MAR WILEJ RIGHT BATCH

Top of text

Top of CT
Temptation and Resistance:
An Integrated Model of
Consumption Impulse
Formation and Enactment
Utpal M. Dholakia
State University of New York at Buffalo

ABSTRACT

Research pertaining to the consumption impulse is sparse in the

literature. To address this lacuna, the author presents and validates
a detailed theoretical framework explicating the consumption
impulse formation process, and examining the role played by
cognitive and volitional processes in its resistance or enactment. The
model makes the distinction between consonant (harmonious)
impulses and dissonant (conflicting) impulses and elaborates on the
role of the impulsivity trait, situational variables, and constraining
factors in enactment or resistance of the consumption impulse. The
results of two studies provide support to the general working of this
theoretical framework. This research has the potential to inform
many critical issues surrounding consumer behavior, such as
regulating consumption impulses in retail and on-line shopping
environments, and developing interventions for prevention of
harmful consumer behaviors such as addictions. 䉷 2000 John Wiley
& Sons, Inc.

Impulsive consumption behaviors comprise a large class of consumer

behaviors that have received disproportionately little attention by con-
sumer researchers (however, see Rook’s research and Puri, 1996). Mar-
keting practitioners, on the other hand, have long realized the impor-
tance of impulsive consumer actions. For instance, much of the retail Base of text

Psychology & Marketing

䉷 2000 John Wiley & Sons, Inc. Vol. 17(11):955–982 (November 2000)
955 Base of DF
MAR WILEJ LEFT BATCH

Top of text
emphasis in packaged goods focuses on trying to get the consumer to Base of text
impulsively purchase products at the point of sale. Indeed, store and
web site layouts, product packaging elements, and in-store promotions
all focus on promoting and regulating such impulsive purchases. Much
of early academic marketing research focused on understanding the in-
fluence of different merchandising stimuli on impulsive purchase be-
haviors as well (e.g., Cox, 1964; Patterson, 1963). A second category of
impulsive consumption pertains to compulsions and addictions such as
smoking, overspending, or overeating (Nataraajan & Goff, 1991). Issues
pertaining to such consumption have been a major focus of clinical psy-
chiatrists (e.g., Plutchik & van Praag, 1995), with emphasis on design-
ing effective intervention strategies to mitigate the harmful conse-
quences of such actions (Puri, 1996; see Bütz & Austin, 1993, for a
review).
One reason for this lacuna with regard to impulsive phenomena is
that many of the dominant consumer behavior paradigms such as be-
havioral decision theory and attitude research focus on goal-directed
and intentional consumer behaviors. This emphasis limits the applica-
bility of these rich bodies of research to issues of consumer impulsivity,
its drivers, mediating mechanisms, and consequences. A second reason
is that the extant psychological research on this topic has focused more
on hedonic aspects of impulsive behaviors (e.g., Weinberg & Gottwald,
1982, but see Rook & Fisher, 1995, for an exception), ignoring for the
most part the role played by cognitive and volitional processes in the
enactment or dissipation of consumption impulses. In addition, rela-
tively little attention has been focused on trying to understand mecha-
nisms of impulse resistance, that is, when and how the consumption
impulse is successfully thwarted by the consumer. A third and final
limitation of existing impulsive consumption behavior research pertains
to an inordinate focus on the purchase of low-cost, frequently purchased
products (e.g., Rook & Hoch, 1985), characterizing such product cate-
gories as “impulse,” and the underlying appetitive urge as the “buying
impulse” (Rook, 1987). This emphasis is limiting, providing a somewhat
narrow view of psychological processes surrounding impulsive con-
sumption. The term consumption impulse (CI) is used in this article
instead, to explicitly acknowledge that the urge can arise in the context
of other behaviors of interest to consumer researchers, such as product
use, disposal, etc. in addition to purchase of products.
A primary thesis of this research is that much like other behavior
types, impulsive consumption behaviors are preceded by distinct psy-
chological processes that can be explicated by a theoretical framework.
In this article, an integrated model of consumption impulse formation
and enactment (CIFE) is presented, elaborating the psychological pro-
cesses and factors leading to formation of the consumption impulse, and
its subsequent dissipation or enactment. The CIFE model describes the short
role of cognitive and volitional psychological mechanisms during the standard

956 DHOLAKIA Base of RF

MAR WILEJ RIGHT BATCH

Top of text
impulse enactment process in great detail. Through understanding the Base of text
psychological processes that govern the formation and enactment of con-
sumer impulses, the framework presented here has the potential to in-
form many critical issues surrounding such consumer behaviors. From
understanding how to better regulate consumer impulses in retail and
on-line shopping environments for attainment of specific marketing ob-
jectives, to developing interventions for prevention of harmful con-
sumption behaviors such as addictions, understanding how and when
consumer impulses translate into behavior is likely to be extremely im-
portant for a slew of marketing strategies in both profit and nonprofit
contexts.

A CONCEPTUAL UNDERSTANDING OF IMPULSIVE

CONSUMPTION BEHAVIORS

Because Murray’s (1938) description of impulsivity as the tendency to

respond quickly and without reflection, issues pertaining to impulsive
actions have been studied by researchers in a number of social science
disciplines, including social work, psychiatry, social psychology, and
cognitive biopsychology. Within the diverse treatments of impulsive
consumption, many common threads can be found. First, impulsive con-
sumption behaviors are viewed to occur as the result of a sudden, often
powerful, and persistent urge to consume, often without much deliber-
ation. As noted before, the consumption impulse may pertain to product
purchase (e.g., buying a candy bar at the supermarket checkout
counter), product use (e.g., eating a piece of cheesecake when watching
late-night television), or even product disposal (e.g., trading in a ser-
viceable car for a new model).
A second characteristic of impulsive behaviors is that such behaviors
are associated with an extremely short time span between formation of
the CI and its enactment (or dissipation). For instance, biopsychologists
Barrett and Patterson (1983) view impulsive behavior to be character-
ized by actions without adequate reflection, by spur-of-the-moment re-
actions, and by getting things done quickly. Similarly, education re-
searchers Lurr and Wunderlich (1985) view the bipolar component of
responding immediately to a stimulus vs. planning before making a
move to be a major constituent of impulsivity. Indeed, impulsive behav-
iors have frequently been characterized as unplanned behaviors by con-
sumer researchers (e.g., Cobb & Hoyer, 1986).
A third interesting and important point is that the experience of the
CI does not necessarily result in behavior. Individuals can, and often
do, resist the CI successfully. For instance, psychiatrists studying ag-
gressive behaviors recognize that aggressive impulses may not get
transformed into overt actions because of what they label as attenuating short
countervailing forces such as a timid personality style or concern about standard

TEMPTATION AND RESISTANCE 957 Base of RF

MAR WILEJ LEFT BATCH

Top of text
family well-being (Plutchik & van Praag, 1995; also see Nataraajan & Base of text
Goff, 1991, for a similar view in the case of compulsive consumer be-
haviors). The theoretical framework presented here considers such non-
behavior following the experience of a CI explicitly.
A fourth characteristic of the CI pertains to frequency of occurrence.
A CI may occur either once (e.g., on encountering a fetching scarf in the
clothing store) or more often (e.g., the recurring urge to eat calorific foods
for someone on a diet) for the same consumer. Finally, researchers have
also noted that psychological conflict often surrounds enactment of im-
pulsive behaviors (e.g., Emmons, King, & Sheldon, 1993). The CIFE
framework elaborates on how and when conflict occurs, and its impact
on impulsive behavior enactment.
It is also of value to understand impulsive behaviors in the context
of other types of behaviors such as goal-oriented, and mindless or au-
tomatic behaviors (e.g., Bagozzi & Dholakia, 1999). Goal-oriented and
impulsive behaviors share some interesting characteristics. First, the
construct of consumption impulse can be thought of as similar to the
intention construct for goal-directed behaviors in that both have moti-
vational and volitional content. Second, both goal-directed and impul-
sive actions follow similar stages. For goal-directed behaviors, goal set-
ting (or selection of the goal) occurs first, followed by goal striving, that
is, effortfully maintaining the goal, planning the best way to attain the
goal, and executing this plan of action (Bagozzi & Dholakia, 1999). Im-
pulsive behaviors unfold similarly in that experience of the CI occurs
first, followed by its enactment or resistance through execution of ac-
tions by the consumer.
But these behaviors have some fundamental differences as well. The
first distinction between goal-oriented and impulsive behaviors pertains
to the role played by the consumer’s goal in enactment of behavior. For
goal-oriented behavior, the goal to be attained is explicitly chosen and
plays a central role in subsequent motivational and volitional processes
for formation and enactment of implementation plans. On the other
hand, goals play at most a marginal role in the impulse formation and
enactment process. Broad consumption goals (e.g., to have a good time)
may make some consumers susceptible to behaving impulsively, but for
the most part, intensity of the experienced CI, and the presence or ab-
sence of impediments to action are the primary drivers of impulsive
behavior enactment.
A second distinction pertains to the amount of time available for pro-
cessing information, and forming and implementing the plan of action.
For goal-directed behavior, information is collected and processed sys-
tematically, with consideration of all available goal alternatives. Usu-
ally, there is a significant time period from the time goal setting is ini-
tiated to enactment of the behavior for goal attainment. For this reason,
such behaviors have been called reflective behaviors (Doob, 1990). For short
impulsive behavior, on the other hand, the entire process of impulse standard

958 DHOLAKIA Base of RF

MAR WILEJ RIGHT BATCH

Top of text
formation and enactment occurs quickly. To put it succinctly, whereas Base of text
intention formation and realization occur deliberately, impulse forma-
tion and enactment occur rapidly. A third distinction between the two
behavior types pertains to intention and CI valence. An intention may
have positive or negative valence in that the consumer may form an
intention not to purchase a particular brand as readily and justifiably
as an intention to purchase the brand. A CI on the other hand, connotes
primarily positive valence, indicating a proclivity to engage in the con-
sumption.
It is also valuable to distinguish between impulsive and automatic or
mindless behaviors (Bargh & Barndollar, 1996). When behaving mind-
lessly, the consumer interacts with the environment in a passive, reac-
tive fashion and is cognitively inert (Langer & Imber, 1980). Mindless
behaviors are characterized as being involuntary, effortless (i.e., not
consuming the individual’s information processing capacity), and occur-
ring outside of awareness (Bargh, 1989), and are often those for which
the consumer has a well-rehearsed script, that is, habitual behaviors.
On the other hand, when behaving impulsively the consumer is cog-
nitively alert and not averse to processing pertinent information. Even
when the individual is sorely tempted to buy some product, she or he
has an unmistakable awareness of positive as well as negative conse-
quences of this action. Indeed, psychological conflict arising from incon-
gruence between emotional (i.e., desires) and cognitive (i.e., evaluations
of long-term consequences) preferences has been noted as a frequent
characteristic of impulsive consumption behaviors (Emmons et al.,
1993; Rook, 1987). The two types of behaviors are related in that if an
impulsive action is performed repeatedly (e.g., purchasing a candy bar
from the checkout aisle at every supermarket visit), rather than enact-
ment through a CI, it is likely to become a behavioral tendency (Oullette
& Wood, 1998). It is then likely to get enacted immediately and reflex-
ively on sensory pickup of pertinent cues from the environment and
eventually become mindless. In this case, the impulsive behavior be-
comes habitual because of frequent enactment over time.

AN INTEGRATED MODEL OF CONSUMPTION IMPULSE

FORMATION AND ENACTMENT (CIFE)

The discussion so far suggests that impulsive consumer behaviors con-

stitute a unique class of behaviors that, though not goal-directed in the
strict sense, involve motivational, volitional, and cognitive psychological
processes. Moreover, these behaviors follow a certain structure with the
experience of the consumption impulse (CI) as a central organizing con-
struct. These ideas and knowledge are integrated with recent advances
in the social psychology literature on motivational and volitional short
processes, and a detailed account of impulsive consumption behaviors standard

TEMPTATION AND RESISTANCE 959 Base of RF

MAR WILEJ LEFT BATCH

Top of text
Base of text

Figure 1. The CIFE framework.

is presented in this section. This is labeled the integrated model of con-

sumption impulse formation and enactment (CIFE). The CIFE theoret-
ical framework is graphically presented in Figure 1.
The natural starting point is the impulse-formation process, and con-
sequently, the CIFE model begins with antecedents of the CI. Three
antecedents to the CI are posited in this model. Many consumers ex-
perience the CI on visual exposure to the product (Hoch & Loewenstein,
1991). Product presentation aspects such as attractive displays, tempt-
ing graphics or copy, or associated sales promotions may also play an
important part. For instance, the role of physical proximity in impulse
experience has been clearly documented in experiments on delay of grat-
ification (Mischel & Mischel, 1983). Physical proximity may activate
positive memories associated with the product’s consumption in the
past, and may also serve to kindle desire for it. Similarly, temporal prox-
imity is also a key stimulus in formation of the CI (Loewenstein, 1990).
The immediate availability of the product for purchase or consumption
may activate a latent need for some consumers, thereby triggering the
CI. In the CIFE model, these factors are called marketing stimuli, be-
cause marketers can control the presentation of such stimuli to the con-
sumer. Marketing stimuli represent the first antecedent factor for for-
mation of the CI. short
A second antecedent of the CI is environmental, as well as personal standard

960 DHOLAKIA Base of RF

MAR WILEJ RIGHT BATCH

Top of text
and social factors surrounding a particular consumption occasion. This Base of text
antecedent is labeled as situational factors (Belk, 1975). Situational fac-
tors may increase or decrease the propensity of the consumer to expe-
rience the consumption impulse. One important category of situational
variables pertains to environmental conditions surrounding the partic-
ular consumption occasion. For instance, an individual receiving his or
her paycheck may be more susceptible to experiencing the CI. A second
category of situational factors pertains to the consumer’s current mood
state. A considerable body of research within social psychology has
shown that a positive mood state increases the risk-seeking propensity
of individuals, and makes them less amenable to systematically process
information (see Schwartz & Bohner, 1996, for a review). These tend-
encies may make the consumers more susceptible to experiencing the
CI.
The third antecedent factor in the CIFE model is the impulsivity trait
of the consumer, which has been extensively studied in the social sci-
ences, in a variety of contexts (e.g., Plutchik & van Praag, 1995; Watson
& Clark, 1993). The impulsivity trait has been defined as the tendency
to respond quickly and without reflection, and characterized by rapid
reaction times, absence of foresight, and a tendency to act without a
careful plan (Doob, 1990). Psychometric analysis has shown that im-
pulsivity is characterized well by a single dimension (cf. Plutchik & van
Praag, 1995) and is associated with other personality characteristics
such as acquisitiveness (Belk, 1985), need for variety (Hirschman,
1980), and risk aversion. In the CIFE model, the impulsivity trait is
viewed as an important antecedent factor in the formation of the CI.
This antecedent factor is examined in greater detail in the experimental
studies.
The presence of one or more of these three antecedents to a sufficient
degree culminates in formation of the consumption impulse, an irre-
sistible urge to consume. The CI is viewed as occurring automatically
(Isen & Diamond, 1989), because it cannot be blocked from occurring
when the antecedents are present with adequate strength. Moreover, it
is useful to think of the consumption impulse experience as having an
intensity dimension, rather than as occurring or not occurring. The in-
fluence of each of the three factors in initiating the consumption impulse
may vary by individual, as well as on different occasions for the same
individual consumer. For instance, the impulsive trait aspect of Jane’s
personality may cause her to experience the consumption impulse on
encountering a beautiful dress in a mall store. In contrast, Sue, who is
less impulsive, may experience the consumption impulse mainly on ac-
count of a half-off sale for the very same dress. In addition to the main
effects of these antecedent factors, the three factors may interact posi-
tively as well. Thus, the impulsive trait of Jane on combining with an
attractively packaged candy bar on a particular visit to the supermarket short
may magnify strength of the impulse experienced. standard

TEMPTATION AND RESISTANCE 961 Base of RF

MAR WILEJ LEFT BATCH

Top of text
When the consumer experiences the CI, mental responses are auto- Base of text
matically triggered to evaluate the presence of possible constraints to
enactment (Loewenstein, 1990). This view is consistent with the idea
held by social psychologists that thoughts regarding action occur spon-
taneously and primarily out of necessity when action execution involves
more than just reflexive responding to cues (Vallacher, 1993). Con-
straints to CI enactment generally fall into one of three categories. First,
the consumer may realize that there are current impediments to smooth
enactment of actions consistent with the CI. For instance, the consumer
may not have adequate time or money, or may be reminded of implicit
but stringent behavioral rules used to guide behavior in the past. The
second category of constraints pertains to consideration of long-term del-
eterious consequences of enacting the behavior. For instance, an over-
weight person may consider the future gain of weight from eating a
particularly high-calorie, if toothsome, dessert as a constraining factor.
Finally, anticipatory emotions, defined as affective states resulting from
anticipation of future consequences of the chosen course (Bagozzi,
Baumgartner, & Pieters, 1998) may also work to constrain the CI en-
actment. For instance, the consumer may imagine the positive emo-
tional experience from successfully resisting the urge, or the negative
affect from enacting the impulse. These three categories of impediments
to impulse enactment are labeled as “constraining factors” in the CIFE
model.
Many research findings in the extant literature are consistent with
the view of constraining factors presented here. First, constraining fac-
tors are similar to the construct of an interrupt (Bettman, 1979; Hoch
& Loewenstein, 1991) that is experienced by consumers and alerts them
to the need for cognitive deliberation. Consistent with this idea, many
of Rook’s (1987) subjects evaluated a particular consumption impulse
negatively, explicitly recognizing that they were breaking budgetary or
dietary rules. For these subjects, consideration of the rule served as an
interrupt or constraining factor. Constraining factors are also similar
to countervailing forces proposed by psychiatrists studying aggressive
impulsive behaviors. This line of research suggests that on experiencing
an aggressive impulse, one or more factors may serve to attenuate its
effect and prevent enactment of violent behavior for the individual. Such
factors, examples of which include appeasement from others or the
memory of a previous injury, are called “attenuating countervailing
forces” in the psychiatry literature (Plutchik & van Praag, 1995) and
serve to diminish effects of the aggressive impulse, much like constrain-
ing factors.
If no constraining factors are identified, the CI may be viewed by the
consumer as harmonious with his or her goals, resources, and situation.
The consumer then responds reflexively through an expression of his or
her CI, without further deliberation or hesitation. This type of behav- short
ioral scenario can be called the consonant CI condition. In this condition, standard

962 DHOLAKIA Base of RF

MAR WILEJ RIGHT BATCH

Top of text
the strength and interaction of antecedent factors drive formation and Base of text
subsequent enactment of the CI, and cognitive evaluation of the impul-
sive behavior or its consequences is minimal.
On the other hand, if constraining factors are identified (as is more
likely to be the case), the consumer experiences conflict and ambivalence
(Dickman, 1990; Rook, 1987). This experience has been framed in the
literature as a psychological conflict between the desire (or the con-
sumption impulse in the CIFE framework) and the strategy of volitional
control or willpower of the individual (Hoch & Loewenstein, 1991; Kuhl,
1987). To characterize this conflict, such a behavioral scenario is labeled
the dissonant CI condition. This distinction between consonant and dis-
sonant CIs follows the distinction made in most motivation theories be-
tween motivational systems concerning positive or pleasurable experi-
ences and negative or painful experiences (Higgins, 1996).
According to the CIFE model, the experience of a dissonant CI results
in a more thought-based evaluation of the consequences of enacting the
CI, and movement from a hedonic impulse-dominated mode of function-
ing to a more evaluative mode. In some sense, the consumer weighs the
pros and cons of behaving in accordance with the CI. This process, which
occurs quickly, resulting in a positive or negative evaluation of impul-
sive behavior enactment, is critical in providing guidance to the con-
sumer (Rook & Fisher, 1995). If the evaluation is positive, the consumer
may view the constraining factor as not significant enough, and may
proceed with enactment of the CI. In this case, the consumer may feel
some amount of psychological uncertainty or conflict, but is likely to
enact the CI anyway. On the other hand, if cognitive evaluation of the
impulsive behavior is negative, the volitional system of the consumer
comes into play (Bagozzi & Dholakia, 1999). The most important func-
tion of the volitional system is to harness different resistance strategies
to counter the CI.
Resistance strategies can be thought of as volitional mechanisms used
by the consumer to effortfully fight back the dissonant CI after the be-
havior is evaluated negatively. Strategies of resistance often involve
self-regulation of mental states and processes by the individual through
different mechanisms. Research on control of actions (Kuhl, 1987) and
on self-control (e.g., Gilbert, 1993; Thaler & Shefrin, 1981) has identified
several such mechanisms, which provide rich insight into the working
of the volitional system. One important resistance strategy pertains to
formulation and learning of explicit rules (Thaler & Shefrin, 1981),
which can then be used as guidelines to resist CIs. For instance, a saving
or a dieting program may be monitored through a weekly budget, finan-
cial or dietary. Explicit rules are particularly useful for frequently oc-
curring CIs such as those associated with addictive or compulsive con-
sumption behaviors.
A second resistance strategy, selective attention (Beck & Emery, 1985; short
Kuhl, 1987) refers to a person’s tendency to attend to information sup- standard

TEMPTATION AND RESISTANCE 963 Base of RF

MAR WILEJ LEFT BATCH

Top of text
porting the volitionally supported course of action, and to ignore com- Base of text
peting information. Such closed-mindedness may encourage a focus
away from the tempting CI. Much empirical support exists in the lit-
erature for the operation of this process. For instance, Mischel and Mis-
chel (1983) found that children learn to maintain an experimentally
induced intention against a more desirable, competing action alterna-
tive by avoiding visual contact with the source of distraction. This has
also been labeled a distraction strategy by impulsive-behavior research-
ers (Hoch & Loewenstein, 1991). A third resistance strategy is encoding
control, which facilitates the protective function of the volitional system
by encoding selective features of the environment that are related to
fighting the CI.
A related strategy is exposure control (cf. Gilbert, 1993), where the
consumer actively manipulates his environment to reduce effect of the
CI-causing stimuli, or to increase effectiveness of other resistance strat-
egies. For example, on experiencing a CI to eat dessert, Fred may
quickly remove himself from the proximity of the dessert tray at a sump-
tuous buffet. Such a resistance strategy has also been called a distanc-
ing strategy (see Rook & Hoch, 1985, for a detailed discussion).
Parsimonious information processing is the fifth resistance strategy
that relates to how and when the consumer stops processing informa-
tion. To resist a CI successfully often requires minimizing length of the
cognitive evaluation process after one or more constraining factors have
been identified, to prevent overlong deliberation regarding the CI. If
further cognitive evaluation is viewed as jeopardizing resistance to the
CI, the process of cognitive evaluation may be brought to a halt. Par-
simonious information processing is especially useful when new infor-
mation supporting the CI constantly becomes available, as when allur-
ing marketing stimuli keep assaulting the consumer’s senses.
Further, research supports the view that failure to resist dissonant
CIs successfully may also arise from negative affective states such as
dissatisfaction, guilt, or anger at oneself (Kanfer, 1996), or from expo-
sure to sadness-causing stimuli (Mischel, Ebbesen, & Zeiss, 1972). Pos-
itive emotional states, on the other hand, promote CI resistance. In this
regard, control of emotions represents the sixth resistance strategy used
to resist CIs. It refers to a set of self-regulatory skills aimed at inhibiting
negative emotional states, and avoiding intrusion of emotional task-
related thoughts that might undermine efficiency of the protective func-
tion of volition (Salovey & Mayer, 1990). Emotion control also enables
generation of positive emotions that may be conducive to successful CI
resistance.
The seventh mechanism, motivation control, refers to a set of self-
regulatory activities aimed at bringing high levels of attention and effort
to bear on the task of CI resistance. For instance, Jane may believe that
she has successfully fended off her CI to buy a new dress, and let her short
standard

964 DHOLAKIA Base of RF

MAR WILEJ RIGHT BATCH

Top of text
guard down, only to have the urge return, and result in an impulsive Base of text
purchase. Motivation control prevents this from happening while the CI
is active, and includes strategies such as substitution (Hoch & Loew-
enstein, 1991) where the consumer may give herself a small immediate
reward for successfully resisting a larger and more deleterious CI (e.g.,
celery sticks in lieu of a sack of potato chips). Motivation control is es-
pecially salient for resisting pernicious dissonant CIs such as breaking
dieting plans, smoking, drinking, etc. In many cases, this strategy may
operate through making the self-efficacy, that is, the conviction that one
can resist the CI successfully (Bandura, 1991) salient to the individual.
These resistance strategies constitute what social psychologists have
termed tactical mental control (Wegner & Erber, 1993), because the con-
sumer uses some specific scheme to suppress his or her natural state of
mind (the impulse), and promote the volitionally chosen state (resis-
tance). The volitional system and its associated resistance strategies are
activated when the behavior is evaluated negatively for a dissonant CI,
and remains so until the window of opportunity for action enactment
closes. In some cases however, in spite of dissonance, the CI proves ir-
resistible and is enacted. In such cases, the volitional system may fa-
cilitate coping with failure, as the consumer detaches himself or herself
from the unsuccessful CI resistance attempt by inhibiting activation of
the resistance elements, and through preventing excessive ruminations
about the CI (e.g., “If only I had controlled myself instead of going on
an eating binge!”).
When the consumer experiences a dissonant CI and evaluates the
behavior negatively, an outcome expectancy, that is, the subjective like-
lihood of successfully resisting the CI, given continued effort, is derived.
This expectancy is directly influenced by the strength of the CI and the
valence of the negative cognitive evaluation, which determines the mo-
tivational available for resistance. If expectancy for successful resis-
tance is sufficiently favorable, the volitional system continues to guide
the individual toward resistance of the CI until the window of oppor-
tunity for CI enactment closes. The consumer may use one or more re-
sistance strategies to resist the impulse. On the other hand, if the ex-
pectancy is sufficiently unfavorable, resistance to the CI dwindles,
resulting in the eventual enactment of the dissonant negatively evalu-
ated CI. In this case, the consumer may experience conflict during CI
enactment, and negative emotions such as guilt or regret. This mecha-
nism of outcome-expectancy evaluation is consistent with the Carver and
Scheier (1990) views in the control-theory – based model of self-regula-
tion.
As the volitional resistance processes operate, the strength of the CI
eventually dwindles as the antecedent factors responsible for the CI
recede from the consumer’s attention. In this case, the CI dissipates
without enactment of the behavior. The CIFE model represents a par- short
standard

TEMPTATION AND RESISTANCE 965 Base of RF

MAR WILEJ LEFT BATCH

Top of text
simonious, yet psychologically rich account of the process through which Base of text
consumption impulses are formed and get enacted or resisted by con-
sumers.

STUDY 1: IMPULSIVE PRODUCT PURCHASE

In the rest of the article, the key ideas underlying the CIFE model are
tested. In particular, the focus is on the differential drivers of behavior
enactment in the consonant and dissonant CI conditions, and on the
operation of the volitional system. A key idea in the CIFE framework is
that in the consonant CI condition, when the consumer does not identify
any constraining factors, the CI gets enacted into behavior smoothly,
without much cognitive deliberation. In this case, the CIFE model sug-
gests that strength of the antecedent factors should determine the in-
tensity of the CI, which should then drive the impulsive consumption.
For tractability and given its relative importance in the existing re-
search (e.g., Dickman, 1990; Rook & Fisher, 1995), this research focuses
on the impulsivity trait of the consumer as the antecedent factor of pri-
mary interest. It is expected that the impulsivity trait of the consumer
will drive behavior enactment when the CI is consonant. At the same
time, the CIFE model suggests that the cognitive evaluation should not
play a role in enactment of the CI, because the consumer responds re-
flexively without much deliberation. Following the reasoning above, the
research hypothesis can be articulated as follows:

H1: In the consonant CI condition, trait impulsivity is the more sig-

nificant predictor, and cognitive evaluation of behavior is the less
significant predictor of impulsive behavior enactment.

Figure 2 shows these hypothesized relationships for the consonant CI

condition.
On the other hand, if the consumer identifies one or more constrain-
ing factors following the experience of the CI, the CIFE framework pos-
its that a cognitive evaluation of the behavior occurs. For such a disso-
nant CI, the consumer is more likely to enact the CI if the evaluation is
positive and vice versa. The CIFE model thus predicts that valence of
the cognitive evaluation is more likely to drive impulsive behavior,
whereas antecedent factors such as trait impulsivity will play a rela-
tively unimportant role in its enactment. Following this reasoning, the
research hypothesis is stated:

H2: In the dissonant CI condition, cognitive evaluation of behavior is

the more significant predictor, and trait impulsivity is the less
significant predictor of impulsive behavior enactment. short
standard

966 DHOLAKIA Base of RF

MAR WILEJ RIGHT BATCH

Top of text
Base of text

Figure 2. Consonant impulse condition.

Figure 3 graphically depicts the relationships presented in Hypoth-

esis 2.
To summarize, it is hypothesized that the type of CI (consonant or
dissonant) plays a moderating role in the influence of antecedent factors
such as the impulsivity trait and cognitive evaluation on enactment of
the impulsive behavior. Additionally, the CIFE framework indicates
that when the evaluative process results in a negative judgment of be-
havior for a dissonant CI, the consumer actively resists the CI. Relative
to the case where the consumer evaluates the dissonant CI positively,
it is expected that the CI is less likely to be enacted when the cognitive
evaluation of the dissonant CI is negative. The following hypothesis
formalizes this reasoning:

H3: The likelihood of impulsive behavior enactment is significantly

less for the dissonant-CI – negative cognitive evaluation condition
than for either the dissonant-CI – positive cognitive evaluation
condition, or the consonant CI conditions.

In this first study, the research hypotheses are tested in the context
of an impulsive product purchase. The experiment conducted to test the
hypotheses is described next. short
standard

TEMPTATION AND RESISTANCE 967 Base of RF

MAR WILEJ LEFT BATCH

Top of text
Base of text

Figure 3. Dissonant impulse condition.

METHOD

Sample and Data Collection

As a partial fulfillment of a course requirement, 101 undergraduate
business students at a large midwestern university participated in the
study. The basic framework of the study used an approach similar to
that of Rook and Fisher (1995). The subjects were first asked to select
one of a set of purchase alternatives in a hypothetical buying scenario.
Following the selection of the behavior, the subjects responded to mul-
tiple measures eliciting cognitive evaluation of behaving impulsively, as
well as the Rook Impulsivity Scale (Rook, 1987; Rook & Fisher, 1995).
The details of these measurements are described next.

Measurement of Impulsive Purchase Behavior. The shopping sce-

nario used by Rook and Fisher (1995) was adapted for this study. Sub-
jects were randomly assigned to either the consonant CI condition, or
the dissonant CI condition. Subjects assigned to the consonant CI con-
dition were given the following scenario: “Mary (Bob) is a 21-year-old
college student. She (He) is well off and generally has plenty of spending
money. Mary (Bob) needs to buy a pair of warm socks for an outdoor
party this weekend. After work, she (he) goes with her (his) friend Susan
(John) to the mall to purchase the socks. As they are walking through
Mervyn’s, Mary (Bob) sees a great-looking sweater on sale for $75 and short
standard

968 DHOLAKIA Base of RF

MAR WILEJ RIGHT BATCH

Top of text
falls in love with it on first sight. Please select which of the following Base of text
five purchase-decision alternatives Mary (Bob) should make.”
On the other hand, subjects in the dissonant CI condition were given
the following scenario: “Mary (Bob) is a 21-year-old college student with
a part-time job. It is two days before she (he) gets the next paycheck
and has only $25 left for necessities. In addition to food, Mary (Bob)
needs to buy a pair of warm socks for an outdoor party this weekend.
After work, she (he) goes with her (his) friend Susan (John) to the mall
to purchase the socks. As they are walking through Mervyn’s, Mary
(Bob) sees a great-looking sweater on sale for $75 and falls in love with
it on first sight. Please select which of the following five purchase-
decision alternatives Mary (Bob) should make, by circling the appro-
priate number.”
Subjects in both groups were asked to select one of five purchase-
decision alternatives. These choice alternatives, similar to Rook and
Fisher (1995), represented varying levels of impulsiveness. From low to
high impulsiveness, these alternatives were (1) buy the socks only; not
even think about the sweater, (2) buy the socks only; want the sweater
but not buy it, (3) decide not to buy the socks and buy the sweater in-
stead, (4) buy both the socks and the sweater with a credit card, and (5)
buy both socks and sweater plus matching slacks and a shirt to complete
the outfit. To control for the possible effects of the stimulus, half of the
subjects were randomly assigned to a scenario with a male protagonist,
Bob. An ANOVA conducted subsequently to evaluate the role of protag-
onist gender found no significant main or interaction effects. Conse-
quently, this variable was dropped from the analysis and is not dis-
cussed further. Subjects were also asked to state their agreement with
the statement: “Mary (Bob) experienced an impulse to purchase the
sweater” on a 5-point scale. This is called the “impulse experience” mea-
sure.

Measurement of the Impulsiveness Trait. Much research in the per-

sonality literature has found the impulsiveness trait to be unidimen-
sional and measurable by relatively short scales (Plutchik & van Praag,
1995). In the context of consumption, Rook and Fisher (1997) developed
a scale, focusing on the individual’s impulsiveness trait with respect to
purchase and use of products. In keeping with the current focus on con-
sumption, this scale was used to measure the subject’s impulsivity trait.
A reliability analysis found the Cronbach alpha for the scale to be 0.89,
with consistently high interitem correlations, indicating a high level of
reliability. This result is consistent with earlier use of this scale (Rook
& Fisher, 1997).

Measurement of Cognitive Evaluation. After subjects had made

their behavioral choice for the scenario, they were told that the protag- short
standard

TEMPTATION AND RESISTANCE 969 Base of RF

MAR WILEJ LEFT BATCH

Top of text
onist chose the most impulsive behavior (i.e., bought both the socks and Base of text
the sweater, as well as matching slacks and shirt to complete the new
outfit). They were asked to evaluate the protagonist’s behavior on a
number of measures, designed to tap the cognitive content of their eval-
uation. A total of 10 bipolar adjective pairs were administered in the
semantic-differential format. These included “rational” – “crazy,” “ac-
ceptable” – “unacceptable,” “agreeable” – “disagreeable,” and “weak” –
“strong.” This constitutes a scale measuring the cognitive evaluation of
impulsive behavior for the subject. An evaluation of scale reliability in-
dicated a Cronbach alpha of 0.92. The cognitive evaluation scale can
therefore be deemed as sufficiently reliable.

RESULTS

As discussed before, both trait impulsivity and cognitive evaluation of

the impulsive behavior were measured with multiple items. Separate
factor analyses were conducted on these two sets of measures to obtain
parsimonious measures of the two antecedent constructs. Both cognitive
evaluation and Rook’s impulsivity yielded single-factor solutions with
the use of the eigenvalue-greater-than-1 criterion. Factor scores were
generated for each variable with the Thompson procedure (Johnston,
1998), and used in subsequent analysis. Given the focus on teasing apart
the effects of trait impulsivity and cognitive evaluation, multicollinear-
ity between the two factor scores is a potential problem. To ensure that
this is not an issue, the bivariate Pearson correlation coefficient between
the factor scores of the impulsivity trait and cognitive evaluation of the
impulsive behavior was computed. It was found to be 0.29, which,
though significant, is moderate, which indicates that multicollinearity
is not problematic to subsequent interpretation of results.
An assumption made in this analysis is that the impulsivity trait of
the consumer is the primary determinant of the consumption impulse
experience for the purchase scenario used. Consequently, a high level
of correspondence between the trait impulsivity and experience of the
CI is a necessary condition to test the research hypotheses as well as to
verify the working of the CIFE model. To ascertain this the bivariate
Pearson correlation was computed for the trait impulsivity factor score
and the impulse experience measure. It was found to be 0.75, indicating
that subjects high in trait impulsivity were more likely to experience
the consumption impulse, and vice versa.
To test the research hypotheses, a general factorial analysis of vari-
ance was carried out, with one factor (CI type: consonant, dissonant)
and two independent variables (trait impulsivity and cognitive evalua-
tion). The dependent variable in the analysis was degree of impulsivity
of the behavior chosen by the subject, measured with the use of the short
ordered 5-point scale discussed before. Results indicated that the main standard

970 DHOLAKIA Base of RF

MAR WILEJ RIGHT BATCH

Top of text
effect of the consonant – dissonant CI factor was significant at F(1,93) ⫽ Base of text
6.33; p ⫽ .014. The trait impulsivity, at F(1,93) ⫽ 5.56, p ⫽ .020, as well
as the cognitive evaluation effects, at F(1,93) ⫽ 23.129, p ⬍ .001 were
significant as well. In addition, the consonant – dissonant CI ⫻ cognitive
evaluation interaction was significant, at F(1,93) ⫽ 4.742, p ⫽ .032)
whereas the trait impulsivity – cognitive evaluation interaction ap-
proached significance, at F(1,93) ⫽ 3.422, p ⫽ .067). The equations for
subjects in the two impulse conditions are as follows.
For the consonant CI condition:

Behavior ⫽ 3.306* ⫹ 0.267* (Cognitive Evaluation)

⫹ 0.417* (Impulsivity)
⫺ 0.07 (Cognitive Evaluation ⫻ Impulsivity). (1)

For the dissonant CI condition:

Behavior ⫽ 2.822* ⫹ 0.709* (Cognitive Evaluation)

⫹ 0.054 (Impulsivity)
⫺ 0.385* (Cognitive Evaluation ⫻ Impulsivity). (2)

The asterisks indicate statistical significance at the ␣ ⫽ 0.05 level of

significance.
The statistical analysis provides strong support to the research hy-
potheses and to the CIFE model presented before. In the consonant CI
condition, the standardized parameter estimates of Eq. (1) indicate that
impulsivity plays a more important role. Similarly, although the influ-
ence of cognitive evaluation on impulsivity of the behavior is significant,
it is less so than the impulsivity trait. These results provide strong sup-
port to Hypothesis 1, and suggest that when the consumer does not
detect constraints to enacting the CI, the strength of antecedent factors
predominates in driving the CI enactment. The more impulsive the con-
sumer, the more likely he or she will be to enact the CI. In this case,
cognitive evaluation of impulsive behavior plays a marginal role in in-
fluencing behavior enactment.
The results are different for the dissonant CI condition. In this case,
standardized parameter estimates obtained from the general factorial
analysis of variance [Eq. (2)] show that although cognitive evaluation
of the impulsive behavior is a significant predictor of behavior enact-
ment, trait impulsivity is not significant. These results provide strong
support to Hypothesis 2. As predicted by the CIFE framework, in the
dissonant CI condition, the cognitive evaluation of the behavior almost
completely drives the enactment of the product purchase CI. Thus, even
consumers who are very impulsive by nature may resist the CI success-
fully in the presence of constraining factors, if they evaluate the behav-
ior negatively. short
Hypothesis 3 aimed to validate the existence of the volitional system standard

TEMPTATION AND RESISTANCE 971 Base of RF

MAR WILEJ LEFT BATCH

Top of text
Base of text

Figure 4. Proportion of subjects choosing impulsive behavior (Study 1).

in the CIFE model, positing that when the dissonant CI is accompanied

by a negative cognitive evaluation of enacting the behavior, the voli-
tional system will resist the CI, resulting in less likelihood of CI enact-
ment. The significant consonant – dissonant CI ⫻ cognitive evaluation
interaction provides strong support to this hypothesis. Consistent with
this, across levels of impulsivity, less than 10% of subjects in the dis-
sonant-CI – negative cognitive evaluation condition engaged in the im-
pulsive behavior, which is significantly lower than any of the other con-
ditions.
In addition, to corroborate these results, the trait impulsivity and
cognitive evaluation scores were categorized into two-level factors. Sub-
jects were assigned to the high-impulsivity level or the low-impulsivity
level depending on whether the subject’s factor score on the Rook’s im-
pulsivity scale fell above or below the sample median. A similar proce-
dure was followed to classify subjects as having positive or negative
cognitive evaluation of the impulsive behavior. In addition, the depen-
dent variable was divided into either impulsive (if the subject chose 3,
4, or 5 in the scenario) or nonimpulsive (if the subject chose 1 or 2). The
proportion of subjects in each cell choosing an impulsive behavior was
computed and is graphically summarized in Figure 4.
The figure visually summarizes the discussion indicating that al-
though trait impulsivity drives impulsive behavior in the consonant CI short
condition, cognitive evaluation is the primary driver in the dissonant CI standard

972 DHOLAKIA Base of RF

MAR WILEJ RIGHT BATCH

Top of text
condition. Moreover, the role of the volitional system is clearly evident Base of text
as well. In general, these results strongly support the research hypoth-
eses, and the CIFE framework for impulsive product purchase behav-
iors.

STUDY 2: EMOTION-LADEN IMPULSIVE CONSUMPTION

The first study tested research hypotheses derived from the CIFE
framework in the context of a product purchase. The constraint here
was a practical one: inadequate money to make the purchase. In this
case, for the dissonant CI condition, it could be argued that the very
nature of the constraint may have made cognitive evaluation necessary,
given the need to consider feasibility of action enactment (e.g., making
a purchase without adequate cash or credit). What happens when the
constraint is less objective, and more emotion laden? This issue arises
in the context of self-control when the consumer has to make a choice
between alternatives providing positive consequences in the short term
but with deleterious long-term impact, and more currently unpleasant
but long-term positive outcome alternatives. Issues of dieting, con-
sumption of alcohol, cigarettes and drugs, as well as engaging compul-
sively in behavior such as gambling, chronic shopping, etc., all fall in
this category of impulsive consumption (see Nataraajan & Goff, 1991,
for a detailed discussion). As discussed before, this impulsive consump-
tion is qualitatively different from the impulsive purchase of products
and services.
In this second study, applicability of the CIFE framework to this type
of impulsive consumption is examined. The researcher’s thesis is that
even when constraining factors are less objective and are laden with
emotion, on perceiving the constraint, the consumer will engage in cog-
nitive evaluation and the volitional system will be activated if this eval-
uation is negative. The cognitive and volitional resistive processes will
then act to dampen the consumption impulse. On the other hand, if no
constraining factors are identified, the strength of the antecedent fac-
tors will determine enactment. The three research hypotheses pre-
sented earlier are tested here in the context of emotion-laden impulsive
consumption. The study is now described.

METHOD

Sample and Data Collection

In exchange for a free tee-shirt, 218 web surfers participated in this
study. These subjects were older (mean age ⫽ 34.2 years) and had short
higher average incomes (M ⫽ $47,000) than the sample for the first standard

TEMPTATION AND RESISTANCE 973 Base of RF

MAR WILEJ LEFT BATCH

Top of text
study. As in the first study, subjects were first asked to select one of a Base of text
set of purchase alternatives in a hypothetical buying scenario. Following
the selection of the behavior, the subjects completed the detailed cog-
nitive evaluation measures, and the Rook impulsivity scale.

Measurement of the Impulsive Behavior. Subjects were randomly

assigned to either the consonant or the dissonant CI conditions. Subjects
assigned to the consonant CI condition were given the following sce-
nario: “Joan is a 26-year-old graduate student. She enjoys exercising
and running and likes to eat health food. On a weekday morning, Joan
decides to have a healthy salad for lunch. After a busy and productive
morning at work, in which she gets a lot accomplished, she goes with
her friend Lynn to Amer’s, her favorite deli. As she is looking though
the display, she sees a mouth-watering tray of strawberry cheesecake,
her favorite dessert, and feels a strong craving for it immediately. Please
select which of the following five options Joan should choose.”
On the other hand, subjects assigned to the dissonant CI condition
were given the following scenario: “Joan is a 26-year-old graduate stu-
dent. She is overweight and likes to eat healthy food in order to keep
her weight in control. On a weekday morning, Joan decides to have a
healthy salad for lunch. After a busy and productive morning at work,
in which she gets a lot accomplished, she goes with her friend Lynn to
Amer’s, her favorite deli. As she is looking though the display, she sees
a mouth-watering tray of strawberry cheesecake, her favorite dessert
and feels a strong craving for it immediately. Please select which of the
following five options Joan should choose.”
Subjects in both groups were asked to select one of five purchase de-
cision alternatives. From low to high impulsiveness, these alternatives
were (1) get the healthy and low-calorie salad for lunch; not even think
about the cheesecake, (2) get the healthy and low-calorie salad for lunch;
want the cheesecake but not get it, (3) decide not to get the salad but
get the cheesecake instead, (4) get both the salad and the cheesecake,
and (5) get both the salad and the cheesecake plus a chicken sandwich
to complete the meal.
The impulse experience measure was elicited here as well by asking
subjects to state their agreement with the statement “Joan (John) ex-
perienced an impulse to get the cheesecake” on a 5-point scale.

Measurement of the Impulsiveness Trait. The 9-item Rook and

Fisher (1995) buying impulsiveness scale was used to measure the im-
pulsiveness trait. The reliability of the scale was sufficiently high, with
a high Cronbach alpha of 0.89, consistent with earlier results.

Measurement of Cognitive Evaluation. After the subjects had made

their behavioral choice for the scenario, they were told that the protag- short
onist chose the most impulsive behavior (i.e., bought the salad and the standard

974 DHOLAKIA Base of RF

MAR WILEJ RIGHT BATCH

Top of text
cheesecake plus a chicken sandwich to complete the meal). They were Base of text
asked to evaluate the protagonist’s behavior using a cognitive evalua-
tion measure, similar to those used in Study 1. Sufficiently high reli-
ability (␣ ⫽ 0.91) was obtained for these measures.

RESULTS

Separate factor analyses were conducted on these two sets of measures

to obtain parsimonious measures of the two antecedent constructs. Both
cognitive evaluation and Rook’s impulsivity yielded single-factor solu-
tions with the eigenvalue-greater-than-1 criterion. Factor scores were
generated for each variable with the use of the Thompson procedure
(Johnson, 1998), and used in subsequent analysis. As in Study 1, the
correlation between the factor scores of the impulsivity trait and cog-
nitive evaluation of the impulsive behavior was computed to evaluate
potential problems of multicollinearity. It was found to be 0.098, and
not statistically significant, suggesting that multicollinearity was not
an issue. As before, the correspondence between trait impulsivity and
the experience of the consumption impulse was evaluated as well. In
this case, the Pearson correlation coefficient for the trait impulsivity
factor score and the impulse experience measure was found to be 0.77,
suggesting that trait impulsivity corresponded highly with experience
of the consumption impulse.
A general factorial analysis of variance was conducted, with one fac-
tor (CI type: consonant, dissonant) and two independent variables (trait
impulsivity and cognitive evaluation). The dependent variable was de-
gree of impulsivity of the behavior chosen by the subject in the scenario.
The main effect of the CI condition (consonance vs. dissonance) factor
was found to be significant, at F(1,204) ⫽ 52.563, p ⬍ .001, as was the
cognitive evaluation at F(1,204) ⫽ 29.193, p ⬍ .001, and the trait im-
pulsivity effect at F(1,204) ⫽ 11.134, p ⫽ .001. In addition, the behavior-
type factor ⫻ cognitive evaluation interaction was also significant, at
F(1,204) ⫽ 4.237, p ⫽ .01. The predictor equations for subjects in the
two impulse conditions are as follows.
For the consonant CI condition:

Behavior ⫽ 3.363* ⫹ 0.238* (Cognitive Evaluation)

⫹ 0.268* (Impulsivity)
⫹ 0.01 (Cognitive Evaluation ⫻ Impulsivity). (3)

For the dissonant CI condition:

Behavior ⫽ 2.349 ⫹ 0.531* (Cognitive Evaluation)

⫹ 0.200* (Impulsivity) short
⫹ 0.046 (Cognitive Evaluation ⫻ Impulsivity). (4) standard

TEMPTATION AND RESISTANCE 975 Base of RF

MAR WILEJ LEFT BATCH

Top of text
Base of text

Figure 5. Proportion of subjects choosing impulsive behavior (Study 2).

The asterisks indicate statistical significance at the ␣ ⫽ 0.05 level of

significance.
The standardized coefficients in Eqs. (3) and (4) above are compara-
ble. In the consonant CI condition, trait impulsivity has the larger co-
efficient, whereas cognitive evaluation has the larger impact on impul-
sive behavior for the dissonant CI condition.
In addition, to further accentuate the results described above, sub-
jects were divided into high trait impulsivity and low trait impulsivity
with the use of a median split on the Rook impulsivity factor score.
Similarly, subjects were assigned to either the positive cognitive eval-
uation or the negative cognitive evaluation cells, based on a median split
along the cognitive-evaluation factor score. Finally, the dependent vari-
able (impulsive behavior) was divided into either impulsive (if the sub-
ject chose 3, 4, or 5 in the scenario) or nonimpulsive (if the subject chose
1 or 2). The proportions of subjects in each cell choosing impulsive be-
havior were computed for each of the eight cells, and are graphically
presented in Figure 5.
The statistical analysis and the graphical summary resoundingly
support the research hypotheses and the main ideas underlying the
CIFE model in the context of emotion-laden impulsive consumption. As
mentioned before, for the consonant CI condition, the standardized pa-
rameter estimates of Eq. (3) from the general factorial analysis of var- short
iance indicate that the individual’s trait impulsivity is the stronger standard

976 DHOLAKIA Base of RF

MAR WILEJ RIGHT BATCH

Top of text
driver of impulsive behavior than cognitive evaluation, providing sup- Base of text
port to Hypothesis 1, for emotion-laden impulsive consumption. In this
case, essentially, strength of the CI will drive subsequent impulse en-
actment.
On the other hand, for the dissonant CI condition, the effects of the
two factors are reversed. The general factorial ANOVA standardized
parameter estimates in Eq. (4) indicate that here, cognitive evaluation
is the more significant driver of impulsive behavior, whereas trait im-
pulsivity is a secondary influence. To test Hypothesis 3, a one-way
ANOVA was conducted with the dissonant-impulse – negative cognitive
evaluation as one factor level, and the other conditions as the second
factor level. The dependent variable was impulsiveness of the behavior.
Results of the ANOVA indicated that subjects in the dissonant-CI – neg-
ative cognitive evaluation condition (M ⫽ 1.81) were engaged in signif-
icantly less impulsive behaviors [F(1,217) ⫽ 89.78; p ⬍ .001] than those
in the other experimental conditions (M ⫽ 3.25). Only about 10.8% of
the subjects in this condition engaged in impulsive behavior, whereas
65.4% of subjects in the other conditions engaged in impulsive behavior.
These results provide strong support to Hypothesis 3.
In general, these results strongly support the CIFE model for emo-
tion-laden impulsive consumption such as that involving in overeating,
drinking, and other compulsive or addictive behaviors.

GENERAL DISCUSSION

Impulsive consumption has attracted attention for the past several dec-
ades because of its pervasiveness, as well as the important practical
implications stemming from it. Identifying certain product attributes,
packaging elements, or display designs as making consumers more sus-
ceptible to purchase impulsively, on which much attention has focused
has great value, as has the identification of correlates of the impulsivity
trait, but these lines of research fail to provide a satisfactory explana-
tion of when and how impulses are formed and enacted by consumers.
Who among us has not experienced the irresistible urge to grab a bar
of chocolate at the supermarket checkout line? Such urges are resisted
successfully by some consumers, but not by others. Even for the same
consumer, the urge may sometimes result in consumption, but may be
resisted successfully at other times. In this research, the researcher
sought to provide a psychologically rich account of processes determin-
ing these outcomes. In two experimental studies, the key postulates of
the theoretical framework were tested and verified in the context of
impulsive product purchases, as well as emotion-laden impulsive con-
sumption.
The theoretical framework presented here has great value in under- short
standing impulsive consumption for several reasons. First, it provides standard

TEMPTATION AND RESISTANCE 977 Base of RF

MAR WILEJ LEFT BATCH

Top of text
a sound basis for conducting theoretical and practical research in this Base of text
important area of consumer research. Indeed, a number of specific op-
portunities for refining and extending the scope of the CIFE model are
discussed later in this section. Second, through articulating the role of
cognitive, motivational, and volitional processes in impulse resistance
and enactment, this research places impulsive behavior on an equal
footing with the more well-researched, goal-directed, and mindless be-
haviors. Third, this research has enormous practical significance as
well. Purveyors of products and services may focus on one or more of
the antecedent factors outlined in the CIFE model to increase the like-
lihood of occurrence of the consumption impulse. Indeed, an important
insight from the CIFE framework is that influence of antecedent factors
is a matter of degree as well as interaction, in formation of the CI. In
other words, an extremely impulsive trait alone may be just as effective
in enabling experience of the consumption impulse as the interaction of
a much less impulsive trait with effective marketing stimuli. Similarly,
multiple marketing stimuli may have the same effect on the consumer.
The practical implication is that modifying multiple antecedent factors
slightly may result in much higher levels of impulsive consumption at
the aggregate level. Similarly, an understanding of resistance strategies
is also useful from a practitioner standpoint. If retailers understand how
and when specific resistance strategies work, they may design specific
tactics to break down the operation of even the most effective resistance
strategies. For instance, a retailer may counter the exposure control
strategy of a resisting consumer by placing the same enticing merchan-
dise at different locations in the store, making it more accessible (and
difficult to get away from).
The theoretical findings and the formulation presented here are use-
ful to consumers as well. For consumers interested in resisting impulses
as well as for social scientists concerned with devising intervention
strategies for addictive and injurious behaviors, this framework pro-
vides specific guidance to enhance the effectiveness of resistance, or to
prevent experience of the consumption impulse in the first place.
Some future directions stemming from the research described here
must also be pointed out. This psychological account of impulse forma-
tion and enactment or resistance is parsimonious for elegance of pre-
sentation, as well as to make its exposition and validation tenable. Such
parsimony necessarily results in less attention on specific aspects. For
instance, one natural area where greater focus may beneficially be be-
stowed is the working of the volitional system and self-resistance strat-
egies. In this article, the focus was on working of these strategies and
their success, and less attention was given to aspects of failure, when
these strategies fail to resist the impulse. Additional research should
seek to examine the short-term and long-term behavioral as well as
psychological consequences of resistance failure. Research on mental short
control failures (e.g., Wegner & Erber, 1993) as well as coping (e.g., standard

978 DHOLAKIA Base of RF

MAR WILEJ RIGHT BATCH

Top of text
Nolan-Hoeksema, 1990) has the potential to enlighten this investiga- Base of text
tion. Expanding on the role of anticipatory emotions, additional re-
search should consider the role of emotions and moods in other aspects
of the impulse formation and enactment process. For instance, given
that a positive mood state is associated with a greater preference for
risk, it is likely that a positive mood may accentuate the impulse-
formation potential of the different antecedent factors.
In the empirical studies, the focus was on the impulsivity trait, one
antecedent of the consumption impulse. Future research would greatly
advance the model presented here through empirically testing the hy-
pothesized relationships through explicit consideration of other ante-
cedent factors. Here, it was demonstrated that consumers indeed inhibit
and resist dissonant impulses when they are evaluated negatively.
However, more attention also needs to be given to the specific operation
of the volitional system and the resistance strategies articulated in the
CIFE model. For instance, individual studies may be designed to study
the working of specific resistance strategies, resulting in a richer un-
derstanding of the impulse resistance process. To conclude, the CIFE
framework represents a starting point and opens up many opportunities
to deepen understanding of a very pervasive and practically important
domain of consumer behavior.

REFERENCES

Bagozzi, R. P., Baumgartner, H., & Pieters, R. (1998). Goal-directed emotions.

Cognition and Emotion, 12, 1 – 26.
Bagozzi, R. P., & Dholakia, U. M. (1999). Goal-setting and goal-striving in con-
sumer behavior. Journal of Marketing, 63, 19 – 32.
Bandura, A. (1991). Social cognitive theory of self-regulation. Organization Be-
havior and Human Decision Processes, 50, 248 – 287.
Bargh, J. A. (1989). Conditional automaticity: Varieties of automatic influence
in social perception and cognition. In J. Uleman & J. Bargh (Eds.), Unin-
tended thought (pp. 3 – 51). New York: Guilford Press.
Bargh, J. A., & Barndollar, K. (1996). Automaticity in action: The unconscious
as repository of chronic goals and motives. In P. Gollwitzer & J. Bargh (Eds.),
The psychology of action: Linking cognition and motivation to behavior (pp.
457 – 481). New York: Guilford Press.
Barrett, E., & Patterson, J. H. (1983). Impulsivity: Cognitive, behavioral and
psychophysiological correlates. In M. Zuckerman (Ed.), Biological bases of
sensation seeking, impulsivity and anxiety (pp. 77 – 116). Hillsdale, NJ:
Lawrence Erlbaum.
Beck, A. T., & Emery, G. (1985). Anxiety disorders and phobias: A cognitive
perspective. New York: Basic Books.
Belk, R. W. (1975). Situational variables and consumer research. Journal of short
Consumer Research, 12, 157 – 164. standard

TEMPTATION AND RESISTANCE 979 Base of RF

MAR WILEJ LEFT BATCH

Top of text
Belk, R. W. (1985). Materialism: Trait aspects of living in the material world. Base of text
Journal of Consumer Research, 12, 265 – 280.
Bettman, J. R. (1979). An information processing theory of consumer choice.
Reading, MA: Addison-Wesley.
Bütz, M., & Austin, S. (1993). Management of the adult impulsive client: Iden-
tification, timing, and methods of treatment. In W. McGown, J. Johnson, &
M. Shore (Eds.), The impulsive client: Theory, research and treatment (pp.
323 – 344). Washington, DC: American Psychological Association.
Carver, C. S., & Scheier, M. S. (1990). Principles of self-regulation: Action and
emotion. In E. T. Higgins & R. Sorrentino (Eds.), Handbook of motivation
and cognition: foundations of social behavior (Vol. 2, pp. 3 – 52). New York:
Guilford Press.
Cobb, C. J., & Hoyer, W. D. (1986). Planned versus impulse purchase behavior.
Journal of Retailing, 62, 67 – 81.
Cox, K. (1964). The responsiveness of food sales to shelf space changes in su-
permarkets. Journal of Marketing Research, 1, 63 – 67.
Dickman, S. J. (1990). Functional and dysfunctional impulsivity: Personality
and cognitive correlates. Journal of Personality and Social Psychology, 58,
95 – 102.
Doob, L. W. (1990). Hesitation: Impulsivity and reflection. New York: Green-
wood Press.
Emmons, R. A., King, L. A., & Sheldon, K. (1993). Goal conflict and the self-
regulation of action. In D. Wegner & J. Pennebaker (Eds.), Handbook of men-
tal control (pp. 528 – 551). Englewood Cliffs, NJ: Prentice Hall.
Gilbert, D. T. (1993). The ascent of man: Mental representation and control of
belief. In D. Wegner & J. Pennebaker (Eds.), Handbook of mental control (pp.
57 – 87). Englewood Cliffs, NJ: Prentice Hall.
Higgins, E. T. (1996). Ideals, oughts and regulatory focus: Affect and motivation
from distinct pains and pleasures. In P. Gollwitzer & J. Bargh (Eds.), The
psychology of action: Linking cognition and motivation to behavior (pp. 91 –
114). New York: Guilford Press.
Hirschman, E. C. (1980). Innovativeness, novelty-seeking, and consumer cre-
ativity. Journal of Consumer Research, 7, 283 – 295.
Hoch, S. J., & Loewenstein, G. F. (1991). Time-inconsistent preferences and
consumer self-control. Journal of Consumer Research, 15, 492 – 507.
Isen, A. M., & Diamond, G. A. (1989). Affect and automaticity. In J. Uleman &
J. Bargh (Eds.), Unintended thought (pp. 124 – 154). New York: Guilford
Press.
Johnson, D. E. (1998). Applied multivariate methods for data analysts. Pacific
Grove, CA: Duxbury Press.
Kanfer, R. (1996). Self-regulatory and other non-ability determinants of skill
acquisition. In P. Gollwitzer & J. Bargh (Eds.), The psychology of action:
Linking cognition and motivation to behavior (pp. 404 – 423). New York: Guil-
ford Press.
Kuhl, J. (1987). Action control: The maintenance of motivational states. In F.
Halisch & J. Kuhl (Eds.), Motivation, intention and volition (pp. 279 – 291).
Berlin: Springer.
Langer, E. J., & Imber, L. (1980). The role of mindlessness in the perception of
deviance. Journal of Personality and Social Psychology, 39, 360 – 367. short
standard

980 DHOLAKIA Base of RF

MAR WILEJ RIGHT BATCH

Top of text
Loewenstein, G. F. (1990). Reference points in intertemporal choice. Unpub- Base of text
lished working paper, Center for Decision Research, University of Chicago.
Lurr, M., & Wunderlich, R. A. (1985). A measure of impulsiveness and its re-
lation to extraversion. Education and Psychological Measurement, 45, 251 –
257.
Mischel, W., Ebbesen, E. G., & Zeiss, A. R. (1972). Cognitive and attentional
mechanisms in delay of gratification. Journal of Personality and Social Psy-
chology, 21, 204 – 218.
Mischel, H. N., & Mischel, W. (1983). The development of children’s knowledge
of self-control strategies. Child Development, 54, 226 – 254.
Murray, H. (1938). Explorations in personality. New York: Oxford University
Press.
Nataraajan, R., & Goff, B. G. (1991). Compulsive buying: Toward a reconcep-
tualization. In F. W. Rudmin (Ed.), To have possessions: A handbook of own-
ership and property. Journal of Social Behavior and Personality (Special Is-
sue), 6, 307 – 328.
Nolan-Hoeksema, S. (1990). Sex differences in depression. Stanford, CA: Stan-
ford University Press.
Oullette, J. A., & Wood, W. (1998). Habit and intention in everyday life: The
multiple processes by which past behavior predicts future behavior. Psycho-
logical Bulletin, 124, 54 – 74.
Patterson, L. W. (1963). In-store traffic flow. New York: Point-of-Purchase Ad-
vertising Institute.
Plutchik, R., & van Praag, H. M. (1995). The nature of impulsivity: definitions,
ontology, genetics and relations to aggression. In E. Hollander & D. Stein
(Eds.), Impulsivity and aggression (pp. 7 – 24). New York: John Wiley & Sons.
Puri, R. (1996). Measuring and modifying consumer impulsiveness: A cost-ben-
efit accessibility framework. Journal of Consumer Psychology, 5, 87 – 113.
Rook, D. W. (1987). The buying impulse. Journal of Consumer Research, 14,
189 – 198.
Rook, D. W., & Fisher, R. J. (1995). Normative influences on impulsive buying
behaviors. Journal of Consumer Research, 22, 305 – 313.
Rook, D. W., & Hoch, S. H. (1985). Consuming impulses. In E. Hirschman &
M. Holbrook (Eds.), Advances in consumer research (Vol. 12, pp. 23 – 27).
Chicago, IL: Association for Consumer Research.
Salovey, P., & Mayer, J. D. (1990). Emotional intelligence. Imagination, Cog-
nition and Personality, 9, 185 – 211.
Schwarz, N., & Bohner, G. (1996). Feelings and their motivational implications:
Moods and the action sequence. In P. Gollwitzer & J. Bargh (Eds.), The psy-
chology of action: Linking cognition and motivation to behavior (pp. 119 –
145). New York: Guilford Press.
Thaler, R. H., & Shefrin, H. M. (1981). An economic theory of self-control. Jour-
nal of Political Economy, 89, 392 – 406.
Vallacher, R. R. (1993). Mental calibration: Forging a working relationship be-
tween mind and action. In D. Wegner & J. Pennebaker (Eds.), Handbook of
mental control (pp. 442 – 472). Englewood Cliffs, NJ: Prentice Hall.
Watson, D., & Clark, L. A. (1993). Behavioral disinhibition versus constraint:
A dispositional perspective. In D. Wegner & J. Pennebaker (Eds.), Handbook
of mental control (pp. 506 – 527). Englewood Cliffs, NJ: Prentice Hall. short
standard

TEMPTATION AND RESISTANCE 981 Base of RF

MAR WILEJ LEFT BATCH

Top of text
Wegner, D. M., & Erber, R. (1993). Social foundations of mental control. In D. Base of text
Wegner & J. Pennebaker (Eds.), Handbook of mental control (pp. 36 – 56).
Englewood Cliffs, NJ: Prentice Hall.
Weinberg, P., & Gottwald, W. (1982). Impulsive consumer buying as a result
of emotions. Journal of Business Research, 10, 43 – 57.
The author would like to thank Kalpesh Desai, Dennis Rook, and Rajan Na-
taraajan for insightful comments on earlier drafts of this article.
Correspondence regarding this article should be sent to: Dr. Utpal M. Dholakia,
Marketing Department, School of Management, Jacobs Management Center,
SUNY at Buffalo, Buffalo, NY 14260 (utpal@adelphia.net).

short
standard

982 DHOLAKIA Base of RF