Content uploaded by Ofer Zellermayer
Author content
All content in this area was uploaded by Ofer Zellermayer on Aug 05, 2015
Content may be subject to copyright.
The Pain of Paying
A dissertation submitted to the department of Social and Decision Sciences
for the degree
Doctor of Philosophy
in
Behavioral Decision Making and Marketing
by
Ofer Zellermayer
Committee Chair: Professor George Loewenstein
Department of Social and Decision Sciences
Carnegie Mellon University
December 1996
All rights reserved
© copyright 1997
by
Ofer Zellermayer
In memory of my missed cousin
Ori Haalman
Acknowledgments
Throughout my stay in the Social and Decision Sciences Department at Carnegie
Mellon University, and especially while I endeavored to produce this dissertation, my
work has benefited from the generous help of many, whom I wish to thank:
My advisor, George Loewenstein, who saw (and oversaw) me pursuing this work
from conception to ultimate completion, for his accessibility and his relentless insistence
that there is always something to improve (an approach which, in retrospect, I can
appreciate). I consider myself privileged to have been guided by such a creative mind
and dedicated advisor.
My committee members Robyn Dawes, Paul Fischbeck, and Baruch Fischhoff,
whose exceptional knowledge and combined advice sometimes overwhelmed me, yet
always made this project better. To Paul, an inspiring teacher, I am also indebted for
stretching my presentation skills to their peak potential.
The entire S.D.S. community and particularly its staff members Barbara Collins,
Michele Colon, Carole Deaunovich, Marge Melenyzer, Ruth Silverman, and Rosa
Stipanovic, who — partly by presenting bureaucratic regulations as if they were friendly
advice — made my time here an experience for which I already feel a great deal of
nostalgia.
My fellow students, especially my cohort members and most greatly my office
mates over the years who occasionally had to re-visit their language-acquisition stage.
Special thanks to Dean Behrens, Stephanie Byram, Diana Gant, Wagner Kamakura, Gary
Katzenstein, Diane Lowenthal, Srinivasa Prakhya, Melissa Rosenstock, Ron Rymon, Ned
Welch, and Oni Wu, for their contribution to this dissertation at different stages of its
evolution.
Sally Sleeper, the best office mate anyone could have, for being not only the best
of friends, but also an amazing just-in-time dictionary and the most eloquent and helpful
thesaurus; you didn't deserve this abuse — thank you!
My friend and colleague Talia Rymon, for many fruitful discussions and
comments and mostly for priceless mentoring on the do's and don'ts of dissertation
writing.
Martin Horn and Lois Welch of DDB Needham Chicago for providing supportive
evidence to this dissertation's results.
Galit, my wife, who has always granted me her love (which I have always taken
for granted), for her never-ending support (especially when the committee's feedback and
advisor's comments made such support essential). Her steady high spirits evened out the
day-to-day fluctuations inflicted by graduate school.
Ziv, my daughter, who was born about the same time that this dissertation was,
and who allowed me to enjoy her smiles and many rapid developments, at times when
this dissertation seemed to not develop quite as rapidly.
And finally, my parents Arie and Dvora and my siblings Hagit and Gal, for their
unconditional encouragement, and their tolerance of our being away for so long while I
climbed this mountain.
Abstract
The Pain of Paying
Ofer Zellermayer
Professor George Loewenstein (advisor)
Making payments in exchange for goods or services is a widespread practice in
modern society. Yet, despite its apparent routine nature, the paying act is far from being
a dispassionate affair. For instance, some payments cause consumers to agonize while
others, perhaps ironically, evoke joy. Though certainly not a new phenomenon, these
feelings about payments have been a largely ignored topic in economics, which
traditionally regards paying as simply a rational act. While recent studies have begun
acknowledging the role of emotions in consumer behavior, they have not focused on the
paying act specifically. Consequently, important questions about it still remain: What
determines the emotions consumers experience when making payments, and how do
those emotions affect their behavior.
This work introduces the "pain of paying" — the notion that a consumer who
pays for a product or service experiences emotions associated with the act of paying.
Arguing that emphasis on the payment side of transactions will advance our
understanding of real consumer behavior, a theory is developed to explain how the pain-
of-paying emotion depends on different characteristics of a given purchase. In addition,
five studies address the following questions: Given a choice, when would consumers like
to pay for their purchases, before or after consumption is made?; What kind of payment
method would they choose for making those payments?; Which behavior is more
common, compulsive buying or compulsive reluctance to spend?; If the latter, how can
marketers combat that reluctance?
Table of Contents
Page
1. Introduction 01
2. Preliminary Studies 06
2.1 In-depth interviews 06
2.2 Open-ended surveys 07
3. Causes of the pain of paying: Experimental evidence (Study One) 10
4. Related literature 18
4.1 Overview 18
4.2 Decision making, Economics, and Marketing 20
4.3 Contextual influences in corporal pain 27
5. Theory of the pain of paying 32
5.1 Characteristics of acquisitions 33
5.2 Predictions 38
6. Causes of the pain of paying: Experimental evidence (Study Two) 44
7. Implications: Consumers payment-preferences 52
7.1 Motivation 52
7.2 Time-preference study (Study Three) 53
7.3 Mode-preference study (Study Four) 61
8. Implications: Compulsive reluctance to buy (Study Five) 72
9. Conclusion, Limitations, and Contribution 83
9.1 Conclusion 83
9.2 Limitations 84
9.3 Potential Contribution 86
Bibliography 88
Appendices
The Pain of Paying — dissertation
Ofer Zellermayer
Chapter 1: Introduction
Mr. A plays tennis twice a week. Recently he discovered a small crack in
his racquet. Although the crack doesn't prevent him from playing, he feels
that it impairs the level of his game, so last weekend he drove to the mall
to buy a replacement. After spending more than an hour in each of two
stores, and looking at several racquets that he would like to own, Mr. A
was still unable to let go of the money. So he left the mall without buying
a racquet.
If Mr. A's anguish over spending money evokes your empathy, you are not alone.
This scenario was presented to 73 subjects who compared it with a matched scenario,
describing Mr. B who purchased a racquet without really needing one. The subjects' task
was to mark on a five-point scale whose behavior resembled theirs more closely, Mr. A's
or Mr. B's. Forty seven respondents identified more closely with Mr. A; only 11
identified with Mr. B; the rest of the subjects found both scenarios equally appealing. In
another survey, other subjects reviewed 50 expenses and were asked directly how
pleasurable or painful paying each bill felt. The responses indicate that the subjects
found paying distressing, even for goods which they obviously require or appreciate,
such as, "heating gas," "a work/school book," and "a credit card's annual fee" (9 on an
11-point scale). Clearly, these strong responses to the prospects of making certain
payments were emotional in nature rather than intellectual.
Despite its intuitive appeal, this distress, referred to here as the pain of paying, is
not accounted for by most formal economic models. The modern theory of consumer
choice over time, as first developed by Irving Fisher (1930), views consumers as
emotionlessly trading off consumption at different points in time, including the present.
Spending is allocated so as to maximize the sum of utilities, with some allowance
typically made for time discounting. Deciding whether to spend or withhold spending, in
2
this model, is simply a matter of allocating resources efficiently, and is seen as
qualitatively no different from, say, a choice between apples and oranges. The cost of
increasing spending in the present is simply the discounted foregone utility from the
next-best use of the money, which will almost always be some act or acts of consumption
that would have otherwise occurred in the future.
While this paradigm might present an ideal to be striven for, even Fisher himself
recognized that it is not psychologically realistic. It is not an adequate description of
consumers' utilities in part because consumers derive direct and immediate displeasure
or pain from the act of making a payment, as alluded to above. This pain is
distinguishable from the future decline in utility that will ultimately result from the
diminishing of the consumer's wealth. Rather, this pain is the psychological, or hedonic,
vexation connected with spending money. To illustrate, if the reason that Mr. A did not
buy a racquet were that he feared the resulting decline in his future buying power, he
would not have driven to the mall in the first place. Since it was the pain that arose at the
prospect of spending the money, Mr. A experienced it only when he was about to pay.
Although costly from a hedonic perspective, the pain of paying plays an adaptive
role. By providing the consumer with an instant emotional signal about the payment's
potential negative ramifications, it impedes excessive immediate indulgence. That pain
would be of little use to an "economic man" or to Star-Trek's Mr. Spock, but it is crucial
to most consumers, for it solves three interrelated problems that arise from the
motivational propensities and cognitive limitations of real consumers.
Unlike paying by bartering goods, paying with money entails temporal and
sporadic separation between a payment and its associated forgone utility. This
separation, combined with myopic behavior, create the first, motivational, problem that
the pain of paying solves. As encapsulated in the notion of "hyperbolic time
3
discounting," people place disproportionate weight on the immediate present relative to
all points in the future (Ainslie, 1975, 1992; Akerlof, 1991; Benzion et al., 1989; Elster,
1985; Laibson, 1995). The result is a type of "immediacy effect," (Prelec &
Loewenstein, 1991) which manifests itself in many ways, e.g., the difficulty of jumping
into a pool of cold water on a hot summer day increases just before one is about to jump;
the anxiety of giving oneself an injection grows just before injecting; and, the temptation
of a bakery intensifies when looking at the luscious desserts. Given the great weight
consumers place on the present, they need a mechanism to help them resist the strong
attraction of immediate gratification. The pain of paying does just that by moving the
cost of immediate indulgence into the present, so consumers experience it at the same
time that they experience the indulgence itself.
A second and related motivational problem is that consumers tend to underweight
opportunity costs; experienced losses are given greater weight than forgone gains. The
feeling of giving up acquiring a tangible wanted good is disturbing here and now. On the
other hand, it is difficult to value goods that are "mere possibilities" as a loss. To a
consumer facing a good to be purchased, this good is much more vivid than alternative
goods that could potentially be purchased in the future. As Nisbett and Ross (1980) have
shown, decision makers are affected more strongly by vivid information than by pallid or
abstract information. While that work examined subjects' likelihood judgments,
subsequent work demonstrates that greater vividness produces stronger emotional
responses (Miller et al., 1987). Thus, consumers tend to find immediately considered
goods, whose consumption they can easily imagine, more emotionally interesting and
exciting than the hypothetical goods that will have to be given up in the future as a result
of the indulgence. The pain of paying counteracts this under-weighting of opportunity
4
costs by converting them into an immediate tangible pain, in contrast to forgone
consumption which is difficult to perceive.
The third, cognitive, problem arises because consumers cannot grasp and
calculate the tradeoffs among all the goods that the marketplace offers, nor think through
the enormous complexity of intertemporal consumption decisions. There are too many
products, shopping outlets, promotional gimmicks, payment times, and payment options
for consumers to optimize their utility over time given limited cognitive capabilities
(Simon, 1955). Research has shown that most consumers lack basic information about
capital markets, or the effect of inflation on interest compounding (Gruen, 1989), and that
most people are unable to solve even the simplest stylized intertemporal maximization
problem (Fehr & Zych, 1994). Given that a systematic calculation is impractical,
consumers need a tool to help them decide on the spot whether they can afford to make
different purchases. The pain of paying is such a tool. It helps consumers allocate
resources more rationally by summarizing the costs associated with the payment in a
manner that they can easily digest. In fact, consumers do not have to make an effort to
understand this summary, because it is an emotional condensation rather than a cognitive
tabulation. The pain happens automatically when the consumer is about to pay.
Therefore, it eliminates the requirement for formal calculation, yet yields a heuristic
cost/benefit analysis where an explicit one is impossible: "if it hurts too much to pay for
it, then don't buy it."
In sum, the pain of paying solves three interrelated problems consumers face by
forcing them to consider the costs and benefits associated with a payment simultaneously.
The sudden rush of emotion frees people from ceaselessly having to heed the tradeoffs
between a given purchase and all its possible alternatives. Instead of mentally
performing the impossible formal cost/benefit analysis, the pain of paying provides an
5
emotional synopsis of the costs concomitant with the purchase. When the negative
outcomes connected with a given payment come into mind, however fleetingly, the
consumer experiences an unpleasant gut feeling (Damasio, 1994). This is a signal to the
consumer: "letting go of the money is bad." By providing this signal, the pain of paying
guards people from capriciously spending their money.
Yet, unfortunately for consumers, the pain of paying is not a wholly beneficial
mechanism. Resolution of the motivational and cognitive problems described above
imposes a penalty of dual payment; that is, every time they purchase a good, consumers,
in effect, pay twice. The payment that the economic model accounts for is that of
forgone utility, which consumers experience in the future when they cannot afford a
desired alternative good. The additional, and yet understudied, payment is hedonic. It is
the distress that consumers feel when they actually make the payment and experience the
pain of paying.
There is a second, potentially negative, consequence of the pain of paying.
Although the pain of paying serves as an emotional synopsis of the costs associated with
the purchase, it is at best a crude representation of these costs. The pain of paying tends
to be highly unstable because contextual factors influence its intensity. As a result, at
times it protects consumers too stringently, and at other times does not protect them
sufficiently from splurging. If the pain of paying is too intense, consumers may not
spend when spending would benefit them. Similarly, when contextual influences
mitigate the pain of paying, consumers may find themselves spending against their best
interest.
The goal of this dissertation is to explore both the determinants of the pain of
paying and its consequences. In the following nine chapters, the research will entertain
six questions associated with the pain of paying: 1) What purchase characteristics
6
influence the pain of paying? 2) What contextual attributes influence the intensity of the
pain of paying? 3) How does the pain of paying affect when consumers prefer to pay for
goods or services? 4) How does the pain of paying affect how consumers prefer to pay
for goods or services? 5) Does the pain of paying affect consumers aggregate level of
purchasing? 6) What theoretical and applied contribution can be made by understanding
the pain of paying? To answer these questions, several preliminary studies have been
conducted which are reported next. Then an explanatory model of the pain of paying is
developed. Subsequent sections describe studies that attempt to establish the contextual
factors that affect the pain of paying and to discover how the pain of paying affects
consumer behavior.
7
Chapter 2: Preliminary studies
The pain of paying is a new topic with no particular stream of research directly
addressing it, nor information which identifies the factors that affect it. However, there
are several disciplines studying phenomena that are potentially tangential to the pain of
paying, and that may be useful to review in creating a theoretical framework of the pain
of paying. Accordingly, two qualitative studies were conducted to focus the investigation
on the relevant bodies of literature, and to gain insight about attributes that influence the
pain of paying.
2.1 In-Depth Interviews
In the first study, eight consumers in face-to-face interviews were asked to think
of specific purchases and expenditures whose payments felt especially emotional to them.
The subject were instructed to talk aloud about payments that made them feel either good
or irritated. They were given sufficient time to contemplate and recall payments about
which they had intense emotions. One insight that emerged early in the study is that the
interviews revealed striking parallels between the situational characteristics that influence
the pain of paying and those that influence physical pain. For example, revealing the
importance of control, some subjects did not like paying their winter heating-gas bill
because they felt that, while it was usually high, there was not much they could do to
substantially lower it. This is notably akin to the grim pain reported by subjects who
have no control over the painful stimuli in corporal-pain experiments. Others did not like
paying their long-distance phone bills because, while that bill tended to be high too, at
the time of paying these subjects did not really know where the money had gone, and had
8
difficulty justifying the high expense. Again, as with physical pains, while "no pain-no
gain" is an often acceptable rule of thumb, "pain without gain" is utterly agonizing.
Several subjects did not like paying for house-maintenance or renovations because they
felt that "no matter how much you pay, there will always be more expenses, so it's a
never-ending ordeal." This too can be paralleled with physical pain, where an enduring
pain usually feels more distressing than a temporary pain. Finally, as an example of a
pleasant payment, most subjects felt good about paying for a present for a good friend or
for paying a craftsman for a job well-done (although, interestingly, only if it was a job
they could not perform themselves).
The interviews elucidated that subjects were quite involved with their payments
and could easily introspect and report about them. Subjects' purchases did not end with
receiving a product or a service, rather, the subjects demonstrated they could have strong
feelings about payments as well. Noticeably, subjects tended to elaborate about many
different aspects of the payments and the transactions, with the surprising exception of
the expense's amount. Subjects' feelings appeared to depend on various aspects of the
transaction, most notably, whether the they had control over the type and amount of the
expense and whether the expense was a money drain. Another interesting finding was
that subjects termed some expenses as affordable and others as unaffordable, but the
reasons they provided for why a given expense was affordable or unaffordable were
usually not its cost. To learn more about these reasons a second study was conducted.
2.2 Open-Ended Surveys
In the second study, 28 subjects — individuals awaiting their jury duty at a
Pittsburgh courthouse — completed a written questionnaire in which they generated lists
of purchases and expenses that they considered affordable or unaffordable/extravagant.
9
Next to each product or expense, subjects were asked to list the reason(s) why they found
the expense affordable or prohibitive. Subjects usually took 20 to 30 minutes to
complete the survey, and on average listed about 10 items. Once more, their responses
revealed patterns that bore a remarkable resemblance to those that physical pain would
produce. For example, in the same way that physical pain tends to worsen when
additional pain awaits the patient, one subject described a car as unaffordable because of
the "maintenance price, plus all repairs." Other subjects drew an opposite conclusion
from a similar reasoning. They listed cars as affordable because "a new vehicle saves
gasoline, [and has] low maintenance cost" and "because you should feel safe when
driving so it is worth spending money on a newer vehicle." Another example which
demonstrates an expectation-of-benefits effect was given by a subject who advocated
"television and video equipment [because] it is a small investment for the amount of use
you receive from them." (As if the pain of paying is mitigated by the anticipation of a
long period of product enjoyment.) Demonstrating a reference-price effect, some
subjects rejected items for which less expensive substitutes readily exist, e.g., expensive
makeup because "cheap stuff is just as good." One subject found a vacation unaffordable
because it is "not a necessity of day-to-day living," while another subject found a
(summer) vacation affordable because of the possibility of "saving throughout the year."
The same logic was used by yet another subject to support the purchase of a VCR, an air-
conditioner, and a CD player since "you can 'save-up' for these items."
While subjects have differed on the particular items that they supported or
rejected, their reasoning was similar. When compiling the reasons subjects provided for
defining an expense as affordable or unaffordable, 12 common reasons emerged. These
reasons are:
1. Investment vs. Consumption: is the purchase expected to appreciate or depreciate?
10
2. Enjoyment length: Is there going to be a continuos stream of benefits form the
expanse?
3. Duration: How long-lasting is the product or service?
4. Control: how much control does one have over the type and amount of the expense?
5. Routine-Non Routine: how normal is this expense among one's personal payments?
6. Social Acceptance: how easily could you justify the expense to others?
7. Visibility: how apparent is the "reward" for making the payment?
8. Fairness: Is it fair or unfair that one faces this bill?
9. Self vs. other: Is the bill for an expense made for oneself or another person?
10. Temporality: Typically, is the payment done before or after consumption occurs?
11. Ramifications: How bad will it be if one fails to make the considered bill?
12. Immediate vs. Drawn-out: will the payments "take care of the problem," or require
additional payments to enjoy the ones already made?
The 12 reasons listed above can be thought of as 12 preliminary attributes which
may influence the pain of paying. To grasp the pain of paying, the most important of
these attributes need to be identified, and their theoretical underpinnings should be
exposed. The information gathered up to this point is useful for identifying the important
literature to review, as well as for paving the course for continuing the empirical
investigation. In terms of the relevant literature, subjects reports indicate that; 1) It is
important whether the expense has been incurred fairly, by some objective measure; 2)
The relationship between payment and consumption, the order in which they occur, and
the amount of the payment that is cushioned by a yet-to-be-experienced consumption are
all important; 3) It is important whether the expense is made under voluntary or coercive
conditions, i.e., does the consumer have control over it; 4) Regarding its dependence on
contextual factors, the pain of paying has a lot in common with physical pain. Chapter
Four reviews the literature relevant to all these assertions.
Before embarking on a literature review, however, the 12 attributes above need to
be tested formally to establish their importance in determining the pain of paying.
Additionally, the most important attributes should be selected and explored further. The
next chapter reports a structured study aimed at focusing the empirical examination. The
11
study begins with the 12 attributes stated above, and ends with a subset consisting of the
attributes that appear to be the most important determinants of the pain of paying.
12
Chapter 3: Causes of Pain of Paying — Experimental evidence
Study One
The empirical investigation to this point was designed to gain preliminary insight
about the variables that affect the pain of paying. It consisted of open-ended interviews
with eight consumers who described attributes of purchases whose payments felt
especially emotional to them. An additional 28 subjects completed a written
questionnaire in which they catalogued expenses that they considered affordable or
unaffordable, and listed the reason(s) why they found each expense affordable or
prohibitive. These initial responses served as inputs for the structured questionnaire
employed by the study reported here.
The study reported in this chapter had three objectives. The first objective was to
experimentally investigate the importance of contextual attributes, particularly those
collected in the open-ended studies described earlier. A second objective was to discover
additional attributes that influence the pain of paying. A final objective was to examine
the effect of the size of an expense on the pain of paying. That is, while it is obvious to
expect that pain of paying will increase with the dollar amount a consumer has to pay, the
relative effect of the dollar amount compared with the effect of contextual attributes is
not clear.
Method
Subjects: Subjects were 34 SDS staff members and SDS and EPP graduate
students, 14 males and 20 females, between the ages of 22 and 58.
13
Procedure: 1) Assessment of pain of paying — The subjects received a 50-item
questionnaire describing different bills that an individual may face over the course of a
month. (The questionnaires were completed during subjects' free time.) Each item was
followed by an 11-point scale, on which subjects were asked to indicate how painful or
pleasurable it felt to pay for a given item. In addition, subjects were requested to specify
the typical dollar amount that a given expense would cost.1 For example, subjects saw
the following item:
A bill for your share of a group-present to a coworker you don't really like.
How painful is it? -5 -4 -3 -2 -1 0 1 2 3 4 5
Painful
Pleasurable
Typical dollar amount $_____
2) Elicitation of attributes of pain of paying — On the last page of the questionnaire, in
addition to demographic information, subjects were asked to list factors (i.e., attributes)
that they found important in determining the pain/pleasure of paying (see Appendix 1).
Over all subjects, a total of 20 unique attributes were listed as important in determining
the pain/pleasure of paying; 12 attributes that were mentioned by more than one subject
closely matched those revealed in the preliminary studies.
1. The order of the pain-of-paying and dollar-amount questions could potentially influence the strength of
the dollar-amount effect on pain of paying. Since both questions are on the same page and in close
proximity, most subjects probably saw them at the same time, and thus an order effect is not expected to be
strong. However, because the main interest of the study was in the effect of contextual attributes, the dollar
amount question was asked second.
14
3) Rating the bills with respect to the elicited attributes — Four "raters" evaluated each
of the 50 questions on each of the 12 attributes (i.e., the 12 attributes which were
mentioned by more than one subject). The average level of each attribute for each
expense across the four raters was computed. To check for inter-rater consistency, a
Cronbach's Alpha reliability test was conducted (across all 50 items) and alpha level
varied between 0.58 and 0.82.
15
Results
Pain-of-paying rating spanned the entire -5 to +5 range, and across all subjects
and all items the average pain rating was -0.43. There were 25 painful expenses, 23
pleasurable expenses, and two neutral expenses (i.e., the average rating for these items
across all subjects was 0; see Appendix 2). The most painful item was "a bank fee bill
for a returned check" which across all subjects averaged -4.43. The most pleasurable
item, "a bill for a gift to a person you liked a lot," averaged +3.18. Looking across bills,
for 28 subjects the average rating for all bills was negative, indicating pain, and for six
subjects it was positive, indicating pleasure of paying. The most "suffering subject" had
an average pain rating of -1.46, and the most "pleased subject" had an average pleasure
rating of 0.96.
The average expenditure amount across all subjects and items was $115. The
most demanding item was "a bill for a wall-to-wall carpet," for which the average subject
paid $1,028. The cheapest item was "a bill for your share of a group-present to a
coworker you don't really like," for which the average subject paid $7. The subject who
was most careful with money spent, on average, for all the 50 items, $65; the most
generous subject spent $214.
Subjects' gender had no effect on either pain rating or dollar expenditure (Pain t32
= 0.12, p = 0.909, ns; Dollar t32 = 0.44, p = 0.66, ns). The correlation between subjects'
age and average pain across items was not significant (r = 0.156, Z = 0.86, p = 0.39, ns).
There was a small but significant positive correlation between subjects' age and the
average
spending across items (r = 0.398, Z = 2.31, p = 0.021), indicating that older subjects
tended to spend more than younger ones.
16
17
To measure the relationship between the pain of paying and the different
attributes, a stepwise regression, with the "pain of paying" (averaged across the 34
subjects) as the dependent variable, was estimated.2 Fourteen variables, including the
four raters' average of each of the 12 attributes, as well as the average and variance of the
typical dollar amount, were used as potential independent variables. Overall, using the
inclusion-exclusion criteria (p < 0.05, p > 0.1), the final model included seven variables:
six attributes plus the dollar amount. Together they explained 83% of the variance in
predicting the average pain of paying (adjusted R square = 0.832; F7, 42 = 35.7, p <
0.001). To address the issue of a potential ecological fallacy, whereby a prediction which
is true at a group (or average) level of analysis does not hold for individual level analysis,
a second conservative analysis was performed for each subject separately. Thirty four
separate regressions were run, one for each subject, using the individual subject's pain-of-
paying ratings for a given expense as the dependent variable. Each subject's indication of
typical dollar amount for a given expense, as well as the raters' average for the six
attributes from the previous analysis, were used as independent variables. This resulted
in 34 R square coefficients. The average adjusted R square across the 34 subjects was
0.497 (SE = 0.018). Thus, while this conservative analysis resulted in a smaller
correlation coefficient, that coefficient was again robust and significant.
The role of different attributes: Table 1 provides the statistical characteristics of
seven variables the final model included — six attributes plus typical dollar amount. A
theoretical discussion of the attributes is provided in Chapter Five. In brief, the theory
2. A factor-analysis approach revealed that the discovered attributes could not be reduced to a few
representative factors, and thus a regression analysis approach was adopted. To test for multicollinearity, a
correlation matrix of the 12 attributes was computed; none of the bivariate correlation coefficients
approached 0.8, and hence multicollinearity was not considered a reason for concern (Kennedy, 1992) (in
fact, for variables that were included in the final model, all of the bivariate correlation coefficients were
below 0.6).
18
suggests that a consumer who is making a payment contemplates the question: "What
does this payment 'buy' me?" The more a payment buys, the better the explanation for
the expense, and the less the pain of paying. Contrary to the implication of traditional
economic theory, there is a big emotional difference between money that is lost and
money that is spent to achieve something; consumers experience strong pain of paying
when their need to feel that they are buying something is not satisfied. Satisfying the
consumers' need to feel the they are really buying something depends on the attributes
listed in the table.
Table 1
Attribute b coefficient t statistic p value < Theory Root
Fairness (high to low) -0.625 -8.55 0.001 Transaction
&
Coe
r
c
i
o
n
Investment vs. Consumption -0.349 -4.75 0.001 Buffering
Immediate vs. Drawn-out -0.237 -3.12 0.005 Buffering
Self vs. Other 0.194 2.98 0.005 Transaction
Temporality
p
a
y
ment before or after ex
p
ense -0.174 -2.95 0.01 Buffering
Control (high to low) -0.159 -1.95 0.05 Coercion
Dollar Amount -0.151 -2.53 0.05
Recall that high negative responses indicate high pain. Thus, the results in the
table reveal the following information: Fairness, Investment-Consumption, Immediate—
Drawn-out, Self-other, Temporality, and Control, are all attributes that significantly
influence the pain of paying in following direction:
19
The more a transaction is perceived as
A1: fair
A2: an investment (as opposed to pure consumption)
A3: an immediate payment (as opposed to drawn out payments)
A4: an expense for the sake of another person (as opposed to oneself)
A5: under one's control
A6: a payment-before-consumption sequence (rather than visa versa)
the weaker the pain of paying.
Additional considered attributes that subjects indicated as potentially influential did not
reach statistical significance.
The role of dollar amount: One could argue that what really makes paying painful
is how much one pays. Therefore, the average of "typical dollar amount" was included in
the regression model. Although statistically significant, this coefficient ranked last
among the seven final independent variables. The variance of the expense's amount
(approximated by the variance for a given item across subjects) did not reach statistical
significance.
To further examine the claim that "it's only the size (of the dollar amount) that
counts" two correlational analyses were run. The first analysis examined the relationship
between the dollar amount of an expense and the pain of paying across subjects for a
given bill. Do people who pay more tend to feel more pain? A positive correlation
means that costlier bills are less painful. The Pearson correlation coefficients were
positive for 24 items and negative for 26. The average was -0.043 (S.E. = 0.034). Only
in 19 of the 50 items the Pearson correlation coefficient exceeded 0.3; of these, for eight
bills it was positive and for 11 negative.3
3. Possibly, a dollar-amount effect in this analysis may be masked by a reverse causality effect, namely,
subjects who feel less pain tend to pay more. If this happened, pain would not be expected to vary with
20
The second analysis examined similar relationships for a given subject across
items. That is, for a specific individual, do more expensive expenses tend to produce
more pain? The Pearson correlation coefficients were positive for 16 subjects and
negative for 18. The average was -0.018 (S.E. = 0.025). Of the 34 subjects, two had
correlations that exceeded 0.3, with one being positive and the other negative. (Also,
only 14 of the subjects had correlations that exceeded 0.1; of those, six were positive and
eight were negative.)
Clearly, both analyses corroborate the regression result that the dollar amount
played, at most, a minor role in determining the pain subjects felt when paying for the
listed items. Other things being equal, price is important. In reality, however, things are
not equal: Payments that consumers make vary considerably on a variety of attributes
that affect the pain of paying more than does price.
Discussion
The study's results provide empirical support for the proposal made about the pain
of paying in the introductory chapter. First, they support the general assertion that
context has an important influence on the pain of paying. The variables included in the
regression model explained 83 percent of the variance in pain of paying. Moreover, to
test the net power of the contextual variables in explaining the pain of paying, when the
dollar amount. Usually, such reverse causality would be expected where there is a large variance in
subjects wealth because richer individuals plausibly feel less pain than poorer ones when spending, and
thus spend more. Because in this study all the subjects were university staff and graduate students, a
sizable wealth variance is unlikely. Additionally, even if richer individuals are less sensitive to the
expenses' amount relative to poorer individuals, across different purchases they too should reveal such
sensitivity. That is, even rich consumers are expected to feel more pain when they spend a lot of money
than when they spend little money. The next analysis looked for that sensitivity.
21
dollar-amount variable was excluded from the model the remaining variables still
explained more than 81 percent of the pain-of-paying variance. Clearly, context matters.
Second, the results nominate six specific attributes as important variables. The
coefficients of the six attributes in the final model were significant, and explained most of
the pain-of-paying variance. Furthermore, additional attributes that some subjects
suggested as potentially important determinants of the pain of paying did not reach
statistical significance. This means that not only is it possible to do a fair job of
predicting which factors determine the pain of paying, but also that this can be done with
a rather small number of factors. This is not to say that additional factors that subjects
somehow ignored cannot exist or are necessarily unimportant. However, of the factors
which came to subjects' minds in this study, six sufficed to explained the bulk of the
pain-of-paying variance.
A possible limitation of this study is its being correlational. Although the
attributes were based on several levels of pretesting, and were elicited directly from
subjects, one cannot ignore their possible weaknesses due to the post hoc nature in which
they were treated in the study. The study reported in Chapter Six addresses this issue by
focusing on one of the attributes, Temporality, and manipulating it directly.
The next two chapters are devoted to developing a greater understanding of the
pain of paying and its dependence on Fairness, Investment-Consumption, Immediate—
Drawn-out, Self-other, Temporality, and Control. Chapter Four reviews the relevant
literature, while Chapter Five composes a theory of the pain of paying and generates
predictions based on the six attributes established here.
22
Chapter 4: Related Literature
4.1 Overview
As the introductory chapter argues, and the experimental studies confirm, the pain
of paying is more than a combination of the pleasure derived from consuming a good and
the displeasure of not being able to consume a different good due to budget constraints.
It is a visceral sensation that depends on the condition under which the purchase is made.
Evidence to that effect emerges from research across several disciplines. The first stream
of pertinent studies arises from research on decision-making, economics, and marketing,
particularly from new developments in these fields. Recall that one of the motivating
forces for the current work was the lack of respect by traditional economic approaches
for emotional influences on consumer behavior. The same cannot be said about
Behavioral Economics — a relatively recent subfield of Economics — which advocates
testing the realism of assumptions, and emphasizes introspection as a source of
behavioral insight over impetuous synthetic formulas. An especially relevant concept to
the pain-of-paying notion that this subfield has produced is Richard Thaler's formulation
of Transaction Utility Theory (Thaler, 1980, 1985). His theory suggests that buyers are
motivated by more than just the acquisition utility associated with obtaining and using a
product. They are also motivated by the transaction utility associated with the difference
between the actual price paid and the buyer's reference price for the product. How
reference prices evolve and their implications for the pain of paying are discussed in the
next section.
A second line of inquiry highly relevant to the pain of paying hails from a recent
development in decision making research. This research, which builds up on mental
23
accounting, yet another idea of Thaler (1980, 1985), has been termed buffering (Prelec &
Loewenstein, 1996). Buffering research studies the interactions between payment and
consumption, and focuses on their interactions with respect to a time line. It submits that
consumption can cushion payments that are associated with it, and that paying can taint
the delight of consumption. In particular this research examines when people prefer to
consume a certain product prior to their paying for it, when they prefer an opposite order,
and when they prefer some mixture of consumption and payments streams.
Demonstrating that consumers' preferences are often opposed to normative prescription
of financial rationale, this work helps to explain some of the influences on the pain of
paying.
A third contributing area consists of a few works in marketing, primarily in
consumer behavior. Systematic research on the effects of emotions (induced or
otherwise) in consumer behavior was almost nonexistent until quite recently (Cohen &
Areni, 1991), let alone research on such effects vis-a-vis payments. Nevertheless, the
few emotionally-focused studies that do exist add to our understanding of the pain of
paying because they reveal the influence of contextual attributes on consumer behavior.
Description of these studies concludes section 4.2.
The final area to be reviewed is corporal pain. Beyond the specific factors
propounded by the studies reported earlier, the list of factors reveals something more
important — the context dependence of pain of paying. This suggests a potential analogy
between pain of paying and ordinary corporal pain, since it is well established that
corporal pain is context dependent. Although the parallelism between corporal pain and
the pain of paying is not perfect, the resemblance is conspicuous and telling. To
accentuate the pain of paying's susceptibility to contextual influences, section 4.3
24
describes the evolution of corporal-pain research and our growing understanding of its
physical, as well as, nonphysical components.
4.2 Decision-making, Economics, and Marketing
Transaction utility theory
As argued earlier, economists and market researchers have generally focused on
the trade-offs that buyers make between the utility they get and the price they pay for a
product. Transaction utility theory includes variables that are independent of the inherent
properties of the service or product for which the consumer pays. Obviously, other
things being equal, the more one pays for equivalent goods, the more annoying the
payment. More interestingly, as captured by the notion of transaction utility (Thaler,
1980, 1985), the pain of paying is also influenced by the price the consumer has to pay in
comparison to some fair or subjective "reference" price (Biswas & Burton, 1994;
Kahneman, 1992; Nagle & Holden, 1995). Elements that Thaler identified as influencing
the consumer's reference price are the cost to the seller and, to a lesser extent, the market
price.
To demonstrate the effect of the cost to the seller on consumers' willingness to
pay, Thaler (1985) asked subjects to imagine that they were lying on the beach, thirsty
for a beer, and that a willing friend would bring the beer from a close vendor, if the price
was acceptable. Half the subjects were asked for the maximum price after being told that
the vendor was a "fancy resort hotel"; the other half were told it was a small, "run-down
grocery store." The median subject was willing to pay $2.65 for the hotel's beer, but only
25
$1.50 for the beer from the grocery store. One apparent possible explanation is that
subjects "allowed" the hotel to incorporate some of its high overheads into the price of
the beer. Subjects were not willing to pay as much to a place with clearly low overheads.
They preferred not to get the beer than to let a low-key operation overcharge them: The
pain of paying a "ripping-off" vendor outweighed the consumption utility of satiating the
thirst.
Another study conducted by Thaler, in collaboration with Daniel Kahneman and
Jack Knetsch (1986), establishes the effect of cost to the seller as influencing the
perception of fairness. These authors described to subjects two scenarios in which, upon
a shortage in a popular car model, a car dealer is charging $200 above what he had been
charging before the shortage had developed. In one scenario, the dealer, who used to sell
the cars at a $200 discount, eliminated that discount and now sells them at list price; the
dealer in the other scenario used to sell the cars at list price, and now is charging $200
above it. Forty two percent of the subjects perceived the first dealer as unfair; seventy
one percent perceived so the second dealer (Kahneman, Knetsch & Thaler, 1986). Both
dealers used the shortage to raise prices by the same amount. In percentage terms, there
was a larger increase on the part of the dealer who eliminated the discount and now sells
at list price. Yet, more subjects attributed unfairness to the dealer who used the shortage
to boost profits above what is perceived as an acceptable level.
Consumers also feel unfairness when charged a high price for a necessity versus a
luxury. Nagle and Holden (1995) submit that people object to high prices for medical
care — since they feel they should not have to pay to be healthy — but would buy
jewelry, a car, or a vacation without objection to high prices or price increases. Their
explanation of this objection is that the high price for luxuries simply reduces the amount
26
of gain (size of diamond, quality of car, length of vacation) that one can obtain, rather
than reducing the quality of life that the buyer has come to perceive as "necessary."
Market price, the other determinant of the consumer's reference price, is formed
according to many inputs in the market environment, such as last price paid,
going/current price, favorite brand price, average prices of similar goods, and expected
future price (Morris & Morris, 1995). An illustration of the effect of current prices to
which a buyer is exposed is provided by Simonson and Tversky (1992), who asked
subjects to choose among different models of microwave ovens. Half the subjects were
asked to choose between two models: a low-end Emerson, and a mid-range Panasonic;
the other half were asked to choose from among three models, the same Emerson and
Panasonic plus a high-end Panasonic. In a clear manifestation of the so called "attraction
effect," the top-end model drew only 13 percent of the subjects, but had a large impact on
the now mid-priced Panasonic, which gained additional 17 percent in market share. The
point seems clear. Context matters: subjects considered more than the benefits of the
product and the cost of buying it. The mid-range Panasonic became more valuable when
it appeared inexpensive in a relative sense.
Another demonstration of the effect of prices paid for similar products or in
similar situations is observed in the pharmaceuticals industry (Nagle & Holden, 1995).
Patients, and even doctors, object to pills costing $1 or more per dose, even when the
reduction in patient suffering and other medical expenses clearly justify the price in value
terms. Since most oral medicines cost only a few cents each, it is painful to pay for the
few medications that cost much more.
In essence, Transaction utility refers to the fact that consumers do not like to be
"suckers" and therefore experience displeasure when having to pay an unjust price. Thus,
paying an expensive price for a delivered pizza is more painful when buying it from a
27
pizza franchise than when buying it from a gourmet restaurant (whose operating costs are
perceived to be higher).
Buffering
Buffering refers to the extent to which consumption of a purchased item cushions
its payments (Prelec & Loewenstein, 1996). That is, a consumer's consumption
enjoyment and pain of paying are interdependent. Typically, making a payment taints the
delight of consumption, while at the same time being able to consume an item mitigates
the pain of paying. However, the amount of the payment "charged" to the item's
consumption and the amount of the pain alleviated by this consumption varies both
across individuals and purchase occasions. When the consumer can clearly impute a
specific consumption to a corresponding payment, that consumption buffers the payment,
i.e., the consumer senses that the payment is really buying something. On the other hand,
if the consumer cannot associate the payment with any particular consumption, then the
payment feels just like throwing away money rather than buying.
Buffering is influenced by the degree of coupling (i.e., the strength of the dyadic
relationship) between payment and consumption. Closely coupled payment and
consumption usually results in a weak pain of paying, while loosely coupled payment and
consumption results in a strong pain of paying (Arkes & Blumer, 1985; Soman &
Gourville, 1995 - draft). To illustrate, it may be less painful to pay for a specific long-
distance calling-card call to a close friend than to pay the same amount for many long-
distance calls where it is difficult to discern how much of the payment to "charge" to
each call. Similarly, it may be less painful to pay for the close-friend call soon after
conducting it than waiting until long after the call has been forgotten and the payment
feels exposed (i.e., not buffered by consumption).
28
Evidence that consumers engage in hedonic editing to achieve coupling is
provided by a series of experiments by Hirst, Joyce, and Schadewald (1994). These
researchers presented subjects with scenarios in which they engaged in a purchase of two
assets; a short-lived asset, such as a vacation, and a longed-lived one, such as a piece of
furniture. Subjects had to finance one item and pay cash for the other, but could choose
which item to finance and which to pay for with cash. The majority of the subjects (82%
in experiment One) preferred to finance the long-lived asset. In another experiment, the
authors asked subjects to divide a work bonus between two loans for two, uninsured,
financed goods (a sailboat and a snowmobile), where one of the goods had been stolen.
Subjects apportioned the bulk of the bonus towards the stolen good (66% and 72% when
the sailboat and the snowmobile were stolen, respectively). A retrospection of one
participant in the experiment is particularly telling: "Because he is no longer receiving
benefit from it, it will be painful to continue pay on it."
Buffering is also based on the idea that people commonly engage in prospective
mental accounting, namely, that mental accounting functions track only benefits and
payments that will occur in the future (Prelec & Loewenstein, 1996). This stresses the
importance of temporal considerations on payment-making. Paying for a used-up good is
much more painful than paying for a good whose consumption lies in the future.
Similarly, making a payment with the expectation that it will be followed by additional
costs (e.g., rent) is more painful than paying under the expectation of a single or final
payment. Least painful should be a payment made with the expectation that it will
generate additional utility (e.g., writing a check for a Christmas savings plan). To
demonstrate, when considering consumption utility, the enjoyment of a prepaid Club-
Med vacation will not be tainted by the already-completed payment. Likewise,
29
considering payment utility, the pain of paying a credit card bill for a vacation taken last
month will not be mitigated by the prospect of future pleasure.
A survey conducted by Prelec and Loewenstein (1996) supports the prospective
accounting intuition. Given a choice between pre-financing a vacation or still owing on
it after returning, 63% of the subjects opted for the prepayment while forfeiting potential
interest gains. Given an option to be paid before or after doing some work, 60% to 66%
of the same subjects (depending on the particular description of the work involved)
preferred to receive payment upon completing the work. Because mental accounts only
track future costs and benefits, anticipating the future, people prefer a situation in which
they can freely enjoy a paid-for vacation to facing a payment stream upon returning from
this vacation. They similarly prefer to face a stream of payments for a completed work
they had performed to owing someone the work for which no future payments will be
received.
Another factor affecting Buffering, and consequently the pain of paying, is the
source of the money. Mental accounting procedures tend to result in earmarked sums
which are tagged at their origins as dedicated to particular expenses. Money accumulated
toward a child's education may leave one's pocket or bank account less painfully toward
that end than, say, toward buying a car or a house. Especially little pain is felt when
paying with a "windfall gain" (Thaler, 1985). Such money has escaped the tagging or
earmarking process, and thus is free from having to serve for any particular expense.
Spending it on luxuries or self-indulgence produces the pure joy of the anticipated
consumption untainted by the usual pain of paying. This lack of pain of paying helps to
explain the "house money effect," whereby many gamblers lose in casinos. They set
aside a certain amount of money with which they play, and quite disciplined they stop
after losing that amount. They have, however, no equivalent rule for stopping gambling
30
when they are ahead, and because the odds favor the house, they eventually lose (Dawes,
1988). Behaving cautiously with their "own money" is helped by the pain of paying.
Such a visceral sensation is attenuated when playing with the "house money." Hence
losing it is easier.
Emotions and consumer behavior
Primary research attention has been directed to advertising and situationally
induced moods rather than more central motivating factors and stronger emotions (Cohen
& Areni, 1991). For example, consumers who rated a store environment as pleasant and
arousing also reported favorable purchase intentions (Donovan & Rossiter, 1982).
Similarly, Sherman and Smith (1987) found a positive relationship between consumers'
moods, the number of items they bought, and amount of money they spent. Being a
correlational study, is impossible to infer whether the mood caused increased shopping or
resulted from the enhanced shopping experience. Milliman (1982) found that the pace of
background music affected shoppers' walking speed as well as money spent. Over a two-
month period, consumers who shopped to the sounds of music at 60 beats per minute
spent 38 percent more money than ones who listened to music at 108 beats per second.
Restaurants patrons behaved similarly (Milliman, 1986).
The aforementioned studies clarify that shopping is more than the utility of a
product and its dollar cost. They also generally clarify that emotions have a role in the
purchase experience. (While it is uncertain if consumers' emotions in Milliman studies
mediated between the manipulated factors and those consumers' behaviors, at a minimum
these studies provide further evidence to the importance of context effect.) The two
studies reported next are of greater direct relevance to the pain of paying. The
contribution of these works are at a conceptual and an experimental level, respectively.
31
At the conceptual level, Hirschman and Holbrook (1982) suggested that being
emotionally stimulated is an important end-goal for consumers, apart from the enjoyment
of the product purchased. This notion, which they termed "hedonic consumption," is
important for pain of paying research, which is also an hedonic sensation. In the same
way that a stimulating shopping environment can "inject" positive affect into the
evaluation of products, contextual variables can inject pain or pleasure into the evaluation
of payments.
At the experimental level, Isen et al. (1978) had a confederate, posing as a
company representative, give half of those shopping in a store a small gift. The other
half of the shoppers were not approached, and did not receive a gift. A second
experimenter, blind to the experimental manipulation, asked shoppers to evaluate the
performance of various products unrelated to the gift item. Shoppers who previously
received a gift reported higher performance levels. The parallelism with the pain of
paying research is salient. Just as products can "become better" by a contextual
manipulation, payments can feel better or worse depending on the context of the
purchase.
While research on the hedonics of consumption and shopping is sparse, there is
practically no research on the hedonics of payment. In the same way that a product's
evaluation is enhanced by an unrelated gift, an evaluation of a payment as painful or
pleasurable should depend on external variables; at this point we have no scientific
knowledge about the effect of such variables. Taken as a whole, however, the results
from the experiments above provide us with every reason to pursue the study of such
variables in an attempt to understand the effect of the purchase context on the pain of
paying.
32
4.3 Contextual influences in corporal pain
Motivation
The motivation for including a section on corporal pain — ostensibly a topic of
little relevance to a discussion about paying and payments — is that it clarifies the effect
of contextual variables on what may appear as an objective phenomenon. Our
understanding of corporal pain, and particularly its susceptibility to context effects, was
less clear in the past than it is today. At different points in time researchers of corporal
pain emphasized different aspects of the experiences, sometime in an extreme fashion.
Examining some of the historical views of corporal pain stresses the importance of
attending to all aspects of a studied phenomenon, rather than focusing only on particular
parts of it. Thus, this section begins with a description of the evolution of corporal pain
research and some of the trends it saw before a clearer understanding of it was achieved.
This is not a formal historical review, but rather a selection of some of the extreme slants
that corporal pain research has taken.
At different points in history, explanations for pain ranged from primarily
psychological, to strictly physiological, to a combination of both (Gamsa, 1994a). In
some early descriptions before the end of the 19th century, pain was frequently accounted
for by psychological reasons. Emotional pain, according to one antiquated perspective, is
displaced into the body where it may be more bearable. Emotionally-generated pain is
most colorfully illustrated by the following:
A terrified butcher who, on trying to hook up a heavy piece of meat,
slipped, was suspended by the arm on the hook, and when taken to a
chemist, said he suffered acute agony. The hook had only traversed his
coat, the arm was uninjured, and yet he cried out with "excessive pain"
33
when the sleeve was cut off in order to allow the arm to be examined
(Tuke, 1884; p. 168).
From the second half of the 19th century through the 1950s, pain was viewed in
purely physiological terms, and formulated as "a specific sensation with its own
structural, functional, and perceptual properties" (Hardy, Wolff & Goodell, 1952).
According to this view, pain had a "measurable threshold of perception . . . identical to
all humans regardless of sex, age, or race" (Zborowski, 1969). As pain research evolved,
more inclusive definitions have been formulated, suggesting that the neurophysiological
analysis of the sensation explained only part of the total phenomenon (e.g., Beecher,
1952; Sternbach, 1968).
In contrast to older perspectives, which reviewed pain as largely physiologically
determined, modern research on pain exposed the importance of other variables. Current
research gives greater emphasis to non-neurophysiological aspects of pain. As argued by
Chapman (1994), "experience with awake, suffering patients inevitably convinces even
the staunchest purist that human pain is intimately linked with emotion, powerfully
altered for better or worse by a patient's expectations and beliefs, and imperfectly
correlated with tissue injury." Variables that alter the pain's intensity, and the patient's
experience of it, are diverse, and include fault, expectations of benefits, cultural norms,
and control. Each of these is discussed in turn.
34
Fault
Fault is the belief that the pain and suffering result from the intent or negligence
of some identifiable person or persons. Though empirical evidence is still sparse,
suffering appears to be intensified when one's pain is not merely seen as an unfortunate
risk of living, but rather is seen as the result of lack of caution, consideration, or concern
by others (e.g., Bigos & Battie, 1987). For example, DeGood and Kiernan (1995) sorted
patients who got injured on the job according to whether they faulted the employer, some
other person, or no one. The employer-fault group demonstrated the most distress, the
no-fault group the least, and the other-fault group fell between the extremes.
Expectations of benefits
Expectations of benefits is the sense that the injury is not in vain and that future
gains are anticipated in consequence of the immediate suffering. In a study of wounded
soldiers behavior in battle conditions, Beecher (1956) found that only one in four soldiers
with serious injuries complained of pain severe enough to require analgesics. Beecher
explained the finding by claiming that there is an effect of expected benefits on
experienced pain intensity. He argued that the injuries provided soldiers with a
respectable reason to leave the battlefield; those injuries were invested with positive
meaning, and thus hurt less.
Cultural norms
Cultural norms in pain research are reflected by the similarity in pain reports
among individuals belonging to the same civilization. A number of early studies of pain
phenomena have observed uniformities in response to pain among individuals who were
members of identifiable groups characterized by certain common traits (Zborowski,
35
1969). For instance, Libman (1926) found that prizefighters, Blacks, and American
Indians, as groups, failed to react to noxious stimuli of intensity great enough to induce
reaction of discomfort in the average white city dwellers. More recently, in a study of
labor pain memory, Niven and Brodie (1995) compared labor pain memories of women
who gave birth with the "memories" of women who had not been pregnant. (The authors
sought to resolve whether there will be a difference in the reports which are based on
actual memory and those based on non-experiential knowledge about this widely talked-
about pain.) The descriptions were very similar providing a vivid demonstration that at
least for "common pain," one can anticipate and report it without experiencing the
physiological trauma associated with it.
Control
Control is the power that the patient has over the type and degree of experienced
pain. Thompson (1981) provides abundant evidence of the relationship between control
over an aversive event and the amount of stress or pain the event produces. For example
subjects' behavioral control was found to increase their willingness to endure noxious
stimuli. In pain experiments (e.g., Glass, Singer & Friedman, 1969; Glass, Reim &
Singer, 1971; Glass, Singer, Leonard, Krantz, Cohen & Cummings, 1972) subjects in the
"high control" condition (are led to believe that they) are given the option to terminate
the painful stimulus if they so desire. Often, the experimenter asks subjects not to opt to
stop the pain, while making sure they understand that they have the option of doing so.
Usually, subjects in that condition endure the painful stimulus (e.g., cold water) for the
duration of the experiment. Furthermore, these subjects typically report less pain than
those subjects who did not have control over the painful stimulus. For instance, Litt
(1988) manipulated perception of instrumental control by leading subjects to believe that
36
termination of a cold-pressor trial would depend on either their ability to keep their hands
warm in the cold water (high-perceived-control condition) or on reaching an unknown
time limit set by the experimenter (low-perceived-control condition). There was a
significant difference of almost one minute in the average cold-pressor tolerance in favor
of the high-control condition.
Measurement issues
Contemporary pain research struggles to improve methodological rigor and
conceptual clarity, particularly with respect to the role of psychological factors (Gamsa,
1994b). As a subjective phenomenon, pain measurement and assessment pose a daunting
task for scientists who strive for objectivity. Most tools for measuring pain involve self-
reports or (mostly in experimental settings) pain tolerance. In a manual describing the
process of pain assessment and reassessment (1996), the Roxane Pain Institute lists
several measurement tools that have been found useful for assessment of pain.
Noticeably, they are all based on some form of patient self-report. According to the
National Institute of Health (1987), the single most reliable indicator of the existence and
intensity of pain is the patient's self-report. Neither behavior nor vital signs can
substitute for a self report (Beyer, McGrath & Berde, 1990), as patients may be
experiencing excruciating pain even while smiling and using laughter as coping
mechanisms (Fritz, 1988).
Corporal pain and pain of paying
While physical pain research has not surmounted all the methodological and
conceptual hurdles it faces, research on the pain of paying hardly exists. Our
understanding of the hedonics of payment resembles that of physical pain several decades
ago. Economic research interests lie in the "neurophysiological" aspect of payments; that
37
is, what is gained by a given purchase, and what is the "injury" or damage that the
purchase entails in terms of reduced future buying power. However, looking at the
everyday experiences of most consumers should convince even the staunchest economist
that when consumers pay, their feelings are powerfully altered for better or worse by
their expectations and beliefs, and by the context in which the payment is made.
38
Chapter 5: The "need to feel that we are buying something" theory
To understand the pain of paying, a unifying theory has to integrate the contextual
variables and show how they influence a consumer who faces the prospect of making a
payment. Traditionally, payments were assumed only to be an apparatus for resource
allocation, and paying was postulated simply to inform consumers about their reduced
ability for future consumption. A natural extension of the "economic man" theory
implies that losing money is just as painful as using it to make a payment. But, as
physical pain is more than a neurophysiological analysis of the damaged tissues, paying
can be more or less painful (or pleasurable) than losing money; reasons for loss of money
matter. The literature just reviewed helps to shed light on why the factors revealed in
Chapters Two and Three might be important. Whether it is corporal pain or pain of
paying, people need to be able to explain and justify to themselves why they are in pain.
When the explanation is satisfactory, the pain is reduced; when no satisfactory
explanation is available, the pain is increased. Contextual variables alter the explanations
available to the individual who suffers either corporal pain or pain of paying. To give
just one example, in the same way that cold water has been shown to feel less painful
when a subject has control over enduring it, paying for an expensive hotel room is
expected to be less painful when one decides to stay in it voluntarily (perhaps to indulge
oneself) than when one has to stay at the expensive hotel due to lack of vacancy in
cheaper hotels. When a person chooses to stay at an expensive hotel, there is a
satisfactory explanation for the steep expense — to have fun. When the steep expense is
involuntary, the only available explanations — I did not reserve in advance; I am
39
unlucky; the clerk at the cheaper hotel did not try hard enough for me — make the pain
feel worse.
Relying on the information gathered in the studies and literature review, the
model developed in this section suggests that a consumer who is making a payment
contemplates the question: "What does this payment 'buy' me?" The more a payment
buys, the better the explanation for the expense, and the less the pain of paying. In other
words, when spending money, we need to feel that we are buying something with it, not
just throwing it away. Contrary to the implication of traditional economic theory, there is
a big emotional difference between money that is lost and money that is spent to achieve
something; the pain of paying is less in the second case.
The theory developed in this section attempts to bring contextual variables into
the scope of payment research. In so doing, a pain-of-paying picture is presented that
better describes what real consumers go through when paying. For conceptual simplicity
it is convenient to lump the effect of the different contextual variables into three
categories that characterize acquisitions: Transaction, Buffering, and Coercion. Each of
these characteristics alters the pain of paying, for better or worse, by enhancing or
harming consumers' perception that the payment actually buys something for them.
Next, the consequences of these characteristics is explained. They are then used to
generate predictions about factors that make purchases feel more or less painful.
5.1 Characteristics of acquisitions
Transaction characteristics
Transaction characteristics influence the pain of paying by adding to, or
subtracting from, the acquisition utility. When a consumer pays more than a just price,
40
there is a feeling that part of the payment goes towards the product or service, while
another part is devoted to "subsidizing" an overcharging merchant, or to compensating
for inadvertent circumstances. The need to feel that the payment buys something is then
only partially fulfilled. Consequently the pain of paying is strong.
This transaction effect on the pain of paying can be closely compared with a
similar effect on physical pain. To a racquet-ball or squash player, being hit by the game
ball is less painful if one believes it was an unintentional fumble than if the player
believes that the opponent used the shot as a "persuasion" technique for opening the mid-
court. When interpreting the pain's intensity, the hurt player does not consider only the
physiological sensation at the injury, but also the entire "transaction" that caused that
pain. The pain feels less severe when facing an apologetic opponent than when facing a
contemptuous one. In quite a similar manner, to a consumer, being hit with a steep
payment is less painful if one believes that the cost is just (in terms of production costs or
market prices) than if one believes the money went to "support" an insolent merchant.
Bargains, thus, feel especially pain-free. A consumer who purchases an item at a
bargain price, i.e., below its subjective reference price, faces favorable transaction
characteristics. The need to feel that the money buys something is more than fully
satisfied. It may actually feel good to pay for an item knowing that the money buys more
than it ordinarily would have.
Buffering characteristics
Recall that Buffering refers to the extent to which consumption of a purchased
item cushions its payments. Usually, making a payment taints the delight of
consumption, and consuming an item mitigates the pain of paying. When the consumer
can clearly impute a specific consumption to a corresponding payment, that consumption
41
buffers the payment, i.e., the consumer senses that the payment is really buying
something. On the other hand, if the consumer cannot associate the payment with any
particular consumption, then the payment feels just like throwing away money rather than
buying.
Strong buffering (and hence reduced pain of paying) occurs when the consumer
can directly connect the costs in terms of money with the benefits in terms of product or
service. The ability to associate costs and benefits is just as important in physical pain.
To a dental patient, an aching tooth may feel less painful if told that the tooth will be
cleaner and shinier after the pain fades than if told the pain is an inconsequential
byproduct of the treatment. Contemplation of the aching tooth tranquilizes the pain when
one can expect benefits in the form of a beautiful future smile. Conversely, such
contemplation exacerbates the current pain when one cannot visualize future benefits as a
result of it. Likewise, to a consumer who is making a payment, contemplation of this
payment will feel more or less painful depending on how much product or service can be
directly associated with it. The need to feel that the money buys something is easily
satisfied when there is a clear buffering consumption to impute to that money.
Alternatively, if there is no such buffering consumption, the money feels as if it buys
nothing.
Coercion characteristics
The third characteristic measures the control that the consumer has over the type
and amount of a given expense. Consumers who make a payment voluntarily feel that
they use their money to buy something while consumers who have to make a payment
because that is what the circumstances dictate feel that the money is taken from them.
When the expense is a result of necessity or coercion (e.g., when having to check into an
42
expensive hotel due to lack of vacancy in cheaper ones), the consumer may feel extorted,
as opposed to being in a transaction situation. Consequently, that consumer is likely to
experience high pain of paying. The same good can be emotionally easier to pay for
when the consumer chooses it, rather than being compelled to have it.
The physical pain literature attests that corporal pain is subjected to similar
influences. To a gymnast, a sore muscle may be less painful if it results from a controlled
intense session of working-out than if it results from an insufficient warm-up or the use
of malfunctioning equipment. Aching after a voluntary intense physical labor may even
feel good, as if signaling to the gymnast that the session was useful and that the planned
drills for the session were achieved. When the pain is unplanned, one may feel
frustration (at oneself for not warming up sufficiently), or anger (at the club operators for
having malfunctioning equipment), which add insult to injury and intensify the pain.
As physical pain is altered by whether or not one had control over acquiring it, so
is the pain of paying. A speeding ticket may always be annoying, but it is probably much
more so when the driver momentarily lost track of the speed limit signs than when
mindfully choosing to assume the risk of speeding. When paying after speeding
intentionally, the money pays for the thrill of driving fast or for the potential time saved;
the need to feel that the money buys something is satisfied. This need is not satisfied,
however, when one has to pay after speeding unintentionally.
Integrating the characteristics
The emotions associated with making a payment can be negative or positive
depending on these three characteristics. They are negative when, in addition to the
disutility caused by contemplation of forgone future consumption, the payment's
characteristics are hedonically unfavorable. For example, when paying a comprehensive
43
car-insurance bill (especially after relocating to a new state), all the characteristics
indicate a high pain of paying. The payment amount often seems unreasonably high
price for the coverage; it is a bad deal in terms of transaction utility. Unless there is an
accident, one cannot easily sense what the money really buys because it is difficult to
imagine a specific consumption that would buffer the payment. In other words, there is
no cushioning of the payment by any vivid consumption. Finally, because it is illegal to
drive without insurance (and because most of us are not wealthy enough to self-insure),
one has low control over the expense's amount.
Alternatively, paying can feel good. For instance, when prepaying next month's
ski clinic, the expense is planned, so the consumer can shop around for a good deal. The
relationship between the payment and the anticipated consumption is clear, and the
contemplation of this consumption is easy. On top of enjoying the skiing with a
professional instructor, there is a self-investment aspect to the transaction which makes
the consumer who engages in prospective mental accounting feel good, i.e., the portion of
unused future consumption is still large. Finally, the decision to pay for a ski clinic is
done freely and not due to situational impositions.
The difference between paying for the car insurance and the ski clinic is in the
emotions the consumer experiences when answering the question: "What benefit has my
money bought?" In the first case, one does not feel that anything actually was purchased.
All that the payment does is keep the insurance company in business and promise some
status-quo assurance (i.e., that if there is an accident the damaged car will be returned to
its pre-accident condition). Comprehensive car insurance for newcomers to a state tend
to be unjustly expensive (negative transaction characteristics), there is no explicit
consumption to impute to the payment (negative buffering characteristics), and it must be
purchased (negative coercion characteristics). A consumer who pays for the ski clinic,
44
on the other hand, feels that the money truly "buys" something (e.g., enjoyment, better
physical shape, improved skiing skills). When a payment is voluntary and planned, it is
within the consumer's expectations, and the residual consumption is large and clearly
traceable to the payment, the "need to feel that we are buying something," is satisfied,
and the payment is emotionally easy to make. On the other hand, when a consumer has
to make a payment of an unjust amount for a used-up or forgotten item, paying becomes
painful.
5.2 Predictions
The Transactions, Buffering, and Coercion characteristics enable the generation
of hypotheses about many attributes of purchases that will make payments feel more or
less painful. Recall that the insight for the conception of these characteristics was gained
both deductively, from general principles found in the decision-making, marketing, and
corporal-pain literatures, and inductively, from subjects' responses in interviews and
surveys. The responses to these interviews and surveys also implored six specific
attributes as important determinants of the pain-of-paying's intensity. The next section
outlines how these six attributes emanate from the combined effect of the transaction,
buffering, and coercion characteristics.
1. Fairness — A consumer often faces an unfair payment when the chain of events that
led to the purchase appears undeserved or unordinary (in the spirit of Kahneman &
Miller's norm theory (1986), distinct from Thaler's, where unfairness is experienced when
the price is too high relative to some reference price). A consumer in such a situation
may have the well-known frustration emotion: "Why did it happen to me?"; "Why did I
forget to call the airline 21 days in advance and now have to get a pricier ticket?"; "Why
45
did they close the discount clothing store just after I moved here and now I have to buy
everything at the mall?" and so on. In the same way that injuries perceived as an
employer's fault result in increased pain (DeGood & Kiernan, 1995), payments made in
"unfair" situations are predicted to be painful because consumers may feel unlucky or ill-
treated. They have to pay more than would be expected (e.g., a high-fare ticket when
discounted tickets are sold out) or purchase more than they wanted (e.g., a business-class
ticket when coach tickets are sold out). When a consumer buys or pays more than would
be expected, the purchase's transaction characteristics are unfavorable; the payment
exceeds the consumer's reference price. This happens when one just missed a sale, has to
purchase an item in a hurry, or when prices go up. When buying more than one wants or
needs, the coercion characteristics are unfavorable, making the consumer feel forced
rather than voluntarily making a payment. This happens when regular items are sold-out
and one has to buy premium items instead, or when social pressures sway an individual
to patronize an expensive retailer. The greatest pain will be experienced when both
characteristics operate in combination to make the situation feel unfair, for example,
when coach tickets are sold out and business-class tickets cost more than usual.
2. Investment vs. Consumption — Consumers consider a purchase an investment when
they expect the good to generate future utility for them in addition to consumption utility.
For example, paying for a vacation that improves one's skiing skills will be less painful
than paying for a vacation taken for the sake of enjoyment per se. This is because, while
the memory of the fun vacation will decay with time, the improved quality of one's skiing
ability is expected to create additional pleasure, which buffers pain of paying. Similarly,
and somewhat more straightforwardly, paying for an oriental rug should be hedonically
less painful than paying for wall-to-wall carpet, because the rug is expected to appreciate
46
or to maintain its value, while the carpet will simply depreciate over time. In the physical
domain, it has been argued that contemplation of future benefits eases and even numbs
the sense of pain, e.g., in bodily injuries (Beecher, 1956) and childbirth (Salovey et al.,
1991). Similarly, "investment" purchases are concomitant with a large residual positive
consumption and result in a lower overall pain of paying. This result can be derived
directly from the purchase's buffering characteristics, which in the case of an investment
are favorable — there is positive prospective consumption to look forward to and thus the
pain of paying is eased.
3. Immediate vs. Drawn-out payments — This is the flip side of the previous attribute,
and it too stems from prospective accounting. Just as paying for an investment purchase
eases the pain, paying for a purchase that is expected to generate additional financial
costs, aggravates it. According to traditional economic models (e.g., Fisher, 1930),
future expenditures should only affect the consumer in the future, when the expense
reduces the consumer's buying power. Hedonically, however, drawn-out purchases
generate immediate pain due to the consumer's anticipation of their future financial drag.
Therefore, while purchases that are "done for" may have no residual consumption to
mitigate the pain of paying, drawn-out purchases are accompanied by "negative" future
consumption, experienced as a pain of paying. For example, to a consumer who
deliberates between buying a high-end VCR or a laser-disk-player, the thought of paying
for the laser-disk-player may be more deterring because of the high expected future
expense to purchase or rent disks. Likewise, paying for a blouse that can be machine-
washed will be less painful than paying for a dry-clean-only blouse where the purchase's
payment will have to be followed by additional future payments. Again, the analogy
47
with physical pain is straightforward: the pain in an aching organ feels more intense if its
bearer believes it will continue.
4. Self versus Other — In the 50-item study reported in Chapter Three, subjects typically
experienced less pain when paying for items that others would consume than when
paying for items for themselves (for some subjects, up to 9 points difference, on an 11-
point pain-to-pleasure scale). Clearly, when paying for a present to be given to a good
friend, one can experience a warm glow imagining the friend's enjoyment, i.e., the
contemplation of the other's consumption enjoyment boosts the payer's pleasure.
Contrarily, paying for a gift for a disliked colleague (perhaps because "it is the right thing
to do") only boosts the pain created by the payment's forgone consumption utility. The
self-versus-other attribute is intuitively enlightening because when paying for a good that
goes to others (either liked or disliked), the customer who makes the payment
experiences no direct consumption utility from the purchase. That these types of
payments, which are not coupled with any consumption by oneself, can invoke strong
emotions of pain or pleasure (averages of responses to such items ranged from 2 to 9 on
an 11-point pain-to-pleasure scale) attest to the importance of the pain-of-paying concept;
a payment can fell better or worse even when no direct consumption is considered.4
Two of the characteristics can help understand why paying for the sake of another
is less painful then spending money on oneself. First, when paying for gifts, people may
draw from a different set of reference prices; gifts are expected to be exquisite and
4. The result that the subjects in the current study preferred to buy things for others than for themselves, can
be an artifact of the particular items that the study employed, such as paying for a sweatshirt for oneself
versus for one's sweetheart. Surely, the pain of paying depends on the type of self-other relationship (e.g.,
in compensation allocation see Loewenstein, Thompson & Bazerman, 1989). Hence, had the study
employed different items, a different result could have emerged, where subjects would feel more pain of
paying when spending on others. The important point though is that it makes a difference: the pain of
paying is affected by attributes that are consumption-free.
48
expensive. The transaction characteristics are more favorable then. When spending a
certain amount of money on a gift, the same amount feels less expensive and painful than
if the money is spent on oneself, where lower reference prices are used for comparison.
Second, some people may see gifts as an investment in relationships. In this case the
buffering characteristic is favorable. This helps explain why paying for a gift to a
disliked colleague is so painful. There is no "investment-in-friendship" aspect to the
expense, and thus it is not buffered by any contemplation of future benefits.
5. Temporality — Of the six attributes, temporality has received the largest theoretical
and empirical attention. This was done by researchers who studied the prospective
accounting aspect of buffering. According to prospective accounting, mental accounts
are affected primarily by future cost and benefits. Most notably, consumers do not like to
owe money (Prelec & Loewenstein, 1996). When a consumer completes a payment
before consumption is made, and assuming the payment amount is acceptable for the
purchased item, the hedonic anticipation of the future consumption utility buffers the
payment entirely; it feels good to pay for an item knowing that its enjoyment will not be
tainted by a burdensome debt. Moreover, even apart from paying, assuming that one
looks forward to consuming the good or service, the anticipation is psychologically
uplifting in a way that increases the consumer's utility. Likewise, paying for an already
consumed good is painful because one contemplates the forgone future utility without
being able to compensate for it with consumption utility (which has already been
exhausted). The residual payment is large relative to its corresponding consumption, and
the consumer feels that the payment simply maintains the status-quo and does not buy
anything new.
49
With more complicated payment schemes, such as installment plans, there are
exceptions to the general rule of preferring to pay before consuming. Importantly, and
specifically because people are averse to debt, making the last payments in a series may
be pleasurable, since it clears the balance on the debt account.
6. Control — People do not like to be forced into situations, and paying situations are no
different. The control attribute accounts for that fact. It is directly derived from the
coercion characteristics discussed earlier, and it answers the consumer's self-posed
question: "How much control do I have over the type and amount of this expense?" In
the same way that cold water feels less painful when a subject has control over enduring
it, paying is expected to be less painful when a consumer has control over making it.
When the purchase is the result of a free choice, e.g., when choosing whether and how
much to tip at a restaurant, consumers feel that they really are buying something. When
the purchase is imposed, e.g., when having to pay tax on the restaurant's meal, consumers
feel that they are not really buying anything, and the pain of paying is similar to the pain
felt when losing money.
Figure One illustrates the relationship between the theoretical concepts presented
in the previous section and the predication regarding the six attributes.
50
Figure 1
Purchase characteristic Prediction
Transaction
Buffering
Coercion Fairness
Investment-Consumption
Immediate-Drawn-out
Temporality
Control
Self-Other
Surely, the six attributes above do not represent the entire spectrum of factors that
moderate the pain of paying which could be predicted from the transaction, buffering,
and coercion characteristics of a purchase. However, they suffice to demonstrate the
crucial command that contextual variables hold over the emotions that consumers
experience when they pay. Future research may identify additional attributes or explore
their interactions. The remainder of this work, however, will focus on: 1) empirically
corroborating a subset of the proposed attributes; 2) using the identified attributes to
predict consumers temporal preference (when one wants to pay) and modal preference
(how one wants to pay) for paying; 3) understanding the linkage between the pain of
paying and compulsive reluctance to buy; and, 4) demonstrating the attributes'
applicability to consumer shopping behavior.
51
Chapter 6: Causes of Pain of Paying — Experimental evidence
Study Two
So far, the discussion has relied on theoretical premises and literature in
neighboring domains to drive home the message that contextual attributes have an
important influence on the pain of paying. First, decision-making and marketing studies
were used to argue that, just as context affects consumption utility, it should also affect
the pain of paying. Second, by way of analogy, it was argued that just as physical pain
depends on the context in which the pain originated, so does the pain of paying.
On the empirical side, to gain preliminary insight about the variables that affect
the pain of paying, consumers in open-ended interviews described attributes of purchases
whose payments felt especially emotional to them. An additional 28 subjects completed
a written questionnaire in which they catalogued expenses that they considered
affordable or unaffordable, and listed the reason(s) why they found each expense
affordable or prohibitive. These initial responses served as inputs for the structured
questionnaire employed by the study reported in Chapter Three. Indeed, that study
established six attributes as important determinants of the pain of paying. However,
because it was a correlational study, causality is impossible to infer. For example, it is
claimed that consumers who feel they have no control over the type and amount of a
certain bill will feel more pain of paying than consumers who have such control and pay
the bill voluntarily. However, causality may have worked in a reverse order, namely,
different pain-of-paying levels may have caused subjects to feel little or ample control
over an expense rather than vice versa. Before endorsing the pain-of-paying theory, a
prudent strategy requires an experimental manipulation of the attributes which are
52
hypothesized to affect pain of paying to examine their effect. In the current study this is
done for the Temporality attribute.
The objective of the study is to establish the importance of Temporality (whether
payment is made before or after consumption occurs) and ratify it as significantly
influencing the degree of pain of paying that consumers experience.
Conceptually, it is straightforward to manipulate Temporality by changing the
consumption-payment order between experimental conditions. Temporality is interesting
to explore because subjects' behavior is hypothesized to contrast with economic
prescription and financial rationale. From an economic perspective, paying late should
always be preferred to paying early due to the potential interest accrued by adhering to
this principle. Alternatively, based on the theoretical arguments laid down in Chapter
Five, the expectation is that consumers will experience more pain of paying when
payment follows consumption (Loewenstein & Prelec, 1996). As implied by the
buffering characteristic, when they have to pay for an already-consumed item, consumers
have no consumption to buffer the payment, which consequently feels painful.
Consumers who pay before they consume feel that the payment actually buys future
consumption for them and are thus expected to experience less pain of paying.
Procedurally, manipulating temporality can be tricky. How can a brief laboratory
experiment clearly reverse the payment-consumption order across experimental
conditions? Consider using a durable good, e.g., a mug. Whether subjects pay the
experimenter before they are handed the mug or immediately afterward will not result in
very different pain levels across experimental conditions, because subjects in both
conditions can expect future use of the mug. If there is an effect, it will take a huge
number of subjects to uncover it. A viable alternative is using a perishable good, e.g.,
candy. Subjects who have to pay for the candy after eating it are expected to feel more
53
intense pain over letting go of their money than subjects who have the candy's
consumption to look forward to. However, since it is impossible to force them to
consume the candy at the experimental site, some subjects may choose to consume it
later, thereby again hazing the expected difference between experimental conditions.
To grapple the aforementioned problems with most durable and perishable goods,
the current experiment uses a lottery ticket as the good for which subjects need to pay.
While it is perfectly licit as well as feasible to require subjects to scratch a lottery ticket
at the experimental site, whether subjects have to pay for a scratched versus an non-
scratched ticket is hypothesized to significantly influence the experienced pain-of-
paying's intensity. Paying for a non-scratched ticket is buffered by future consumption
— the hope of getting rich, the excitement of revealing the unknown. Paying for an
already-scratched loser ticket feels like throwing away money that does not buy anything.
Normatively though, these arguments should make no difference.
In terms of measuring subjects' pain of paying, two dependent variables are
examined. The first, behavioral, measure is the number of subjects who actually demand
back the money associated with purchasing the lottery ticket by asking for it from the
experimenter. It can be thought of as a proxy of pain tolerance, a measure used
extensively in corporal-pain studies. Receiving money back is assumed to somewhat
compensate for pain felt when purchasing that ticket. Therefore subjects who feel strong
pain should be more likely to demand their money back than those who experience weak
pain. The second dependent variable is subjects' introspective reports of the pain they
felt when paying for the lottery ticket, which they indicate on a scale at the end of the
experiment.
Rationally, in both conditions, all subjects should demand their money back.
According to the experimental hypothesis such requests are to occur more often for the
54
payment-before-consumption condition. That is, subjects who pay before the lottery is
played should be less likely to demand their money back, as well as report less pain of
paying, than subjects who have to pay after their loss is known.
Method
Subjects: Subjects were 120 adult (41 male, 79 female; age 18 to 83) visitors to
the Pittsburgh Aviary . They were solicited to participate in the experiment by a sign
promising them $2 for their participation.
Stimuli: As a reason for receiving the $2 subjects had to perform a filler task —
an unrelated questionnaire about social judgment. The filler task was a survey in which
subjects answered 78 general-knowledge and self-descriptive questions, as well as
provided their estimates of the majority responses of randomly sampled people to the
same questions. The average subject took about 15 minute to complete the survey. In
addition subjects received a Pennsylvania state lottery ticket for which they had to pay
$1. Each ticket had a purchase value of $1 and a maximum prize value of $5,000, which
was stated on the ticket's face. Finally, subjects received a short form on which they
indicated if they wanted reimbursement for the money associated with purchasing the
lottery ticket. That form also included demographic questions.
Procedure: All subjects — Signs at the entrance to the aviary informed visitors
that they could earn $2 for participating in a CMU study. Visitors interested in
participating sat at a table. An experimenter, blind to the experimental hypothesis,
informed subjects that, after reconsidering the difficulty of the task, it has been decided
that they will receive $3 for their time and effort. After completing an informed-consent
form, the experimenter handed each subject three $1 bills and asked the subjects to put
55
their money in their wallets. The experimenter then handed the subjects the unrelated
task, and asked them to hand it back after they finish it. When they returned the survey,
subjects were alternately allotted to one of the two experimental conditions, unless
several subjects completed the survey at the same time, in which case they were allotted
to the same condition so as not to raise their suspicion about the experimental
manipulation.
Payment Before consumption condition — Upon completion of the unrelated task,
the experimenter told subjects that they have to use $1 of their earnings to buy a
Pennsylvania lottery ticket. The experimenter then collected back $1 from the subjects.
(If subjects resisted paying for the lottery, the experimenter reminded them that they were
only promised $2 and that the lottery was part of the experiment.) Subjects could then
scratch the lottery ticket or keep it, and the experimenter gave them a white poker chip to
keep. (See App. 2A.)
Payment After consumption condition — The procedure is identical, except that
subjects paid for the lottery after they scratched it, already knowing whether they won or
lost. Then the experimenter gave subjects a blue poker chip to keep. (See App. 2B.)
Both conditions — After the lottery transaction was completed, the experimenter
asked all the subjects to stop by the experimental booth again after they finished their
tour of the aviary to complete another very short form. When they made the second stop,
subjects received a form with the following note: "Because we cannot really sell lottery
tickets, we have to give you an opportunity to be reimbursed for the dollar you paid to
purchase the lottery ticket. If you demand your money back, please sign below and the
experimenter will return your money." Subjects who wanted reimbursement signed
below a line that read: "I want to be reimbursed for the money I paid to purchase the
lottery ticket."
56
After they decided whether to demand reimbursement or relinquish it, subjects
turned the page over. On the back of the form subjects used a -5 (painful) to +5
(pleasurable) scale to rate their pain of paying for buying the lottery ticket, after reading
the following instructions: "When you are paying for expenses, different aspects of the
expense influence your emotional reaction to the payment. For example, you may
actually feel pleasure when paying for a long awaited vacation, and feel pain or
aggravation when you have to make a monthly payment for a product that turned out to
be a 'lemon'. Think about paying the $1 in order to purchase the lottery ticket. How did
that payment feel? Please try to introspect so as to avoid automatic reaction. That is, try
to recall the moment at which you parted with the $1 bill: how did it feel?"
In addition, subjects had to circle whether they found the study fair or unfair,
indicate if they won anything on the lottery and how much, and provide demographics.
There was also a space for subjects' comments.
When they completed the pain-rating form, subjects handed it back to the
experimenter together with the poker chip. Depending on the chip's color the form was
filed to a "before" or "after" file.
Results
Thirty-four subjects won their lotteries and were excluded from the analysis. Of
the remaining subjects, 45 were in the payment-before-consumption condition and 41
were in the payment-after-consumption condition. Overall, 40% of the subjects
demanded their money back, the average pain of paying was -0.36, and at the 0.1 level
was significantly different from zero (t81 = -1.47, p = 0.072). The rank-correlation
57
coefficient between the two dependent measures, money-back-demand and pain-of-
paying, was in the expected direction (recall that -5 stands for strong pain; rs = -0.17, Z =
-1.55, p = 0.06).
There was no significant difference between experimental conditions in either
money back demand (Proportion Before = 45%, Proportion After = 36%, χ21 = 0.61, p =
0.43, ns) or pain-of-paying reports (Mean Before = -0.63, Mean After = -0.08, t80 = -1.11,
p = 0.14, ns). The lack of significant differences was consistent with respect to both
gender and age.
Discussion
The results failed to support the experimental hypothesis; paying for an already
scratched lottery ticket was more or less painful than paying for an unscratched lottery
ticket. (If one can think of beautiful results as a consumable good, then introspecting
about the money, and other resources, spent to obtain the current results produces intense
pain of paying.) Although disappointing in terms of hypothesis testing, a silver lining
appears with respect to the results' ramification to the pain-of-paying concept as a whole.
Subjects were able to reflect on their pain-of-paying experience, and their introspective
reports were positively correlated with their behavior, as measured by their demand for
reimbursement. So, pain of paying receives convergent validation from multiple
measurement methods. Furthermore, having to pay for the lottery ticket generally caused
subjects to experience pain of paying, as indicated by the average subject's reported pain
that was significantly different from zero. Commenting to the experimenters, one subject
depicted the pain-of-paying sensation particularly colorfully when he indicated that, "The
experiment was unfair for it caused me to feel a pinch when I had to forgo the money I
58
just earned." The implication is that the lottery manipulation worked in as much it
caused subjects to experience pain of paying. The manipulation failed in that the pain-of-
paying intensity did not vary between experimental conditions.
As for understanding the failure to reject the null hypothesis, the availability of
post-hoc explanations is of course myriad. One explanation that appears plausible is
based on verbal and written comments that subjects made in reaction to the request to
purchase a lottery ticket. Quite a few subjects indicated that they had never bought a
lottery ticket and were against doing so. Though, except for two instances of blunt
refusal, all subjects eventually purchased the lottery ticket, some may have done so
despite themselves. For such subjects, paying for a scratched ticket may have actually
been a relief if they were saying to themselves, "It's not a winner, so I didn't really
gamble, no harm done." On the other hand, gambling-averse subjects who paid for an
unscratched lottery ticket could not employ such rationale to improve their self-
perception, and thus may have felt annoyed at performing the disgraceful activity. That
annoyance may have diffused into subjects pain-of-paying reports and boosted the
average pain of paying in the payment-before-consumption group. Consequently, both
groups experienced similar pain's intensity.
The second explanation is based on reexamining the particular procedure that the
experiment followed. Subjects in the payment-after-consumption condition may have
felt as if they were playing with the "experimenter's lottery," and consequently may have
detached themselves from it emotionally. When the disappointing outcome was revealed
those subjects felt bad for having to pay for a loser ticket, but they may have focused
more on the game aspect than on the payment aspect of the transaction, as reflected by a
comment of one subject in that group: "I was happy to buy the lottery ticket because it
was a surprise — 'maybe it's my lucky day' kind of feeling." Subjects who bought the
59
ticket prior to playing it may have felt greater attachment to it. When they found it was a
loser, it was their ticket which lost, not the experimenter's. Following such a thought-
process creates cognitive dissonance, which may have increased the pain-of-paying of
subjects in that group, and equaled it with the pain-of-paying intensity in the payment-
after-consumption group.
To address these two issues a future experiment should have subjects in both
conditions hold on to the ticket for a period of time before they scrape it. This way,
subjects in both experimental conditions will feel as emotionally attached to the ticket.
Secondly, only subjects who have no objection to lottery-playing should be included in
the study.
60
Chapter 7: Implications — Consumers Payment-Preferences
"For some of the nation's biggest banks, the main event at Atlanta's Centennial Olympics
was an ambitious consumer test of the 'smart card,' the plastic cash substitute that is billed
as the first step on the road to electronic money. Yet despite an Olympic-size promotion,
the cards appear to have failed so far to score with Atlanta consumers."
Wall Street Journal, August 5, 1996.
7.1 Motivation
Study One used a 50-item survey to explore the relationship between the pain of
paying and the potential underlying factors on which the 50 items varied. This chapter
describes two studies that used the same 50 items to explore the link between the pain of
paying and consumers' payment-preferences. Undoubtedly, the studies are somewhat
artificial because in the real world consumers do not always have the option of choosing
when and how to pay for goods they purchase. Technological improvements, however,
greatly increase consumers' flexibility in selecting payment plans and payment vehicles.
As these new payment options become available in the marketplace, consumer
preferences — free of technical complications — become increasingly important.5
Plausibly, the pain-of-paying emotion serves as a moderating variable between the
identified attributes and the consumer's payment-preferences. Understanding if and how
the pain of paying affects consumers' payment-time and payment-mode preferences will
allow marketers to better cater to those preferences, which will result in enhanced
shopping experiences (e.g., retailers will create payment plans that consumers will find
less painful). Ignoring consumers' emotions and preferences can result in creating
payment vehicles that back-fire, as the aforementioned "smart card" anecdote illustrates.
5. For a review of the some innovative possibilities see, The New York Times Magazine "Dead as a
Dollar."
61
The next two studies explore the pain of paying influence on consumers' payment-time
and payment-mode preferences. Each study develops a theoretical framework as to why
the pain of paying should affect those preferences, and then empirically tests the
theoretical predictions.
7.2 Study Three: Time-preferences study
The objective of this study is to explore the relationship between the pain of
paying and consumers' preferences regarding when they want to pay for various
expenses. The strong prediction is that items associated with intense pain or pleasure
emotions will lead consumers to choose one of two extreme options: pay immediately or
defer payment for as long as possible. Such bimodal preferences were found by
Loewenstein (1987), who examined subjects temporal preferences of experiencing very
negative, mildly negative, mildly positive, and very positive outcomes. Loewenstein
proposed that people would prefer to get (severely) negative outcomes over with
quickly, as long as the outcome is short-lived. For (very) positive outcomes, he
suggested that people would prefer to delay the experience if the outcome is
"savourable," because they enjoy contemplating it. (The results for the mild outcomes
elicited weaker preferences, which generally fitted the economic prediction.)
The rationale behind these predictions is that people would like to defer fleeting
outcomes that are positive in order to get a chance to experience the enjoyment of
anticipation. Thus, if one has a chance to kiss one's favorite movie star, and can either
get the kiss immediately or with some delay, the delay will actually enhance the overall
utility. The "consumption utility" is the same whether the kiss is immediate or deferred,
but the delay allows for some anticipal utility. If, on the other hand, a negative outcome
is to occur, the anticipation adds fear and dread to the negative consumption utility.
62
Therefore, as long as the outcome is transient and short-lived (e.g., an electric shock), an
individual who must experience it may prefer to get it over with as quickly as possible.
For prolonged positive outcomes no delay is needed for savourabillity to be
experienced. If a person has won a new car at a television show and can choose between
receiving it immediately or with some delay, no delay is "hedonically required." That
person can drive the car immediately and still enjoy the anticipation of using it again in
the future. For negative outcomes that are prolonged, expediting the experience will not
alleviate future suffering. Thus one might as well defer the deterioration of quality of life
as long as possible.
Similar (i.e., bimodal) predictions can be derived from Linville and Fischer's
(1991) Renewable Resources model. According to this model, people possess limited
loss-buffering resources that are consumed in the process of coping with negative events,
but that are naturally renewable over time. Likewise, people possess limited gain-
savouring resources that are consumed in the process of experiencing positive events, but
that are also naturally renewable over time. Although Linville and Fischer developed
their model to explain people's choices between two outcomes (e.g., two rejected
manuscripts), its implications can be extended to predicting preferences for a single
outcome. For a negative outcome, as long as people have the needed resources, they will
prefer to experience it immediately; otherwise, they will prefer to defer it until a time
when the resources are renewed. Regarding a positive outcome, people will prefer to
experience it sooner rather than later, as long as they are not satiated; if they are satiated,
they will opt for delaying the experience.
Based on the findings of Loewenstein and of Linville and Fischer, it is
hypothesized that for items they feel strongly about, consumers (and subjects) will opt
either to pay immediately or defer paying for as long as possible. Distinguishing
63
between these two options depends on which force, delaying or expediting, is stronger.
Consider a painful payment, e.g., "paying for an expensive dinner to which you went
with a boring date." Some subjects may choose to pay immediately, and by so doing
close the "account" and eliminate the pain of paying associated with it. Alternatively,
other subjects, who perhaps lack the psychological resources to deal with thinking about
that miserable evening, may choose to defer paying for as long as they can. By so doing,
these subjects do not eliminate the pain of paying, but at least they save themselves from
experiencing it immediately. As memories of the bad evening decay, the "negative
consumption" fades away, and the payment, which is no longer exacerbated by that
negative consumption, becomes less painful. Similarly, consider a pleasurable payment,
e.g., "paying for an expensive dinner to which you went with a great date." Although
paying is not painful, and therefore nothing argues against an immediate clearance of the
account, consumers may actually decide to defer paying so as to enjoy having the
payment as a mnemonic device to help them remember the great evening.
A final prediction concerns the difference between the intensity of pain versus
pleasure of paying. In an absolute sense, making a "negative" payment is expected to be
much more painful than making a "positive" payment pleasurable.6 This is reminiscent
of what is now an established result, that losses loom larger than gains (Kahneman &
Tversky, 1979). The idea is that negative events are more emotionally influential than
positive ones of the same objective magnitude. For example, writing a check for your
part of a group present to a coworker that you do not like will have a greater emotional
impact than if the gift is for a coworker that you like a lot. Consequently, people will
have stronger preferences regarding when to pay for negative than for positive payments.
6. In Study One, the average pain of paying, across all subjects, ranged from +2.88 to -4.38, and the
medians ranged from +3 to -5. Although this result could be an experimental artifact, since the study
employed 50 different items, it is reasonable to assume that negative payments hurt more than positive ones
comfort.
64
It is thus hypothesized that opting for the two extreme options (paying immediately and
delaying payment as long as possible) will be more frequent for painful than for
pleasurable payments.
Method
Subjects: Fifty-five travelers (29 female and 26 male) awaiting flights at the
Pittsburgh International Airport participated in the study. Several features recommend
the airport as a data collection site: It has a large "captive population" who can afford
(and often enjoy) the distraction of participating in the study; it features numerous
shopping outlets, which plausibly facilitates subjects' introspection when asked about
payments and spending; and, there is expected to be significant income variability among
subjects sampled (which may prove important if wealth plays a role in determining the
pain of paying). Completing the questionnaire took about 10 minutes.
Stimuli: Subjects received the 50-item questionnaire that Study One employed,
describing bills that an individual may face over the course of a month. Each item was
followed by the 11-point scale, on which subjects indicated how painful or pleasurable
paying for a given item feels. In addition, subjects chose their preferred time of paying
each bill among three listed options (see Appendix 3):
As early as possible At a normal time (or don't care) Delay as long as possible
Results
Pain of paying reports: Pain of paying responses spanned the entire -5 to +5
(pain to pleasure) scale and the averaged 0.85 across all 50 all subjects and bills.
Somewhat surprisingly, only 28% of the responses indicated pain (answers below 0)
65
while 57% indicated pleasure of paying. None of the rank-order correlations between the
average pain of paying (for each subject across all expenses) and subjects' demographics
(age, education, income, and the subject's proclivity to carry a credit-card balance)
reached statistical significance.
Payment-time preferences: The modal response was "pay at a normal time (or
don't care)." Of the 2,674 time-choices that subjects made (after excluding missing
values), paying as early as possible was chosen 193 times, delaying payment as long as
possible was chosen 785 times, and paying at a normal time (or don't care) was chosen
1,696 times.
Pain of paying — pay-time relationships: To test how well pain of paying
explains pay-time choices, three separate pairs of binary logistic regressions were run.
Each model tested the ability of pain of paying to predict choice of one payment-time
alternative (e.g., pay as soon as possible) over the other two alternatives (e.g., pay at a
normal time and delay payment). The first analysis tested the study's main hypothesis
that paying at a normal time will become less likely as pain or pleasure intensifies. To do
this the "pay at a normal time" responses were compared with the combined "pay as soon
as possible" (hereafter pay asap) responses and "delay as long as possible" responses.
The other two analyses looked at the other two possible comparisons (because
independence of responses cannot be assumed in these analyses, p-values were adjusted
by a Bonferroni procedure, reported as pB). To control for subject effects, dummy
variables were introduced for each subject. (Additionally, all analyses were run both
with and without dummy variables for questions. Since there were no directional
differences in the pain-of-paying coefficients due to the questions' dummy variables, the
results are reported once for the analyses with the dummy variables for questions
included.)
66
The hypothesis that paying at a normal time will be affected by pain of paying is
supported by the first analysis: pay at a normal time vs. rest (B = 0.073, p = 0.033). This
result indicates that as pain increases, paying at a normal time becomes less likely. The
pain-of-paying coefficient reached statistical significance in one of the other two
analyses: delay as long as possible vs. rest (B = 0.112, p = 0.007 [pB = 0.021]). This
result indicates that as pain of paying decreases, the probability of opting to delay
payment increases, lending some support to the "savourabillity hypothesis." The third
analysis, pay asap vs. rest, did not reach statistical significance (B = -0.026, p = 0.656 [pB
= 1], ns).
Though not appropriate for model estimation purposes, to get a feel for pain-of-
paying levels under the different pay-time options, the independent and dependent
variables were switched, and a one-way ANOVA was run. The pain-of-paying means
under the three pay-time options were 0.425, 0.772, and 1.124, providing further
indication that painful responses tend to fall in the pay "pay asap" category, while less
painful (i.e., more pleasurable) ones tend to belong to the "delay as long as possible"
category. Using a Scheffe F-test procedure, the means' difference was significant for the
"pay asap" vs. "delay as long as possible" (F2, 976 = 5.04, p = 0.01) and "pay at a normal
time" vs. "delay as long as possible" (F2, 2479 = 4.43, p = 0.02) comparisons, but not for
the "pay asap" vs. "pay at a normal time" comparison (F2, 1887 = 1.39, p > 0.1, ns). (A
second ANOVA, which was run with dummy variables for each question and each
subject, to account for question and subject effects, produced the same results' pattern).
To look for particular associations between time of payment and pain-of-paying
levels, the responses were collapsed into three pain-of-paying levels: painful responses
(P.o.P rating below -2), "neutral" responses (P.o.P rating between -2 and 2), and
pleasurable responses (P.o.P rating above 2; see Table 2). Overall, paying asap was
67
chosen about 7.2% of the times. However of painful responses, it was chosen 10.7%.
Deferring payment was opted for 29.4% of the time, yet it constituted 41.5% of
pleasurable responses.
Additionally, observing the table lends support to the hypothesis that for painful
items there will be a bimodal distribution of pay-time responses, namely, paying at a
normal time was chosen less often for painful payments than for neutral payments. This
observation was corroborated by a Chi-Square test. For painful and neutral responses,
comparing the "pay at a normal time" responses with the other two types of pay-time
responses combined, resulted in a Chi-Square of 38.1 (independence of responses cannot
be assumed, so a significance test is inappropriate). Another test further supported the
savourabillity hypothesis. Delay paying was chosen more often for pleasurable items
than for painful and neutral items (Chi-Square = 78.8)
Table 2: Pay-time responses by Pain level
P.o.P
Pay
ASAP
Normal
paying
Delay
paying
Totals
-5 to -3 10.7% 55.2% 34.0% 100%
-2 to 2 6.8% 71.6% 21.6% 100%
3 to 5 6.3% 52.2% 41.5% 100%
Totals
7.2%
63.4%
29.4%
100%
Discussion
The goal of Study Three was to answer the third question posed in the
introductory chapter: How does the pain of paying affect consumers' payment-time
68
(when to pay) preferences? It was hypothesized that a pain-of-paying effect will lead to a
decrease in the paying-at-a-normal-time choices. The results support this hypothesis by
indicating that as pain increases, paying at a normal time becomes less likely.
It was also hypothesized that the most painful items will have a higher likelihood
of being paid either immediately, to get them over with, or as late as possible, to avoid an
immediate pain. This hypothesis is not fully supported by the results because, when
examining all the responses, an increase in pain is not associated with a significant
increase in the likelihood of either paying immediately or delaying payment. The
hypothesis receives some support from the descriptive analysis that concludes the results'
section. Comparing across different pain levels, paying asap is opted for almost twice as
often for painful responses than for benign or pleasurable responses. At least in a relative
sense, then, the tendency to pay immediately and close the account appears stronger
when feeling pain than when feeling pleasure or indifference.
Additionally, it was hypothesized that paying for pleasurable items will be
deferred so as to savour the experience (Loewenstein, 1987). This hypothesis is
supported by the results which indicate a significant increase in the likelihood of delaying
payment as long as possible as the pain of paying declines. Support for this hypothesis is
also evident in the descriptive analysis. Comparing across different pain levels, delaying
payment was associated about 30% more often with pleasurable responses than with
benign or painful responses. Thus here, again in a relative sense, the tendency to delay
(and perhaps savour) paying appears stronger when feeling pleasure than when feeling
pain or indifference.
Finally, it was hypothesized that extreme pay-time choices will occur more often
under painful than pleasurable items. The results in table 2 do not support this
hypothesis. In comparison with benign payments, the pay at a normal time option lost
69
about the same response-share under painful and pleasurable payments (16.4% and
20.6%, respectively).
The reason for the results' provision of only partial hypotheses' support may be
better understood by considering that most responses did not reveal extreme pain (or
pleasure) of paying. Consequently, effects that depend on strong pain or pleasure
responses are difficult to detect. In addition, possibly due to imperfect wording, the
middle option on the pay-time scale ("pay at a normal time, (or don't care)") attracted the
bulk of the responses (e.g., in retrospect it is clear the defining one option as "normal"
should cause it to get most of the responses). As a result, effects that depend on polar
payment-time choices also are difficult to detect. That said, what becomes surprising is
that hypotheses that subsist on the combination of strong emotional responses and polar
payment-time choices received any support at all — evidently, subjects emotions and
preferences were powerful enough to be revealed despite the study's limitations. A future
study may achieve significant results to support all the hypotheses by providing subjects
with only the two extreme payment-time options. Presumably, given the success of the
current study in showing relationships despite including the "attractive" pay at a normal
time option, achieving statistical significance with an improved questionnaire should not
be difficult.
7.3 Study Four: Mode-preferences study
The objective of this study is to explore the relationship between the pain of paying
and consumers' preferences regarding how they want to pay for various expenses (i.e.,
Cash, Check, Credit Card, Bank auto-deduction). Making specific predictions about how
consumers would prefer to pay is an even more difficult task than predicting when they
70
would like to pay, because each of the four payment tools have some features that attenuate
pain of paying and others that exacerbate it.
Cash immediately eliminates the pain of paying which makes it a clear candidate
for painful items. The idea is that paying with cash allows consumers to instantly rid
themselves of painful payments. For instance, say last week you borrowed lunch-money
from a colleague who is now nagging you to pay back. A likely response in this situation
would be to give the colleague cash on the spot and get the unpleasant affair over with.
Paying with any other mean but cash will just be a reminder for the already nagging event.
On the other hand, cash also maximizes the coupling between consumption and payment,
which sometimes can make paying with it painful. In some situations in which painful
items are considered, consumers would prefer to minimize the consumption-payment
coupling. For example, paying cash at a gas station for filling up a gas-guzzling car
emphasizes the fact that driving the car costs too much. A different payment tool may thus
be chosen.
A check blatantly makes consumers feel as if they are paying twice — once while
writing it and then again when the money is withdrawn from the account (or when the
bank statement or the copy of the canceled check arrives). It is therefore plausible that
painful items will not be paid by a check. But, checks also provide consumers with some
verification and revocation power, because they can use it as proof of payment or even
void it if things turn ugly. This may be the reason that certain consumers pay with a
check for some, usually small, expenses, even though paying cash would incur less
hassle. Examples include paying the local news boy for the monthly delivery, or paying
various administrative fees.
Credit cards are probably the most difficult to analyze from a pain-of-paying
perspective. In addition to the lack of knowledge about how they affect consumers
71
hedonically, credit cards, ubiquitous as a payment tool, offer benefits that add to the
Acquisition utility. Hence, at least two reasons invoke paying via a credit card as the
preferred option for painful items. First, because many consumers already associate
paying by a credit card with extra benefits, they may choose it so as to balance the pain
of paying with the perks that the card provides (e.g., airline frequent-flyer miles). This
effect has not escaped the eye of savvy marketers, as reflected in the VISA advertisement
depicting the introspection of buying an expensive car: "Was it a midlife crisis? Or was it
the miles?"
The second reason driving credit cars as tools for painful payments is that in cases
of negative consumption, such as the expensive dinner with a boring date, a credit card,
which minimizes the consumption-payment coupling, may serve to make the payment
feel more generic, and thus less painful. Indeed, there is evidence that consumers spend
more with a cash-card and with a credit card than with cash (Borgen, 1976; Galanoy,
1980; Hirschman, 1979; Huck, 1976; The New York Times Magazine, 1977, 1996).
Feinberg (1986), demonstrated that, similarly to the "weapons effect" — where the
presence of a weapon triggered aggressive responses, there is a "credit-card effect." In a
series of four experiments, Feinberg's subjects reported higher reservation prices for
different products if a credit card was present than if it was not.
Alternatively, paying with a credit card can, in a way, multiply the pain of paying.
Especially for consumers who are meticulous about reviewing their credit-card statement
at the end of each billing cycle, using a credit card for making painful payments can be
torturous. Such consumers may still use this payment vehicle for the reasons stated
above, yet they may experience strong pain when they contemplate the payment.
Finally, an automatic bank deduction frees the consumer almost completely from
dealing with the pain, so this "out of sight, out of mind" tool may be chosen for painful
72
items. Paying for insurance is a case in point. Retrospectively, insurance payments are
pure waste of money; if no catastrophe happened, the insurance payment had not bought
the consumer any benefit. Paying it would most certainly be painful, but paying it
electronically perhaps less so.
While, as indicated above, it is difficult to lay down specific hypotheses about
payment-mode preferences, reports of many consumers suggest that there might be
inconsistencies between the payment vehicles available to the consumers and the ones
they would prefer given a choice. For example, in a 1996 survey of 3,000 US consumers,
which has been conducted annually by DDB-Needham advertising agency for the last 22
years, 60% of female and 63% of male consumers responded affirmatively to the
question: "I like to pay cash for everything I buy." However, most of the respondents
(81% female, 82% male) also indicated that they have used a major credit card (DDB
Needham Annual Lifestyle Survey 1974-1996). Undoubtedly, there is room for a
focused investigation on how consumers prefer to pay for different types of expenses.
The following study examines consumers' payment-vehicle preferences, and submits that,
similarly to the payment-time preferences, pain of paying is plausibly an important
moderator.
73
Method
Subjects: Fifty-five travelers (26 female and 29 male) awaiting flights at the
Pittsburgh International Airport participated in the study. (For an elaboration on the
reasoning for choosing this population, refer to Study Three.)
Stimuli: The questionnaire is identical to that used in Study Three, except that, in
addition to the painful/pleasurable responses, subjects chose their preferred vehicle of
paying each bill among four listed options (see Appendix 4):
Cash Check Credit Card Bank auto-deduction
Results
Pain of paying reports: Pain of paying responses spanned the entire -5 to +5
(pain to pleasure) scale, and the average pain of paying across all 50 choices that all
subjects made was -0.01. Thirty-five percent of the responses indicated pain (answers
below 0) and 38% indicated pleasure of paying. None of the rank-order correlations
between the average pain of paying (for each subject across all expenses) and subjects'
demographics (age, education, income, and the subject's proclivity to carry a credit-card
balance) reached statistical significance.
Payment-mode preferences: The modal response was "pay by a Check." Of the
2,586 mode-choices that subjects made (after excluding missing values), Cash was
chosen 445 times, Check 1355 times, Credit Card 557 times, and Bank-Deduction 229
times.
Pain of paying — pay-mode relationships: Four separate pairs of binary logistic
regressions were run. In each model, one type of pay-mode responses was compared
74
with the other pay-mode responses (all the p values were adjusted by a Bonferroni
procedure). Dummy variables were included for subjects but not for questions, because
particular transactions may be associated with a particular payment vehicle and with an
extreme pain of paying, an observation that would not be detected if question effect is
controlled for. (Adding question dummy variables did not result in a sign change in any
of the coefficients, though not all reached statistical significance.)
The pain-of-paying coefficient reached statistical significance in all four analyses
: Cash vs. rest (B = 0.128, pB < 0.001); Check vs. rest (B = -0.143, pB < 0.001); Credit
Card vs. rest (B = 0.15, pB < 0.001); Bank-Deduction vs. rest (B = -0.126, pB = 0.0187).
These results suggest that check and bank-deduction usage increases with an increase in
the pain of paying, while cash and credit card usage increases with a reduction in the pain
of paying.
Again, to get a feel for pain-of paying levels under the different pay-mode
options, the independent and dependent variables were switched, and a one-way ANOVA
was run. The pain-of-paying means under the four pay-mode options were: Cash 0.46,
Check -0.35, Credit Card 0.51, Bank-Deduction -1.135. Using a Scheffe F-test
procedure, except for Cash vs. Credit Card, the two "pleasuring" payment-modes, all the
pair-wise comparisons of means' differences were significant: Cash vs. Check (F2,1798 =
11.21, p = 0.001); Cash vs. Credit Card (F2, 1000 = 0.031, p > 0.1, ns); Cash vs. Bank-
Deduction (F2, 672 = 19.6, p = 0.001); Check vs. Credit Card (F2, 1910 = 14.88, p = 0.001);
Check vs. Bank-Deduction (F2, 1582 = 6.15, p = 0.002); Credit Card vs. Bank-Deduction
(F2, 784 = 22.37, p = 0.001).
To look for particular associations between mode of payment and pain-of-paying
levels, the responses were collapsed into three pain-of-paying levels: painful responses
(P.o.P rating below -2), "neutral" responses (P.o.P rating between -2 and 2), and
75
pleasurable responses (P.o.P rating above 2; see Table 3). Most interestingly, two clear
trends appear in the table: for cash and credit card, relative frequency increase with
pleasure of paying; for check and bank deduction relative frequency increases with pain
of paying.
Table 3: Pay-mode responses by Pain level
P.o.P
Cash
Check Credit
Card
Bank
Deductio
n
Totals
-5 to -3 10.9% 61.4% 13.6% 14.2% 100%
-2 to 2 18.1% 51.5% 22.0% 8.6% 100%
3 to 5 21.4% 44.8% 29.6% 4.2% 100%
Totals
17.2%
52.4%
21.5%
8.9%
100%
Discussion
The goal of Study Four was to answer the second part of the third question posed
in the introduction chapter: How does the pain of paying affect consumers' payment-
mode (how to pay) preferences? Though, as explained in the introduction to this study,
no strong operational hypotheses were raised, it was generally expected that high-pain
responses will be associated with payment methods that clear the account quickly (i.e.,
Cash) or that put it out of sight and mind (e.g., Bank-Deduction). A further expectation
was that such painful responses will not be associated with methods such as Check and
Credit Card, which emphasize the degree of consumption-payment coupling and/or
require consumers to attend to the expense beyond what is absolutely necessary (i.e.,
after writing a check consumers often see it again on a bank statement, and after signing a
76
credit-card slip consumers are reminded of the expense when going over the credit card
company statement).
The logistic regressions verified some of these expectations, and they received
additional, though moderate, support from the qualitative analysis. Bank-Deduction,
which was chosen about 9% of the times over all, represents 14% of the painful
responses, and Credit Card, which account for 22% of all choices, represents only 14% of
the painful responses but 30% of pleasurable ones. Additionally, cash, which maximizes
payment-consumption coupling, was relatively chosen most frequently under pleasurable
responses. There were also two counter-expectations results; while Check held its own
for all pain levels, it is somewhat more prevalent for painful than for pleasurable items;
and Cash, which accounts for 17% of the total responses, represents only 11% of the
painful ones.
Explaining the particular preferences that the results suggest is difficult even post
hoc. However, looking back at the questionnaire and reexamining the data may elucidate
why these preferences do not perfectly match the a priori expectation. Check was the
preferred payment-mode more often than the rest of the methods combined, and Bank-
Deduction was chosen less than 10% of the times. These data suggest that, in this study,
subjects' payment-habits and pragmatic considerations drove their responses more than
did their emotions. Although the questionnaire requested subjects to "introspect so as to
avoid automatic responses," it did not instruct subjects to deliberately ignore pragmatic
considerations. Choosing when to pay an expense is a relatively unique task that requires
subjects to actually look inward and reveal their preferences. Choosing how to pay, on
the other hand, is highly routine and likely to be more determined by habit than by
immediate hedonic considerations. Especially because different payment methods are
ubiquitous, subjects may have revealed their memorized or programmed responses by
77
associating each expense with its typical, rather than emotionally-preferred, payment
mode. A pain-of-paying effect, then, ceases to exist.
Furthermore, in some cases pragmatism requires consumers to choose payment
tools in a counter-hedonic way. For example, as speculated earlier, people tend to choose
paying by check when they need to keep records of their payments. Typically, though
not always, expenses for which keeping records is important are unpleasant; for instance,
tax payments. Thus, in reality paying by check may be associated with painful items
more often than it would have been in the absence of pragmatic constraints.
To try overcome this automaticity problem, the logistic regression analyses were
repeated on a 25-item subset of the 50 items, which produced the largest variation in
subjects' choices.7 Although this procedure cannot guarantee that individual subjects'
habits do not affect payment-mode choice, it at least targets transactions in which there is
less evidence of inter-individual unanimity about accepted payment mode. All the
coefficients were in the same direction as in the analyses on the full set of questions, but
now, employing the Bonferroni procedure, only the coefficient for Bank-Deduction
reached statistical significance (B = -0.229, pB = 0.0044). This result indicates that as
pain of paying increases, the tendency to choose bank-deduction as a payment method
increases. Recall that bank-deduction was advocated as a payment tool for painful items
because it puts the payment out of sight and out of mind. Furthermore, unlike with the
three other payment vehicles, there are no apparent arguments supporting bank-deduction
as a payment method for pleasurable items. This result is important because it essentially
informs that when a single sided predication can be made based on pain of paying, and
the transactions are characterized by a great variability across consumers (i.e., there is no
7. This was done by ranking the 50 expense items according to their entropy statistic computed as:
- ∑ pi*Log pi
78
consensus due to pragmatic considerations), pain of paying becomes a good predictor for
payment-mode.
There are several ways for future research to further battle the problem of subjects
who provide memory-based, rather than emotionally- or preferentially-based, responses.
One straightforward way is to present subjects with unusual and unfamiliar expenses.
Generalizing the results from such a study will not be a trivial effort, but it will constitute
a first step in establishing the importance of emotions in the choice of payment vehicles.
General Discussion
Beyond hypotheses-testing and expectations-corroborating, the payment-time and
payment-mode studies submit that the pain-of-paying emotion varies with when and how
consumers pay. In both studies different payment times and payment methods were
associated with different pain-of-paying levels. This has theoretical as well as applied
implications. Theoretically, it further substantiates the importance of emotions in
general, and the pain of paying in particular, in influencing what are apparently the
"coldest" decisions. Ostensibly, when and how to pay "should" be decided according to
financial rationale and budgetary convenience. Studies Three and Four, however,
suggest that this is not the way people make those decisions, at least not entirely.
Consumers bring "hot" feelings to their financial decision making. Studying these
emotions and their financial implications will undoubtedly advance our ability to explain
and predict consumers' preferences and behaviors.
Understanding consumers' preferences vis-a-vis when and how to pay has far-
reaching applied ramifications as well. Some payment tools attract consumers, in spite of
79
having financial disadvantages. For instance, SPRINT is offering a prepaid calling card
(which is in vast use in other countries). One of the reasons the company believes in the
card's potential is that it "allows consumers to budget their long-distance usage by
purchasing only the amount of calling that they can afford." In spite of obvious loss to
consumers (in term of opportunity cost, i.e., potential interest gains), in the second
quarter of 1995, the card has achieved a 58% growth in distribution. Students, young
adults (18-29), business travelers, and military personnel are among the biggest users
(SPRINT press releases, June 27,1995). Separately, SPRINT offers cellular phone
customers the option to prepay for their plan. Again, one reason the company initiated
the prepay method is "there were still some people who felt cellular service was beyond
their means . . . with the prepay plan, everyone could afford cellular services . . . ."
Customers get a tone indicating their account balance is near depletion, at which point
they can make another payment (SPRINT press release, July 20,1995). In another
domain, a novel college-financing venture that allows consumers to prepay up to four
years of tuition, at a state college at a locked-in price, has already 400,000 participating
families. The plan mostly attracts middle income people who earn $35,000-$70,000
(New York Times, Aug. 6, 1995). In all these programs payment is made before
consumption occurs.
Although these examples have an applied flavor, tying them back to the pain-of-
paying concept is trivial. Paying for a consumed good or service is painful; paying a
priori is less so. To minimize pain of paying, consumers do not like to owe money, and
thus programs that allow them to prepay have an inherent emotional advantage for
consumers.
Just as it is useful in explaining success stories (admittedly, post hoc), the pain of
paying can help us understand payment-plan fiascoes. Recall the aforementioned "smart-
80
card," or "money-replacer" card, described in the opening of this chapter, which failed to
allure Atlanta's consumers. With such a tool consumers can amasses expenses, without
having a clear sense of how to impute a given payment to any particular expense. There
is, therefore, no buffering of payment by consumption, leading to strong pain of paying.
Added to that the fact that the smart-card does not carry the amenities that credit cards
do, it is easy to see why it struggled.
Moreover, the pain of paying can explain the failure of programs which are
clearly financially advantageous. For example, a new credit card that allows customers
to borrow from their retirement saving accounts and thus pay lower interest rates, has not
caught on. The card has been suggested by Franco Modigliani et al. who were granted a
patent on it in 1993 (New York Times, Aug. 27, 1995). Having to pay for credit items
after they have been consumed, knowing that you depleted your saving account, is
especially painful because at the time of payment there will be no consumption to buffer
that payment. Anticipating this pain, consumers have not been attracted to Modigliani's
credit card.
Yet another perplexing phenomena emerges when examining spending behaviors
across different nations. In some countries the typical consumer constantly overdrafts
heavily (more than their monthly income). In other countries such behavior is rare. Do
the indebted in one country have a different comprehension of basic financial rules than
their fully-paid counterparts?; are they "victims" of a "slippery slope" effect?; or, are they
insensitive to pain (pain of paying, that is)?
The results from Studies Three and Four submit that to understand the behaviors
described above, and to design successful payment vehicles, being familiar with
economic and financial principles do not suffice. Psychological and hedonic principles
appear to be at least as important.
81
Chapter 8: Implications
Compulsive Reluctance to Buy and the Pain of Paying
"Money is a good servant but a dangerous master."
Dominique bouhours
Husband: "I want you to teach my wife the beauty of thrift."
Psychiatrist: "Do you mean the necessity or advisability of thrift?"
Husband: "Not at all. I mean the beauty of thrift. There is something
nice and sentimental about thrift. A dollar is a dollar."
Psychiatrist: "That is mathematically correct."
Husband: "There is more to it in thrift. Beauty, do you hear me, real beauty..."
Edmund Bergler, Money and Emotional Conflicts.
Economics and Marketing descriptions of buying behavior typically assume that
consumers act emotionlessly so as to maximize their long-term (subjective) utility.
Attempts to describe consumers' deviations from this normative model have concentrated
on situations in which consumers buy too much, too often, too easily or too frivolously.
These kinds of buying behavior, which fail to serve the consumer's long-term preferences
due to uncontrollable profligacy, have been termed "compulsive shopping." This
behavior has received ample attention by the popular press (Time, 1987; Washington
Post, 1994), as well as academic consideration (Dittmar, Beattie & Friese, 1995; Faber &
O'Guinn, 1992; O'Guinn & Faber, 1989; Olshavsky & Granbois, 1979; Rook, 1987; Rook
& Fisher, 1995; Rook & Hoch, 1985; Stern, 1962). Moreover, compulsive shopping has
been officially certified as a distinct malady in the directory of psychiatric disorders.
While compulsive shopping is inherently interesting to study, an equally
intriguing phenomenon, as yet largely unexplored by researchers, is "compulsive
reluctance to buy." This behavior is manifested as consumers' (emotion-driven)
reluctance to purchase an item even when purchasing would comply with their long-term
82
preferences. This chapter uses the case of Mr. A again as an example of compulsive
reluctance-to-buy behavior and its drawbacks to the consumer. Reconsider the following
scenario from Chapter One:
Mr. A plays tennis twice a week. Recently he discovered a small crack in his
racquet. Although the crack doesn't prevent him from playing, he feels that it
impairs the level of his game, so last weekend he drove to the mall to buy a
replacement. After spending more than an hour in each of two stores, and looking
at several racquets that he would like to own, Mr. A was still unable to let go of
the money. So he left the mall without buying a racquet.
After losing some more games due to playing with a cracked racquet, Mr. A. will
presumably drive to the mall again, and eventually, after more torturous thoughts about
spending the money, purchase the racquet he needs. Given such a presumption, the first
trip to the mall, in which he could not bring himself to buy the racquet, was a waste of
time and money (e.g., gas), and overall an unnecessary aggravating experience for Mr. A.
But, as much as this hypothetical scenario may evoke empathy, the magnitude and
intensity of the compulsive reluctance-to-buy phenomenon is yet unknown.8
In addition to the discomfort and waste of resources that it causes to the
consumer, reluctance-to-buy should also be of concern to marketers (e.g., retailers).
While compulsive shoppers are a reliable source of income to merchants, "compulsive
misers" merely waste marketers' resources. By wandering through stores without buying,
such customers squander salespersons' time and consume attention that would be more
8. Possibly, other reasons have led Mr. A to defer buying the racquet. He may be an extremely risk averse
consumer, who has to examine many models before making a purchase, or he may be a "bargain hunter,"
who tenaciously searches for the best possible deal (Schindler, 1989). Alternative explanations for this
particular scenario not withstanding, Study Five, as well as field evidence, suggest that in many cases the
pain of paying plays at least a partial role in impeding purchases. To consider only one additional example
here, many drivers can report useless prolonged parking-search experiences, when the thought of paying for
garage parking was unbearable. Since there is little variation in the quality of the "treatment" one's car gets
while parked, risk aversion cannot explain the reluctance-to-pay-for-parking phenomenon. Since most
garages commonly charge similar (high) rates, bargain hunting is not a likely explanation either.
83
lucrative if directed at customers who are ready to spend. Hence, both consumers and
merchants would benefit from knowledge about this phenomenon, its emotional and
behavioral aspects. The study reported next was conducted to obtain such knowledge.
That is, to lay open the presence of compulsive reluctance to buy — impeded spending
even when not spending is against the consumer's best interest — and explore its
relationship with the "pain of paying" — people's emotional distress over spending
money.
Method
Subjects: Fifty-two travelers (20 female and 32 male) awaiting flights at the
Washington D.C. and Pittsburgh International Airports participated in the study. (For an
elaboration on the reasoning for choosing this population, refer to Study Three.)
Stimuli: Subjects received a questionnaire consisting of four parts. The first part
generally asked subjects about their difficulties to spend and to limit their spending. The
second part did the same for twelve specific consumption categories (e.g., gifts, phone-
calls). The third part presented subjects with three pairs of scenarios. In each pair, the
scenarios describe two consumers who generally have quite similar preferences.
However, in one scenario the consumer behaves as a spendthrift; the consumer in the
matching scenario exhibits spending-aversion. (The story about Mr. A's cracked racquet
is, in fact, the one that subjects received as Scenario 1A.) Subjects indicated the
consumer with whom they could more easily identify and (separately) the consumer
whose actions they found more typical of most consumers. The questionnaire's fourth
part asked subjects about their behavior and feelings before, during, and after going
84
shopping. The questions addressed both compulsive shopping and compulsive reluctance
to buy, as well as specifically the pain of paying. Overall, the questionnaire contained 28
questions plus demographics (Appendix 5).
Procedure: An experimenter requested travelers at the airport to fill out the
questionnaire. Those who agreed returned the questionnaire to the experimenter upon
completion of it. Subjects took about 10-15 minute to complete the questionnaire.
Results
The presentation of the results follows the topical organization of the
questionnaire, namely, general money-spending behavior, classified money-spending
behavior, scenarios, and before/during/after shopping behavior and feelings. In each
question the scale range and the average answer are reported. Where appropriate (i.e.,
where the scale has a natural middle point), a two-tails t-Test compares the sample
average to what would have been expected had there been no significant one-sided
tendency. For example, the question "Which of the following descriptions fits you
better?" was answered on an 11-point (tightwad-to-spendthrift) scale, on which 6
corresponded to "about the same or neither." Assuming a normal distribution, if subjects
did not feel that one of the anchors describes them better than the other, the average
answer would not significantly differ from 6. Also, where appropriate a within-subjects
paired two-tails t-Test compares the answers to two related questions (e.g., "Which of the
following descriptions fits you better?" versus "Which of the following descriptions fits
the typical consumer better?"). A summary of the results is presented in Table 4. The
85
relationship between the pain-of paying and the reluctance-to-spend results in the table is
analyzed later in this section.
86
Table 4
General difficulties to spend and to limit spending
q# Question description Scale Mean 1-Sample
t-Test
Paired t-Test
1 Have trouble limit spending 1-never to 5-always 2.4
2 Have trouble spending 1-never to 5-always 2.6 (q1-q2) -1.04
3 Seeing oneself as Tightwad vs. Spendthrift 1-tightwad to
11-spendthrift
5.7 -1.53
4 Seeing others as Tightwad vs. Spendthrift 1-tightwad to
11-spendthrift
7.4 6.51*** (q3-q4) -5.24***
Self reported under- and over-spending for specified consumption categories
q# Type of expense Scale (all items):
1-don't spend when should to
11-spend when shouldn't
Mean 1-Sample
t-Test
5 Dry cleaning 4.6 -5.42***
6 Fragrances & Colognes 5.2 -3.07**
7 Gifts 6.7 2.08*
8 Life-insurance 5.4 -1.95
9 Long-distance phone calls 6.3 1.10
10 Lunches 5.98 -0.07
11 Nice restaurant dinners 6.3 0.92
12 Parking 4.5 -5.09***
13 Professional clothing 4.7 -3.88***
14 Taxies 4.5 -4.04***
15 Vacations 5.8 -0.73
16 Favorite hobby 6.4 1.00
Scenario pairs: who are you (most people) more similar to, A or B?
q# Scenario description Scale Mean 1-Sample
t-Test
Paired t-Test
17 Mr. A "can't" spend money on a racquet he needs
Mr. B buys a racquet he doesn't really need
1-Mr. A to 5-Mr. B 2.4 -3.3**
18 Who do you think most people are more alike? 1-Mr. A to 5-Mr. B 3.6 4.39*** (q17-18) -
5.48***
19 Ms. C buys unneeded things because they're on sale
Ms. D "can't" buys needed things even on sale
1-Ms. C to 5-Ms. D 3.3 2.17*
20 Who do you think most people are more alike? 1-Ms. C to 5-Ms. D 2.5 -3.48** (q19-20) 4.17***
21 Miss E. "can't" pay for parking
Miss F. pays for a garage so as not to waste time
1-Miss E to 5-Miss F 3.0 0.00
22 Who do you think most people are more alike? 1-Miss E to 5-Miss F 3.4 2.87** (q21-q22) -1.85
Before, during, and after shopping behavior and feelings
q# Question description Scale Mean 1-Sample
t-Test
Paired t-Test
23 Before shopping: Excitement about anticipated
products vs. concern about money to be spent
1-Anticipation to
11-Concern
5.9 -0.31
24 When buying "on sale" you're thinking about 1-$spent to 11-
$saved
6.6 1.54
25 Compromising quality so as not to over-spend 1-never to 5-always 2.5
26 Avoiding buying because paying feels painful 1-never to 5-always 2.9
87
27 Regretting not buying an item 1-never to 5-always 3.0
28 Regretting buying an item 1-never to 5-always 2.7 (q27-q28) 2.47*
*p < 0.05; **p < 0.01; ***p < 0.001. Italics indicate a significant result in the sought direction.
88
The general description questions (q1-4) support the supposition that spending
aversion behavior really exists. While 28 subjects indicated that they sometimes or often
have difficulty spending money, only 23 subjects indicated similar difficulty with
limiting their spending; yet, the difference in the average response did not reach
statistical significance. Twenty-two subjects described themselves more as tightwads
(answered lower than 6-"about the same or neither") and only 16 opted towards the
spendthrift end. Possibly, social desirability could drive this result; however, an informal
pretest found that, if anything, tightwad has a more negative connotation than spendthrift,
which makes a social desirability explanation unlikely.9 While the average response did
not significantly differ from 6 for the entire sample (in a two-tail t-Test), it did for
females; they were more likely to describe themselves as spending averse (Mean = 5.1,
t19 = -3.02, p = 0.071; the male's sample mean was 6.03, ns). Finally, 39 subjects
described the typical consumer as a spendthrift (answered greater than 6) and only five as
a tightwad; the average answer significantly differed from 6, and from the average
answer to the self description question (see Table). (There was no directional difference
between males and females.)
The spending-aversion intuition, supported by the general description questions,
received further corroboration from answers to the specific consumption questions (q5-
16). Subjects reported they under-spend for eight of the 12 categories, and the result was
significant for five of those. In only one category, "gifts," the opposite tendency, over-
9. In a follow-up study, Drazen Prelec, George Loewenstein, and I tested the social explanation directly by
asking a different group of subjects "which of the two descriptions fits them better," and "where they would
like to be on the tightwad-spendthrift scale." The difference in the responses to these question was not
statistically significant, but if anything, respondents wanted to be more spendthrift than they are (Means =
4.16 & 4.22, t63 = 0.4, p= 0.69, ns).
89
spending, reached statistical significance. Gender had a significant directional effect in
two items, "life insurance" (Females' Mean = 6.03, ns; Males' Mean = 4.89, t29 = -2.89, p
= 0.0074), and "favorite hobby" (Females' Mean = 4.75, t19 = -2.73, p = 0.014; Males'
Mean = 7.39, t30 = 3.44, p = 0.0017), indicating that men feel they spend too little on life
insurance and too much on their hobbies, while women feel the opposite.
The responses to the scenario pairs — coupled narrative descriptions of one
consumer who spends impulsively and another who behaves miserly — provide further
evidence for compulsive reluctance to spend/buy. In the "racquet" and "sale" scenarios,
subjects found themselves more similar to the penny pincher than to the prodigal. (In the
"meter" scenario there was a null effect, with the responses below the mid-point perfectly
mirroring those above it.) Subjects perceived other people as easy spenders, which is the
opposite of how they perceived themselves. These results are significant for all three
scenarios, as are the differences in responses between self and other within each pair for
the "racquet" and "sale" scenarios. Subjects' gender had no effect.
Responses to the final part of the questionnaire provide some evidence about the
other concept of interest — the pain of paying. Thirty subjects reported that they
sometimes or often "buy items of lesser quality than they 'know' they should, just because
they didn't want to over-spend." More strikingly, 39 subjects reported that they
sometimes or often "avoid buying an item because the thought of paying for it feels too
'painful'" (italics in the original); only two subjects reported that they never do it.
Finally, 31 subjects reported that they sometimes or often regret buying things when they
retrospect about it, and 37 subjects reported regret about not buying. The pairwise
difference in responses is significant (see Table 4).
The study's final goal was to explore the association between the two concepts of
interest, reluctance to buy and pain of paying. For that, responses to the pain-of-paying
90
question (q26) were correlated with those to the reluctance-to-buy questions. Of the 18
correlations run, 15 were in the anticipated direction — the higher the pain of paying, the
greater the reluctance to buy — though the coefficients were usually small. Three
correlation coefficients that significantly differed from zero were with the two tightwad
descriptions question (q2 & q3, r = 0.3, p = 0.03) and with the description of the tightwad
person in the "sale" scenario (q19, r = 0.34, p = 0.015). The coefficients for the three
items that correlated against the hypothesized direction were practically zero (r < 0.07).
Thus, at a minimum, the results provide a mild indication for a positive relationship
between the two concepts.
Finally, it is interesting to consider whether rich people feel more or less pain of
paying than poorer ones. Since income was reported on a 15-unequal-steps scale, a
Spearman correlation examined its association with the pain of paying. Contrary to the
prediction of the economic theory, by what is known as the "wealth effect," the results
hint that the richer a person is, the more painful paying feels, or vice versa (rS = 0.32, Z =
2.26, p = 0.024; rS corrected for ties = 0.23, Z= 1.72, p = 0.083).
Discussion
This study looked for the existence of the compulsive-reluctance-to-buy
phenomenon and found it. What's more, between compulsive shopping and compulsive
reluctance to buy, the latter appears stronger, at least according to the subjects who filled
out the survey. Overall, in 13 out of 18 possible unilateral comparisons (q1-3, 5-17, 19,
21), and in the direct pairwise comparison, a majority of subjects' responses emphasize
their difficulty to spend money over the difficulty to conserve it. Given these results,
91
why is there a greater fascination with people who lose control over their money than
with those whose out-of-control behavior is expressed by self-defying frugality?
A tentative answer lies in the differences between intra- and inter-individual
perception. Relative to themselves, in all four self-other comparisons, subjects perceived
their peers — whether denoted "the typical consumer" or "most people" — as looser with
money. Across the four questions, the average difference in self-other perception was 16
percent. Plausibly, most consumers, most of the time, see themselves as doing okay; that
is, spending "correctly" with respect to their means and aspirations. To the extent the
results from this study are generalizable to the population at large, if people (writers,
journalists, and researchers included) see themselves as more cautious with money than
they see the next guy, this could explain the tagging of others as compulsive shoppers.
(Indeed, people often use the term casually to describe acquaintances who are far from
deserving it clinically.) But why the difference in perception?
Arguably, the differential perception is partially due to our limited accessibility to
other people's behavior, and partially to our even more limited accessibility to their
emotions. When observing themselves, consumers "see" both spending and conservation
behaviors. On the other hand, when consumers have an opportunity to observe others,
they are more likely to see those others spend than conserve. For instance, when other
consumers pay for merchandise, we see them buy and spend; we have no idea of all the
things that they wanted to buy but did not. Similarly, when we see others dine at a
restaurant, we see them consume (and spend); we do not, however, see them all those
other times, when they have dinner at home (perhaps to save for that one dinner when we
do see them eat out). Even given the opportunity to randomly sample from others'
behaviors, it is unlikely that the bias in perception will be completely overcome because
one cannot experience another's feelings, at least not perfectly.
92
Throughout this work — and corroborated by this study's results — it has been
argued that the pain of paying plays an important role in impeding people's expenditures.
Accordingly, a consumer needs to sense the pain of paying in order to curb an abutting
expense. When reporting about themselves subjects may have done just that. Namely,
sensing the pain, they may have "curbed" their hypothetical expenses, or alternatively the
pain memory may have served as a reminder for times when it led them to curb their
expense in actuality. However, when analyzing another's behavior, subjects did not and
could not experience the other's emotions since it is difficult to empathize with pain.
Tooth, ear, and head aches are notoriously painful, yet try to imagine having a headache;
most people cannot. Luckily, we can anticipate, or perhaps gut-feel, many pains just
before they are about to hit us, and thus often avoid them; but we cannot mentally force
ourselves to experience a virtual pain in a pain-free context (Loewenstein, 1996,
Proposition 7; Scarry, 1985). For themselves, subjects had access to the emotions that
led them to avoid spending, or to the memory of that resultant avoidance behavior. For
others, at most, they had the mere facts or dry descriptions. Consequently, subjects
reported less self-spending than they did for others.
This study leaves some interesting questions unanswered. Because it engaged
consumers in an introspective task rather than monitored their spending behavior, it sheds
dim light on what really happens out there. How many people really spend too little?10
Unlike with extreme compulsive shopping, which may lead to bankruptcy, or even
mental hospitalization, at least ostensibly compulsive miserliness does not represent a
problem for the consumer. After all, who would not want to "suffer" from having too
much money. Thoroughly answering this question is also left for future researchers (and
10. Especially interesting is this question considering individual differences. Recall that men felt they spend
too little on life insurance and too much on their hobbies, while women felt the opposite. Whether these
reports correlate with actual behavior is an empirical question that is left for future research.
93
psychoanalysts, and writers ala Moliere), but the data suggest that the ostensible answer
above should be reexamined. More subjects reported regretting not buying things than
buying them, and the difference between the averages of the responses was significant
(where are all the compulsive buyers and impulse shoppers?; are they happy with all their
purchases and regret they hadn't bought even more?). Other novel research on consumer
regret (Frederick, 1995) also points in the same direction, that when looking backwards,
consumers are unhappy when they under-spend. That study found that consumers more
often regret buying cheap items cheaply than expensive ones expensively.
Additionally, there is evidence that the trends identified in the current study are
not unique to the study's particular pool of subjects, but rather characterize consumers at
large. In a 1996 survey of 3,000 US consumers, most respondents did not agree with
items that describe them as careless spenders. 74% of the respondents denied the claim:
"I pretty much spend for today and let tomorrow bring what it will"; 72% denied the
claim: "I frequently buy things when I can't afford them"; and 64% denied the claim:
"I'm an impulse buyer." In contrast, most of the surveyed consumers embraced items
describing thoughtful spending and shopping behaviors. 60% of the consumers (50%
male, 71% female) responded affirmatively to the question: "Before going shopping, I sit
down and make out a complete shopping list"; 73% (65% male, 81% female) concurred
the claim: "I always check prices even on small items"; and, 86% affirmed the claim:
"When buying expensive items for my home I usually take a lot of time to examine all
the choices before making a purchase." (DDB Needham Annual Lifestyle Survey 1974-
1996.)
In summary, the goal of Study Five was to answer the fifth question posed in the
introduction chapter: Does the pain of paying impact consumers' buying behavior? As
hypothesized, consumers who tend to anticipate high pain of paying reported more
94
"miserly" behavior than consumers who are relatively free of such pains. Although
somewhat suggestive due to reliance on retrospective reports, the results indicate that
compulsive reluctance to buy is a phenomenon at large. As such, it is consequential to
both consumers and marketers, which leads to the conclusion that directing research
efforts to studying this behavior is a promising investment.
95
Chapter 9: Conclusion, Limitations, and Contribution
9.1 Conclusion
Summarizing what has been achieved in this dissertation can be done by
reviewing the answers to the questions posed at the outset of this work:
1) What are the purchase characteristics that influence the pain of paying?
2) What are the contextual attributes that determine the intensity of the pain of paying?
3) How does the pain of paying affect consumers' payment-time/mode preferences?
4) Does the pain of paying impact consumers' reluctance to spend?
To answer the first two questions about the pain-of-paying's underpinnings,
Chapter Four draws a close analogy between the pain of paying and corporal pain,
showing how the latter, too, acutely depends on the context in which the injury or ache
was acquired. Chapter Five then develops a theory proposing that three characteristics of
a purchase — transaction, buffering, and coercion — work together to make a payment
feel more or less painful. It subsequently generates six predictions about specific
attributes of purchases that impact the intensity of the pain of paying. Chapter Six
empirically examines a prediction of the theory that paying before consuming is less
painful than vice versa.
Concerning Question Three, while the variety of payment vehicles continues to
increase, it is unclear that developers of such vehicles are attuned to consumer
preferences. That is, in lieu of empirical research, payment tools are constructed based
on solely financial and pragmatic considerations, ignoring consumers' psychological
dispositions. In some cases, pragmatism may reside in harmony with psychological
96
propensities, but not always. For example, as Loewenstein and Prelec (1996) show,
people who are about to perform a job for another person typically prefer to do the work
before collecting compensation for it, although such a preference is financially
"irrational." By addressing the issues of payment-time and payment-mode preferences
directly, the current research provides better understanding of consumers' predilections
for when and how to pay. Developers and providers of payment tools can utilize this
increased understanding to fine-tune their endeavors to comply with consumers'
tendencies, or at least not contradict them.
Question Four essentially asks whether the pain of paying affects actual
purchasing behavior. By answering this question Chapter Eight makes the leap from
revealing that payments are painful to showing that they are consequential; in fact,
painful enough to impede consumers' spending.
9.2 Limitations
By introducing the pain of paying as an emotion deserving of research, this work
draws an important distinction between two aspects of purchases, paying and consuming.
This distinction then raises two quandaries regarding consumers' abilities to separate,
identify, and report their emotions that are purely associated with paying.
The first quandary is whether consumers can isolate a payment's utility from the
entire transaction utility. For example, when paying for a dinner with a boring date, can
one separate the pain of paying from the disutility of the entire transaction? Consider, for
instance, a scenario in which the person could charge the dinner to an expense account.
The dinner with the boring date would probably still be rated as a negative experience,
although not due to the pain of paying. The studies reported in the current work do not
97
provide a clear answer to this question because it is possible that when responding to an
introspective questionnaire, subjects did not (and perhaps could not) purely isolate the
feeling they expected to experience when paying. This empirical limitation does not by
any means impair the consumption-payment distinction at a conceptual level. That is, as
Chapter Five shows, regardless of any experimental shortcomings, at a theoretical level
the distinction is clear and robust. Furthermore, the consumption-payment linkage is
central in defining the pain of paying. Had this work attempted to treat payments and
consumption as foreign entities, adequate study of the pain of paying would have been
impossible.11
The second quandary is about consumers' abilities to integrate the pain of paying.
When paying with cash, pain integration is not an issue because, if there is pain, it is felt
at once upon parting with the cash. When using other payment methods, it is
questionable at what point consumers experience the pain of paying. For example, when
paying by a check, it is unclear whether consumers experience the pain of paying when
signing the check, when the check clears, or when reviewing the monthly bank statement.
It is likely that some pain is felt at each of these occasions, but which carries the bulk of
it remains an open question that the current work does not directly address.
Risking tautology, an argument can be made that most (or at least much) of the
pain of paying is collapsed into the initial payment episode, e.g., at the moment of
signing a check. In fact, just as the difficulty of jumping into a pool of cold water on a
hot summer day increases right before one is about to jump, the pain-of-paying is felt
11. To address the empirical issue and circumvent subjects' requirement to introspect, future research could
explore the pain of paying of consumers who pay with their own money versus using someone else's
dollars, such as an expense account. For example, photographs of facial expressions of consumers who pay
out of their own pocket may convey more anguish than corresponding expressions of consumers who
splurge at their boss's expense. (The suggestion for this experiment is based on a discussion with Danny
Kahneman.)
98
right before one is about to pay. Otherwise, the pain of paying could not play the
adaptive sentinel role that it is hypothesized to play. (If consumers only feel the pain of
paying when looking at their bank statements, it will be too late to prevent careless
profligacy.) Yet, empirical validation of this intuition is left for future research.
99
9.3 Potential contribution
The focus of this dissertation is on the pain of paying, its determinants, and its
effects on spending behavior. To my knowledge, this is the first piece of research
focusing specifically on the payment side of the transaction. Beyond this focused
exploration, however, the dissertation will hopefully make two broader contributions.
First, it seeks to provide a broader conception of consumer behavior that
incorporates the role of emotions. Contrary to the economic perspective, spending is an
emotional affair which depends both on personal and situational factors. Consumers
differ in the degree to which paying is painful. The spendthrift has difficulties
controlling spending despite himself. The tightwad has difficulties parting with the
money even when it is in his best interest. As highlighted in Chapter Eight, most of the
research has focused on the former category — consumers who spend their money too
freely and against their better judgment — neglecting consumers at the other extreme.
The pain of paying provides a theoretical basis for understanding individual differences
and for directing research to the unexamined area of unwilling frugality. Once the
connection between the pain of paying and the reluctance to buy is made, future research
could look for ways to help consumers who fall in that category find ways to fight the
pain of paying. Albeit, as argued throughout this work, the pain of paying is largely a
positive emotion that assists consumers against splurging, it can have negative
consequences to those who experiences it too intensely. Similarly to many other
emotions (fear turning to phobia, love turning to obsession), such consequences can be
avoided by better understanding of the pain of paying.
Additionally, the pain of paying depends on situational factors; contextual
influences which are ostensibly and economically irrelevant. By showing that the
100
amount of a payment often plays a smaller role than contextual influences, this work
highlights the general importance of attending to such factors. In other words, this work
stresses the importance of adopting a realistic view of consumer payment behavior and
suggests an alternative to economic models which disregard the emotional aspects of that
behavior.
Second, broadening the scope even more, the dissertation seeks to contribute to
basic research on the role of emotions in decision making. It presents a unique attempt to
understand the crucial role that emotions play in what is seemingly a cold process. Many
studies have researched affective aspects of decision making (Isen, Daubman & Nowicki,
1987; Isen & Geva, 1987; Isen & Patrick, 1983; Johnson & Tversky, 1983; Salovey &
Birnbaum, 1989; Wright & Bower, 1992). Typically, though, emotions are portrayed as
interfering with the rational and logical decision making process: smokers cannot resist a
cigarette after deciding to quit; obese persons cannot fight the temptation of a dessert
after deciding to go on a diet; passionate individuals cannot resist sexual temptation after
deciding to abstain; and, consumers cannot resist a sale after deciding to save. By
contrast, this dissertation is consistent with the minority of works which present emotions
in a positive light (e.g., Frank, 1988, Goleman, 1995). For most consumers, most of the
time, the pain of paying is a good thing, a decision aid for the times when cold cognitive
processes could not be employed. While it is a crude mechanism that may occasionally
result in faulty decisions, the pain of paying also serves as an emotional decision support
system. Potentially, this dissertation will encourage future research to explore the
adaptive role of emotions in other decision making domains. After all, a person facing a
dessert tray cannot mathematically weigh the cost in calories and fat versus the benefit in
joy of eating a rich desert. Hence "visceral decision" makes perfect sense as a heuristic,
even if for some people some of the time it results in an error.
101
102
References
Ainslie, G. (1975). Specious Reward: A Behavioral Theory of Impulsiveness and
Impulse Control. Psychological Bulletin, 82, 463-509.
Ainslie, G. and Haslam, N. (1992). Hyperbolic Discounting. In G. Loewenstein, and J.
Elster, (Eds.), Choice Over Time. (pp. 57-92). Russell Sage Foundation.
Akerlof, G. A. (1991). Procrastination and obedience. American Economic Review, 81,
1-19.
Arkes, H. R. and Blumer, C. (1985). The Psychology of Sunk Cost. Organizational
Behavior and Human Decision Processes, 35, 124-140.
Beecher, H. K.. (1952). Experimental Pharmacology and the Measurement of the
Subjective Response. Science, 157-162.
Benzion, U., Rapoport, A. and Yagil, J. (1989). Discount Rates Inferred from Decisions:
An Experimental Study. Management Science, 35 270-284.
Bergler, E. (1959). Money and Emotional Conflicts. International Universities Press,
Inc.
Beyer J. E., McGrath P. J., Berde CB. (1990). Discordance between self-report and
behavioral pain measures in children aged 3-7 years after surgery. J Pain Symptom
Manage, 5(6):350-6.
Bigos, S. J. and Battie, M. C. (1987). Acute Care to Prevent Back Disability. Clin.
Ortho. Rel. Res., 221, 212-130.
Biswas, A. and Burton, S. (1994). An Experimental Assessment of Effects Associated
With Alternative Tensile Price Claims. Journal of Business Research, 29, 65-73.
Borgen, C. W. (1976). Learning Experiences in Retailing: Text and Case, New York:
Goodyear.
Chapman, C. R. (1994). Assessment of Pain. In W. Nimo and G. Smith (Ed.),
Anesthesia. Blackwell. England.
Cohen, J. B. and Areni, C. S. (1991). Affect and Consumer Behavior. In Robertson, T.
S. and Kassarjian, H. H. Ed. (1991). Handbook of Consumer Behavior. Prentice-Hall,
NJ.
103
Dawes, R. M. (1988). Rational Choice in an Uncertain World. Harcourt Brace
Jovanovich, Inc. San Diego.
Damasio, A. R. (1994). Descartes' Error: Emotion, Reason, and the Human Brain. New
York: Putnam.
DDB Needham Annual Lifestyle Survey (1974-1996), Chicago: DDB Needham
Worldwide.
DeGood, D. E. and Kiernan, B. (1996). Perception of Fault in Patients with Chronic
Pain. Pain, 64, 153-159.
Dittmar H., Beattie J. and Friese S. (1995). Gender Identity and Material Symbols:
Objects and Decision Considerations in Impulse Purchases. Journal of Economic
Psychology, 16, 491-511,
Donovan, R. J., and Rossiter, J. R. (1982). Store Atmospherics: an Environmental
Psychology Approach. Journal of Retailing, 58, 34-57.
Elster, J. (1985). Weakness of Will and the Free-Rider Problem. Economics and
Philosophy, 1 231-265.
Elster, J. and Loewenstein, G. (1992). Choice Over Time. New York: Russell Sage
Foundation.
Faber, R. J. and O'Guinn, T. C. (1992). A Clinical Screener for Compulsive Buying.
Journal of Consumer Research, 19(3), 459-469.
Fehr, E. and Zych, P. K. (1994). The Power of Temptation: Irrationally Myopic Excess
Consumption in an Addiction Experiment. Working paper #3/94, Institute of Economics,
Technical University of Vienna.
Feinberg, R. A. (1986). Credit Cards as Spending Facilitating Stimuli: A Conditioning
Interpretation. Journal of Consumer Research, 13, 348-356.
Fisher, I. (1930). The Theory of Interest. New York: Macmillian, .
Frank, R. H. (1988). Passions Within Reason: The Strategic Role of Emotions. Norton.
New York.
Frederick, S. (1995). Consumer Regret. Working paper, Social and Decision Sciences
Dept. Carnegie Mellon University.
104
Fritz, D. J. (1988). Non-invasive pain control methods used by cancer outpatients
(meeting abstract). Oncology Nursing Forum, Suppl. 108.
Galanoy, T. (1980) Charge it: Inside the Credit Card Conspiracy, New York: G. P.
Putnam Sons.
Gamsa, A. (1994a). Clinical Reviews; The Role of Psychological Factors in Chronic
Pain. 1. A Half Century of Study. Pain, 57, 5-15.
Gamsa, A. (1994b). The Role of Psychological Factors in Chronic Pain. 2. A critical
appraisal. Pain, 57, 5-15.
Glass, D. C., Reim, B. and Singer, J. E. (1971). Behavioral Consequences of Adaptation
to Controllable and Uncontrollable Noise. Journal of Experimental Social Psychology, 7,
244-257.
Glass, D. C., Singer, J. E., Leonard, H. S., Krantz, D. Cohen, S. and Cummings, H.
(1973). Perceived Control of Aversive Stimulation and the Reduction of Stress
Responses. Journal of Personality and Social Psychology, 41, 577-595.
Goleman, D. (1995). Emotional Intelligence. Bantam Books, New York.
Gruen, D. W. R. (1989). What People Know and What Economists Think They Know:
Survey on Ricardian Equivalence. Working paper, Research Bank of Australia.
Hardy, J. D., Wolff, H. G. and Goodell, H. (1952). Pain Sensation and Reaction.
Williams and Wilkins. Baltimore.
Hirschman, E. (1979). Differences in Consumer Purchase Behavior of Credit Card
Payment Systems, Journal of Consumer Research, 6, 58-66.
Hirschman, E. C., and Holbrook, M. B. (1982). Hedonic Consumption: Emerging
Concepts, Methods and Propositions. Journal of Marketing, 46, 92-101.
Hirst, D. E., Joyce, E. J. and Schadewald, M. S. (1994). Mental Accounting and
Outcome Contiguity in Consumer-Borrowing Decisions. Organizational Behavior and
Human Decision Processes, 58, 136-152.
Isen, A. M., Daubman, K. A. and Nowicki, G. P. (1987). Positive Affect Facilitates
Creative Problem Solving. Journal of Personality and Social Psychology, 52, 1122-
1131.
Isen, A. M. and Geva, N. (1987). The Influence of Positive Affect On Acceptable Level
of Risk: The Person With The Large Canoe Has a Large Worry. Organizational
Behavior and Human Decision Processes, 39, 145-154.
105
Isen, A. M. and Patrick, R. (1983). The Effect of Positive Feelings On Risk Taking:
When the Chips Are Down. Organizational Behavior and Human Performance, 31, 194-
202.
Isen, A. M., Shalker, T. E., Clark, M., and Karp, L. (1978). Affect, Accessibility of
Material in Memory, and Behavior: A Cognitive Loop? Journal of Personality and
Social Psychology, 36, 1-12.
Johnson, E. J. and Tversky, A. (1983). Affect, Generalization, and the Perception of
Risk. Journal of Personality and Social Psychology, 45, 20-31.
Kahneman, D. (1992). Reference Points, Anchors, Norms, and Mixed Feelings.
Organizational Behavior and Human Decision Processes, 51. 296-312.
Kahneman, D., Knetsch, J. L. and Thaler, R. H. (1986). Fairness as a Constraint to Profit
Seeking: Entitlement in the Market. American Economic Review, 76, 728-741.
Kahneman, D. and Miller, D. T. (1986). Norm Theory: Comparing Reality to Its
Alternatives. Psychological Review, 93 (2), 136-153.
Kahneman, D. and Tversky, A. (1979). Prospect Theory: An Analysis of Decision Under
Risk. Econometrica, 47, 263-291.
Kennedy, P. (1992). A Guide to Econometrics. MIT Press, Cambridge, MA.
Laibson, D. (1995). Hyperbolic Discount Functions, Undersaving, and Savings Policy.
Working paper, Harvard University.
Libman, E. (1926). Observations on Sensitiveness to Pain. Transactions of the
Association of American Physicians, 41, 305.
Linville, P. W. and Fischer, G. W. (1991). Preference for Separating or Combining
Events. Journal of Personality and Social Psychology, 60 (1), 5-23.
Litt, M. D. (1988). Self-Efficacy and Perceived Control: Cognitive Mediators of Pain
Tolerance. Journal of Personality and Social Psychology, 54(1), 149-160.
Loewenstein, G. (1987). Anticipation and the Valuation of Delayed Consumption. The
Economic Journal, 97, 666-684.
Loewenstein, G. (1996). Out of Control: Visceral Influences on Behavior.
Organizational Behavior and Human Decision Processes, 65(3), 272-292.
106
Loewenstein, G. F. Thompson, L. and Bazerman, M. H., (1989). Social Utility and
Decision Making in Interpersonal Contexts. Journal of Personality and Social
Psychology, 57 (3) 426-441.
Miller, G. A. Levin, D. N., Kozak, M. J., Cook, E. W., McLean, A., Jr. and Lang, P. J.
(1987). Individual Differences in Imagery and the Psychophysiology of Emotion.
Cognition and Emotion 1, 367-90.
Milliman, R. E. (1982).. Using Background Music to Affect the Behavior of Supermarket
Shoppers. Journal of Marketing, 46, 86-91.
Milliman, R. E. (1986). The Influence of Background Music on the Behavior of
Restaurant Patrons. Journal of Consumer Research, 13, 286-289.
Morris, M. H. and Morris, G. (1995) Market Oriented Pricing: Strategies for
Management. NTC Business Books, IL.
Nagle, T. T. and Holden, R. K. (1995). The Strategy and Tactics of Pricing. Prentice
Hall, New Jersey.
National Institutes of Health Consensus Development Conference (1987). The Integrated
Approach to the Management of Pain. J Pain Symptom Manage, 2(1):35-44.
Nisbett, R. E. and Ross, L. (1980). Human Inference: Strategies and Shortcomings of
Social Judgment. Englewood Cliff, N. J.: Prentice-Hall.
Niven, C. A. and Brodie, E. E. (1995) Memory for Labor Pain: Context and Quality.
Pain, 64, 387-392.
O'Guinn, T. C. and Faber, R. J. (1989). Compulsive Buying: A Phenomenological
Exploration. Journal of Consumer Research, 16(2), 147-157.
Olshavsky, R. W. and Granbois, D. H. (1979). Consumer Decision Making — Fact or
Fiction? Journal of Consumer Research, 6, 93-100.
Prelec, D. and Loewenstein, G. (1991). Decision Making Over Time and Under
Uncertainty: A Common Approach. Management Science, Vol. 37. No. 7. 770-786.
Prelec, D. and Loewenstein, G. (1996). The Red and the Black: Mental Accounting of
Saving and Debt. Working paper, MIT/CMU.
Rook, D. (1987). The Buying Impulse. Journal of Consumer Research. 14. 189-199.
Rook, D. and Fisher, J. (1995). Normative Influences on Impulsive Buying Behavior.
Journal of Consumer Research, (22) 305-315.
107
Rook, D. and Hoch, S. J. (1985). Consuming Impulses. In Advances in Consumer
Research, (12) ed. Hirschman, E. C. and Holbrook, M. B. UT: Association for Consumer
Research, 23-27.
Roxane Pain Institute www home-page (1996). Acute Pain Management: Operative or
Medical Procedures and Trauma — Process of Pain Assessment and Reassessment.
http://www.roxane.com/Roxane/RPI/HealthcareProfessionalLibrary/AHCPR/apmc3.html
#Head12
Salovey, P. and Birnbaum, D (1989). Influence of Mood on Health-Relevant Cognitions.
Journal of Personality and Social Psychology, 57, 539-551.
Salovey, P., Sieber, W. J., Smith, A. F., Turk, D. C., Jobe, J. B., and Willis, G. B. (1991).
Laboratory research on response error in health interview survey questions: judgment and
recall of pain episodes. Draft final report, National Center for Health Statistics. DHHS
200-88-7001.
Scarry, E. (1985). The Body in Pain. Oxford, England: Oxford University Press.
Schindler, R. (1989). The Excitement of Getting a Bargain: Some Hypotheses
Concerning the Origins and Effects of Smart-Shopper Feelings. Advances in Consumer
Research, (16) 447-453.
Sherman, E., and Smith, R. B. (1987). Mood States of Shoppers and Store Image:
Promising Interactions and Possible Behavioral Effects. Advances in Consumers
Research, 251-254.
Simon, H. A. (1955). A Behavioral Model of Rational Choice. Quarterly Journal of
Economics. (69) 99-118.
[Soman, D. and Gourville, J. T. (1995). Transaction Decoupling: When Costs and
Benefits Become Losses and Gains. Draft, Grad. School of Bus. Univ. of Chicago.]
Stern, H. (1962). The significance of Impulse Buying Today. Journal of Marketing. 26.
56-62.
Sternbach, R. A. (1968). Pain: A Psychophysical Analysis. Academic Press. New York
and London.
Thaler, R. (1980). Toward a Positive Theory of Consumer Choice. Journal of Economic
Behavior and Organization. 1 39-60.
Thaler, R. (1985). Mental Accounting and Consumer Choice. Marketing Science. (4). 3.
199-214.
108
Thompson, S. C, (1981). Will It Hurt Less If I Can Control It? A Complex Answer to a
Simple Question. Psychological Bulletin, 90 (1), 89-101.
Time (1987). Fighting the Urge to Splurge. December 14.
The New York Times Magazine (1977). Merchants of Debt. Feb. 28.
The New York Times Magazine (1996). Dead as a Dollar. June 16.
Tuke, D. H. (1984). Illustrations of the Influences of the Mind upon the Body in Health
and Disease Designed to Elucidate the Imagination, 2nd edition., Henry C. Lea's,
Philadelphia, PA.
Wall Street Journal (1996). The Smart Money is on 'Smart Cards,' But Electronic Cash
Seems Dumb to Some. August 5.
Washington Post (1994). Stop Me Before I Shop Again; We Seem to Have a Serious
Conflict Between Mental and Economic Health. September 14.
Wright, W. F. and Bower, G. H. (1992). Mood effects on Subjective Probability
Assessment. Organizational Behavior and Human Decision Processes, 52, 276-291.
Zborowski, M. (1969). People In Pain. Jossey-Bass Inc., Publishers. San Francisco.
