Time Window as a Self-Control
Denominator: Shorter Windows Shift
Preference toward Virtues and Longer
Windows toward Vices
RAFAY A. SIDDIQUI
FRANK MAY
ASHWANI MONGA
When planning future consumption, individuals are known to opt for large virtue quantities and small vice quantities as a means of self-control. We argue that such planning
may also involve the time window within which a given quantity needs to be consumed
because the final objective is to plan for a desired consumption rate (i.e., quantity/time
window)—a high virtue rate and a low vice rate. Five studies reveal that, when quantity
is held constant, a short window (i.e., high rate) nudges individuals toward virtues, and
a long window (i.e., low rate) toward vices. We find this effect for hypothetical and real
virtue-vice choices, preferences, and willingness to extend a time window. Furthermore,
these effects are mediated by the pursuit of long-term health goals, and are moderated
such that the effect of time windows is stronger for those who need more help in meeting their self-control goals—that is, impulsive individuals. While these effects are consistent with self-control, we discuss a blend of mechanisms that may be working in conjunction, particularly at the stage that we focus on: planning rather than consuming. Our
results offer strong theoretical implications and important consequences for the marketplace where expiration periods and other time windows are ubiquitous.
Keywords: time, virtue, vice, self-control
C
onsumers often encounter a time window within
which the units of a good need to be consumed. For
instance, a grocery store may sell packs of salad boxes,
cupcakes, and other products that need to be consumed
within a given period. We propose that time windows have
asymmetric consequences for virtues versus vices: a
shorter window increases preference for virtues (e.g., salads), but a longer window increases preference for vices
(e.g., cupcakes).
Our findings add to the self-control literature about a farsighted planner trying to control a myopic doer’s actions
(Benabou and Pycia 2002; Schelling 1984; Thaler and
Shefrin 1981). The planner tries to nudge the doer self toward virtues, and away from vices (DellaVigna and
Malmendier 2006; Prelec and Loewenstein 1998; Schelling
1992; Trope and Fishbach 2000). In the domain of products, self-control plans often involve opting for a large
quantity of virtues and a small quantity of vices (Mishra
and Mishra 2011; Wertenbroch 1998). We argue that such
planning may also involve the consumption time window
because the final objective is not simply to control the
Rafay A. Siddiqui (rafay.siddiqui@polyu.edu.hk) is assistant professor
of marketing at the Faculty of Business, Hong Kong Polytechnic
University, Hung Hom, Kowloon, Hong Kong; Frank May (fmay@vt.edu)
is assistant professor of marketing at the Pamplin College of Business,
Virginia Tech, Blacksburg, VA 24061; and Ashwani Monga
(amonga@business.rutgers.edu) is professor of marketing at Rutgers
Business School, Rutgers University, Newark and New Brunswick, NJ
07102. All authors contributed equally. Corresponding author: Rafay A.
Siddiqui.
Vicki Morwitz served as editor and Rebecca Hamilton served as associate
editor for this article.
Advance Access publication October 24, 2016
C The Author 2016. Published by Oxford University Press on behalf of Journal of Consumer Research, Inc.
V
All rights reserved. For permissions, please e-mail: journals.permissions@oup.com Vol. 43 2017
DOI: 10.1093/jcr/ucw064
932
SIDDIQUI, MAY, AND MONGA
quantity, but to consume virtues at a high rate and vices at
a low rate. Given our perspective of a consumption rate for
self-control (i.e., quantity/consumption time window),
prior research has focused on just the numerator (i.e., quantity). Five studies reveal how, keeping the numerator constant, the denominator (i.e., time window) has an
asymmetric effect on plans for virtue versus vice
consumption.
We also contribute to the literature on the psychology of
time periods. One stream of research has examined the effect of switching time periods, such as on aspirational behaviors, mental accounting, and employee motivation (Dai,
Milkman, and Riis 2014; Kaur, Kremer, and Mullainathan
2010; Soster, Monga, and Bearden 2010). Another stream
has explored the time period available to complete activities, such as redeeming coupons and mail-in rebates
(Inman and McAlister 1994; Krishna and Zhang 1999).
Because the cost of spending time looms large (Zauberman
and Lynch 2005), consumers procrastinate time-consuming
activities such as visiting museums, writing class reports,
and completing rebate forms and saving plans (Ariely and
Wertenbroch 2002; Gourville and Soman 2011;
O’Donoghue and Rabin 1999; Shu and Gneezy 2010;
Soman 1998). Complementing this research on timeconsuming activities, we show how the time windows
associated with products have an asymmetric influence on
virtues versus vices.
Our findings also have strong managerial and public policy implications because time windows are ubiquitous
across food products as well as other contexts, such as payper-view programs available for a few hours, and amusement park tickets valid for a weekend. In the general discussion, we will elaborate further on the implications for
both theory and practice.
SELF-CONTROL: PLANNING MORE
VIRTUE AND LESS VICE CONSUMPTION
People face a variety of self-control challenges. From
the perspective of multiple selves within a single individual, the planner self wants to maximize utility over the
long run, while the doer self wants to maximize utility for
just the immediate time period (Benabou and Pycia 2002;
Schelling 1984; Thaler and Shefrin 1981). Self-control involves the farsighted planner trying to control the myopic
doer’s actions. This dynamic, which can also be seen as a
desire-willpower conflict pitting short-term desires against
long-term interests (Hoch and Loewenstein 1991), is integral to virtues and vices.
Virtues offer immediate costs, but benefits are accrued
in the long run; one may dislike eating bland salads, but
the eventual outcome will be a boost to one’s health.
Conversely, vices offer immediate benefits, but costs are
incurred in the long run; one may enjoy eating delicious
933
desserts, but the eventual outcome will be added inches to
the waistline. Individuals understand that their long-term
interest is better served when they consume more virtues
and fewer vices (Read, Loewenstein, and Kalyanaraman
1999; Wertenbroch 1998). They want to embrace virtues
and their positive self-signals, and avoid vices and their
negative self-signals (Benabou and Tirole 2004; Bodner
and Prelec 2003; DellaVigna and Malmendier 2004; Dhar
and Wertenbroch 2012; Khan and Dhar 2006, 2007; Kivetz
and Zheng 2006). Despite these good intentions, individuals frequently lack the willpower to consume a distasteful
virtue, or deny themselves a tempting vice.
Knowing that willpower may give way in the face of desires, a planner often makes self-control plans so that the
future doer consumes more virtues and fewer vices
(Benabou and Tirole 2004; Thaler and Shefrin 1981).
Taking a virtue example, individuals may not have the
self-control to go ahead with a painful medical procedure
that provides long-term benefits. Therefore, they impose
penalties on themselves if they fail to undergo the procedure (Trope and Fishbach 2000). Taking a vice example, individuals may not have the self-control to avoid frivolous
expenditures. To restrict future spending, they proactively
resort to mental accounting—they restrict the inherent fungibility of money by creating a preset budget for categories
such as “entertainment” (Gourville and Soman 1998;
Heath and Soll 1996; Prelec and Loewenstein 1998; Thaler
1980, 1985). Also, to discourage future drug use, addicts
deter themselves by writing self-incriminating letters that
others may see in the event of a relapse (Schelling 1992).
Thus, individuals make plans that help them align future
actions with self-control goals.
ROLE OF QUANTITY IN PLANNING
VIRTUE AND VICE CONSUMPTION
The current research is focused on the domain of products. In this domain, individuals are known to prefer certain quantities, depending on whether the goal is to
promote virtue consumption or limit vice consumption. For
virtues, because there is a long-term benefit to increasing
the intake, consumers prefer large quantities. Mishra and
Mishra (2011) observe that for products perceived as virtues (e.g., raisins), individuals forgo a discount and favor a
bonus pack that offers a larger quantity. In the case of virtue magazines (e.g., Foreign Policy), readers pay more for
a subscription of all issues through the year, even though it
is cheaper to purchase single issues; that is, subscription
prices are higher than corresponding individual newsstand
rates (Oster and Morton 2005). And at gyms, it is often
cheaper to just pay for visits than have a monthly membership. However, individuals sign up for a membership to encourage themselves to visit the gym more often
(DellaVigna and Malmendier 2006).
934
For vices, because there is a long-term benefit to
decreasing the intake, consumers prefer small quantities.
Wertenbroch (1998) shows that smokers favor small packs
of cigarettes over price discounts on cartons. Thus, they
opt for the economically inferior option to restrict their cigarette quantity. Similarly, Mishra and Mishra (2011) observe that for tasty chocolates, which are perceived as
vices, individuals forgo a bonus pack and opt for a discounted regular pack.
We should clarify that doers may or may not comply
with planners. Considering virtues, one may buy more raisins but not eat them, receive Foreign Policy through the
year and never read an issue, and sign up for gym memberships but not take advantage of the unlimited gym visits.
Indeed, members do not visit the gym frequently enough to
even recoup their membership expense. To use the title of
DellaVigna and Malmendier’s (2006) article, buying gym
membership is akin to “paying not to go to the gym.”
Similarly, considering vices, buying small packs of cigarettes or tasty chocolates does not ensure reduced consumption because one is still free to consume the small packs
quickly and then buy more packs. Thus, it is not critical to
our theory whether or not doers consume as planners intend. What is critical is that planners lean toward consumption opportunities that entail higher quantities of virtues
and lower quantities of vices.
ROLE OF TIME WINDOW IN PLANNING
VIRTUE AND VICE CONSUMPTION
Our thesis is that just planning the quantity to be consumed does not offer a comprehensive plan to align future
consumption with self-control objectives. A large quantity
of virtue spread over a very long time would be ineffective,
just as a small quantity of vice compressed in a very short
time period would be harmful. A comprehensive selfcontrol plan requires one to control the rate at which future
consumption occurs—consume virtues at a high rate and
vices at a low rate. Considering gym visits as a virtue product, individuals would meet the goal of good health not
just by signing up for more visits, but by completing those
visits within a time window that yields a high rate of gym
visits. Similarly, considering cigarettes as a vice product,
individuals would meet the self-control goal of good health
not just by buying fewer cigarettes, but by smoking those
cigarettes within a time window that yields a low rate of
cigarette smoking. Given our perspective of a self-control
consumption rate (i.e., quantity/consumption time window), prior research has focused on just the numerator
(i.e., quantity). We argue that because the denominator
(i.e., time window) can alter the consumption rate, it would
also factor into individuals’ planning of virtue and vice
consumption.
JOURNAL OF CONSUMER RESEARCH
Consider the case of virtues. When quantity is held constant, a shorter time window promotes a higher consumption rate. Therefore, for an individual trying to exert selfcontrol over future consumption (i.e., aim for higher consumption rate), a shorter time window would be preferable.
As an example, think of a consumer who sees wheatgrass
juice as a virtue that offers an unpleasant taste in the short
run, but health benefits in the long run. When considering
a package of four drinks that last for a long time (e.g., twoweek expiration), the individual knows that because she
finds wheatgrass juice distasteful, she would stretch the
consumption over two weeks rather than consume quickly.
This low rate of consumption would be misaligned with
her self-control goal of a high rate of virtue consumption.
Now assume that the drinks last for a shorter time (e.g.,
one week). A one-week period would be adequate to consume all four drinks, but provide less consumption flexibility than a two-week period. Precisely because of this
restricted flexibility, a planner knows that stretching out
juice consumption will not be an option, and the consequent high consumption rate will yield a positive impact
on health. It is possible that the doer still procrastinates and
ends up never consuming the drinks. However, for the
planner, a one-week (vs. two-week) period offers the promise of a higher rate of consumption. Thus, a one-week
period is more attractive for someone genuinely aiming for
self-control of virtues (i.e., elevating future rate of consumption). Moreover, this shorter period also provides a
more favorable self-signal (Benabou and Tirole 2004;
Bodner and Prelec 2003; Dhar and Wertenbroch 2012).
Specifically, if self-control were not of any concern, one
would prefer a two-week (vs. one-week) period that provides an excuse to consume the distasteful drinks later rather than sooner. However, given self-control
considerations, a one-week period proves a more favorable
self-signal of aiming for a high rate of virtue consumption.
Overall, therefore, the consumer will be more inclined toward the drinks package if it were offered with a one-week
(vs. two-week) expiration period; that is, shorter time windows will increase preference for virtues.
Now consider the case of vices. When quantity is held
constant, a longer time window promotes a lower consumption rate. Therefore, for an individual trying to exert
self-control over future consumption (i.e., aim for lower
consumption rate), a longer time window would be preferable. As an example, think of a consumer who sees cupcakes as a vice that offers a great taste in the short run, but
adverse health consequences in the long run. When considering a package of four cupcakes that last for a short
time (e.g., one-week expiration), the consumer knows that
he would not let the tempting cupcakes go to waste and,
therefore, a high rate of consumption is assured. This high
rate of consumption would be misaligned with his selfcontrol goal of a low rate of vice consumption. Now assume that the consumer considers cupcakes that last for a
SIDDIQUI, MAY, AND MONGA
longer time (e.g., two weeks). As a planner, the consumer
knows that he will now be able to stretch out cupcake consumption over two weeks, and the consequent low consumption rate will limit the negative impact on health. It is
possible that the doer still consumes all cupcakes within
the first week. However, for the planner, a two-week
period provides a greater opportunity to consume at a
lower rate than what a one-week period would permit.
Thus, a two-week period is more attractive for someone
genuinely aiming for self-control of vices (i.e., curbing future rate of consumption). Moreover, this longer period
also provides a more favorable self-signal (Benabou and
Tirole 2004; Bodner and Prelec 2003; Dhar and
Wertenbroch 2012). Specifically, if self-control were not
of any concern, one would prefer a one-week (vs. twoweek) period that provides an excuse to consume the
tempting cupcakes sooner rather than later. However, given
self-control considerations, a two-week period provides a
more favorable self-signal of aiming for a low rate of vice
consumption. The literature on guilt mitigation and justification (Mishra and Mishra 2011) is also consistent with
this notion. While not consuming vice at all, in any quantity, may be ideal from a self-control perspective, a smaller
quantity is in the right direction because it promises to inflict less harm than a larger quantity. Analogously, a longer
time window offers the promise of less harm. Overall,
therefore, the consumer will be more inclined toward the
cupcakes package if it were offered with a two-week (vs.
one-week) expiration period; that is, longer time windows
will increase preference for vices.
OVERVIEW OF PREDICTIONS AND
STUDIES
Following the above discussion, we predict an asymmetric effect of time windows on virtue versus vice products.
A shorter time window should increase preference for virtues because the self-control goal is a high consumption
rate. However, a longer time window should increase preference for vices because the self-control goal is a low consumption rate. Taken together, a shorter time window
should nudge preference toward virtues and away from
vices. Five studies support our predictions.
In study 1, we find that choice of virtue rather than vice is
higher when the time window is shorter. In study 2, we replicate this effect in a choice between real products, where
we also track consumption. In study 3, we employ a preference, rather than choice, format. This format reveals not
only that shorter windows nudge people toward virtues and
away from vices, but also teases apart the virtue versus vice
effects; shorter time windows increase preference for virtues, whereas longer windows increase preference for vices.
In study 4, we test our theory by asking participants
whether they would invest nominal effort for a longer time
935
window, or just accept a short window. In line with our
other results, people opt for a longer window more in the
case of vices rather than virtues. We also show that, consistent with self-control, the pursuit of long-term goals mediates the influence of time windows. Specifically, time
windows influence perceptions of achieving good health
differently for virtue versus vice products, which then leads
to differences in the inclination to opt for a longer time
window.
In study 5, we test our theory via moderation. Our argument is that because people anticipate self-control challenges, they adopt plans that help them exert self-control.
However, not all individuals are aware of their self-control
issues, and there is individual variation in the extent to
which people view themselves as impulsive (O’Donoghue
and Rabin 1999; Puri 1996). Making plans to facilitate future self-control is more likely for those who need greater
help with self-control—those who are high on impulsivity.
For instance, Wertenbroch, Soman, and Nunes (2001) find
that high (vs. low) impulsives self-impose constraints on
their borrowing, so that the reduced liquidity deters them
from shopping. Similarly, in research more central to our
theorizing, Wertenbroch (1998) observes that high (vs. low)
impulsives are more likely to rely on quantity as a means of
self-control. Our argument for time windows follows a similar logic. A low-impulsive planner knows that the doer is
less likely to falter on self-control at the point of consumption, and so there is a lesser need to rely on time windows to
control consumption. Conversely, a high-impulsive planner
knows that the doer is more likely to falter on self-control,
and so there is a greater need to rely on time windows. In
line with this, study 5 shows that the effect of shorter time
windows shifting preference toward virtues and away from
vices is stronger for high (vs. low) impulsives.
We should note that the results that we observe for planners are consistent with self-control but do not necessarily
implicate genuine self-control as the sole underlying mechanism. Indeed, whether self-control is truly exerted or not
would be evident only when doers consume. At the planning stage, several mechanisms may be working in conjunction, such as a genuine desire for self-control, selfsignaling, and justification. Later, in the General
Discussion, we will elaborate on these finer processes and
offer some tentative suggestions to explore them further.
For now, in the studies that follow, we will limit our interpretations to how our results are broadly consistent with
self-control.
STUDY 1
In study 1, we test whether shorter time windows increase the choice share of virtues relative to vices.
Specifically, we present participants with a scenario of a
choice between a relative vice (i.e., dessert subscription)
936
and a relative virtue (i.e., salad subscription) while varying
the time window in which the vice or virtue is to be consumed. We keep the subscription quantity constant at eight
servings of either dessert or salad. Half the participants
choose between the two when the servings are spread over
a short time window (four weeks), and half make this
choice when the servings are spread over a long time window (four months). We predict that a shorter time window
will increase the choice share for a salad (vs. dessert)
subscription.
Two aspects of this study are worth highlighting. First,
we spread out consumption over fixed time intervals. In
the short-time-window condition (four weeks), we stipulate
two servings every week, and in the long-time-window
condition (four months), we stipulate two servings every
month. Note that time intervals are not manipulated: they
are present in both conditions, and the number of intervals
is four for both. Hence, any choice effects that we observe
cannot be attributed to the intervals, but only to the time
window. The only reason to include intervals is to convey
to participants that the time window would be strictly followed. When the salad or dessert arrives at fixed intervals,
one would be bound to a rate of consumption. Thus, a participant choosing dessert in a specific time-window condition would do so only if he or she genuinely accepts that
rate of dessert consumption. Therefore, the current setting
offers a strict test of our theorizing. However, in later studies, we also examine whether our results replicate when
this constraint of intervals is relaxed.
Second, although the use of dessert as vice and salad as
virtue is common in prior research (Kivetz and Keinan
2006; Kivetz and Zheng 2006), we used pretests to confirm
two aspects: (a) the two products differ in terms of vicevirtue perceptions, and (b) the two products are consumed
similarly in participants’ regular lives. For the first test, 59
participants from the Mechanical Turk online panel (41%
female, Mage ¼ 33) responded to a question adapted from
Khan and Dhar (2007). Specifically, we informed the participants, “A virtue is something that is not very tempting
now but may be more beneficial later on. A vice is something tempting that may have fewer benefits later on.”
Participants then rated either eating a salad or a dessert
on a scale from 0 to 100 (0 ¼ vice; 100 ¼ virtue). (The
numerical values on the slider scale were not visible to
participants.) As expected, in a relative sense, a salad
was more virtue and a dessert was more vice (MSalad ¼
79.86, SD ¼ 17.53, MDessert ¼ 14.60, SD ¼ 15.64; F(1, 57) ¼
227.98, p < .001).
To test whether participants consume the two products
similarly in their regular lives, we told a separate group of
59 undergraduate students from the University of South
Carolina (68% female, Mage ¼ 21) the following: “We are
interested in learning about how much you consume certain types of food. Think about the number of times you
consume desserts and salads in a typical month. Using the
JOURNAL OF CONSUMER RESEARCH
slider scale below, please indicate whether you, on average, eat salads more times a month than desserts, desserts
more times a month than salads, or you eat them an equal
number of times.” Participants provided their answers on a
scale from 0 to 100 (0 ¼ Eat desserts more times a month;
50 ¼ Eat salads and desserts an equal number of times a
month; 100 ¼ Eat salads more times a month). The mean
response did not deviate significantly from the midpoint of
50, confirming that there was no significant difference between the rate of consumption of salads and desserts (M ¼
55.85, SD ¼ 33.10, t(58) ¼ 1.36, p > .1). In sum, both
salad and dessert are consumed similarly by participants,
but salad (vs. dessert) is viewed as more of a virtue rather
than vice product.
Method
Two-hundred twenty-two undergraduate students (46%
female, Mage ¼ 21) at the University of South Carolina successfully participated in the study. (Eight responses were
excluded because participants incorrectly answered an
attention-filter question: “This question is just to make sure
that you are paying attention to this survey. Please mark
‘5’ as your response.”) A single factor (time window: short
vs. long) between-subjects design was employed.
Participants in the short time window (i.e., four weeks)
condition were asked to imagine the following scenario:
Imagine that a “Food of the Week” club is offering a promotion on two of its offerings: desserts and salads. 4-week subscriptions to both desserts and salads are on sale for $20
each.
The dessert subscription covers a variety of great tasting
desserts. If you buy the 4-week subscription for desserts,
you will receive two servings of a unique dessert every
week for 4 weeks. The salad subscription covers a variety of
healthy salads. If you buy the 4-week subscription for salads, you will receive two servings of a unique salad every
week for 4 weeks. Which one would you buy?
In the long-time-window condition (i.e., four months),
the scenario was similar to the one above, and the number
of servings was identical. However, the eight servings of
“Food of the Week” (i.e., two servings four weeks) were
presented as eight servings of “Food of the Month” (i.e.,
two servings four months). Participants’ choice of subscription served as the dependent variable. Finally, participants answered standard demographic questions.
Results
To test whether the number of people choosing the dessert subscription over the salad subscription varied as a
function of the time window, we conducted a binary logistic regression with the time window serving as the independent variable (Short ¼ –1, Long ¼ 1) and the choice
between the two subscriptions serving as the dependent
SIDDIQUI, MAY, AND MONGA
variable (0 ¼ Salad subscription, 1 ¼ Dessert subscription). We found that the salad (vs. dessert) subscription
was chosen significantly more often when the time window
was four weeks rather than four months (50.9% vs. 37.5%;
b ¼ .27, SE ¼ .136, z ¼ 2.01, 95% CI (.003, .54), p < .05,
R2 ¼ .02).
Discussion
Consistent with our predictions, we found that a shorter
time window shifted preference away from the vice (i.e.,
dessert) and toward the virtue (i.e., salad). Thus, more virtuous consumption results when we increase the rate of
consumption by shrinking the denominator (i.e., time window), while keeping the numerator (i.e., quantity of servings) constant.
STUDY 2
Study 2 builds upon study 1 in two ways. First, we test
whether our effects emerge for choices between real products as opposed to hypothetical ones. Instead of using different products (salad vs. dessert), we rely on a single
product category, granola bars, but employ a decadent
chocolate version as a vice, and a nutritional fruit version
as a virtue. In a pretest with 33 undergraduate students
from Virginia Tech using the scale described in study 1,
we confirmed that, in a relative sense, fruit granola bars are
seen as a virtue, and chocolate granola bars as a vice (Mfruit
¼ 59.82, SD ¼ 32.33, Mchocolate ¼ 31.06, SD ¼ 30.71, F(1,
31) ¼ 6.85, p < .05).
Second, we remove the constraint of fixed time intervals; that is, we give people an overall time window (two
days vs. seven days) and do not specify how they should
spread out the granola bars that they choose. We predict
that our study 1 results will replicate, such that the choice
share of virtue will be higher when the time window is two
(vs. seven) days.
One possibility here is that participants may choose
more vice in the seven-day condition (as we predict), but
only as an excuse to choose the vice, without having any
actual self-control intention to spread out consumption.
Therefore, we track consumption to get some initial empirical insights. We cannot make any a priori prediction about
whether the doer will comply with the planner (i.e., consume as stipulated by the time window), but want to empirically observe whether compliance for virtues versus vices
is similar. If we observe that vices are consumed faster
than virtues, then two possibilities exist: either self-control
intentions were honest but then faltered at the time of consumption; or self-control intentions were never honest in
the first place—vice choosers planned faster consumption
all along. In contrast, if we observe that vices are consumed at the same pace as virtues, then there is only one
possibility: self-control intentions were genuine for vices
937
as much as for virtues. Note that we are not making a theoretical prediction because the actual pace of consumption
would depend on several factors, such as how tempting a
product is. We merely contend that for our particular stimuli, if virtue and vice choosers consume similarly, then
there is no reason to suspect that vice choosers planned to
ignore the stipulation of time windows, while virtue choosers did not. We can infer that all choosers were honestly
aiming for self-control. While this would be only an empirical inference, we will later explore self-control more theoretically, using mediation and moderation.
Method
Eighty-six undergraduate students (52% female, Mage ¼
21) at Virginia Tech participated in the study for extra
course credit. Under the guise of a study to understand
their opinions of granola bars, students were asked to pick
the type they wanted, and randomly assigned to either the
short-window or long-window condition. Given the
Monday-Wednesday class schedule, choice data was collected on Monday and consumption data on Wednesday.
To ensure that students did not become aware of the different time-window conditions, they were asked not to converse with each other regarding the study, as that would
influence opinions. (Probes at the end of the study confirmed that students were unaware of assignments to
conditions.)
On a Monday, the students were provided the cover
story about understanding opinions of granola bars. They
then read the following: “We would like to give you some
food products today (as soon as you are done with this survey). In later classes, we will ask you for your opinions
about these products. You can choose which products you
want to give opinions on: 2 decadent Nature Valley chocolate granola bars or 2 nutritional Nature Valley fruit granola bars.” Right after this, students were told the time
window within which they should consume the two bars
(i.e., two days or seven days). Participants then chose the
bars they wanted. (They could choose to opt out and not receive any bars; no one did that.)
We then verified that, in their regular lives, participants
consumed chocolate and fruit granola bars at a comparable
level. Specifically, when asked to indicate the extent to
which they consume fruit relative to chocolate bars in a
week (1 ¼ Eat fruit bars more times, 4 ¼ Eat both equally,
7 ¼ Eat chocolate bars more times), the mean response
was not significantly different from the midpoint of 4 (M
¼ 3.91, SD ¼ 1.57, t(85) ¼ –.55, p > .5). Next, participants
completed standard demographic questions.
The follow-up survey was conducted two days later, on
Wednesday, to assess the number of bars that participants
had consumed. Note that this two-day mark was the critical
point for us to assess differences between participants who
were in the two-day versus seven-day conditions (i.e., by
938
the seven-day mark, both groups were expected to consume all the bars anyway). Once participants provided their
responses, they were debriefed and the study was
terminated.
Results
Choice. We conducted a binary logistic regression with
time window serving as the independent variable (Short ¼
–1, Long ¼ 1) and choice as the dependent variable (0 ¼
Fruit granola bars, 1 ¼ Chocolate granola bars). The analysis revealed that more participants chose the fruit bars
when the time window was short (two days) rather than
long (seven days), indicating that, as predicted, the shorter
window encouraged choice of virtue over vice (% choosing
fruit: Short ¼ 67%, Long ¼ 42%; b ¼ .53, SE ¼ .22, z ¼
2.35, 95% CI (.09, .96), p < .05, R2 ¼ .09).
Consumption. Out of the 86 participants who made
choices on Monday, 81 were available for the follow-up
study at the two-day mark on Wednesday. (For the five
who were not available for the follow-up, drop-out numbers were very similar across different groups; one participant each for fruit–two-day, fruit–seven-day, and
chocolate–seven day; and two participants for chocolate–
two day.)
The results revealed that the time window influenced the
rate of consumption. More bars were consumed in the twoday (vs. seven-day) condition (Mshort ¼ 1.80, SD ¼ .79,
Mlong ¼ 1.29, SD ¼ .78, F(1, 79) ¼ 8.42, p < .01; Cohen’s
d ¼ .65). More important, the possibility that participants
may choose chocolate bars in the seven-day condition but
consume them right away (but not do so for fruit bars) does
not hold. In the seven-day condition, whether participants
had chosen chocolate or fruit bars, the mean consumption
was identical (Mchocolate ¼ 1.29, SD ¼ .81, Mfruit ¼ 1.29,
SD ¼ .77, F(1, 39) ¼ .00, p > .9; Cohen’s d ¼ 0). Thus, it
seems unlikely that participants used the seven-day window as simply an excuse to pick the vice, only to consume
all bars quickly. Similarly, in the two-day condition, it did
not matter whether participants had chosen chocolate or
fruit bars (Mchocolate ¼ 1.67, SD ¼ .65, Mfruit ¼ 1.86, SD ¼
.85, F(1, 38) ¼ .48, p > .4; Cohen’s d ¼ .25).
Discussion
Consistent with the results of study 1, a shorter time window encouraged choice of virtue over vice. We observed
this for real choices in which actual consumption was
anticipated. Moreover, participants complied with the time
window, in that the consumption rate was higher in the
two-day (vs. seven-day) condition, with no difference between virtue and vice participants. We cannot interpret
these results as evidence that consumption plans will always carry over to actual consumption. After all, our theory is only about planners, not doers, and these empirical
JOURNAL OF CONSUMER RESEARCH
results may change from one product to another. However,
these results do help us draw an inference. Given that virtue and vice doers consume the products similarly, vice
planners could not have thought of long time windows as
just an excuse to choose vice, while actually intending to
consume early. If they had thought so, then their later
selves, vice doers, would have consumed faster than virtue
doers. Thus, we can make an empirical (though not theoretical) inference that doers were indeed aiming for selfcontrol, and not using the time window as just an excuse to
cater to their impulses. We provide more evidence in the
next three studies, which is consistent with self-control. In
study 3, we rely on preference (rather than choice), which
lets us clearly observe the asymmetry—the pattern that
emerges for vices is opposite to the pattern for virtues. In
study 4, consistent with self-control, we show mediation
via pursuit of long-term goals. Finally, in study 5, we show
moderation via impulsivity, such that our results are stronger for those who need more help with self-control.
STUDY 3
Studies 1 and 2 provide evidence for the effect of shorter
time windows leading to a shift toward virtues rather than
vices. However, given the choice format, these studies do
not provide evidence for the predicted separate effects of
vice versus virtue—that a shorter time window elevates
preference for virtues, whereas a longer time window elevates preference for vices. We test these separate effects in
the current study.
Another change from our earlier studies pertains to the
stimuli. In study 1, we employed different product categories (i.e., salad vs. dessert). In study 2, we employed different variants of the same product category (i.e., chocolate
vs. fruit granola bars). In the current study, we employ a
single product, red wine, and simply measure vice-virtue
perceptions. Some people think of red wine as a vice; they
may consider it a tempting product that helps them unwind,
but is short on future benefits and may lead to unpleasant
consequences. Others extoll its virtues; they may consider
it a product that has unwanted immediate consequences,
but delivers future benefits such as improved heart health
due to antioxidants. Depending on how people view red
wine, we expect them to react differently to time windows.
Our setting involves a scenario in which perception of
wine coupons serves as a proxy for wine preference. We
predict that a shorter time window will decrease preference
for those who see red wine as a relative vice, but increase
preference for those who see red wine as a relative virtue.
Because vice-virtue perceptions are measured rather than
manipulated, we cannot conclusively establish that any
observed effects are caused by such perceptions, rather
than other correlated variables. For instance, virtue perception may be correlated with a high consumption rate that
SIDDIQUI, MAY, AND MONGA
makes one prefer the short (vs. long) time window. Or the
cause may be some other unforeseen correlates.
Establishing causality requires an experimental vice-virtue
manipulation of the type we employ in the two preceding
studies and the two that follow. In this study, our goal is
simply to capitalize on natural variation in vice-virtue perceptions of a product, and test if such perceptions are associated with different consequences for short versus long
time windows.
Method
We used the Mechanical Turk online panel to solicit
adult participants who self-report as wine drinkers. One
hundred twenty of them (35% female, Mage ¼ 33) successfully participated in the study. (Four responses were
excluded because participants incorrectly answered the
attention-filter question, as described in study 1.) A 2 (time
window: short vs. long) 2 (product: virtue vs. vice;
measured) between-subjects design was employed.
Participants in the short-time-window condition were
asked to imagine the following scenario:
Imagine that an upscale wine store has recently opened
within walking distance of your house. The store specializes
in red wine. It is offering a free coupon as part of a promotional campaign. With the coupon, you can come into the
wine store and have 1 free glass of wine every day for 7
days. To get the coupon, you must first go online to the
wine store’s website and take a 1 hour survey on beverages.
After the survey is completed, you will be emailed a unique
printable coupon in your name. Since it would be in your
name, only you can use the coupon.
In the long-time-window condition, the scenario was
similar to the above, and the number of glasses was identical. However, the seven glasses in a week (i.e., one glass
seven days) were presented as seven glasses in seven
weeks (i.e., one glass seven weeks); that is, the coupon’s
recipient would receive one glass of wine every week for
seven weeks.
Afterward, participants indicated preference for the coupon by answering the following two questions on 100-point
scale in the short-time-window condition [long-time-window condition]: a) How likely would you be to take the
survey for a coupon which gives you 1 glass of wine every
day for 7 days [every week for 7 weeks]? (0 ¼ Very unlikely, 100 ¼ Very likely), and b) How appealing is the
coupon, which gives you 1 glass of wine every day for 7
days [every week for 7 weeks]? (0 ¼ Not at all appealing,
100 ¼ Very appealing). Indicating higher numbers would
suggest a higher preference for consuming the wine. Next,
to measure the extent to which a participant perceives wine
as a virtue versus vice, we utilized the question that we
used in the pretests of studies 1 and 2, which involved
describing a vice as something tempting that may have
939
fewer benefits later on, and a virtue as being not tempting
but offering later benefits (0 ¼ vice; 100 ¼ virtue). For all
these slider scales, the numerical values were not visible to
participants. Finally, participants answered standard demographic questions.
Results
Our goal was to test whether preference for the wine
(i.e., average of the two coupon measures; a ¼ .88)
changed differently with time window, depending on vicevirtue perceptions about wine. Therefore, using the dependent variable of preference, we investigated the twoway interaction between time window (Short ¼ –1, Long
¼ 1) and perceived type of good (virtue vs. vice; continuous) using the Johnson-Neyman floodlight analysis technique (Johnson and Neyman 1936; Spiller et al. 2013). The
two-way interaction was significant (b ¼ –.33, SE ¼ .13,
t(119) ¼ –2.54, 95% CI (–.58, –.07), p ¼ .01, R2 ¼ .06), and
the predicted pattern emerged. As is clear from figure 1,
when the time window became shorter (i.e., one week instead of seven weeks), there was a preference shift toward
wine seen as virtue, and away from wine seen as vice.
The preference shift was driven by statistically significant shifts for both virtue and vice. For those who perceive
wine as a relative virtue (who were 1.26 SD above the
mean—or higher—of the vice-virtue index), preference
was higher when the time window was one week rather
than seven weeks (bJN ¼ –9.69, SEJN ¼ 4.89, t(119) ¼ –
1.98, 95% CIJN (–19.37, .00), p ¼ .05). In contrast, for
those who perceive wine as a relative vice (who were 1.24
SD below the mean—or lower—of the vice-virtue index),
preference was higher when the time window was seven
weeks rather than one week (BJN ¼ 9.52, SEJN ¼ 4.81,
t(119) ¼ 1.98, 95% CIJN (.00, 19.04), p ¼ .05).
A similar preference shift was evident for preferences
within each time window. When the time window was
long (seven weeks), the slope of vice versus virtue perceptions was not significant; that is, the preference for wine
did not change when wine was perceived to be more of a
virtue than a vice (b ¼ –.26, SE ¼ .184, t(119)¼ –1.42,
95% CI (–.63, .10), p> .1). However, when the time window was short (one week), the slope of vice versus virtue
perceptions was positive, such that preference increased
when wine was perceived to be more of a virtue (b ¼ .39,
SE ¼ .179, t(119)¼ 2.18, 95% CI (.04, .75), p< .05). In
other words, while preference was similar for vice versus
virtue in the seven-week condition, the preference for virtue was higher in the one-week condition. The shorter window shifted preference toward wine seen as a virtue.
Discussion
In the current study, we observed the same effect as we
did in studies 1 and 2, but for preference rather than choice.
940
JOURNAL OF CONSUMER RESEARCH
FIGURE 1
STUDY 3: PREFERENCE FOR WINE COUPON WHEN WINE IS PERCEIVED TO BE A VIRTUE VERSUS VICE (MEANS AT –/þ 1SD)
100
Preference for W ine Coupon
90
80
70
70.67
67.66
55.3
60
52.29
1 week
50
7 weeks
40
30
20
10
0
Virtue
Examining preference helps us delineate how time windows have an effect for both virtues and vices, but in opposite directions. For those who perceive wine as a relative
virtue, preference was higher when the time window was
shorter; but for those who perceive wine as a relative vice,
preference was higher when the time window was longer.
Furthermore, we observed these results not for different
virtue and vice products, but for a single product. While
we forgo causal evidence by not manipulating virtue-vice
perceptions (as we manipulate in our other studies), this
study is helpful in showing that even naturally occurring
variance in virtue-vice perceptions yields an asymmetric
influence of time windows.
STUDY 4
Having observed the predicted effects for both choice
and preference, in the current study we add to the empirical
evidence in three ways. First, in our previous studies, we
tested how participants responded to different time windows. We now want to test whether people would actively
opt for a longer time window over a shorter one. However,
because the long time window offers greater consumption
flexibility, socially desirable responding (Holtgraves 2004)
may arise; all participants may opt for it in order to appear
more rational. To get around this issue, we ask participants
whether they would like to extend the time window in return for a nominal amount of effort. Thus, extending the
time window is portrayed as a benefit in return for a small
cost. Our prediction is that those in the vice (vs. virtue)
condition are more likely to extend the time window. This
is because a longer time window would facilitate the goal
Vice
of decreasing the vice rate, but hinder the goal of increasing the virtue rate.
Second, we further test the extent to which our observed
effects are consistent with self-control. In the previous
studies, we show an asymmetry for vices and virtues,
which is consistent with our ideas about time windows aiding individuals’ self-control efforts in the service of their
long-term goals. We now test this implication more directly via mediation. Specifically, we examine whether time
windows influence perceptions of achieving good health
differently for virtue versus vice products, which in turn
leads to differences in the inclination to opt for a longer
time window.
Finally, instead of using different products (studies 1
and 2), or measuring virtue-vice perceptions (study 3), we
manipulate virtue-vice perceptions of a single product,
shakes, to further establish causality. We pretested this manipulation. Following the measure used in prior studies, we
describe a vice as something tempting that may not offer
many benefits, and a virtue as being not tempting but offering benefits later on (0 ¼ vice; 100 ¼ virtue). Participants
read scenarios in which the shake was described either as a
vice or a virtue (see the Method section below for how the
vice-virtue manipulation was incorporated). The pretest
with 60 Mechanical Turk participants (48% female, Mage
¼ 35) confirmed that, in a relative sense, the shake
described in the virtue condition was seen as a virtue, and
the shake described in the vice condition was seen as a
vice (MVirtue ¼ 76.79, SD ¼ 29.27, MVice ¼ 11.97,
SD ¼ 14.49; F(1, 58) ¼ 120.56, p < .001). Having confirmed this vice-virtue difference, we proceeded to the
main study.
SIDDIQUI, MAY, AND MONGA
Method
One hundred thirty participants from the Mechanical Turk
online panel (41% female, Mage ¼ 34) successfully participated in the study. (Five responses were excluded because
participants incorrectly answered the attention-filter question,
as described in study 1.) A single factor (product: virtue vs.
vice) between-subjects design was employed.
All participants were told about a gift certificate for
eight shakes with a validity period of one month, which
they could extend to four months. Those in the virtue condition read the following:
Please imagine that you have received a gift certificate in
the mail. This certificate is good for 8 healthy shakes at a
new branch of the restaurant “Bob’s Shakes.” This new
branch is located just a block away from your home.
A quick look at Yelp confirms that Bob’s Shakes are indeed
very healthy. For instance, one comment is as follows: “. . .
Bob’s Shakes aren’t the tastiest, but they sure are good for
you! I’ve been drinking them for a while now. The taste
sucks, but I look better in a swimsuit now! That healthy
feeling is definitely worth it.”
As you look closer at the certificate, you notice the fine
print. The fine print lists many things such as how the certificate cannot be shared with others, and how you can redeem only 1 shake on any given day. However, what really
catches your attention is the following:
“The certificate is valid for 1 month. That is, all 8 healthy
shakes must be consumed within 1 month. However, you can
increase the validity of the certificate to 4 months by taking the
unused certificate to the regional office of Bob’s Shakes.”
In the vice condition, the scenario was similar, but the
eight healthy shakes were described as eight tasty shakes.
Also, the Yelp comment was described differently:
A quick look at Yelp confirms that Bob’s Shakes are indeed
very tasty. For instance, one comment is as follows: “. . .
Bob’s Shakes are the tastiest, but they sure aren’t good for
you! I’ve been drinking them for a while now. The taste
rocks, but I look worse in a swimsuit now! That delicious
taste is definitely worth it.”
All participants were then informed that the regional office,
where they can extend the validity, is less than a mile away
from their home. They indicated whether they would make
the trip to extend the validity of the gift certificate (yes/no),
and then answered the following question: “To what extent
would you better achieve your long-term health goals if you
were able to extend the certificate validity period (to 4 months
rather than 1 month)?” (0 ¼ Not at all, 100 ¼ A lot). Finally,
participants answered standard demographic questions.
Results
Choice. To test whether participants chose to increase
the time window differently depending on the virtue-vice
941
manipulation, we conducted a binary logistic regression
with the type of product serving as the independent variable (Virtue ¼ –1, Vice ¼ 1) and the option of increasing
the time window as the dependent variable (0 ¼ No, 1 ¼
Yes). More participants chose to increase the time window
when the product was a vice rather than a virtue (76% vs.
56%; b ¼ .47, SE ¼ .19, z ¼ 2.45, 95% CI (.10, .84), p <
.05, R2 ¼ .06); that is, when we portrayed the opportunity
of a longer time window as a benefit in return for a small
cost of effort, those in the vice (vs. virtue) condition were
more willing to avail themselves of this opportunity.
Health Goal Perceptions. We examined whether participants in the virtue versus vice conditions vary in how they
perceive a longer time window as helping them achieve
their health goals. Specifically, we conducted an ANOVA
with product type serving as the independent variable and
participants’ perceptions of health goal achievement as the
dependent variable. As predicted, participants perceived
that they would better achieve their health goals by extending the time window in the vice condition than in the virtue
condition (MVice ¼ 66.42, SD ¼ 25.68, MVirtue ¼ 54.41,
SD ¼ 33.86, F(1, 128) ¼ 5.23, p < .05, Cohen’s d ¼ .40).
In other words, the longer time window appeared more
beneficial to vice participants than it did to virtue participants. This, as we discuss next, led to a higher tendency to
opt for the longer time window.
Mediation. We ran a mediational analysis using the
PROCESS SPSS macro (Model 4; Hayes 2013; Preacher
and Hayes 2004). In the regression model, the dependent
variable was choice, while the independent variables were
product type and greater perceived achievement of health
goals. The effect of the mediator, greater perceived
achievement of health goals, was significant (b ¼ .03, SE
¼ .007, z ¼ 4.51, 95% CI (.02, .05), p < .01), while that of
product type was not (b ¼ .33, SE ¼ .21, z ¼ 1.54, 95% CI
(–.09, .75), p > .1). Most importantly, a bootstrap analysis
confirmed a significant indirect effect (b ¼ .20, SE ¼ .11,
95% CI (.03 to .46)). Collectively, these results provide
evidence that perceived achievement of health goals mediates the effect of product type on tendency to increase the
time window of consumption.
Discussion
While the first three studies examined choice or preference when consumers are given a specific time window,
the current study focused on whether participants are willing to invest effort in order to extend a time window. We
observe that they are more willing to do so for a vice (vs.
virtue), thus providing further support for our core effect.
We also show that, consistent with self-control, people act
in the service of a health goal. Specifically, vices (vs. virtues) make participants perceive that a longer time window
would better help them achieve health goals, which makes
them opt for a longer time window.
942
JOURNAL OF CONSUMER RESEARCH
STUDY 5
The empirical evidence from the first four studies documents that shorter time windows lead to a relative shift toward virtues rather than vices. Our results are consistent
with self-control in that individuals try to increase the consumption of virtues and decrease consumption of vices. In
this final study, we provide additional support via moderation. As described earlier, just as high impulsives are
more prone to relying on quantity in service of their selfcontrol goals (Wertenbroch 1998; Wertenbroch et al.
2001), they should also be more prone to relying on time
windows. Thus, we predict that the shift in preference that
we observed in our earlier studies would be driven more by
those who are high (vs. low) impulsives. The scenario that
we employed was broadly similar to the one on shakes,
which we used in the previous study.
Method
Two hundred twenty-eight participants from the
Mechanical Turk online panel (34% female, Mage ¼ 32)
successfully participated in the study. (Nine responses
were excluded because participants incorrectly answered
the attention-filter question, as described in study 1.) A 2
(product: virtue vs. vice) 2 (time window: short vs. long)
2 (impulsivity: low vs. high; measured) betweensubjects design was employed.
The scenario of Bob’s Shakes was similar to that of
study 4. The key difference was that the option to extend
the time window for eight shakes was not available.
Instead, the time window was simply manipulated to be either one month or four months. Participants were informed
that they would have to drive three miles to the corporate
office to obtain the free gift certificate, and asked the following question: “How likely are you to make the drive?”
They provided their responses on a 0–100 scale with 0 ¼
very unlikely and 100 ¼ very likely. (The numerical values
on the slider scale were not visible to participants.) This
measure served as a proxy for individuals’ preference to
consume the shake. Next, participants completed Puri’s
(1996) impulsivity scale (a ¼ .92), which assesses the extent to which certain adjectives (impulsive, self-controlled,
etc.) describe an individual. Finally, participants answered
standard demographic questions.
Results
Our goal was to test whether preference for the shakes
(i.e., likelihood to drive to get the gift certificate) changed
differently with time window, depending on vice-virtue
condition and individuals’ impulsivity. Therefore, using
the dependent variable of preference, we conducted a regression with type of good (Virtue ¼ –1, Vice ¼ 1), time
window (Short ¼ –1, Long ¼ 1), and impulsivity as the
independent variables. A significant two-way interaction
emerged between type of good and time window (b ¼ 7.16,
SE ¼ 2.36, t(227) ¼ 3.04, 95% CI (2.51, 11.80), p < .01)
and there was a significant three-way interaction between
type of good, time window, and impulsivity (b ¼ 3.50,
SE ¼ 1.77, t(227) ¼ 1.97, 95% CI (.01, 7.00), p < .05,
R2 ¼ .07).
Before exploring the three-way interaction, let us explore the overall two-way interaction, which was as we
predicted. Specifically, as is evident from the pattern of results (see figure 2), when the time window was shorter
(i.e., one month instead of four months), preference shifted
toward virtue, and away from vice. This preference shift
was driven by statistically significant shifts for both virtue
and vice. For those who view the shake as a relative virtue,
preference was higher when the time window was shorter
(M1month ¼ 64.15, SD ¼ 33.08 vs. M4months ¼ 49.52,
SD ¼ 37.34; F(1, 224) ¼ 4.60, p < .05, Cohen’s d ¼ .41).
In contrast, for those who view the shake as a relative vice,
preference was higher when the time window was longer
(M1month ¼ 56.44. SD ¼ 38.20 vs. M4months ¼ 69.95,
SD ¼ 31.87; F(1, 224) ¼ 4.26, p < .05, d ¼ .38).
A similar preference shift was evident when we compared preference within each time window. When the time
window was long (four months), preference for the vice
significantly exceeded that of the virtue (F(1, 224) ¼ 9.67,
p< .01, Cohen’s d ¼ .59). However, when the time window was short (one month), preference was immune to
whether the shake was viewed as a virtue or vice (F(1,
224) ¼ 1.28, p> .2, Cohen’s d ¼ .22). Thus, although vice
was preferred more than virtue in the long-time-window
condition, this difference dissipated in the short-time-window condition because, as stated above, preference for virtue increased while that of vice decreased.
Next, to investigate the previously reported significant
three-way interaction, we used the Johnson-Neyman floodlight analysis technique (Johnson and Neyman 1936,
Spiller et al. 2013). The two-way interaction between type
of good and time window was significant only for those
who were relatively impulsive—who were at, or higher
than, .44 SD below the mean on the impulsivity scale (bJN
¼5.08, SEJN ¼ 2.58, t(227) ¼ 1.97, 95% CIJN (.00, 10.17),
p ¼ .05). To further explore our results, we examined the
two-way interaction at 1 SD below and above the mean of
the impulsivity scale (see figure 3). At 1 SD below the
mean of impulsivity (i.e., low impulsives), the two-way
interaction was not significant (b ¼ 2.45, SE ¼ 3.35, t(227)
¼ .73, 95% CI (–4.15, 9.06), p > .4; see figure 3, panel A).
At 1 SD above the mean of impulsivity (i.e., high impulsives), the two-way interaction was significant (b ¼ 11.86,
SE ¼ 3.35, t(227) ¼ 3.54, 95% CI (5.26, 18.47), p < .001;
see figure 3, panel B).
In line with these results showing that the significant
interaction for high impulsives is the driving force, we see
that the two-way pattern for high impulsives (figure 3, panel
SIDDIQUI, MAY, AND MONGA
943
FIGURE 2
STUDY 5: PREFERENCE FOR SHAKE GIFT CERTIFICATE WHEN SHAKE IS MANIPULATED TO BE A VIRTUE VERSUS VICE
(OVERALL TWO-WAY PATTERN)
Preference for Shake Gift Certificate
100
90
80
70
60
50
69.95
64.15
56.44
49.52
1 month
4 months
40
30
20
10
0
Virtue
B) is a stronger version of the overall two-way pattern (figure 2). We further explored this significant interaction for
high impulsives. For high impulsives who view the shake as
a relative virtue, preference was higher when the time window was one month rather than four months (b ¼ –11.58,
SE ¼ 4.89, t(227) ¼ –2.37, 95% CI (–21.21, –1.95), p <
.05). But for high impulsives who view the shake as a relative vice, preference was higher when the time window was
four months rather than one month (b ¼ 12.15, SE ¼ 4.59,
t(227) ¼ 2.64, 95% CI (3.09, 21.20), p< .05). A similar
preference shift was evident when we compared preference
of virtue to that of vice. When the time window was long
(four months), preference for the vice significantly exceeded
that of the virtue (b ¼ 12.37, SE ¼ 4.48, t(227) ¼ 2.76, 95%
CI (3.54, 21.21), p< .01). But when the time window was
short (one month), preference for the virtue significantly exceeded that of the vice (b ¼ –11.35, SE ¼ 4.99, t(227)¼ –
2.28, 95% CI (–21.18, –1.52), p< .05). Thus, higher preference for vice in the four-month condition reversed to higher
preference for virtue in the one-month condition because, as
explained above, preference for virtue increased but that of
vice decreased.
Discussion
Our earlier effects replicated. For those seeing shakes as
a relative virtue, preference was higher when the time window was shorter. However, for those seeing shakes as a
relative vice, preference was higher when the time window
was longer. Moreover, consistent with self-control, the effect of time windows was stronger for those who need
Vice
more help in meeting their self-control goals—that is, high,
rather than low, impulsives.
GENERAL DISCUSSION
Consumers often face a consumption time window, a time
limit within which to consume multiple units of a good. We
argue that time windows have consequences for consumers’
self-control goals of managing virtue and vice consumption
in service of their long-term interests. Our key premise is
that consumers are focused on controlling not just product
quantity in isolation, as previous research has shown, but
rather the rate of consumption (i.e., quantity/time window).
As the denominator shrinks (i.e., time window becomes
shorter), virtues become more attractive because the selfcontrol goal is to elevate the consumption rate of virtues.
Conversely, as the denominator increases (i.e., time window
becomes longer), vices become more attractive because the
self-control goal is to restrict the consumption rate of vices.
Consequently, shorter time windows nudge people toward
virtue products, and away from vice products. This shift is
robust. It emerges in all five studies irrespective of whether
the dependent measure is choice or preference, whether consumption is spread out over fixed intervals or not, whether
participants are students or adult consumers, whether a product is mentioned as being free or at a price, and whether the
choice is real or hypothetical. Additionally, this effect arises
regardless of how virtue-vice differences are operationalized:
by relying on different products that differ on virtue-vice perceptions, measuring such perceptions for a single product, or
manipulating such perceptions for a single product.
944
JOURNAL OF CONSUMER RESEARCH
FIGURE 3
STUDY 5: PREFERENCE FOR SHAKE GIFT CERTIFICATE WHEN SHAKE IS MANIPULATED TO BE A VIRTUE VERSUS VICE
(PATTERNS FOR LOW VERSUS HIGH IMPULSIVES)
Panel A: Low Impulsives (Means at -1 SD of impulsivity)
Preference for Shake Gift Certificate
100
90
80
70
60
62.15
60.14
54.97
47.17
50
1 month
4 months
40
30
20
10
0
Virtue
Vice
Panel B: High Impulsives (Means at +1 SD of impulsivity)
Preference for Shake Gift Certificate
100
90
80
76.36
74.77
70
60
51.61
52.07
1 month
50
4 months
40
30
20
10
0
Virtue
As we discussed in our theorizing, several mechanisms
may be working in conjunction to yield our effects, particularly in the case of vices. Longer time windows may be
more attractive because one genuinely wants to reduce the
harm from vices (i.e., genuine self-control), one wants to
tell oneself that he or she is a good person trying to reduce
the harm from vices (i.e., self-signaling), or one uses the
reduced harm from vices as a rationale to consume at least
some vice (i.e., justification). It is important to distinguish
Vice
between these mechanisms (as we discuss later under
Future Research). However, it is also important to appreciate that, given our focus on the planning stage, these different mechanisms lead to the same effect: individuals prefer
vices more when the time window is longer (and virtues
more when the window is shorter). Furthermore, the overall evidence is broadly consistent with self-control to the
extent that study participants are making efforts to serve
long-term goals, as opposed to looking for excuses to cater
SIDDIQUI, MAY, AND MONGA
to their short-term impulses. Three pieces of evidence indicate that our results are consistent with self-control.
The first piece of evidence is simply the overarching
virtue-vice difference that we observe in all studies. When
people are keen to forsake long-term goals and embrace
short-term impulses (i.e., be “bad”), they look for excuses
to do so. Thus, one may find excuses to prefer virtues with
longer time windows (to allow the option to delay consumption) and vices with shorter time windows (to enable
earlier consumption). Conversely, when people are keen to
engage in self-control (i.e., be “good”), they are likely to
prefer the opposite: virtues with shorter time windows (to
promote earlier consumption) and vices with longer time
windows (to promote delayed consumption). Our results do
not reflect the “bad” outcomes associated with excuses but
rather the “good” outcomes associated with self-control.
Another perspective related to excuses is also countered
by our evidence. This perspective is as follows: even
though individuals may make plans in line with selfcontrol predictions, they are actually using the time window as just an excuse (e.g., long window as an excuse to
opt for vices), while actually intending to cater to shortterm impulses at the consumption stage (i.e., still consume
vices earlier). Our theoretical counterargument is that this
perspective is insufficient to explain the entire pattern; it
may sometimes work for vices (i.e., a long window leaves
open the option of consuming vices early), but can never
work for virtues (i.e., a short window closes the option of
delaying virtue consumption). Moreover, we observed the
predicted results even when participants were not free to
consume vices early in studies 1 and 3 because fixed intervals were in place; that is, even when the option of early
indulgence was closed, vices with long time windows were
still preferred. Furthermore, the evidence at the consumption stage does not suggest that individuals are finding excuses to indulge (study 2).
The final piece of evidence pertains to our mediation
and moderation results, which are consistent with selfcontrol to the extent that individuals are trying to serve
long-term goals rather than short-term impulses.
Specifically, in study 4, we showed mediation via perception of achieving one’s health goals. In the case of vices
(vs. virtues), participants perceived that a longer time window would help achieve health goals and, consequently,
they opted for a longer window. Then, in study 5, we
showed moderation via impulsivity. Because high (vs. low)
impulsives need more help in meeting their self-control
goals, they are known to rely more on self-control devices
at the planning stage (Wertenbroch 1998; Wertenbroch
et al. 2001). Consistent with this self-control perspective,
the effect of time windows was stronger for high
impulsives.
Overall, therefore, the current research reveals a new effect—the asymmetric influence of time windows on virtues
versus vices—with our results being broadly consistent
945
with a self-control perspective of pursuing long-term goals
over short-term impulses. These results afford interesting
implications.
Theoretical and Practical Implications
People often face a self-control challenge in the intertemporal domain. The planner self wants to maximize utility over the long run, while the doer self wants to
maximize utility for just the immediate time period
(Benabou and Pycia 2002; Schelling 1984; Thaler and
Shefrin 1981). Therefore, the planner often makes advance
plans to control the doer’s actions (DellaVigna and
Malmendier 2006; Prelec and Loewenstein 1998; Schelling
1992; Trope and Fishbach 2000). In the domain of products, the planner opts for large quantities for virtues and
small quantities for vices (Mishra and Mishra 2011;
Wertenbroch 1998). Our key insight is that self-control
goals for products may be pursued independently of controlling quantity because the final objective is not simply
to control quantity. For a comprehensive self-control plan,
one needs to control the rate at which future consumption
occurs, with the end goal of consuming virtues at a high
rate and vices at a low rate. Given our perspective of a
self-control consumption rate (i.e., quantity/consumption
time window), prior research has focused on just the selfcontrol numerator (i.e., quantity). We demonstrate the influence of the self-control denominator (i.e., time window)
on the preferences and choices of virtue versus vice
products.
Research on quantities (Mishra and Mishra 2011;
Wertenbroch 1998) is part of a broader literature on planning
consumption, which includes diversification and variety seeking (Read et al. 1999; Simonson 1990). The commonality of
this literature is that consumption is planned based on the
characteristics of the product, whereas we reveal the role of
the consumption time period. This influence of time period
adds to the growing literature on the relation of time to selfcontrol. For instance, wait time perception can influence
intertemporal preferences, even when the smaller-sooner and
larger-later options are held constant (Bilgin and LeBoeuf
2010; Kim and Zauberman 2013; May and Monga 2014;
Zauberman et al. 2009). We examine the role of time in a different intertemporal setting in which the costs and benefits
are separated in time (costs earlier for virtues, and benefits
earlier for vices). We show how time windows can influence
preferences, even when the quantity of the product is held
constant.
Our results add to two streams of research that relate to
time windows. The first stream is on how the beginning or
end of a time period may have implications for initiating
aspirational behaviors, changing mental accounts, and
motivating employees to work harder (Dai et al. 2014;
Kaur et al. 2010; Soster et al. 2010). A second stream relates to the procrastination of activities that require time to
946
complete, such as visiting museums and completing rebate
forms (Ariely and Wertenbroch 2002; Gourville and
Soman 2011; O’Donoghue and Rabin 1999; Shu and
Gneezy 2010; Soman 1998). We complement this research
by focusing on the consumption time periods that are an integral part of several products. We show that time windows
have an asymmetric influence on virtue versus vice products because consumers’ self-control efforts are aimed at
increasing the consumption rate of virtues and decreasing
the consumption rate of vices. This reliance on time windows is stronger for those who need more help in meeting
their self-control goals—that is, impulsive individuals.
Our theory of time windows also opens the door to practical implications. A grocery store manager trying to increase consumers’ preference for a virtue (e.g., pack of
salad boxes) would be better off highlighting shelf life when
it is short, but not when it is long. To increase consumers’
preference for a vice (e.g., pack of ice cream bars), it would
be better to highlight shelf life when it is long, but not when
it is short. As another example, managers often invest in
product development efforts to extend the time window
within which their products may be consumed. The implicit
assumption is that a longer time window would make their
products more attractive to consumers. Given our results,
such an investment may be worthwhile only when they are
selling vice products. For virtue products, keeping the time
window short may be a win-win situation. It benefits marketers who are motivated to avoid the product development
costs of extending a time window, and consumers who are
motivated to consume virtues at a high rate.
Just as managerial action may be influenced by our results, public policy may be as well. Consider the example
of the United States Food and Drug Administration (FDA)
examining the claimed shelf life of a food product. It may
be assumed that a longer shelf life, because it improves
consumption flexibility, will enhance consumer welfare.
Our results suggest that an unintended consequence, which
could reduce consumer welfare, is that a longer shelf life
may make virtue products less attractive to consumers, and
vice products more attractive.
Future Research
As discussed in detail earlier, our results are broadly
consistent with self-control to the extent that individuals
seem to be pursuing long-term goals rather than giving in
to short-term impulses. However, we did not disentangle
what part of our results could be attributed to genuine selfcontrol attempts, and what part to other mechanisms such
as self-signaling and justification. The reason was that our
focus was on planners and, regardless of the finer processes
involved, we expected and observed the same effect: preference shifting toward virtues when the time window is
shorter and toward vices when the time window is longer.
For further insights into these mechanisms, more extensive
JOURNAL OF CONSUMER RESEARCH
studies comparing planners to doers are required. More
conceptual clarity is also required, particularly on justification. If one follows the traditional view of justification, as
finding an excuse to engage in “bad” behavior (e.g., find
excuses to opt for vices with short time windows), then the
justification goal (i.e., consume vice at a high rate) clearly
diverges from a self-control goal (i.e., consume vice at a
low rate). However, if we adopt a nuanced view of justification, as relying on reduced harm as a reason to engage in
bad behavior, albeit to a limited extent (e.g., rely on
reduced harm as a reason to opt for vices with long time
windows), then justification seems to morph into selfcontrol because the goal is the same (i.e., consume vice at
a low rate). It is possible that those relying on genuine selfcontrol see the long window as imposing the constraint of
a low vice rate, which helps them become “good.” But
those relying on justification see the long window as allowing the opportunity to consume at least some vice, in return
for becoming “less bad.” We will defer to other scholars
who may unravel the motivational distinctions and moral
equivalencies of these processes.
Our theory is about a farsighted planner trying to control
the myopic doer’s actions (Benabou and Pycia 2002;
Schelling 1984; Thaler and Shefrin 1981; Wertenbroch
1998). However, if the doer’s actions are not myopic, but
hyperopic, the planner may exhibit a precommitment to indulgence (Keinan and Kivetz 2008; Kivetz and Simonson
2002)—preferring large quantity of vices with short time
windows (i.e., high rate of vice consumption) and small
quantity of virtues with long time windows (i.e., low rate
of virtue consumption). This may also happen when consumption is imminent, because impulsive consumers may
want to indulge, and distort perceptions to justify indulgence (Gray 1987; May and Irmak 2014; Sengupta and
Zhou 2007); that is, an individual who faces a consumption
opportunity without having made any self-control plans
may prefer a high consumption rate for vices, and a low
rate for virtues. Apart from individual-specific factors
(e.g., impulsivity), there may also be product-specific factors (e.g., how tempting a product is) that determine the extent to which doers comply with planners.
Future research could also expand on mental accounting
using the ideas that we present in the current research.
Individuals are known to construct mental budgets (Heath
and Soll 1996). Given our perspective of a consumption
rate, individuals trying to adopt a healthier lifestyle may
construct a mental budget that includes not only the quantity, but also the time period (e.g., “I want to restrict myself
to four hamburgers per week”). Would a self-control failure (e.g., eating a fifth hamburger the same week) be construed as a failure to restrict quantity (“I failed to stick to
my four-hamburger quota”), or a failure to comply with the
time window (“I failed to stick to my one-week quota”)? It
would be interesting to examine such mental budgets, and
SIDDIQUI, MAY, AND MONGA
how self-control goals are recalibrated when self-control
failure is attributed to quantity versus time window.
In this research, we focused on how time windows influence vices versus virtues, but future research could examine related dichotomies, such as hedonic versus utilitarian
products. Hedonic products are affective and sought as
goals in themselves, whereas utilitarian products are functional and sought as intermediate steps to higher-end goals
(Botti and McGill 2011; Khan, Dhar, and Wertenbroch
2005). They need not involve a self-control dilemma of
costs and benefits separated in time. For instance, eating
juicy strawberries may be a hedonic experience in that consumption yields pleasure, which in itself is the final goal.
There is no self-control dilemma because the immediate
taste benefits are not countered by any costs in the future,
but rather supplemented by future health benefits. Given
no need for self-control, time windows may not have an influence via the process that we put forth. But they may still
have an influence via other processes, such as diversification and variety seeking (Read et al. 1999; Simonson
1990). For instance, one may prefer a long window for
strawberries (i.e., low consumption rate) if diversifying
with other juicy fruits offers more value than consuming
strawberries alone.
Another interesting research avenue relates to temporal
construal (Trope and Liberman 2010). Because distant (vs.
near) future events are viewed more abstractly, could long
(vs. short) time windows evoke a more abstract construal?
This is an intriguing possibility, but one that we suspect is
unlikely. If long (vs. short) windows evoked a more abstract construal, then that should have led to greater selfcontrol. This is because abstract thinking reduces evaluations of immediate temptations and increases self-control
(Fujita et al. 2006; Pyone and Isen 2012). In other words, a
longer window ought to have shifted preference toward virtue rather than vice. However, we find the opposite, suggesting that abstract thinking is not at play. Perhaps this is
because a long (vs. short) window is not really seen as
being in the distant (vs. near) future. After all, both time
windows start immediately, with one window just being
longer than the other. That said, more research is needed
on whether different time windows elicit different construals, or evoke any other meanings that go beyond what
we propose.
Finally, future research could address whether the size
of the time window has linear effects. While we find that a
longer time window increases preference for vices, taking
this to an extreme could have the opposite effect. For example, two chiropractors in Michigan have put on display
in their office an opened McDonald’s cheeseburger, which
has not decomposed in even two years (Hughes 2015). If a
consumer were aware that a McDonald’s cheeseburger can
last this long, her preference may go down because she
may draw negative inferences regarding the nature of the
ingredients. Thus, while a long time window may increase
947
preference for vices, an inordinately long window may decrease preference. Conversely, extremely short time windows for virtues may lead to positive inferences regarding
the ingredients (e.g., “fresh ingredients that spoil easily”),
and further strengthen the preference for virtues when the
time window is short.
Conclusion
Offering the perspective of a self-control consumption
rate (i.e., quantity/consumption time window), we delineate the influence of the denominator. Shorter windows increase preference for virtues, but longer windows increase
preference for vices. Our findings yield important theoretical and managerial implications. Furthermore, by showing
that closing a time window shifts preferences, we open the
door to new research possibilities.
DATA COLLECTION INFORMATION
The data in this manuscript were collected on
Amazon.com’s Mechanical Turk website (studies 3, 4, and
5; study 1’s first pretest; study 4’s pretest) at the Moore
School of Business, University of South Carolina (study 1,
study 1’s second pretest), and at the Pamplin College of
Business, Virginia Tech (study 2, study 2’s pretest) by research assistants and the authors of the manuscript between
January 2014 and March 2016. The data were analyzed by
the first two authors under the guidance of the third author.
REFERENCES
Ariely, Dan and Klaus Wertenbroch (2002), “Procrastination,
Deadlines,
and
Performance:
Self-Control
by
Precommitment,” Psychological Science, 13 (3), 219–24.
Benabou, Roland and Marek Pycia (2002), “Dynamic
Inconsistency
and
Self-Control:
A
Planner–Doer
Interpretation,” Economics Letters, 77 (3), 419–24.
Benabou, Roland and Jean Tirole (2004), “Willpower and
Personal Rules,” Journal of Political Economy, 112 (4),
848–86.
Bilgin, Baler and Robyn LeBoeuf (2010), “Looming Losses in
Future Time Perception,” Journal of Marketing Research, 47
(3), 520–30.
Bodner, Ronit and Drazen Prelec (2003), “Self-Signaling and
Diagnostic Utility in Everyday Decision Making,”
Psychology of Economic Decisions, 1, 105–26.
Botti, Simona and Ann L. McGill (2011), “When Choosing Is Not
Deciding: The Effect of Perceived Responsibility on
Satisfaction,” Journal of Consumer Research, 33 (2),
211–19.
Dai, Hengchen, Katherine L. Milkman, and Jason Riis (2014),
“The Fresh Start Effect: Temporal Landmarks Motivate
Aspirational Behavior,” Management Science, 60 (10),
2563–82.
DellaVigna, Stefano and Ulrike Malmendier (2004), “Contract
Design and Self-Control: Theory and Evidence,” Quarterly
Journal of Economics, 119 (2), 353–402.
948
——— (2006), “Paying Not to Go to the Gym,” American
Economic Review, 96 (3), 694–719.
Dhar, Ravi and Klaus Wertenbroch (2012), “Self-Signaling and
the Costs and Benefits of Temptation in Consumer Choice,”
Journal of Marketing Research, 49 (1), 15–25.
Fujita, Kentaro, Yaacov Trope, Nira Liberman, and Maya LevinSagi (2006), “Construal Levels and Self-Control,” Journal of
Personality and Social Psychology, 90 (3), 351–67.
Gourville, John T. and Dilip Soman (1998), “Payment
Depreciation: The Behavioral Effects of Temporally
Separating Payments from Consumption,” Journal of
Consumer Research, 25 (2), 160–74.
——— (2011), “The Consumer Psychology of Mail-in Rebates,”
Journal of Product and Brand Management, 20 (2), 147–57.
Gray, Jeffrey A. (1987), “Perspectives on Anxiety and
Impulsivity: A Commentary,” Journal of Research in
Personality, 21 (December), 493–509.
Hayes, Andrew (2013), Introduction to Mediation, Moderation, and
Conditional Process Analysis, New York: Guilford Press.
Heath, Chip and Jack Soll (1996), “Mental Budgeting and Consumer
Decisions,” Journal of Consumer Research, 23 (1), 40–52.
Hoch, Stephen and George Loewenstein (1991), “TimeInconsistent Preferences and Consumer Self-Control,”
Journal of Consumer Research, 17 (4), 492–507.
Holtgraves, Thomas (2004), “Social Desirability and SelfReports: Testing Models of Socially Desirable Responding,”
Personality and Social Psychology Bulletin, 30 (2), 161–72.
Hughes, Ian (2015), “McDonald’s Cheeseburger Left in Doctors’
Surgery for Two Years; Can You Guess What Happens to
It?” http://www.mirror.co.uk/news/weird-news/mcdonaldscheeseburger-left-doctors-surgery-4940927.
Inman, J. Jeffrey and Leigh McAlister (1994), “Do Coupon
Expiration Dates Affect Consumer Behavior?” Journal of
Marketing Research, 31 (3), 423–28.
Johnson, Palmer O. and Jerzy Neyman (1936), “Tests of Certain
Linear Hypotheses and Their Application to Some Educational
Problems,” Statistical Research Memoirs, 1 (January), 57–93.
Kaur, Supreet, Michael Kremer, and Sendhil Mullainathan (2010),
“Self-Control and the Development of Work Arrangements,”
American Economic Review, 100 (2), 624–28.
Keinan, Anat and Ran Kivetz (2008), “Remedying Hyperopia:
The Effects of Self-Control Regret on Consumer Behavior,”
Journal of Marketing Research, 45 (6), 676–89.
Khan, Uzma and Ravi Dhar (2006), “Licensing Effect in Consumer
Choice,” Journal of Marketing Research, 43 (2), 259–66.
——— (2007), “Where There Is a Way, Is There a Will? The Effect
of Future Choices on Self-Control,” Journal of Experimental
Psychology: General, 136 (2), 277–88.
Khan, Uzma, Dhar Ravi, and Wertenbroch Klaus (2005), “A
Behavioral Decision Theory Perspective on Hedonic and
Utilitarian Choice,” in Inside Consumption: Frontiers of Research
on Consumer Motives, Goals, and Desires, ed. S. (Ratti)
Ratneshwar and David Glen Mick, London: Routledge, 144–65.
Kim, B. Kyu and Gal Zauberman (2013), “Can Victoria’s Secret
Change the Future? A Subjective Time Perception Account
of Sexual-Cue Effects on Impatience,” Journal of
Experimental Psychology: General, 142 (2), 328–35.
Kivetz, Ran and Anat Keinan (2006), “Repenting Hyperopia: An
Analysis of Self-Control Regrets,” Journal of Consumer
Research, 33 (2), 273–82.
Kivetz, Ran and Itamar Simonson (2002), “Self-Control for the
Righteous: Toward a Theory of Precommitment to
Indulgence,” Journal of Consumer Research, 29 (2), 199–217.
JOURNAL OF CONSUMER RESEARCH
Kivetz, Ran and Yuhuang Zheng (2006), “Determinants of
Justification and Self-Control,” Journal of Experimental
Psychology: General, 135 (4), 572–87.
Krishna, Aradhna and Z. John Zhang (1999), “Short- or LongDuration Coupons: The Effect of the Expiration Date on the
Profitability of Coupon Promotions,” Management Science,
48 (8), 1041–56.
May, Frank and Caglar Irmak (2014), “Licensing Indulgence in
the Present by Distorting Memories of Past Behavior,”
Journal of Consumer Research, 41 (3), 624–41.
May, Frank and Ashwani Monga (2014), “When Time Has a Will
of Its Own, the Powerless Don’t Have the Will to Wait:
Anthropomorphism of Time Can Decrease Patience,”
Journal of Consumer Research, 40 (5), 924–42.
Mishra, Arul and Himanshu Mishra (2011), “The Influence of
Price Discount Versus Bonus Pack on the Preference for
Virtue and Vice Foods,” Journal of Marketing Research, 48
(1), 196–206.
O’Donoghue, Ted and Matthew Rabin (1999), “Doing It Now or
Later,” American Economic Review, 89 (1), 103–24.
Oster, Sharon and Fiona Morton (2005), “Behavioral Biases Meet
the Market: The Case of Magazine Subscription Prices,”
Advances in Economic Analysis and Policy, 5 (1), 1–30.
Preacher, Kristopher J. and Andrew F. Hayes (2004), “SPSS and
SAS Procedures for Estimating Indirect Effects in Simple
Mediation Models,” Behavior Research Methods,
Instruments, & Computers, 36 (4), 717–31.
Prelec, Drazen and George Loewenstein (1998), “The Red and the
Black: Mental Accounting of Savings and Debt,” Marketing
Science, 17 (1), 4–28.
Puri, Radhika (1996), “Measuring and Modifying Consumer
Impulsiveness: A Cost-Benefit Accessibility Framework,”
Journal of Consumer Psychology, 5 (2), 87–113.
Pyone, Jin Seok and Alice M. Isen (2012), “Positive Affect,
Intertemporal Choice, and Levels of Thinking: Increasing
Consumers’ Willingness to Wait,” Journal of Marketing
Research, 48 (3), 532–43.
Read, Daniel, George Loewenstein, and Shobana Kalyanaraman
(1999), “Mixing Virtue and Vice: Combining the Immediacy
Effect and the Diversification Heuristic,” Journal of
Behavioral Decision Making, 12 (4), 257–73.
Schelling, Thomas (1984), “Self-Command in Practice, in Policy,
and in a Theory of Rational Choice,” American Economic
Review, 74 (2), 1–11.
——— (1992), “Self-Command: A New Discipline,” in Choice over
Time, ed. George Loewenstein and Jon Elster, New York:
Russell Sage Foundation, 167–76.
Sengupta, Jaideep and Rongrong Zhou (2007), “Understanding
Impulsive Eaters’ Choice Behaviors: The Motivational
Influences of Regulatory Focus,” Journal of Marketing
Research, 44 (2), 297–308.
Shu, Suzanne and Ayelet Gneezy (2010), “Procrastination of
Enjoyable Experiences,” Journal of Marketing Research, 47
(5), 933–44.
Simonson, Itamar (1990), “The Effect of Purchase Quantity and
Timing on Variety-Seeking Behavior,” Journal of Marketing
Research, 27 (2), 150–62.
Soman, Dilip (1998), “The Illusion of Delayed Incentives:
Evaluating Future Effort-Money Transactions,” Journal of
Marketing Research, 35 (4), 427–37.
Soster, Robin, Ashwani Monga, and William Bearden (2010),
“Tracking Costs of Time and Money: How Accounting
SIDDIQUI, MAY, AND MONGA
Periods Affect Mental Accounting,” Journal of Consumer
Research, 37 (4), 712–21.
Spiller, Stephen, Gavan Fitzsimons, John Lynch Jr., and Gary
McClelland (2013), “Spotlights, Floodlights, and the Magic
Number Zero: Simple Effects Tests in Moderated
Regression,” Journal of Marketing Research, 50 (2), 277–88.
Thaler, Richard (1980), “Toward a Positive Theory of Consumer
Choice,” Journal of Economic Behavior and Organization, 1
(March), 39–60.
——— (1985), “Mental Accounting and Consumer Choice,”
Marketing Science, 4 (3), 199–214.
Thaler, Richard and Hersh M. Shefrin (1981), “An Economic
Theory of Self-Control,” Journal of Political Economy,
392–406.
Trope, Yaacov and Ayelet Fishbach (2000), “Counteractive SelfControl in Overcoming Temptation,” Journal of Personality
and Social Psychology, 79 (4), 493–506.
949
Trope, Yaacov and Nira Liberman (2010), “Construal-Level Theory of
Psychological Distance,” Psychological Review, 117 (2), 440–63.
Wertenbroch, Klaus (1998), “Consumption Self-Control by
Rationing Purchase Quantities of Virtues and Vice,”
Marketing Science, 17 (4), 317–37.
Wertenbroch, Klaus, Dilip Soman, and Joseph Nunes (2001),
“Debt Aversion as Self-Control: Consumer SelfManagement of Liquidity Constraints,” Working Papers—
INSEAD R & D, Fontainebleau, France.
Zauberman, Gal, B. Kyu Kim, Selin Malkoc, and James Bettman
(2009), “Discounting Time and Time Discounting:
Subjective Time Perception and Intertemporal Preferences,”
Journal of Marketing Research, 46 (4), 543–56.
Zauberman, Gal and John G. Lynch Jr. (2005), “Resource Slack
and Propensity to Discount Delayed Investments of Time
versus Money,” Journal of Experimental Psychology:
General, 134 (1), 23–37.
Copyright of Journal of Consumer Research is the property of Oxford University Press / USA
and its content may not be copied or emailed to multiple sites or posted to a listserv without
the copyright holder's express written permission. However, users may print, download, or
email articles for individual use.
Purchase answer to see full
attachment