is nonrefundable. There is a total of 10 weekly classes. Now, after 2 weeks, you feel
very bored and really want to drop the course. What would you do?
I would decide to…
a. immediately stop taking any more classes
b. take 1 or 2 more classes and then decide
c. complete all the remaining classes
2A. You go out to dinner at a restaurant. After dinner, the waiter tells you the desserts are free
because it is the restaurant’s one-year anniversary, and you order a dessert. It is wonderful but
very rich, and after 2 bites, you feel very full. However, you don’t want to take it home. What
would you do?
2B. You go out to dinner at a restaurant. After dinner, you order a dessert, which cost you $18. It
is wonderful but very rich, and after 2 bites, you feel very full. However, you don’t want to take
it home. What would you do?
I would decide to…
114
a. immediately stop eating the dessert
b. eat a few more bites of the dessert and then decide
c. finish eating all of the dessert
3A. For the past week, you have been playing a free online video game with several friends.
There are 50 levels in total. You are currently at level 10 and have started to lose interest and feel
exhausted while playing the game. What would you do?
3B. For the past week, you have been playing an online video game with several friends. You
have paid $120 for all necessary upgrades for the game. There are 50 levels in total. You are
currently at level 10 and have started to lose interest and feel exhausted while playing the game.
What would you do?
I would decide to…
a. immediately stop playing the game
b. play another 5 to 10 levels and then decide
c. attempt to complete all 50 levels of the game
4A. Due to the holiday season, you decide to donate some old clothes to charity. While cleaning
out your closet, you find 4 old cotton shirts that were gifts from friends and that still look pretty
good. You haven’t worn the shirts for at least 2 years as they don’t fit you anymore. What would
you do?
4B. Due to the holiday season, you decide to donate some old clothes to charity. While cleaning
out your closet, you find 4 old silk shirts that you bought a long time ago. Each cost you about
$100. They still look pretty good. You haven’t worn the shirts for at least 2 years as they don’t fit
you anymore. What would you do?
I would decide to…
115
a. donate all 4 shirts
b. keep my favorite shirt and donate the other 3
c. keep all 4 shirts
5A. You recently borrowed a best-selling book on self-improvement from a nearby library. You
have read only 2 of the 10 chapters. However, the story is really not as interesting as you had
hoped it would be. Now you are easily distracted by other things while reading it. What would
you do?
5B. You recently bought a best-selling book on self-improvement from Amazon, which cost you
$26. You have read only 2 of the 10 chapters. However, the story is really not as interesting as
you had hoped it would be. Now you are easily distracted by other things while reading it. What
would you do?
I would decide to…
a. immediately stop reading the book
b. read 1 or 2 more chapters and then decide
c. finish reading the book
6A. You used to run outside every day. As your neighbor moved to another state several months
ago, the neighbor left his treadmill to you as a gift. The treadmill is pretty new and contains
preprogrammed workouts designed to reach specific fitness goals. You started using the
treadmill and stopped running outside. However, after 2 months, you now have lost interest in
running at home on a treadmill. You miss running outside. What would you do?
6B. You used to run outside every day. Several months ago, you bought a treadmill that contains
preprogrammed workouts designed to reach specific fitness goals. It cost you $500. You started
116
using the treadmill and stopped running outside. However, after 2 months, you now have lost
interest in running at home on a treadmill. You miss running outside. What would you do?
I would decide to…
a. immediately stop using the treadmill and run outside
b. continue using the treadmill for a few more weeks and then decide
c. continue using the treadmill to run at home
7A. To challenge yourself, you registered for a local marathon 6 months ago and have been
training regularly for it. In order to encourage employees to do more sports, your company paid
your entry fee. However, two days before the marathon, you have a fever. On the day of the race,
you feel very fatigued at the starting line. What would you do?
7B. To challenge yourself, you registered for a local marathon 6 months ago and have been
training regularly for it. The entry fee cost you $185. However, two days before the marathon,
you have a fever. On the day of the race, you feel very fatigued at the starting line. What would
you do?
I would decide to…
a. not begin the marathon and go back home to rest
b. walk for the first 5 to 10 minutes of the marathon course and then decide
c. attempt to run as much of the marathon as possible, as planned
8A. You like to listen to music. Your cousin recently bought a very fancy smartphone speaker
for you as a birthday gift. The speaker provides great-sounding music. However, when using it,
you always need to connect the phone to the speaker. After using it for 2 weeks, you are very
tired of switching the phone back and forth for receiving calls, texting, and checking emails
etc.… What would you do?
117
8B. You like to listen to music. You recently bought a very fancy smartphone speaker, which
cost you $160. As a final-sale item, the speaker cannot be returned. The speaker provides great-
sounding music. However, when using it, you always need to connect the phone to the speaker.
After using it for 2 weeks, you are very tired of switching the phone back and forth for receiving
calls, texting, and checking emails etc.… What would you do?
I would decide to…
a. immediately stop using the speaker
b. continue using the speaker for another week or two and then decide
c. continue listening to music with the speaker
9A. You are staying in a hotel on vacation. Because it is raining outside, you decide to stay
inside for the evening. You start to watch a documentary film on your laptop about a topic that
you have great interest in and always wanted to know more about. However, after 5 minutes, you
are very disappointed with the quality of the film and the depth of information. What would you
do?
9B. You are staying in a hotel on vacation. Because it is raining outside, you decide to stay
inside for the evening. You pay $15 to watch a documentary film on your laptop about a topic
that you have great interest in and always wanted to know more about. However, after 5 minutes,
you are very disappointed with the quality of the film and the depth of information. What would
you do?
I would decide to…
a. immediately stop watching the film
b. watch for 5 to 10 more minutes and then decide
c. continue watching the film until the end
118
10A. At your company’s end-of-the-year party, you were randomly selected to win a 2-day-1-
night ski getaway at a nearby resort. You like skiing and were extremely excited about this. You
are checking in your hotel. You had planned to ski both this afternoon and tomorrow morning.
The reservations cannot be canceled or rescheduled at this late date. However, when you left
home this morning, one of your eyes is red and appears to be infected. Though it doesn’t affect
your vision, it is itchy. The eye tears a lot from either bright light or cold wind. What would you
do?
10B. You got a very good deal from Groupon for a 2-day-1-night ski getaway at a nearby resort,
which cost you $200. You like skiing and were extremely excited about this. You are checking in
your hotel. You had planned to ski both this afternoon and tomorrow morning. The reservations
cannot be canceled or rescheduled at this late date. However, when you left home this morning,
one of your eyes is red and appears to be infected. Though it doesn’t affect your vision, it is
itchy. The eye tears a lot from either bright light or cold wind. What would you do?
I would decide to…
a. not go skiing and stay in the hotel until my eye returns to normal
b. go skiing this afternoon and then decide
c. continue with the ski trip as planned
11A. In the past few months, you have been working on a project to develop a new hobby for
personal development. However, lately you have lost interest in the project. Whenever you work
on the project, you are bored and would rather be doing something else. What would you do?
11B. In the past few months, you have been working on a project to develop a new hobby for
personal development and have spent $280 on supplies. However, lately you have lost interest in
119
the project. Whenever you work on the project, you are bored and would rather be doing
something else. What would you do?
I would decide to…
a. immediately stop working on the project
b. continue working on the project for another couple of weeks and then decide
c. continue with the project until self-development goals are achieved
12A. You got your cell phone for free about 2 years ago after joining a wireless phone plan.
Your phone is still fully functional, but the battery charge is limited. The battery recently cannot
be fully charged and is sealed inside and cannot be replaced. You have to charge it at least 3
times a day now, which is quite inconvenient for you. What would you do?
12B. You have used your cell phone for about 2 years. A few weeks ago, due to a broken screen,
you spent $129 to fix it. Your phone is still fully functional, but the battery charge is limited. The
battery recently cannot be fully charged and is sealed inside and cannot be replaced. You have to
charge it at least 3 times a day now, which is quite inconvenient for you. What would you do?
I would decide to…
a. immediately replace this cell phone
b. continue using this cell phone for a few more weeks and then decide
c. continue using this cell phone until the battery will not charge at all
13A. You are trying to lose weight and increase your level of fitness. You signed up for a free
10-week fitness and weight loss program a few weeks ago. After 4 weeks of following the
program, you find that you still have not lost any weight and your fitness level has not improved.
You are beginning to get discouraged. What would you do?
120
13B. You are trying to lose weight and increase your level of fitness. You signed up for a 10-
week fitness and weight loss program a few weeks ago, which cost you $500. The money is
nonrefundable. After 4 weeks of following the program, you find that you still have not lost any
weight and your fitness level has not improved. You are beginning to get discouraged. What
would you do?
I would decide to…
a. immediately stop the program and search for a new one
b. continue in the program for 1 or 2 more weeks and then decide
c. continue in the program till it ends
14A. After finishing some survey questionnaires online, you luckily were selected to win season
tickets to attend 9 home games of a sports team. However, after 3 games, your favorite player
now is injured and won’t return until next season. You really enjoy watching the player and are
very disappointed that the player won’t return until next season. What would you do?
14B. You spent $325 on season tickets to attend 9 home games of a sports team. However, after
3 games, your favorite player now is injured and won’t return until next season. You really enjoy
watching the player and are very disappointed that the player won’t return until next season.
What would you do?
I would decide to…
a. immediately stop attending games altogether
b. attend 1 or 2 more games and then decide
c. continue attending the remaining games I have tickets for
15A. You have registered for a one-day free workshop this coming Saturday to acquire some
new skills that you have always wanted to have. The workshop includes 2 sessions in the
121
morning and 2 sessions in the afternoon. However, on the Saturday morning, when you get up,
you have a headache and a sore throat and really don’t feel well. What would you do?
15B. You have registered for a one-day workshop this coming Saturday to acquire some new
skills that you have always wanted to have. The workshop includes 2 sessions in the morning
and 2 sessions in the afternoon, and costs you $180. The money is nonrefundable. However, on
the Saturday morning, when you get up, you have a headache and a sore throat and really don’t
feel well. What would you do?
I would decide to…
a. not go to the workshop and stay at home
b. go and attend 1 or 2 of the morning sessions and then decide
c. attend the 1-day workshop as planned
16A. You just went to the doctor’s office in the morning and learned that your cholesterol and
blood pressure were too high. The doctor put you on a low-fat, low-sodium diet. However, due to
a sales promotion, you just got a box of 7 frozen dinners for free last weekend. You only tried
one dinner and really liked it, although the dinners are high in both sodium and fat. What would
you do?
16B. You just went to the doctor’s office in the morning and learned that your cholesterol and
blood pressure were too high. The doctor put you on a low-fat, low-sodium diet. However, you
just bought a box of 7 frozen dinners last weekend, which cost you $42. You only tried one
dinner and really liked it, although the dinners are high in both sodium and fat. What would you
do?
I would decide to…
a. discard the remaining dinners and not eat any of them
122
b. eat 1 or 2 of the remaining dinners and then decide
c. eat all of the remaining dinners in the freezer
17A. You recently won a free 3-month membership at a health club and the membership includes
6 free lessons for an activity of your choice. At the end of 3 months, the free membership and
lessons expire. However, after the second lesson, you develop an injury and it is somewhat
painful to take the lessons. The doctor tells you that the pain will continue for about 3 months.
What would you do?
17B. You recently paid $300 for a 3-month membership at a health club and the membership
includes 6 lessons for an activity of your choice. The fee is nonrefundable. At the end of 3
months, the membership and lessons expire. However, after the second lesson, you develop an
injury and it is somewhat painful to take the lessons. The doctor tells you that the pain will
continue for about 3 months. What would you do?
I would decide to…
a. immediately stop taking any more lessons
b. take 1 or 2 more lessons and then decide
c. continue taking the remaining lessons as planned
18A. You usually walk to work. It takes you about 15 minutes. Today you decide to take the bus
to work instead of walking because you received a week’s worth of bus passes for free and the
bus ride would be more comfortable than walking. The bus is quite unreliable, and it is running
20 minutes behind schedule today. What would you do?
18B. You usually walk to work. It takes you about 15 minutes. Today you decide to take the bus
to work instead of walking because you bought a week’s worth of bus passes for $20 and the bus
123
ride would be more comfortable than walking. The bus is quite unreliable, and it is running 20
minutes behind schedule today. What would you do?
I would decide to…
a. immediately stop using the bus passes and walk to work
b. try the bus tomorrow and then decide
c. continue using the bus passes for the rest of the week
19A. You recently got a box of ginseng tea from your cousin. The tea is supposed to be very
healthful by aiding in digestive disorders, and boosting the immune system etc. However, you
find that the taste and the smell are all very unappealing. The box contains a total of 12 tea bags.
After trying 2 bags, you feel that you really do not enjoy the taste and smell of the tea. What
would you do?
19B. You recently bought a box of ginseng tea online, which cost you $60. The tea is supposed
to be very healthful by aiding digestive disorders, and boosting the immune system etc.
However, you find that the taste and the smell are all very unappealing. The box contains a total
of 12 tea bags. After trying 2 bags, you feel that you really do not enjoy the taste and smell of the
tea. What would you do?
I would decide to…
a. immediately stop using the tea bags
b. continue using a few more tea bags and then decide
c. use all the remaining tea bags
20A. As the 10th person who called in on a radio show, you recently won a free ticket worth
$125 to a concert featuring your favorite group. You have been waiting for the opening night of
the show and are very excited to see it. However, two days before the concert, you realize that
124
you have misplaced the ticket and will have to buy a new one. You are afraid that the tickets will
sell out but also worries that you may find the old ticket after buying a new one. What would you
do?
20B. You recently bought a ticket to a concert featuring your favorite group, which cost you
$125. You have been waiting for the opening night of the show and are very excited to see it.
However, two days before the concert, you realize that you have misplaced the ticket and will
have to buy a new one. You are afraid that the tickets will sell out but also worries that you may
find the old ticket after buying a new one. What would you do?
I would decide to…
a. immediately buy a new ticket
b. look for the ticket until the night before the concert and then decide whether to buy a
new ticket if it is not found
c. look for the ticket until the day of the concert, and purchase the ticket right before the
concert if it is not found
125
Appendix B: Decision Scenarios and Sample Choices Used to Study Delay
Discounting in Making Decisions for Oneself in Chapter 4
Small-Distant Condition
In order to encourage more customers to join a new free rewards membership program, a
grocery store in your area now offers one-time cash rewards to customers who join the program
and meet the minimum shopping requirement. As the store sells almost everything you need in
daily life and you go shopping there almost once a week, the shopping requirement is negligible.
You want to join the free rewards membership program. The store offers two options to receive
the one-time cash rewards for loyalty. One option is to receive $100 after two years. The other
option is to receive a smaller cash reward after one year. Each customer can choose only one
option and cannot change it in the future. However, the exact amount for the 1-year (smaller)
cash reward has not been finalized yet. You are now informed of a few possible amounts for it.
Your task is to decide whether to choose the $100 cash rewards option after two years or a
smaller cash reward one year from now. For each of the following possible amounts, please
indicate whether you would choose the cash rewards option two years from now or a smaller
cash rewards option one year from now.
For my cash rewards option from the grocery store, I would choose…
A. the smaller cash rewards option – $50 one year from now
B. the cash rewards option – $100 two years from now
Small-Near Condition
You recently applied for a new credit card from a bank and got approved. It offers a
welcome bonus - $100 one year from now with a minimum requirement on purchases with this
card. As you use credit cards on purchases for almost everything in daily life, the purchase
126
requirement seems very minor. Besides, the bank also offers a second option for people who
want to receive the welcome bonus immediately. However, the amount for this immediate
welcome bonus will be less than $100. The two welcome bonus options are mutually exclusive.
At this moment, the exact amount for the immediate welcome bonus has not been disclosed, but
you are informed of a few possible amounts. Your task is to decide whether to receive the $100
welcome bonus one year from now or a smaller bonus immediately. For each of the following
possible amounts, please indicate whether you would choose the welcome bonus one year from
now or a smaller bonus immediately.
For my welcome bonus from the bank, I would choose…
A. the smaller welcome bonus – $50 immediately
B. the welcome bonus – $100 one year from now
Large-Distant Condition
You were recently involved in a lawsuit and won the case. You will be paid $10,000 at
the end of two years from now. However, if you want to get your payment earlier, you have an
option to receive a smaller amount one year from now. The exact amount for the earlier payment
has not been determined yet, but the attorney for the defendant suggests you consider a few
possible amounts. Your task is to decide whether to receive the $10,000 payment two years from
now or a smaller payment one year from now. For each of the following earlier payment offers,
please indicate whether you would choose the original payment two years from now or a smaller
payment one year from now.
For my winning case, I would choose…
A. the smaller payment – $5,000 one year from now
B. the original payment – $10,000 two years from now
127
Large-Near Condition
You bought a lottery ticket a week ago and are now informed of winning $10,000 but
receiving it one year from now. However, the lottery has an alternative option to accept a smaller
amount for your winnings immediately. The exact amount of the immediate offer has not been
fixed yet, but you are now informed of a few possible offers. Your task is to decide whether to
receive your lottery winnings one year from now or a smaller amount immediately. For each of
the following offers, please indicate whether you would choose the $10,000 lottery winnings one
year from now, or a smaller lottery winning amount immediately.
For my lottery winnings, I would choose…
A. the smaller lottery winnings – $5,000 immediately
B. the lottery winnings– $10,000 one year from now
Abstract (if available)
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
The cost of missing objectives in multiattribute decision modeling
PDF
Sequential decisions on time preference: evidence for non-independence
PDF
Behabioral and neural evidence of state-like variance in intertemporal decisions
PDF
Social exclusion decreases risk-taking
PDF
Insula activity during safe-sex decision-making in sexually risky men suggests negative urgency and fear of rejection drives risky sexual behavior
PDF
A roadmap for changing student roadmaps: designing interventions that use future “me” to change academic outcomes
Asset Metadata
Creator
Chen, Zhiqin
(author)
Core Title
Choice biases in making decisions for oneself vs. others
School
College of Letters, Arts and Sciences
Degree
Doctor of Philosophy
Degree Program
Psychology
Publication Date
10/17/2018
Defense Date
07/16/2018
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
affect,ambiguity aversion,choice bias,delay discounting,loss aversion,OAI-PMH Harvest,prediction,risk attitude,self-other decisions,suggestion,sunk cost
Format
application/pdf
(imt)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
John, Richard (
committee chair
), Bechara, Antoine (
committee member
), Dehghani, Morteza (
committee member
), Monterosso, John (
committee member
), von Winterfeldt, Detlof (
committee member
)
Creator Email
zhiqin.chen.usc@gmail.com,zhiqinch@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c89-82244
Unique identifier
UC11670732
Identifier
etd-ChenZhiqin-6860.pdf (filename),usctheses-c89-82244 (legacy record id)
Legacy Identifier
etd-ChenZhiqin-6860.pdf
Dmrecord
82244
Document Type
Dissertation
Format
application/pdf (imt)
Rights
Chen, Zhiqin
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
affect
ambiguity aversion
choice bias
delay discounting
loss aversion
prediction
risk attitude
self-other decisions
suggestion
sunk cost