ArticlePDF Available

Residual Effects of Past on Later Behavior: Habituation and Reasoned Action Perspectives

January 2002
Personality and Social Psychology Review 6(2):107-122

January 2002
6(2):107-122

Authors:

University of Massachusetts Amherst

The frequency with which a behavior has been performed in the past is found to account for variance in later behavior independent of intentions. This is often taken as evidence for habituation of behavior and as complementing the reasoned mode of operation assumed by such models as the theory of planned behavior. In this article, I question the idea that the residual effect of past on later behavior can be attributed to habituation. The habituation perspective cannot account for residual effects in the prediction of low-opportunity behaviors performed in unstable contexts, no accepted independent measure of habit is available, and empirical tests of the habituation hypothesis have so far met with little success. A review of existing evidence suggests that the residual impact of past behavior is attenuated when measures of intention and behavior are compatible and vanishes when intentions are strong and well formed, expectations are realistic, and specific plans for intention implementation have been developed.

Content uploaded by Icek Ajzen

Content may be subject to copyright.

Residual Effects of Past on Later Behavior:

Habituation and Reasoned Action Perspectives

Icek Ajzen

Department of Psychology

University of Massachusetts—Amherst

The frequency with which a behavior has been performed in the past is found to ac-

count for variance in later behavior independent of intentions. This is often taken

as evidence for habituation of behavior and as complementing the reasoned mode

of operation assumed by such models as the theory of planned behavior. In this arti-

cle, I question the idea that the residual effect of past on later behavior can be at-

tributed to habituation. The habituation perspective cannot account for residual ef-

fects in the prediction of low-opportunity behaviors performed in unstable contexts,

no accepted independent measure of habit is available, and empirical tests of the

habituation hypothesis have so far met with little success. A review of existing evi-

dence suggests that the residual impact of past behavior is attenuated when mea-

sures of intention and behavior are compatible and vanishes when intentions are

strong and well formed, expectations are realistic, and specific plans for intention

implementation have been developed.

Past behavior, people are often told, is the best pre-

dictor of future behavior. Human beings are said to be

creatures of habit; they tend to persist in doing what

they have become accustomed to. It is well-known that,

with repeated performance, many behaviors become

routine to the point where they can be executed with

minimal conscious control (Schneider & Shiffrin,

1977; Shiffrin & Schneider, 1977). For most of us,

walking, driving a car, brushing our teeth, getting

dressed, and the myriad of other activities we perform

every day become routines that do not require much fo-

cused attention. Even complex behaviors that are ini-

tially guided by explicit intentions and self-regulation

can, with sufficient repetition and practice, habituate

and become more or less automatic in the sense that

they are performed quickly, outside awareness, with

minimal attention, and in parallel with other activities

(Bargh, 1996; Ouellette & Wood, 1998; Posner &

Snyder, 1975).

Faced with uneven and generally modest success in

attempts to account for human social behavior, it is

tempting to look with envy at the reputed ability of

prior behavior to predict future action. At the very

least, we might want to incorporate past behavior as

one of the major predictors in our theories. This possi-

bility has often been examined in the context of the

theory of reasoned action (Ajzen & Fishbein, 1980;

Fishbein & Ajzen, 1975) or its successor, the theory of

planned behavior (Ajzen, 1988, 1991). Briefly, accord-

ing to the theory of planned behavior, human action is

guided by three kinds of considerations: beliefs about

the likely consequences of the behavior (behavioral

beliefs), beliefs about the normative expectations of

others (normative beliefs), and beliefs about the pres-

ence of factors that may facilitate or impede perfor-

mance of the behavior (control beliefs). In their respec-

tive aggregates, behavioral beliefs produce a favorable

or unfavorable attitude toward the behavior, normative

beliefs result in perceived social pressure or subjective

norm, and control beliefs give rise to perceived behav-

ioral control. In combination, attitude toward the be-

havior, subjective norm, and perception of behavioral

control lead to the formation of a behavioral intention.

As a general rule, the more favorable the attitude and

subjective norm and the greater the perceived control,

the stronger should be the person’s intention to per-

form the behavior in question. Finally, given a suffi-

cient degree of actual control over the behavior, people

are expected to carry out their intentions when the op-

portunity arises. Intention is thus assumed to be the

immediate antecedent of behavior. However, because

many behaviors pose difficulties of execution that may

limit volitional control, it is useful to consider per-

ceived behavioral control in addition to intention. To

the extent that perceived behavioral control is veri-

dical, it can serve as a proxy for actual control and con-

tribute to the prediction of the behavior in question.

Personality and Social Psychology Review

2002, Vol. 6, No. 2, 107–122

Lawrence Erlbaum Associates, Inc.

107

I thank Sheina Orbell and Bas Verplanken for providing unpub-

lished data and information about their research, and Kerry Cour-

neya, Bas Verplanken, and Wendy Wood for their comments on an

earlier draft.

Requests for reprints should be sent to Icek Ajzen, Department of

Psychology, Box 3771—Tobin Hall, University of Massachusetts,

Amherst, MA 01003. E-mail: aizen@psych.umass.edu

It is an undisputed fact that the frequency with

which a behavior has been performed in the past can

be a good predictor of later action. Of greater impor-

tance for my purposes, however, is the finding that

the relation between prior and later behavior is not

fully mediated by the constructs that serve as predic-

tors in the theories of reasoned action or planned be-

havior (Ajzen, 1991; Albarracin, Johnson, Fishbein,

& Meuellerleile, 2001; Bagozzi, 1981; Bentler &

Speckart, 1979; Fredricks & Dossett, 1983; see Con-

ner & Armitage, 1998, and Ouellette & Wood, 1998,

for reviews). A study of exercise behavior (Norman

& Smith, 1995) can serve as an illustration. Un-

dergraduate college students completed a theory of

planned behavior questionnaire on two occasions, 6

months apart. Frequency of exercise reported on the

second occasion (later behavior) was regressed on the

variables contained in the theory of planned behavior

and on exercise frequency reported on the first survey

(prior behavior). The results of a hierarchical regres-

sion analysis revealed a significant (β= .32, p< .01)

residual effect of prior exercise on later exercise.

Without the consideration of past exercise, the theory

of planned behavior variables accounted for 41% of

the variance in exercise behavior, with both intentions

and perceptions of behavioral control making signifi-

cant contributions to the prediction. However, adding

past exercise behavior to the prediction equation

raised the proportion of explained variance to 54%, a

highly significant increase.

Findings of this kind have been reported in a variety

of behavioral domains (see Conner & Armitage, 1998;

Ouellette & Wood, 1998), and they appear inconsistent

with the theory of planned behavior. According to the

theory, measures of intention and perceived behavioral

control should fully mediate the effects of earlier expe-

riences on later action. The finding that past behavior

has a significant residual impact on later behavior con-

tradicts this expectation. In this article, I consider sev-

eral alternative explanations for the residual effect of

prior on later behavior and examine these explanations

in light of published research.

Routinization of Social Behavior

Habituation Versus Reasoned Action

One way to look at the process of routinization is to

assume that repeated performance of a behavior pro-

duces habituation. When a habit develops, behavior is

said to come under the control of stimulus cues (Aarts,

Verplanken, & van Knippenberg, 1998; Ouellette &

Wood, 1998; Ronis, Yates, & Kirscht, 1989). On future

occasions, presence in a similar situation is sufficient

to trigger the automatic response sequence. A stable

stimulus context is therefore crucial for habitual be-

havior to occur, and habit has indeed been defined as

the tendency to repeat past behavior in a stable context

(Ouellette & Wood, 1998).

Routinization of behavior, however, is also consis-

tent with a reasoned action perspective. The theory of

planned behavior does not propose that individuals re-

view their behavioral, normative, and control beliefs

prior to every enactment of a frequently performed be-

havior. Instead, attitudes and intentions—once formed

and well-established—are assumed to be activated au-

tomatically and to guide behavior without the necessity

of conscious supervision (Ajzen & Fishbein, 2000).1

However, whereas the habituation perspective asserts

that routinized behavior is under the control of stimulus

cues, the reasoned action perspective postulates that

such behavior is guided by automatically activated or

spontaneous attitudes and intentions. The distinction

between routinization as reflecting habit formation ver-

sus formation of spontaneous attitudes and intentions is

reminiscent of the exchanges between behaviorists who

adhered to Hull’s (1943) reinforcement theory of learn-

ing and those who supported Tolman’s (1938) purposive

behaviorism. Like the modern habituation perspective,

Hull and his disciples thought of learning in terms of au-

tomatic habits established through reinforcement. In

contrast, and similar to the reasoned action perspective

of routinized behavior I advocate here, Tolman and his

followers preferred to think of learning in terms of men-

tal processes, cognitiveevents, and the formation of tacit

hypotheses and expectations. The implications of the

two perspectives, however, are quite similar: So long as

the context remains relatively unchanged, routinized

behavior is performed in a largely automatic fashion

with minimal conscious control.

Deliberate Versus Spontaneous Modes

of Operation

Currently popular dual-mode processing models of

human judgment, attitude change, and the attitude–be-

havior relation (Fazio, 1990; Petty & Cacioppo, 1986;

see Chaiken & Trope, 1999; Smith & DeCoster, 2000)

imply that behavior can be generated in one of two

ways: It can be guided by conscious deliberation or by

automatic reliance on well-established routines (cf.

Bargh, 1989, 1994; Eagly & Chaiken, 1993; Ouellette

108

AJZEN

1Unlikeimplicitcognitiveresponses (Greenwald & Banaji,1995),

which are said to be unavailableto self-report or introspection and that

can be unrelated to explicit responses (Greenwald, McGhee, &

Schwartz, 1998; Kawakami & Dovidio, 2001), the spontaneous atti-

tudes and intentions in the reasoned action perspective are assumed to

be accessible in memory and equivalent to explicitly expressed atti-

tudes and intentions. This is not to deny the importance of implicit atti-

tudes and intentions. However, the reasoned action approach I advo-

cate here deals only with attitudes and intentions that are available to

introspection, whether effortfully brought to mind or invoked in a

more automatic or spontaneous manner.

& Wood, 1998).2The theory of planned behavior is

consistent with this approach. Novel behaviors and un-

familiar situations are said to evoke careful delibera-

tion and controlled production of beliefs, attitudes, and

intentions that direct subsequent behavior. In contrast,

routine behaviors are assumed to be performed sponta-

neously and to be guided by automatically activated at-

titudes and intentions (see Ajzen & Fishbein, 2000;

Ajzen & Sexton, 1999).

There is abundant evidence to show that attitudes,

intentions, and simple acts can indeed be activated au-

tomatically (Bargh & Chartrand, 1999; Bargh, Chen,

& Burrows, 1996; Fazio, Sanbonmatsu, Powell, &

Kardes, 1986; Gollwitzer, 1999). However, most rou-

tines in everyday life, whether going to the movies or

to a restaurant, are best described as semiautomatic re-

sponse patterns that involve controlled as well as au-

tonomous phases (Abelson, 1981; Bargh, 1989; Logan

& Cowan, 1984; Wegner & Bargh, 1998). Moreover,

even in the case of controlled activities, the amount of

deliberation varies depending on motivation and cog-

nitive capacity. At one end of the information process-

ing continuum are novel or rarely performed behaviors

that require deliberation, formation of an explicit in-

tention, and conscious monitoring for their execution.

At the other extreme are familiar behaviors that have

become automatic as a result of frequent performance

and are now guided by spontaneous attitudes and inten-

tions. From a reasoned action perspective, at both ends

of the continuum behavior is controlled by such cogni-

tive factors as beliefs, attitudes, and intentions—ef-

fortfully at the controlled end and automatically at the

spontaneous end. This view contrasts with the habitua-

tion perspective that assumes that routinized behavior

is under the direct control of stimulus cues.

Is Frequency of Past Behavior a Valid

Indicator of Habit Strength?

According to the habituation perspective, the more

frequently a behavior has been performed in the past,

the more it has come under stimulus control, that is, the

stronger the habit. The tendency for social psycholo-

gists to identify past behavior frequency with habit

strength can perhaps be traced to Triandis’s (1977)

model of interpersonal behavior. According to this

model, the probability that an act will be performed is a

function of intentions and habits (moderated by fa-

cilitating conditions), and habit strength “is measured

by the number of times the act has already been per-

formed by the person” (Triandis, 1977, p. 10). Thus,

the strength of a habit is assumed to grow with repeated

performance of the behavior, and its effect on later be-

havior increases accordingly.

In contrast, the reasoned action perspective sees no

necessary link between frequency of past behavior and

its routinization. Just because a behavior has been per-

formed many times does not, by itself, prove habitua-

tion. No matter how often we may have climbed the

same mountain, it is difficult to believe that this be-

havior has become routine in the sense of constituting

an automatic response sequence. Behaviors of this

kind require conscious control, even after they have

been performed many times. Whether a frequently per-

formed behavior has or has not habituated is an empiri-

cal question, and to answer it we need an independent

and validated measure of habit.

Automaticity in motor behaviors versus auto-

matic decisions. Another problem of equating habit

strength with frequency of past behavior is that this

practice fails to distinguish between automaticity in the

execution of a behavior and automaticity in the deci-

sion to take the action in question. A behavioral inten-

tion is the cognitive representation of a decision to per-

form a given behavior. When the same decision is

made repeatedly, the intention becomes spontaneous;

it is readily accessible in memory and is activated au-

tomatically without conscious intervention. However,

even when an intention is constructed mindfully and

with careful deliberation, the subsequent behavior may

be performed in largely an automatic fashion. Thus, a

person may deliberate at length before settling on a

choice in a restaurant, but once the decision is made,

consumption of the meal may proceed automatically,

without much conscious effort.

The habit construct, as developed in psychology by

learning theorists (e.g., Hull, 1943), was originally ap-

plied to relatively simple behaviors by lower animals,

such as a rat running through a T-maze. The rat’s ten-

dency to make a left turn at the junction can be streng-

thened by placing food at the end of the maze’s left

arm, thus reinforcing the left-turn response. It is the de-

cision to turn left that is reinforced in the T-maze, not

the mechanics involved in running along the pathway

to the goal because these mechanics are the same

whether the rat turns left or right. A large number of re-

inforced trials produces a strong habit to turn left, as in-

dicated by a high probability (relative frequency) of the

left-turn response. However, it should be clear that this

learning process also produces at the same time a low

probability of a turn to the right. That is, the rat devel-

ops not only a strong habit to turn left but also a strong

habit not to turn right. If we recorded the frequency

with which the rat turns right, a low frequency would

not indicate a weak habit. Instead, a weak habit would

be indicated by a pattern of behavior that shows no

clear preference between the left and right options.

109

RESIDUAL EFFECTS OF PAST ON LATER BEHAVIOR

2Most dual-mode processing models assume that motivation and

cognitive capacity determine whether the deliberative or spontane-

ous mode dominates. In contrast, the habituation perspective seems

to assume that the mode of operation employed is a function of the

frequency with which the behavior has been performed in the past (in

a stable context).

This discussion shows that frequency of past behav-

ior may not provide a valid indication of habit strength.

First, the fact that a behavior has been performed many

times is no guarantee that it has habituated. Second, even

if habituation occurred, researchers cannot be sure how

habit strength is related to performance frequency. We

would expect skill level to increase with performance,

but if we want to predict the course of action people will

take, that is, which of several alternatives they will pur-

sue, then frequency of past behavior may not be a useful

indicator of the relevant habit. Drivers who always get in

their cars, start their engines, and begin to drive may de-

velop a strong habit to perform this sequence of events.

The fact that they exhibit a low frequency of wearing a

seat belt does not indicate a weak habit (cf. Mittal,

1988). In a decision-making context, therefore, habit

strength may not be directly proportional to frequency

of past behavior. Instead, it may exhibit a U-shaped

function such that habit strength rises when the fre-

quency of a chosen alternative increases as well as when

the frequency of the chosen alternative decreases.

Can Habit Explain the Correlation

Between Prior and Later Behavior?

Past behavior, even when it has been frequently per-

formed, does not directly control performance of the

behavior on future occasions. Instead, investigators

have proposed that the strong and consistent relation

between past and later behavior is due to the process of

habituation (e.g., Aarts et al.,1998; Bagozzi, 1981;

Bentler & Speckart, 1979; Ouellette & Wood, 1998).

With repeated performance, a habit is assumed to de-

velop, and the habitual behavior is triggered automati-

cally in the presence of controlling stimulus events.

Strictly speaking, the observed correlation between

frequency of prior and later behavior is no more (or

less) than an indication that the behavior in question is

stable over time. In the absence of an independent mea-

sure of the habit construct, using habit to explain the re-

lation between prior and later behavior involves circu-

lar reasoning: One infers the existence of a habit from

the behavior’s temporal stability and then uses the in-

ferred construct to explain the observed phenomenon.

Moreover, reasoned as well as automatic processes can

account for temporal stability of behavior. Thus, be-

havioral stability may be attributable not to habituation

but to the influence of cognitive and motivational fac-

tors that remain unchanged and are present every time

the behavior is observed (Ajzen, 1991; Eagly & Chai-

ken, 1993).3In the theory of planned behavior, the cru-

cial factors of interest are intentions and perceptions of

behavioral control. So long as these factors remain un-

changed, the behavior should also remain the same.

Can Habit Explain the Residual

Impact of Prior on Later Behavior?

Like the behavior’s temporal stability, the residual

impact of past behavior on later behavior has also been

attributed to the process of habituation or routinization.

When the behavior habituates, it presumably comes

under the direct control of stimulus cues and, more im-

portant, intentions and other cognitive factors are said

to lose some of their predictive validity (Aarts et al.,

1998; Ouellette & Wood, 1998). As a result, past be-

havior frequency acquires a residual relation to later

behavior. However, the proposition that habit, as indi-

cated by frequency of past behavior in a stable context,

can explain the residual effect of prior on later behavior

faces a number of difficulties.

1. Residual effects of past on later behavior are

not evidence of habit. The observed residual impact

of prior behavior on later behavior has often been

taken as prima facie evidence for the operation of

habit (e.g., Bentler & Speckart, 1979; see Conner &

Armitage, 1998). However, just as behavioral stabil-

ity does not provide evidence for habituation, the re-

sidual effect of past behavior on later behavior, unme-

diated by intention, is also not sufficient to prove the

operation of habit. Any factor that influenced behav-

ior in the past and that continues to exert an effect at

present could explain the residual effect. Indeed,

there is no shortage of candidates. Questioning the

sufficiency of the theory of planned behavior, investi-

gators have suggested a variety of additional determi-

nants of intentions and behavior, including personal

or moral norms (e.g., Gorsuch & Ortberg, 1983), an-

ticipated regret (e.g., Richard, van der Pligt, & de

Vries, 1995), desire to attain a behavioral goal (Peru-

gini & Bagozzi, 2001), self-identity (e.g., Sparks &

Guthrie, 1998), and affect (Manstead & Parker,

1995). If added to the model, these factors—and

many others—could, in theory, mediate the effect of

prior behavior on later behavior. For example, it is

possible that repeated contributions to the American

Cancer Society produce in an actor a sense of com-

mitment to the cause of fighting cancer. This commit-

ment may then be responsible for continued contribu-

tions on future occasions, over and above intentions.

To show that habit is the crucial mediating factor, it is

not sufficient to point to the residual link between fre-

quency of prior behavior and later behavior. Instead,

evidence needs to be provided that is independent of

the behavior that habit is meant to explain.

2. Low-opportunity behaviors have residual effects.

Because habituation requires repeated enactment of a

behavior, it applies only to high-opportunity activities

110

AJZEN

3Temporal stability of behavior may also be in part a function of

self-perception processes whereby people infer their attitudes from

their past behavior and these attitudes then influence their present in-

tentions and actions (Albarracin & Wyer, 2000; Zanna, Olson, &

Fazio, 1981).

that can be performed repeatedly. However, the residual

effect of past on later behavior has been reported not

only with respect to behaviors of this kind but also with

respect to low-opportunity behaviors. In a study of ex-

tra-relationship involvement, for example, the residual

effect of past on later infidelities was found to be strong

and significant (Drake & McCabe, 2000). Intentions

and perceptions of behavioral control together account-

ed for 22% of the variance in frequency of extra-rela-

tionship involvements over a 6-month period. By com-

parison, past frequency of such involvements by itself

explained 56% of the variance, and it significantly im-

proved the prediction when added to the theory of

planned behavior. Participants reported engaging in ex-

tra-relationship affairs with relatively low frequency,

and it would therefore be difficult to arguethat the resid-

ual effect was due to habituation. Residual effects have

also been reported for other low-opportunity behaviors

such as searching for employment (Van Ryn & Vinokur,

1992) and attending medical checkups (Norman & Con-

ner, 1996).

In conclusion, attempts to attribute residual effects

of past behavior to habituation face serious difficulties.

Because no independent indicators of habit strength

are readily available, investigators have had to rely on

measures of past behavior frequency. However, it is not

clear that past behavior is necessarily a measure of

habit. It may reflect the operation of other factors that

could account equally well for its residual effect on

later behavior. In addition, the hypothesized process of

habituation cannot account for residual effects in the

prediction of low-opportunity behaviors.

A Reasoned Action Approach

to the Residual Effect of Prior

on Later Behavior

As long as the stimulus situation remains stable and

intentions remain unchanged, there is no reason for be-

havior to change, and past behavior should be a good

predictor of later behavior. What requires explanation,

however, is why and how past behavior acquires a di-

rect, residual impact on later behavior, independent of

intentions and perceptions of behavioral control. In a

stable context, one would expect intentions as well as

past behavior to be good predictors of later behavior,

and there is no reason to anticipate that past behavior

will account for variance not explained by intentions.

The habituation perspective attributes the residual ef-

fect to a process of habituation in which control over

the behavior shifts from cognitive factors to environ-

mental cues, but this view fails to describe the changes

in intentions or behavior that would be required to

produce the observed residual effects. Why—if the

situation remains stable—should intentions lose their

predictive validity?

The most likely explanation from a reasoned ac-

tion perspective is that intentions have changed and

that the newly formed intentions are relatively poor

predictors of later behavior. There is general agree-

ment that even performance of a high-opportunity be-

havior is at least initially controlled by deliberate in-

tentions. If the stimulus situation, as well as the

cognitive determinants of a repeated behavior, re-

mained stable over time, there would no reason for

intentions or behavior to change. Under such condi-

tions, current intentions would be perfectly consistent

with past behavior, and they would be sufficient to

predict later action. The presence of a residual impact

of past on later behavior implies that something must

have changed: Later behavior is still consistent with

past behavior, but it is no longer completely in line

with intentions. Assuming that the context has re-

mained the same, the only possible explanation is that

intentions have changed and that the new intentions

are less than perfect predictors of later behavior.

Several reasons can be suggested as to why people

may fail to carry out newly formed intentions (see

Ajzen, 1985). Of limited theoretical interest is the pos-

sibility that they simply forget to enact the new inten-

tion. Thus, people who form the intention to put on a

seat belt after driving without one for many years may

sometimes, “by force of habit,” forget to enact their in-

tentions until they have become accustomed to the new

behavior. There is evidence to show that difficulties of

this kind can be overcome by forming an implemen-

tation intention, that is, an explicit plan as to exactly

when and how the behavior is to be enacted (cf. Goll-

witzer, 1999). I return to the question of implementa-

tion intentions later in this article.

A conceptually more interesting possibility is that

the failure to act in accordance with a newly formed in-

tention is due to factors related to the predictors in the

theory of planned behavior. Among the possible rea-

sons are early indications of detrimental unanticipated

consequences, negative reactions from important ref-

erents, underestimation of the behavior’s difficulty,

and lack of resolve or willpower. “Bad habits” are of-

ten attributed to task difficulty or lack of resolve. Thus,

many smokers intend to quit but fail to carry out their

intentions, and people who intend to maintain a diet to

lose weight often have great difficulty doing so. It

should be noted, however, that these failures do not es-

cape the person’s attention. In fact, they are generally

accompanied by a good deal of anguish and regret. It

would be difficult to argue that, in these instances, the

residual impact of past on later behavior represents the

operation of an automatic habit. People are conscious

of their routine response tendencies—that is, their bad

habits—and their behavior exhibits none of the hall-

marks of automaticity: It is not performed quickly, out-

side awareness, with minimal attention, and in parallel

with other activities.

111

RESIDUAL EFFECTS OF PAST ON LATER BEHAVIOR

In sum, forgetting, lack of resolve, unanticipated

detrimental consequences, and other difficulties may

account for people’s failure to carry out their intentions

and may thus help explain the residual impact of prior

on later behavior. However, this does not provide evi-

dence for habituation of behavior because the failure to

carry out an intention is often a conscious process that,

far from proceeding automatically, can greatly inter-

fere with other activities.

Empirical Tests

of the Habituation Perspective

As noted, direct tests of the hypothesis that habit

mediates the relation between prior and later behavior

frequency require an independent measure of habit

strength. However, lacking such a measure, it is still

possible to derive certain implications from the habit-

uation hypothesis that can be submitted to empirical

test. Thus, it has been suggested that if habit is an ac-

tive influence it should moderate the effect of inten-

tions on behavior. When behavior has come under the

direct control of stimulus cues, intentions and other

cognitive factors should be relatively unimportant.

Under these conditions, frequency of past behavior

should be a good predictor of later behavior, but cog-

nitive and motivational factors such as intentions

should lose their predictive validity (Aarts et al.,

1998; Ouellette & Wood, 1998). This hypothesis has

been tested in different ways.

Moderating Effect

of Past Behavior Frequency

One approach was adopted by Verplanken, Aarts,

van Knippenberg, and Moonen (1998). In this study,

car use during a 7-day period—a high-opportunity be-

havior—was predicted from intentions to use the car

and from frequency of past car use (considered a mea-

sure of habit). Using simple slope analysis (Aiken &

West, 1991), Verplanken et al. (1998) found a signifi-

cant interaction between intentions and frequency of

past behavior. The relation between intention and later

car use was, as predicted, weakest for individuals who

had used the car frequently in the past and strongest for

individuals at the lower end of past car use. Individuals

who reported using the car at an intermediate rate fell

in between the other two groups.

The results of Verplanken et al. (1998), however,

run counter to a great deal of other available evidence.

Whatever else it may signify, the frequency with

which a behavior has been performed in the past is a

good indication of the amount of direct experience

with the behavior, and it is well established that the

ability of attitudes and intentions to predict later be-

havior increases with amount of direct past

experience (see Fazio & Zanna, 1981). In one of their

studies, Fazio & Zanna (1978) used attitudes toward

volunteering to participate in psychological research

to predict actual volunteering behavior, and the atti-

tude–behavior correlation was found to be strongest

for individuals who had participated in a large number

of experiments in the past (top third). The attitude–be-

havior correlation in this group was .42 compared

with a correlation of –.03 for individuals who had par-

ticipated in relatively few experiments in the past

(bottom third).

Similar results emerged in a reanalysis of data from

a study of class attendance (Ajzen & Madden, 1986).

College students’ class attendance was recorded over a

period of 16 class sessions. After the first 8 sessions, a

questionnaire was administered that contained mea-

sures of the constructs in the theory of planned behav-

ior. The first 8 sessions thus served as a measure of past

behavior and the final 8 sessions as the criterion. Par-

ticipants were divided into thirds on the basis of their

past attendance frequency and correlations were com-

puted between intentions and the frequency of later

class attendance separately for each subgroup. Con-

trary to the findings of Verplanken et al. (1998), the in-

tention–behavior correlation increased with frequency

of past attendance, from –.21 in the low-frequency

group to .14 in the intermediate frequency group to .31

in the high-frequency group.

The implications of these moderating effects of past

behavior frequency are far from clear, however. When

dealing with a high-opportunity behavior, such as car

use or class attendance, all participants can potentially

develop a strong habit. I noted earlier, however, that

just as high frequency of past behavior may indicate a

strong habit to perform the behavior, low frequency of

performance may indicate a strong habit not to perform

the behavior or to perform an alternative behavior. It

could be argued, therefore, that to examine the effect of

habit in terms of the moderating effect of prior behav-

ior it may be necessary to compare high and low per-

formance frequency with moderate frequency. In con-

trast to high and low frequency, moderate frequency

indicates an irregular pattern of behavioral perfor-

mance and thus, arguably, low habit strength.

This possibility was examined in another reanalysis

of the data from the class attendance study described

previously (Ajzen & Madden, 1986). Students who re-

vealed patterns of high or low past attendance were

compared with those who came to class on an irregular

basis (intermediate level of past attendance). Contrary

to the revised habit hypothesis, prediction of later be-

havior from intentions was greater for students who

had exhibited consistently high or low attendance in

the past (r= .44, p< .01) than for students who had at-

tended inconsistently (r= .14, ns).

112

AJZEN

Moderating Effects

of Behavior Opportunity

Perhaps a better approach would consider the op-

portunities people have to perform a given behavior.

Frequent opportunities to engage in a behavior in a sta-

ble context should produce a strong habit to perform or

not to perform the behavior in question. If taken as a

measure of habit, past behavior frequency should

therefore be a better predictor of high-opportunity as

opposed to low-opportunity behaviors. Conversely, in-

tentions should be better predictors of low- as com-

pared to high-opportunity behaviors because the for-

mer are presumably not as much under the control of

stimulus cues as are the latter.

In a test of these ideas, Ouellette and Wood (1998)

performed a meta-analysis of 15 data sets that they

classified as dealing with behaviors that can be per-

formed frequently (e.g., seat belt use, coffee drinking,

class attendance) or infrequently (e.g., flu shots, blood

donation, nuclear protest). To test the relative predic-

tive power of intentions and habits, the investigators re-

gressed behavior on intentions and past behavior si-

multaneously. The problem with this approach is that it

relies on the unproven equation of past behavior fre-

quency with habit strength. Whether we assume that

frequently performed behaviors have come under the

control of stimulus cues or remain under the control of

deliberate processes, we would expect that high-oppor-

tunity behaviors will exhibit greater stability over time

than low-opportunity behaviors. This expectation fol-

lows from well-established principles of psychological

measurement. An index of behavioral frequency based

on many observations is more reliable than one based

on relatively few observations. There is extensive em-

pirical evidence to show that a measure of behavior ag-

gregated over many occasions is more stable over time

than one that aggregated over fewer occasions (Ep-

stein, 1979, 1980). The habituation and reasoned ac-

tion perspectives differ only in their expectations

regarding the effect of intentions on later behavior. Ac-

cording to the habituation perspective, intentions

should be less accurate predictors of high-opportunity

as opposed to low-opportunity behaviors. The rea-

soned action perspective, in either its deliberate or

spontaneous mode, would anticipate little difference.

The results provided no clear support for habituation.

As expected on the basis of both perspectives, past be-

havior frequency was a more accurate predictor for

high-opportunity (r= .64) than for low-opportunity (r=

.37) behaviors. However, consistent with the reasoned

action perspective, prediction of later behavior from in-

tentions was found to be quite accurate for both types of

behavior (r= .59 and r= .67 for high- and low-opportu-

nity behaviors, respectively). Moreover, as the investi-

gators acknowledged (Ouellette & Wood, 1998), it is

impossible to derive definite conclusions from this

meta-analysis because the high-opportunity behaviors

differed in substance from the low-opportunity behav-

iors. The two types of behavior may have differed not

only in performance opportunities but also in degree of

importance, familiarity, or other properties that could

account for any observed effects.

Moderating Effect of Context Stability

To address this problem, Ouellette and Wood

(1998) reported the results of primary research that

was designed to demonstrate the moderating effect of

contextual stability on the prediction of a target behav-

ior from intentions and prior behavior. The target be-

haviors selected were two high-opportunity activities:

watching TV and recycling. To estimate stability of the

supporting context, participants were asked to list the

activities (if any) they always performed prior to en-

gaging in each of these behaviors. On the basis of their

responses, they were divided into groups of high and

low context stability.

As expected, the results of this investigation re-

vealed a moderating effect of past behavior frequency,

but the pattern of findings again did not conform to pre-

dictions that would be derived from the habituation

perspective. For such high-opportunity behaviors as

watching TV and recycling, a stable context should al-

low strong habits to be formed, whereas an unstable

context should not. Consequently, one would expect a

high correlation between prior and later behavior for

individuals with a stable supporting context and a low-

er correlation for individuals with an unstable context.

The results only partly confirmed these expectations.

The correlations between prior recycling and recycling

assessed 3 weeks later were .96 in the stable context

and .12 in the unstable context. With respect to TV

watching, however, there was no appreciable differ-

ence (r= .59 and r= .56, respectively). This latter find-

ing implies either that watching TV does not require a

stable context or that the measure of context stability

developed in this study lacked construct validity.

Of greater importance, however, were the correla-

tions between intentions and later behavior. The habit-

uation perspective suggests that intentions should be

relatively good predictors of later behavior in an unsta-

ble context, but in a stable context where the behavior

is presumably under direct control of stimulus cues,

their predictive validity should decline. The reasoned

action perspective would not lead us to expect appre-

ciable differences between stable and unstable con-

texts. The results of the study were inconclusive. With

respect to watching TV, the intention–behavior corre-

lation was higher in the unstable context (r= .63) than

in the stable context (r= .46), although a reanalysis

showed that the difference between these two correla-

113

RESIDUAL EFFECTS OF PAST ON LATER BEHAVIOR

tions was not significant (z= .96). Moreover, there was

little difference with respect to recycling. Here, the

prediction of later behavior from intentions was actu-

ally slightly better in the stable context (r= .48) than in

the unstable context (r= .43). Thus, neither the meta-

analysis performed by Ouellette and Wood (1998) nor

their primary research provides clear support for habit-

uation as an explanation of the residual impact of prior

on later behavior.

Independent Measures of Habit

The research reviewed thus far offers no convincing

evidence to suggest that the residual effect of past on

later behavior is due to the operation of habit. The

problem, in part, is due to the fact that most efforts to

examine this issue have made the arbitrary assumption

that habit can be equated with frequency of past behav-

ior. A more appropriate approach would rely on an

operationalization of habit that is independent of the

behavior it is supposed to explain and predict. Like

measures of such constructs as moral obligation or

self-identity, an independent measure of habit could be

added to the theory of planned behavior, and its role as

a mediator between past and later behavior could then

be tested.

One attempt to develop an independent measure of

habit (Verplanken, Aarts, van Knippenberg, & van

Knippenberg, 1994) is based on the idea that the ac-

cessibility of a behavioral alternative increases with

habit strength, and some research based on this ap-

proach has been summarized in a recent review (Aarts

et al., 1998).4The approach adopted is rooted in the

assumption that habits are akin to behavioral scripts

(Abelson, 1981), that is, to schemas that embody

knowledge of stereotyped event sequences. The mea-

sure, developed by Verplanken et al. (1994), confronts

respondents with a set of alternative behavioral choices

(e.g., different travel modes, such as car, bus, train, bi-

cycle) and asks them to indicate, as quickly as possible,

which option they would select in a number of hypo-

thetical situations (e.g., when going to the beach, visit-

ing friends, etc.). Frequency of choice across situations

is assumed to indicate habit strength.

The scripted behavior index results in highly reli-

able scores that tend to correlate well with frequency of

past behavior and to act in other ways similar to mea-

sures of past behavior (Aarts et al., 1998). However, in

his discussion of behavioral scripts, Abelson (1981)

warned against the equation of scripts with habits. He

noted that “the difference between a script and a habit

is that a script is a knowledge structure, not just a re-

sponse program” (Abelson, 1981, p. 722). More im-

portant, the procedure developed by Arts et al. may

have little to do with scripts or habits. Respondents

were asked to indicate their intentions or tendencies to

perform a particular behavior in different hypothetical

situations. Depending on the exact phrasing, the result-

ing measure is best interpreted as a generalized inten-

tion to perform the behavior or as a report of past be-

havior generalized across situations. The justification

for assuming that it may represent something other

than a generalized intention or past behavior is the in-

struction to participants to respond as quickly as possi-

ble. It is an empirical question whether time pressure

has any effect on responses and, if so, whether the mea-

sure obtained under time pressure is in fact an indicator

of habit strength.

If researchers disregard the measure’s low face va-

lidity, they can use it to examine whether habit, as as-

sessed, can help to account for the residual effect of

prior on later behavior. This was done in a study of

travel mode choice (Bamberg, Ajzen, & Schmidt, in

press) that examined bus use among college students.

In support of the theory of planned behavior, students’

use of the bus was predicted quite accurately from in-

tentions and perceptions of behavioral control. How-

ever, consistent with past research, addition of prior be-

havior resulted in a significant increase in explained

variance. A structural equation analysis was used to ex-

amine the mediating effect of the scripted habit mea-

sure. Prior behavior was found to have strong and sig-

nificant impacts on habit (β= 0.64), as well as on

perceived behavioral control (β= 0.58), on intention

(β = 0.55), and on later behavior (β= 0.28). The direct

path from habit to later behavior, however, was weak

and not significant (β= 0.07), indicating that inclusion

of the scripted behavior index failed to support the as-

sumed mediating role of habit.

Empirical Tests of the Deliberate

Reasoned Action Perspective

The previous review of the literature indicates that

attempts to explain the residual effect of prior on later

behavior in terms of habituation have, so far, not been

very productive. Reliance on frequency of past behav-

ior as an indicator of habit strength is of questionable

validity, and neither the stability of behavior over time

nor the residual impact of prior on later behavior can be

taken as prima facie evidence for habituation. More di-

rect tests of the habituation perspective, examining

whether past behavior frequency moderates the effect

of intentions on later behavior, have also yielded in-

conclusive results. In the remainder of this article, I

114

AJZEN

4Other attempts to assess habit directly have asked respondents

to indicate whether they perform the behavior in question by force

of habit (Wittenbraker, Gibbs, & Kahle, 1983), as a matter of habit

or automatically (Orbell, Blair, Sherlock, & Conner, 2001), or

without awareness (Mittal, 1988). However, these investigations

did not try to test whether habit, as assessed, mediated the effect of

prior on later behavior.

consider alternative explanations for the residual im-

pact of prior on later behavior, explanations involving

the deliberate mode of the reasoned action perspective.

Scale Compatibility

Studies that have dealt with repeated behavior have

usually tried to predict reported performance frequency

(e. g., Norman & Smith, 1995; Verplanken, Aarts, van

Knippenberg, & Moonen, 1998). In a comparable fash-

ion,such studies havealso tended to assess past behavior

in terms of the number of times it has been performed. In

contrast, measures of attitudes, intentions, and percep-

tions of behavioral control usually ask about performing

(or not performing) the behavior, without reference to

frequency. Thus, participants in a study on exercising

may be asked to indicate how often they exercised in the

past 2 weeks and, 2 weeks hence, they may be asked the

same question again to assess their subsequent behavior.

The other variables in the theory of planned behavior,

however, might be assessed by means of 7-point graphic

scales asking participants to rate regular exercise in the

next 2 weeks on an evaluative continuum, to indicate

whether other people expected them to exercise regu-

larly, to judge the difficulty of engaging in regular exer-

cise, and to express their intentions to exercise regularly

in the next 2 weeks. The measure of past behavior is

therefore likely to share method variance with the mea-

sure of later behavior, variance not shared by the mea-

sures of attitude, subjective norm, perceived behavioral

control, or intention (see Ajzen, 1991). Thus, the supe-

rior compatibility between the scales used to measure

past and later behavior may lend greater validity to the

measureof past behavioras compared to measures of the

other constructs (Courneya & McAuley, 1993).

A reanalysis of published data (Courneya & McAu-

ley, 1994) confirms the importance of scale compati-

bility. In this study, participants reported the number of

times they had engaged in physical activity in the past 4

weeks and did so again 4 weeks later. At the first inter-

view, they also indicated their intentions to engage in

physical activity during the next 4 weeks. These inten-

tions were assessed on a 7-point likelihood scale rang-

ing from 1 (extremely unlikely)to7(extremely likely)

and on a numerical scale (the number of times respon-

dents intended to exercise in the next 4 weeks). Clearly,

the numerical scale was more compatible with the

measure of behavior than the likelihood scale.

The data obtained by Courneya and McAuley (1994)

revealed a strong correlation (r= .62, p< .01) between

past and later behavior. To test the residual effect of past

on later behavior, each measure of intention was held

constant in a reanalysis of the data. The correlation be-

tween prior and later behavior was reduced to .55 when

the likelihood measure was held constant, but to .34

when the numerical measure was held constant. The dif-

ference between these two partial correlations is statisti-

cally significant (z= 2.42, p< .01), indicating that a

behavior-compatible measure of intention can indeed

weaken the residual effect of past on later behavior.

However, the remaining partial correlation of .34 was

still statistically significant (p< .01), suggesting that the

residual effect of past behavior could not be explained

completely by lack of scale compatibility.5

Strength of Attitudes and Intentions

Strong attitudes and intentions have a number of

interesting qualities (see Petty & Krosnick, 1995, for re-

views). Of greatest interest for my purposes here, strong

attitudes and intentions are expected to be relatively sta-

ble over time and to predict manifest behavior better

than weak attitudes and intentions. Weak attitudes and

intentions with respect to a behavior most likely reflect

some degree of ambivalence, indifference, or uncer-

tainty. Clearly, such uncertain dispositions fail to pro-

vide clear guides to action. In the case of respondents

with weak dispositions, past behavior may be a better in-

dicator of likely future action than are intentions. These

considerations imply that the residual effect of past be-

havior on later behavior should emerge when people

hold relatively weak attitudes and intentions but not

when their attitudes and intentions are strong.

Some support for this hypothesis can be found in a

study (Conner, Sheeran, Norman, & Armitage, 2000)

that tried to predict attendance at a health screening

from variables in the theory of planned behavior with

the addition of past attendance. Theory of planned be-

havior data were collected by mail survey at two points

in time, separated by 1 year. About 4 weeks after com-

pletion of each survey, participants received an invita-

tion to attend a health screening at their local general

practice in the course of the subsequent month. Patient

records were used to establish actual behavior at each

time point. The first observation of attendance served

as a measure of past behavior. It was taken prior to

administration of the second survey, which was then

used to predict attendance at the second screening—

the measure of later behavior. Stability of intentions

over the 1-year period, a measure of intention strength,

was obtained by computing the absolute difference be-

tween intentions assessed in the two surveys.

The results showed that past attendance at the health

screening had a significant direct effect on later behav-

ior even after the variables of the theory of planned be-

havior had been entered into the regression equation

(β= 0.93, p< .05). However, the effect of prior on later

behavior was found to be moderated by the stability of

intentions (β= –1.08, p< .05). Simple slope analysis

115

RESIDUAL EFFECTS OF PAST ON LATER BEHAVIOR

5In the theory of planned behavior, the predictors of later behav-

ior are intentions and perceptions of behavioral control. It is possible

that adding a measure of perceived behavioral control as a mediator

would further reduce the relation between prior and later behavior.

showed that, when intention stability was low (1 SD

below the mean) or moderate (at the mean), past be-

havior strongly predicted later action (β= 1.96, p< .01

and β= 0.93, p< .01, respectively). However, when the

stability of intentions was high (1 SD above the mean),

past attendance at the health screening failed to predict

later attendance (β= –0.10, ns).

Very similar results were reported in an investiga-

tion (Sheeran, Orbell, & Trafimow, 1999) that pre-

dicted the amount of time students spent studying

during their winter vacation. Theory of planned behav-

ior questionnaires were administered approximately 6

weeks prior to the vacation and again in the last week

prior to the vacation. Stability of intentions was mea-

sured by computing within-subjects correlations be-

tween the two waves across the five items used to as-

sess intentions. Past behavior was the amount of time

participants reported studying prior to the vacation,

and subsequent behavior during the vacation was re-

ported after students returned for the second semester.

Past studying behavior had a significant residual effect

on later behavior (β= 0.36, p< .01), but this effect was

again moderated by intention stability (β= –0.36, p<

.05). A simple slope analysis revealed the expected

pattern. When intentions were stable, it made little dif-

ference how much students had studied in the past;

their studying behavior was very well predicted from

intentions. However, when intentions were unstable,

they were relatively poor predictors of later studying

behavior that was instead consistent with prior behav-

ior.6Strong, stable intentions thus appear to be a pre-

requisite for accurate prediction. When intentions are

strong, a measure of past behavior fails to account for

additional variance in later behavior. However, when

intentions are weak, as indicated by relatively low sta-

bility, past behavior can serve as a useful predictor over

and above intentions.

Belief Veridicality

Another possible explanation for the residual effect

of prior on later behavior is suggested by the stipula-

tion in the theory of planned behavior that perceived

behavioral control can contribute to the prediction of

behavior only to the extent that it accurately reflects a

person’s actual control. Clearly, however, people are

not always accurate in their appraisals. Lack of rele-

vant information can produce inaccurate beliefs, and

there is plenty of evidence in the psychological litera-

ture to indicate that people’s beliefs can also be biased

by a variety of cognitive and motivational processes

(for reviews, see Ajzen & Sexton, 1999; Nisbett &

Ross, 1980). For example, related to the veridicality of

perceived control, people are found to be overly opti-

mistic in estimating the amount of time it will take

them to complete a task (Buehler, Griffin, & Ross,

1994). In a series of studies, Buehler et al. showed that

people are inclined to underestimate task completion

times because they tend to focus on possible comple-

tion scenarios rather than on prior experience with task

completion. Estimates were found to become more re-

alistic when participants were instructed to relate their

predictions to past experiences.

Questions of veridicality can be raised not only with

respect to control beliefs but also with respect to behav-

ioral and normative beliefs. Just as people may

underestimate or overestimate the difficulty of perform-

ing a behavior, they may also have unrealistic

expectations regarding the likely consequences of a

behavior, and they may misperceive what important oth-

ers expect of them. To the extent that people overestimate

the ease of performing a behavior and the favorability of

its outcomes, they are likely to form unrealistic intentions

to perform the behavior in question. When confronted

with the actual behavior, two processes may be set in

motion. First, people may reconsider their decisions, be-

come more realistic in their expectations, and modify

their intentions. Support for this process can be found in

a variety of domains. For example, research on contin-

gent valuation of public goods has documented that peo-

ple consistently overestimate the amount of money they

would be willing to pay for a worthwhile purpose (e. g.,

Brown, Champ, Bishop, & McCollum, 1996). When

they know that they will actually have to make a pay-

ment, indications of willingness to pay tend to be much

lower—a tendency known as hypothetical bias (cf.

Cummings & Taylor, 1999). A similar tendency was ob-

served in the study on exercising discussed previously

(Courneya & McAuley, 1994) in which participants’ in-

tentions were found to overstate the frequency with

which they would engage in physical activity in the next

4 weeks. The actual frequency was, on average, 13.65,

but the intended frequency, M= 15.33, was significantly

higher. Respondents appeared to realize that their inten-

tions were overly optimistic, however. When asked to in-

dicate how often they expected to actually exercise in the

next 4 weeks, the average estimate was 13.51, very close

to the reported behavior.

Such findings are also common in other research

on the intention–behavior relation. For instance, in a

study (Linn, 1965) of racial attitudes and behavior,

White female college students were asked to indicate

their willingness to pose for a photograph with a

Black male that was to be used for a variety of pur-

116

AJZEN

6Sheeran et al. (1999) reported figures showing the results of

simple slope analyses but did not provide the regression coefficients

at different levels of intention stability. They did, however, report the

results of subgroup analyses in which the sample was split at the me-

dian level of intention stability. This analysis confirmed the conclu-

sions reached earlier: Past behavior was a significant predictor only

when intentions were unstable (β= 0.37, p< .01), not when they

were stable (β= 0.09, ns). Conversely, intentions predicted later be-

havior when they were stable (β= 0.58, p< .01) but not when they

were unstable (β= 0.08, ns).

poses. On a later occasion, they were asked to sign

releases of the photograph for these same purposes.

The results revealed inconsistencies between inten-

tions and behaviors because a large proportion of par-

ticipants failed to carry out their favorable intentions.

Postexperimental interviews traced these inconsisten-

cies to the fact that, when expressing their intentions,

many participants had disregarded the likely negative

reactions of family and friends, but such consider-

ations became prominent when they were asked to

actually release the photographs.

These examples suggest that confrontation with the

actual behavioral situation can be sufficient to change

people’s intentions prior to engaging in the behavior. In

other instances, the changes may occur gradually as

people begin to engage in the behavior and receive

feedback contradictory to their unrealistic expecta-

tions. Evidence in support of this process comes from a

study (Doll & Ajzen, 1992) on the effects of direct ex-

perience on the predictive validity of attitudes and in-

tentions. Participants were either given direct experi-

ence playing several video games or they received

secondhand information about the games by watching

recorded sessions. It stands to reason that beliefs about

the games based on personal experience (how enjoy-

able, challenging, or difficult the games are) will tend

to be more realistic than beliefs based on indirect infor-

mation. When they were later given an opportunity to

play the different games, participants in the direct ex-

perience condition displayed less change in their atti-

tudes, perceptions of control, and intentions than did

participants in the indirect experience condition. As

would therefore be expected, attitudes and intentions

were better predictors of game-playing behavior in the

direct as opposed to the indirect experience condition.

In sum, people often bring unrealistic expectations

to a behavioral situation with the result that intentions

are less predictive of actual behavior than they would

be if beliefs were veridical. This is especially likely

when individuals have had little direct experience

with the behavior, but it also seems to occur when

past experience could have produced veridical be-

liefs. People seem to find ways to convince them-

selves that they will be able to do what they did not

manage to accomplish in the past or that they will do

what they know they should be doing. Thus, smokers

intend to quit soon even though past attempts have

failed, weight-conscious individuals intend to adhere

to a strict diet despite past failures, students expect

higher grades than they have received in previous

courses, parents intend to spend more time with their

children even though they have not kept past resolu-

tions of this kind, and so forth.

In comparison to beliefs, attitudes, and intentions,

which may be unrealistically optimistic (or pessimis-

tic), the frequency with which people have performed a

behavior in the past can provide a relatively realistic

estimate of a person’s actual abilities and dispositions.

In other words, past performance or nonperformance

of a behavior may provide information about a per-

son’s likely future behavior that is, at least for some

people, more accurate than are expressed intentions

and perceptions of behavioral control. Therefore, to the

extent that beliefs are in fact unrealistic, a measure of

past behavior can account for systematic variance in

later behavior that is not accounted for by the predic-

tors in the theory of planned behavior.

Several lines of research provide indirect evidence

for the importance of veridicality in beliefs for the pre-

diction of behavior. As noted previously, belief veri-

dicality is likely to increase as a result of direct experi-

ence and, indeed, it has been found consistently that

attitudes formed on the basis of direct experience are

better predictors of later behavior than are attitudes

formed on the basis of secondhand information (Doll

& Ajzen, 1992; Fazio & Zanna, 1978; Fazio & Zanna,

1981). In a similar fashion, the predictive validity of at-

titudes and intentions has been reported to increase

with amount of knowledge about the attitude object

(Davidson, Yantis, Norwood, & Montano, 1985) and

with reflection about it (Snyder & Swann, 1976). A

greater amount of information and increased reflection

are likely to be associated with more realistic expecta-

tions, resulting in better prediction of later behavior.

More direct evidence for the mediating role of be-

lief veridicality comes from a reanalysis of data ob-

tained in a study of academic achievement (Ajzen &

Madden, 1986). At the beginning of the semester and

again toward the end of the semester, college students

enrolled in several different classes completed a ques-

tionnaire that assessed the constructs in the theory of

planned behavior with respect to attaining an “A” in the

course. In addition, the students also estimated the

grade they thought they would get. Accuracy of grade

expectations could be assessed by computing the dis-

crepancy between this expected grade and the actual

grade attained at the end of the semester. To examine

the effect of veridicality, participants were divided at

the median discrepancy score into high- and low-accu-

racy groups. Results of the reanalysis are displayed in

Table 1. It can be seen that intentions were much better

predictors of current grades when participants were

relatively accurate in their expectations (r= .61, p<

.01) than when they were inaccurate (r= –.01, ns). Past

academic performance, indexed by grade point aver-

age, correlated somewhat better with present grade at-

tainment in the low-accuracy group (r= .35, p< .05)

than in the high-accuracy group (r= .23, p< .05). Thus,

when students were unrealistic in their grade expecta-

tions, their intentions failed to predict the grades they

actually attained, but their past accomplishments pro-

vided a relatively good indication of their prospects.

When they were more realistic, their intentions pre-

dicted grade attainment very well and past grades had a

117

RESIDUAL EFFECTS OF PAST ON LATER BEHAVIOR

lower, albeit still significant, correlation with grades

attained in the present course.

Most important for my purposes were the results of

a hierarchical regression analysis in which current

grade attainment was predicted from intentions on the

first step and past grades on the second step. As can be

seen in Table 1, past performance had a significant re-

gression coefficient (β= 0.43, p< .05) in the low-accu-

racy subgroup even after intentions had been entered

into the equation; the regression coefficient for inten-

tions (β= –0.01) was not significant. By contrast, in the

high-accuracy subgroup, the regression coefficient for

past behavior was low and not significant (β= 0.13),

but for intentions it was high and significant (β= 0.61,

p< .01). The residual effect of past on later behavior

was thus significant only for individuals who held rela-

tively unrealistic expectations regarding the grade they

could expect in the course. When expectations were re-

alistic, past behavior lost its predictive validity.

Very similar results were obtained in another re-

analysis of the data from Ajzen and Madden (1986), an

analysis that looked at the veridicality of control be-

liefs. Because inaccurate expectations of control over

attaining an “A” are likely to be revised in the course of

the semester, changes in control beliefs can be taken as

an indication of low veridicality of initial beliefs. In the

reanalysis, amount of change in control beliefs from

the beginning to the end of the semester was computed,

and participants were divided at the median score into

high and low belief accuracy subgroups. Past per-

formance predicted grade attainment at r= .28 (p< .05)

in the low-accuracy subgroup and r= .19 (ns)inthe

high-accuracy group, whereas intentions predicted

grade attainment much better for students with high

belief accuracy (r= .50, p< .01) than for students with

low accuracy (r= .17, ns). As can be seen in Table 1, a

multiple regression analysis revealed the expected pat-

tern. In the low-veridicality subgroup, the regression

coefficients were 0.26 (p< .05) for past performance

and 0.06 (ns) for intentions. In the high-veridicality

subgroup, the regression coefficients were reversed: β

= 0.14 (ns) for past performance and β= 0.50 (p< .01)

for intentions. Clearly, the residual effect of past on

later behavior depended on low veridicality of control

over attaining a good grade in the course.

Behaving in Accordance

With One’s Intentions

The theory of planned behavior assumes that for rel-

atively novel behaviors, people engage in deliberation

before they form an intention to perform or not per-

form the behavior in question. After repeated opportu-

nities for performance, deliberation is no longer re-

quired because the intention is activated spontaneously

in a behavior-relevant situation (see Ajzen & Fishbein,

2000). This view would imply that once a positive in-

tention is formed, it is carried out when the appropriate

opportunity arises. However, Gollwitzer (1993; Goll-

witzer & Bayer, 1999) has shown that considerable

work is often required to carry out an intention. It may

be necessary to collect information in preparation for

the behavior, to obtain needed cooperation by other

people, or simply to remember to enact it. These tasks

are facilitated when people have formed an imple-

mentation intention, that is, a specific plan that deter-

mines when, where, and how to carry out their in-

tended actions. Such plans are particularly important

for intentions that can be acted on in different ways or

intentions that are poorly specified (Bargh, 1990). For-

mation of an implementation intention is said to bring

the behavior under the control of stimulus cues such

that it is activated automatically at the designated time

and place (Gollwitzer & Brandstätter, 1997; Sheeran &

Orbell, 1999).7

Clearly then, the formation of an intention to en-

gage in a particular behavior may not be sufficient for

the behavior to occur and intentions may turn out to be

poor predictors of behavior, especially for people who

have not formed an implementation intention. These

individuals may fall back on their common responses

to the situation; that is, their behavior may be more

consistent with prior behavior than with current inten-

tions. However, individuals who have formed not only

an intention to engage in a behavior but also a specific

plan for its implementation should exhibit a different

pattern. These individuals should carry out their inten-

118

AJZEN

Table 1. Predicting of Grade Attainment From Past

Performance and Intentions in Low- and High-Accuracy

Subgroups

Variable rβR

Grade Expectations

Low Accuracy

Past Performance .35* .43*

Intentions –.01 –.01 .39*

High Accuracy

Past Performance .23* 0.13

Intentions .61** .61** .61**

Perceived Behavioral Control

Low Accuracy

Past Performance .28* .26*

Intentions .17 .17 .29*

High Accuracy

Past Performance .19 .14

Intentions .50** .50** .52**

Note: Secondary analysis of data published in Ajzen & Madden

(1986).

*p< .05. **p< .01.

7In contrast to the habituation hypothesis, which requires repeated

performance in a stable context for the establishment of a habit,

Gollwitzer’s (1993) theory stipulates that one-time institution of an

implementation intention is sufficientto produce automatic elicitation

of the intended behavior under the designated circumstances.

tions; their past behavior should make little if any addi-

tional contribution to the prediction of later behavior.

This hypothesis was confirmed in a study of breast

self-examination among female students and adminis-

trative staff (Orbell, Hodgkins, & Sheeran, 1997). A

theory of planned behavior questionnaire was adminis-

tered and participants also reported on breast self-ex-

amination during the preceding month as a measure of

prior behavior. About 1 month later, the participants re-

ported their behavior during the preceding month. This

second self-report constituted a measure of later be-

havior. Immediately following the theory of planned

behavior questionnaire, a subgroup of women in the

intervention condition was asked to write down where

and when they would perform breast self-examina-

tions, that is, to develop implementation intentions.

This intervention was found to be highly effective. At

the end of the 1-month follow-up, 64% of women who

had established implementation intentions reported

that they had performed the self-examinations as op-

posed to only 14% in the no intervention control group.

All women in the implementation intention group who

had indicated clear intentions to perform the examina-

tion reported actually doing so, whereas in the control

group, of the women who intended to perform the ex-

amination, only 53% actually did. Most important for

my purposes, however, is the effect of forming an im-

plementation intention on the correlation between pri-

or and later behavior. After the constructs of the theory

of planned behavior had been entered into the regres-

sion equation, the addition of past behavior improved

prediction of later behavior in the no intervention con-

trol group (β= 1.00, p< .01) but not in the implementa-

tion intention group (β= 0.18, ns). Thus, forming an

implementation intention improved prediction of be-

havior from intentions and reduced the residual effect

of past behavior to nonsignificant. When intentions

were not accompanied by specific action plans they

were less accurate predictors of later behavior and the

measure of past behavior retained a residual impact.

Summary and Conclusions

Human behavior is often remarkably stable over

time to the point of overshadowing the effects of atti-

tudes and intentions. Although intentions are generally

good predictors of behavior, some people fail to carry

out their intentions and instead revert to past patterns of

behavior. The usual explanation for this phenomenon

is that the behavior in question has become habitual,

has come under the control of stimulus cues, and no

longer conforms to intentions. Evidence commonly

cited in support of this idea is the finding that the fre-

quency with which a behavior was performed in the

past tends to account for variance in later behavior,

over and above the influence of attitudes, intentions,

and other variables included in the theory of planned

behavior.

Whether we adopt the habituation or reasoned ac-

tion perspective, we would expect that, so long as the

situation remains stable, a behavior that has been per-

formed frequently in the past is likely to be performed

again. There is also general agreement that frequently

performed behaviors can become habitual or routine

and be enacted without much conscious attention. This

is not to say, however, that the residual impact of prior

on later behavior must necessarily be attributed to ha-

bituation. In fact, “the role of habit per se remains inde-

terminate in this research because of the difficulty of

designing adequate measures of habit” (Eagly & Chai-

ken, 1993, p. 181). It would be erroneous to simply

assume that past performance frequency is a valid mea-

sure of habit. Just because a behavior has been per-

formed many times, it does not follow that it is now au-

tomatically activated or that it occurs below conscious

awareness and with little effort—the usually accepted

criteria of automaticity (Bargh, 1994). Whether or not

a behavior has become routine to the point of being

an automatic habitual response is an empirical ques-

tion, and it requires an independent measure of habit

for confirmation. Although some attempts have been

made to develop such a measure, the evidence thus far

provides little support for the proposition that the re-

sidual effect of prior on later behavior is indeed medi-

ated by habit.

In fact, it is not at all clear how the concept of habit

could account for the residual effect of prior on later

behavior. Generally speaking, when a behavior habitu-

ates, it turns from being cognitively controlled and

guided by deliberate intentions to routine and requiring

little cognitive effort for its initiation and execution.

However, if people’s behavior is initially guided by ex-

plicit intentions and remains consistent over time, then

why should intentions, even if they become spontane-

ous, lose their predictive validity? Once a behavior has

become routine or habitual, the frequency of past be-

havior should be a good predictor of later behavior, but

it should not eclipse the impact of intentions.

Strictly speaking, the association between past be-

havior frequency and frequency of later behavior, by it-

self, merely demonstrates that the behavior in question

is stable over time. The reason for this stability may

simply be that whatever factors determined the behav-

ior in the past continue to exert their influence in the

present (see Ajzen, 1991; Eagly & Chaiken, 1993). In

the context of the theory of planned behavior, these

factors are said to be intention and perceived control

over the behavior. To have a residual effect on later be-

havior, frequency of past behavior must reflect the in-

fluence of factors not adequately captured in measures

of these determinants. In this article, I examined sev-

eral possibilities that, individually or in combination,

may be able to account for the residual effect of prior

119

RESIDUAL EFFECTS OF PAST ON LATER BEHAVIOR

on later behavior. At least in part, the residual effect

may be due to issues of measurement. In most empiri-

cal investigations, measures of prior and later behavior

have common method variance not shared by measures

of intentions and other dispositions. The only relevant

data set available shows that scale incompatibility can

account for some, but not all, of the residual associa-

tion between prior and later behavior.

Other possible explanations attribute the residual

effect of past behavior to factors of a more substantive

nature. First, when attitudes and intentions are held

with a degree of ambivalence, indifference, or uncer-

tainty, they are unstable and fail to provide clear guides

to action. Empirical evidence shows that, under these

conditions, past behavior is a good predictor of later

behavior. The residual effect of prior on later behavior

disappears when attitudes and intentions are strong and

well formed. Second, the expectations embodied in be-

havioral, normative, and control beliefs may not be

particularly accurate. In the theory of planned behav-

ior, these beliefs provide the cognitive foundation for

attitudes, intentions, and behavior, but if they are inac-

curate they will fail the test of reality. As a result, the

predictive validity of intentions and perceived behav-

ioral control will suffer. Past performance frequency

reflects people’s preexisting response dispositions.

That is, it reflects the operation of all factors that

exerted an influence on the behavior in the past. To the

extent that people are insufficiently cognizant of these

factors—that is, to the extent that their beliefs lack

veridicality—a measure of past behavior can provide

the information needed to improve prediction of later

behavior. Empirical evidence supports this proposition

by showing that the residual effect of prior on later be-

havior is eliminated for respondents with relatively ac-

curate control beliefs and realistic intentions. Finally,

carrying out an intention often requires development of

a detailed plan that specifies when, where, and how the

behavior is to be enacted. The available evidence sug-

gests that when such implementation intentions are in-

duced, the residual effect of prior on later behavior is

no longer significant.

In conclusion, empirical evidence for a residual ef-

fect of prior on later behavior, controlling for inten-

tions, has raised a number of important practical and

conceptual issues. From a practical perspective, it im-

plies that we as researchers may want to include a mea-

sure of prior behavior in our models to improve predic-

tion of later action. My review suggests that this will be

particularly valuable when intentions are relatively

weak and unstable, when underlying expectations are

inaccurate, or when people have not developed a clear

plan of action. From a theoretical perspective, how-

ever, past behavior frequency adds little to our under-

standing of a behavior’s determinants. Its correlation

with later behavior merely provides a measure of the

behavior’s temporal stability. The attempt to imbue

past behavior frequency with theoretical significance

by considering it a measure of habit strength has, so far,

been unsuccessful. Rather than treating the residual ef-

fect of prior on later behavior as evidence for habitua-

tion, it may be more useful to take it as an opportunity

to explore the factors involved in going from beliefs to

actions. In this review, I suggest that the limits of rea-

soned action are not the habituation of behavior with

repeated performance but may instead be related to in-

accurate or unrealistic behavioral, normative, and con-

trol beliefs; weak or unstable attitudes and intentions;

and inadequate planning required for successful imple-

mentation of an intended behavior.

References

Aarts, H., Verplanken, B., & van Knippenberg, A. (1998). Predicting

behavior from actions in the past: Repeated decision making or

a matter of habit? Journal of Applied Social Psychology, 28,

1355–1374.

Abelson, R. P. (1981). Psychological status of the script concept.

American Psychologist, 36, 715–729.

Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and

interpreting interactions. Newbury Park, CA: Sage.

Ajzen, I. (1985). From intentions to actions: A theory of planned

behavior. In J. Kuhl & J. Beckman (Eds.), Action-control:

From cognition to behavior (pp. 11–39). Heidelberg, Ger-

many: Springer.

Ajzen, I. (1988). Attitudes, personality, and behavior. Chicago:

Dorsey.

Ajzen, I. (1991). The theory of planned behavior. Organizational Be-

havior and Human Decision Processes, 50, 179–211.

Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and pre-

dicting social behavior. Englewood Cliffs, NJ: Prentice Hall.

Ajzen, I., & Fishbein, M. (2000). Attitudes and the attitude-behavior

relation: Reasoned and automatic processes. In W. Stroebe &

M. Hewstone (Eds.), European review of social psychology (pp.

1–33). Chichester, England: Wiley.

Ajzen, I., & Madden, T. J. (1986). Prediction of goal-directed behav-

ior: Attitudes, intentions, and perceived behavioral control.

Journal of Experimental Social Psychology, 22, 453–474.

Ajzen, I., & Sexton, J. (1999). Depth of processing, belief congru-

ence, and attitude–behavior correspondence. In S. Chaiken &

Y. Trope (Eds.), Dual-process theories in social psychology

(pp. 117–138). New York: Guilford.

Albarracin, D., Johnson, B. T., Fishbein, M., & Meuellerleile, P. A.

(2001). Theories of reasoned action and planned behavior as

models of condom use: A meta-analysis. Psychological Bulle-

tin, 127, 142–161.

Albarracin, D., & Wyer, R. S., Jr. (2000). The cognitive impact of

past behavior: Influences on beliefs, attitudes, and future be-

havioral decisions. Journal of Personality & Social Psychology,

79, 5–22.

Bagozzi, R. P. (1981). Attitudes, intentions, and behavior: A test of

some key hypotheses. Journal of Personality & Social Psychol-

ogy, 41, 607–627.

Bamberg, S., Ajzen, I., & Schmidt, P. (in press). Choice of travel

mode in the theory of planned behavior: The roles of past be-

havior, habit, and reasoned action. Journal of Basic and Applied

Social Psychology.

Bargh, J. A. (1989). Conditional automaticity: Varieties of automatic

influence in social perception and cognition. In J. S. Uleman &

J. A. Bargh (Eds.), Unintended thought (pp. 3–51). New York:

Guilford.

120

AJZEN

Bargh, J. A. (1990). Auto-motives: Preconscious determinants of so-

cial interaction. In E. T. Higgins & R. M. Sorrentino (Eds.),

Handbook of motivation and cognition: Foundations of social

behavior (pp. 93–130). New York: Guilford.

Bargh, J. A. (1994). The four horsemen of automaticity: Awareness,

intention, efficiency, and control in social cognition. In R. S.

Wyer, Jr. & T. K. Srull (Eds.), Handbook of social cognition

(2nd ed., Vol. 1, pp. 1–40). Hillsdale, NJ: Lawrence Erlbaum

Associates, Inc.

Bargh, J. A. (1996). Automaticity in social psychology. In E. T. Hig-

gins & A. W. Kruglanski (Eds.), Social psychology: Handbook

of basic principles (pp. 169–183). New York: Guilford.

Bargh, J. A., & Chartrand, T. L. (1999). The unbearable automaticity

of being. American Psychologist, 54, 462–479.

Bargh, J. A., Chen, M., & Burrows, L. (1996). Automaticity of social

behavior: Direct effects of trait construct and stereotype activa-

tion on action. Journal of Personality and Social Psychology,

71, 230–244.

Bentler, P. M., & Speckart, G. (1979). Models of attitude-behavior

relations. Psychological Review, 86, 452–464.

Brown, T. C., Champ, P. A., Bishop, R. C. & McCollum, D. W.

(1996). Which response format reveals the truth about dona-

tions to a public good? Land Economics,72, 152–166.

Buehler, R., Griffin, D., & Ross, M. (1994). Exploring the “planning

fallacy”: Why people underestimate their task completion times.

Journal of Personality & Social Psychology, 67, 366–381.

Chaiken, S., & Trope, Y. (Eds.). (1999). Dual-process theories in so-

cial psychology. New York: Guilford.

Conner, M., & Armitage, C. J. (1998). Extending the theory of

planned behavior: A review and avenues for further research.

Journal of Applied Social Psychology, 28, 1429–1464.

Conner, M., Sheeran, P., Norman, P., & Armitage, C. J. (2000). Tem-

poral stability as a moderator of the relationships in the theory

of planned behavior. British Journal of Social Psychology, 39,

469–494.

Courneya, K. S., & McAuley, E. (1993). Predicting physical activity

from intention: Conceptual and methodological issues. Journal

of Sport and Exercise Psychology, 15, 50–62.

Courneya, K. S., & McAuley, E. (1994). Factors affecting the inten-

tion-physical activity relationship: Intention versus expectation

and scale correspondence. Research Quarterly for Exercise and

Sport, 65, 280–285.

Cummings, R. G., & Taylor, L. O. (1999). Unbiased value esti-

mates for environmental goods: A cheap talk design for the

contingent valuation method. American Economic Review, 89,

649–665.

Davidson, A. R., Yantis, S., Norwood, M., & Montano, D. E. (1985).

Amount of information about the attitude object and attitude-

behavior consistency. Journal of Personality and Social Psy-

chology, 49, 1184–1198.

Doll, J., & Ajzen, I. (1992). Accessibility and stability of predictors

in the theory of planned behavior. Journal of Personality & So-

cial Psychology, 63, 754–765.

Drake, C. R., & McCabe, M. P. (2000). Extrarelationship involve-

ment among heterosexual males: An explanation based on the

theory of planned behavior, relationship quality, and past be-

havior. Jounal of Applied Social Psychology, 30, 1421–1439.

Eagly, A. H., & Chaiken, S. (1993). The psychology of attitudes. Fort

Worth, TX: Harcourt Brace.

Epstein, S. (1979). The stability of behavior: I. On predicting most of

the people much of the time. Journal of Personality & Social

Psychology, 37, 1097–1126.

Epstein, S. (1980). The stability of behavior: II. Implications for psy-

chological research. American Psychologist, 35, 790–806.

Fazio, R. H. (1990). Multiple processes by which attitudes guide be-

havior: The MODE model as an integrative framework. In M. P.

Zanna (Ed.), Advances in experimental social psychology (Vol.

23, pp. 75–109). San Diego, CA: Academic.

Fazio, R. H., Sanbonmatsu, D. M., Powell, M. C., & Kardes, F. R.

(1986). On the automatic activation of attitudes. Journal of Per-

sonality and Social Psychology, 50, 229–238.

Fazio, R. H., & Zanna, M. P. (1978). Attitudinal qualities relating to

the strength of the attitude-behavior relationship. Journal of Ex-

perimental Social Psychology, 14, 398–408.

Fazio, R. H., & Zanna, M. P. (1981). Direct experience and atti-

tude-behavior consistency. In L. Berkowitz (Ed.), Advances in

experimental social psychology (Vol. 14, pp. 161–202). New

York: Academic.

Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and be-

havior: An introduction to theory and research. Reading, MA:

Addison-Wesley.

Fredricks, A. J., & Dossett, D. L. (1983). Attitude-behavior rela-

tions: A comparison of the Fishbein–Ajzen and the Bentler–

Speckart models. Journal of Personality & Social Psychology,

45, 501–512.

Gollwitzer, P. M. (1993). Goal achievement: The role of intentions.

European Review of Social Psychology, 4, 141–185.

Gollwitzer, P. M. (1999). Implementation intentions: Strong effects

of simple plans. American Psychologist, 54, 493–503.

Gollwitzer, P. M., & Bayer, U. (1999). Deliberative versus im-

plemental mindsets in the control of action. In S. Chaiken & Y.

Trope (Eds.), Dual-process theories in social psychology (pp.

403–422). New York: Guilford.

Gollwitzer, P. M., & Brandstätter, V. (1997). Implementation inten-

tions and effective goal pursuit. Journal of Personality and So-

cial Psychology, 73, 186–199.

Gorsuch, R. L., & Ortberg, J. (1983). Moral obligation and attitudes:

Their relation to behavioral intentions. Journal of Personality &

Social Psychology, 44, 1025–1028.

Greenwald, A. G., & Banaji, M. R. (1995). Implicit social cognition:

Attitudes, self-esteem, and stereotypes. Psychological Review,

102, 4–27.

Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. K. (1998). Mea-

suring individual differences in implicit cognition: The implicit

association test. Journal of Personality and Social Psychology,

74, 1464–1480.

Hull, C. L. (1943). Principles of behavior: An introduction to behav-

ior theory. New York: Appleton-Century-Crofts.

Kawakami, K., & Dovidio, J. F. (2001). The reliability of implicit stereo-

typing. Personality and Social Psychology Bulletin, 27, 212–225.

Linn, L. S. (1965). Verbal attitudes and overt behavior: A study of ra-

cial discrimination. Social Forces, 43, 353–364.

Logan, G. D., & Cowan, W. B. (1984). On the ability to inhibit

thought and action: A theory of an act of control. Psychological

Review, 91, 295–327.

Manstead, A. S. R., & Parker, D. (1995). Evaluating and extending

the theory of planned behaviour. In W. Stroebe & M. Hewstone

(Eds.), European review of social psychology (Vol. 6, pp.

69–95). Chichester, England: Wiley.

Mittal, B. (1988). Achieving higher seat belt usage: The role of habit

in bridging the attitude-behavior gap. Journal of Applied Social

Psychology, 18, 993–1016.

Nisbett, R., & Ross, L. (1980). Human inference: Strategies and short-

comings of social judgment. Englewood Cliffs, NJ: Prentice Hall.

Norman, P., & Conner, M. (1996). Predicting health-check atten-

dance among prior attenders and nonattenders: The role of prior

behavior in the theory of planned behavior. Journal of Applied

Social Psychology, 26, 1010–1026.

Norman, P., & Smith, L. (1995). The theory of planned behaviour

and exercise: An investigation into the role of prior behaviour,

behavioural intentions and attitude variability. European Jour-

nal of Social Psychology, 12, 403–415.

Orbell, S., Blair, C., Sherlock, K., & Conner, M. (2001). The theory

of planned behavior and ecstasy use: Roles for habit and per-

ceived control over taking versus obtaining substances. Journal

of Applied Social Psychology, 31, 31–47.

121

RESIDUAL EFFECTS OF PAST ON LATER BEHAVIOR

Orbell, S., Hodgkins, S., & Sheeran, P. (1997). Implementation in-

tentions and the theory of planned behavior. Personality & So-

cial Psychology Bulletin, 23, 945–954.

Ouellette, J. A., & Wood,W. (1998). Habit and intention in every-

day life: The multiple processes by which past behavior pre-

dicts future behavior. Psychological Bulletin, 124, 54–74.

Perugini, M., & Bagozzi, R. P. (2001). The role of desires and antici-

pated emotions in goal-directed behaviours: Broadening and

deepening the theory of planned behaviour. British Journal of

Social Psychology,40, 79–98.

Petty, R. E., & Cacioppo, J. T. (1986). Communication and persua-

sion: Central and peripheral routes to attitude change. New

York: Springer-Verlag.

Petty, R. E., & Krosnick, J. A. (Eds.). (1995). Attitude strength: Ante-

cedents and consequences (Vol. 4). Mahwah, NJ: Lawrence

Erlbaum Associates, Inc.

Posner, M. I., & Snyder, C. R. R. (1975). Attention and cognitive

control. In R. L. Solso (Ed.), Information processing and cogni-

tion: The Loyola symposium (pp. 55–85). Hillsdale, NJ: Law-

rence Erlbaum Associates, Inc.

Richard, R., van der Pligt, J., & de Vries, N. K. (1995). The impact of

anticipated affect on (risky) sexual behaviour. British Journal of

Social Psychology, 34, 9–21.

Ronis, D. L., Yates, J. F., & Kirscht, J. P. (1989). Attitudes, decisions,

and habits as determinants of repeated behavior. In A. R. Prat-

kanis, S. J. Breckler, & A. G. Greenwald (Eds.), Attitude structure

and function (pp. 213–239). Hillsdale, NJ: Lawrence Erlbaum As-

sociates, Inc.

Schneider, W., & Shiffrin, R. M. (1977). Controlled and auto-

matic human information processing: I. Detection, search,

and attention. Psychological Review, 84, 1–66.

Sheeran, P., & Orbell, S. (1999). Implementation intentions and

repeated behaviour: Augmenting the predictive validity of

the theory of planned behaviour. European Journal of So-

cial Psychology, 29, 349–369.

Sheeran, P., Orbell, S., & Trafimow, D. (1999). Does the temporal

stability of behavioral intentions moderate intention-behavior

and past behavior-future behavior relations? Personality and

Social Psychology Bulletin, 25, 721–730.

Shiffrin, R. M., & Schneider, W. (1977). Controlled and automatic hu-

man information processing: II. Perceptual learning, automatic at-

tendingand a general theory.Psychological Review,84, 127–190.

Smith, E. R., & DeCoster, J. (2000). Dual-process models in social

and cognitive psychology: Conceptual integration and links to

underlying memory systems. Personality & Social Psychology

Review, 4, 108–131.

Snyder, M., & Swann, B. W., Jr. (1976). When actions reflect atti-

tudes: The politics of impression management. Journal of Per-

sonality & Social Psychology, 34, 1034–1042.

Sparks, P., & Guthrie, C. A. (1998). Self-identity and the theory of

planned behavior: A useful addition or an unhelpful artifice?

Journal of Applied Social Psychology, 28, 1393–1410.

Tolman, E. C. (1938). The determiners of behavior at a choice point.

Psychological Review, 45, 1–41.

Triandis, H. C. (1977). Interpersonal behavior. Monterey, CA:

Brooks/Cole.

Van Ryn, M., & Vinokur, A. D. (1992). How did it work? An exami-

nation of the mechanisms through which an intervention for the

unemployed promoted job-search behavior. American Journal

of Community Psychology, 20, 577–597.

Verplanken, B., Aarts, H., van Knippenberg, A., & Moonen, A.

(1998). Habit versus planned behavior: A field experiment.

British Journal of Social Psychology, 37, 111–128.

Verplanken, B., Aarts, H., van Knippenberg, A., & van

Knippenberg, C. (1994). Attitude versus general habit: Ante-

cedents of travel mode choice. Journal of Applied Social Psy-

chology, 24, 285–300.

Wegner, D. M., & Bargh, J. A. (1998). Control and automaticity in

social life. In D. T. Gilbert, S. T. Fiske, & L. Gardner (Eds.), The

handbook of social psychology (4th ed., Vol. 1, pp. 446–496).

Boston, MA: McGraw-Hill.

Wittenbraker, J., Gibbs, B. L., & Kahle, L. R. (1983). Seat belt atti-

tudes, habits, and behaviors: An adaptive amendment to the

Fishbein model. Journal of Applied Social Psychology, 13,

406–421.

Zanna, M. P., Olson, J. M., & Fazio, R. H. (1981). Self-perception

and attitude-behavior consistency. Personality & Social Psy-

chology Bulletin, 7, 252–256.

122

AJZEN

Factors Influencing Tax Declaration Decisions on Social Networks of Online Retailers in Vietnam

Article

Feb 2024

The growth of e-commerce and online retail in Vietnam, particularly on social networks, has presented challenges for the government’s tax authorities. Preventing tax loss through inspections can be difficult due to missing or incomplete taxpayer data from transaction processing servers. Explaining undeclared or incompletely declared revenue in online trading transactions on the server or the Internet can be challenging, especially for retail activities and service provision for businesses and individuals selling on social networking. In Vietnam, the Tax Management Law 2019 established the tax rates online retailers must pay. However, tax collection for social network retailers still depends on tax authorities encouraging them to self-declare their revenue. Current research and practice show that very few sellers on social networks participate in the self-declaration of revenue to pay taxes. Moreover, few studies focus on factors affecting the participation of online retailers in Vietnam in tax declaration and payment, especially those selling on social networks, and most lack a solid theoretical foundation. Through a review of related studies, this research has identified several factors that affect tax payment decisions for retailers on social networks in Vietnam. The study also provides several recommendations to promote and enhance retailers’ tax declaration and payment decisions on social networks in Vietnam.

EMPREENDEDORISMO NA UNIVERSIDADE: COMPETÊNCIAS EMPREENDEDORAS E A INTENÇÃO DE EMPREENDER

Article

Apr 2024

Rodrigo Celso Gheno

A educação empreendedora desenvolve nos indivíduos a capacidade de reconhecer oportunidades, assumir riscos e buscar o novo. Tais capacidades são potencializadas pelo desenvolvimento da autoconfiança e da proatividade, o comportamento empreendedor é uma proposta de solução para o mercado de trabalho em crise, especialmente nos períodos de mudança econômica e social. Com base na Teoria do Comportamento Planejado (TCP), este estudo analisa a influência das competências empreendedoras sobre a intenção empreendedora (IE). A amostra foi composta por 130 estudantes regularmente matriculados no curso de graduação em Administração da Universidade Federal Fronteira Sul. Para tanto, utilizou-se de pesquisa descritiva, com abordagem quantitativa, por meio de survey, para análise dos resultados utilizou-se da análise de regressão linear múltipla. Os resultados indicaram a existência de forte influência das competências empreendedoras na intenção empreendedora. Esse resultado é explicado pela TCP, no qual se observa que as competências e a intenção empreendedora são influenciadas diretamente pelas experiências de vida do indivíduo. A educação empreendedora mostrou-se importante para melhorar as competências empreendedoras, porém não apresentou influência significativa direta na IE. O estudo contribui ao destacar que os estudantes de Administração que possuem maiores competências empreendedoras estão mais propensos a abrir e gerir um novo negócio.

Utility of the Theory of Planned Behaviour for predicting parents' intentions to vaccinate their children against COVID-19

Article

Mar 2024
J HEALTH PSYCHOL

Vaccination is a crucial form of primary prevention, and it is important to understand the factors that influence parents’ decisions to vaccinate their children. The current study examines the utility of the Theory of Planned Behaviour (TPB) and anticipated affect for explaining parents’ intentions to vaccinate their children against COVID-19. Parents ( N = 843) living in the United States completed an online survey. The TPB variables explained 65% of the variability in parents’ intentions. In addition to all three of the TPB antecedents predicting vaccine intentions, both anticipated regret of not vaccinating and anticipated positive emotions of vaccinating were associated with parent intentions. Contrary to predictions, subjective norms were a stronger predictor of intentions when perceived behavioural control was lower compared to higher. These findings help further our understanding of parent-for-child vaccine decisions in the context of novel health threats and inform intervention efforts aimed at encouraging this behaviour.

The utility of a protection motivation theory framework for understanding sedentary behavior

Article

Full-text available

Mar 2016

Multilevel determinants of sedentary behavior (SB), including constructs couched within evidence-based psychological frameworks, can contribute to more efficacious interventions designed to decrease sitting time. This study aimed to: (1) examine the factor structure and composition of sedentary-derived protection motivation theory (PMT) constructs and (2) determine the utility of these constructs in predicting general and leisure sedentary goal intention (GI), implementation intention (II), and self-reported SB. Sedentary-derived PMT (perceived severity, PS; perceived vulnerability, PV; response efficacy, RE; self-efficacy, SE), GI, and II constructs, and a modified SB questionnaire were completed by undergraduate students (n = 596). SE was broken into three psychological (productive, focused, tired), and two situational (studying, leisure) constructs to capture the main barriers to reducing sitting time. After completing socio-demographics and the PMT items, participants were randomized to complete general or leisure GI and II. Based on model assignment, they completed either the general or leisure SB questionnaire one week later. Irrespective of model, exploratory followed by confirmatory factor analysis revealed that the PMT items grouped into eight coherent and interpretable factors consistent with the theory's threat and coping appraisal tenets: PV, PS, RE, and five scheduling SE constructs (tired, productive/focused, TV/video games/computer, studying at home, studying in library/Wi-Fi area). Using linear regression, general and leisure models predicted 5% and 1% of the variance in GI, 10% and 16% of the variance in II, and 3% and 1% of the variance in SB, respectively. Variables that made unique and significant contributions were: RE (general) and SE (leisure) for goal intention; PV and RE (general), PV, RE, and SE (leisure) for implementation intention; and only goal intention (leisure) for SB. Support now exists for the tenability of an eight-factor PMT sedentary model and its utility in predicting II and to a lesser extent GI and behavior.

Analysis of Waste Separation Drivers in Urban Centers Using the Theory of Planned Behavior and the Norm Activation Model

Article

Apr 2024

Urban centers worldwide are grappling with complex waste management challenges, including efficient collection, transportation, processing, and an over-reliance on landfills. A promising approach to mitigate these issues lies in bolstering public participation in waste separation, which could significantly improve recycling efforts. To effectively encourage this practice, it is crucial to understand the underlying factors that motivate community engagement in waste segregation activities. This study utilizes the Theory of Planned Behavior and the Norm Activation Model to identify and analyze determinants influencing individuals' propensity to separate waste in the sampling area of Balikpapan City, Indonesia. Balikpapan, one of the cities in Indonesia, is currently facing several distinct challenges related to waste management. Through the empirical validation of eight hypotheses, it becomes apparent that the presence of market facilitators (H3) and the influence of past behavior (H5) play pivotal roles in shaping the intention to engage in waste separation. The findings suggest that providing accessible, well-maintained market facilities and initiatives designed to enrich the public's waste separation experience are essential strategies. Implementing these strategies could significantly improve waste separation practices within specific urban contexts such as Balikpapan, Indonesia, and other cities facing similar environmental management challenges.

Financial Technology's Role in Attaining Financial Inclusion: Empirical Findings from Egypt

Article

Full-text available

Mar 2024

Somaia Salem

Financial Technology's Role in Attaining Financial Inclusion: Empirical Findings from Egypt

Article

Feb 2024

Expanding the UTAUT2 framework to determine the drivers of mobile shopping behaviour among older adults

Article

Full-text available

Dec 2023
PLOS ONE

Tianyang Huang

In the current severe aging of the population, the problem of "digital divide" of the elderly has become increasingly prominent, and the elderly market represents a vast demographic that is increasingly becoming an important customer segment for mobile shopping in the future. However, there is currently insufficient attention given to the research on mobile shopping behavior among older adults. This study tries to answer what are the driving factors of mobile phone shopping behavior among the elderly? The purpose of this study is to analyze the factors that drive the elderly’s mobile phone shopping behavior, and to establish a mobile phone shopping acceptance model for the elderly to predict the factors of the elderly’s behavioral intention of using smart phones. Based on the second edition of Unified Theory of Acceptance and Use of Technology theory (UTAUT 2), this study proposed a mobile phone shopping acceptance model for the elderly. The study collected valid data from 389 Chinese elderly people through questionnaires and analyzed them using structural equation models. The results showed that utilitarian, anxiety, trust, performance expectancy, effort expectancy, social influence, facilitating conditions and habit directly impact the older adults’ intention to engage in mobile shopping. Additionally, facilitating conditions, habit and the older adults’ intention to engage in mobile shopping act as driving factors for actual use behavior. This study further expands the UTAUT theoretical model, provides a theoretical basis for the research of mobile shopping behavior of the elderly, and enricues the application groups and fields of the UTAUT theoretical model. The results of this study provide inspiration for the development, design and marketing of age-appropriate mobile shopping products, and contribute to the realization and further adoption of age-appropriate mobile shopping, and also contribute to promoting the active aging of the elderly.

Influência da impulsividade, promoção de vendas e momento da equipe na intenção de compra de artigos esportivos de uma equipe de futebol do Sul do Brasil / Influence of impulsivity, sales promotion and team moment on the intention to purchase sporting goods from a football team in southern Brazil

Article

Full-text available

Nov 2023

O comportamento humano tem sido estudado em muitas áreas, e o esporte é uma dessas áreas em que o comportamento é influenciado por diferentes fatores. Considerando o contexto do futebol e os aspectos ligados à compra de produtos do clube, o objetivo deste estudo foi identificar a influência da impulsividade, de aspectos estratégicos (promoções de venda) e situacionais (momentos da equipe) na compra e de artigos esportivos do clube de futebol Grêmio Foot-ball Porto Alegrense. Foi realizada uma pesquisa com 219 respondentes, utilizando a Modelagem de Equações Estruturais (MEE) por meio da qual foi testado o modelo conceitual. Os resultados mostram que os construtos influenciam positivamente a intenção de compra, com destaque para o construto “Satisfação com o momento da equipe”. *** Human behavior has been studied in many areas, and sport is one of those areas in which different factors influence behavior. Considering the context of football and aspects linked to the purchase of club products, the objective of this study was to identify the influence of impulsivity, strategic (sales promotions), and situational (team moments) aspects on the purchase of club sporting goods of football Grêmio Foot-ball Porto Alegrense. A survey was carried out with 219 respondents using Structural Equation Modeling (SEM), through which the conceptual model was tested. The results show that the constructs positively influence purchase intention, with emphasis on the construct “Satisfaction with the team’s moment.

INVESTIGATING DRIVERS OF DIGITAL BANKING ADOPTION OF GEN Z IN INDONESIA

Article

Full-text available

Nov 2023

Despite the fact that digital banks are the most recent financial service that provides online banking services without direct contact with customers, they have yet to transform the whole financial sector. This study was done in a quantitative manner utilizing primary data to investigate drivers of digital banking adoption by Gen Z using the Unified Theory of Acceptance and Use of Technology 2 (UTAUT 2) approach. There was a total of 232 respondents participating in this study. The data was analyzed using Structural Equation Modeling - Partial Least Square (SEM-PLS). The results show that social influence, facilitating conditions, hedonic motivations, perceived value, and habit positively influence the Gen Z’s behavioral intention to use the digital banks; while performance expectancy and effort expectancy show the opposite influence. Further, the behavioral intention also positively influences the Gen Z’s use behavior of digital banks. To promote digital bank customer commitment and satisfaction, digital banking service providers are suggested to increase the value of benefits outweighing the costs borne by users, as well as to improve service quality in terms of user friendliness.JEL: G40, O33.

Direct Experience And Attitude-Behavior Consistency

Chapter

Full-text available

Dec 1981
ADV EXP SOC PSYCHOL

The chapter discusses the role of the manner of attitude formation. It focuses on the development of an attitude through direct behavioral experience with the attitude object and examines whether such attitudes better predict subsequent behavior than attitudes formed without behavioral experience. The chapter provides an overview of the attitude-behavior consistency problem and describes the effect of the manner of attitude formation through the “housing” study, the “puzzle” experiment, and the “subject pool” study. The prior-to-later behavior relation is also discussed in the chapter, wherein it has described the self-perception of past religious behaviors, attitudes and self-reports of subsequent behavior, an individual difference perspective, and a partial correlation analysis. The chapter discusses attitudinal qualities—namely, confidence and clarity, the persistence of the attitude, and resistance to attack. The reasons for the differential strength are also explored in the chapter—namely, the amount of information available, information processing, and attitude accessibility. The chapter briefly describes the attitude-behavior relationship, personality traits, and behavior.

Implementation intentions and repeated behaviour: augmenting the predictive validity of the theory of planned behaviour

Article

Mar 1999
EUR J SOC PSYCHOL

Two experiments based upon Gollwitzer's (1993) concept of implementation intentions are described. In both experiments, attitudes, subjective norms, perceived behavioural control and intentions from Ajzen's (1991) theory of planned behaviour were used to measure participants' motivation prior to an intervention in which participants made implementation intentions specifying where and when they would take a vitamin C pill each day. Behaviours were assessed by self-report and pill count at both 10 days and 3 weeks in Experiment 1, and at 2 weeks and 5 weeks in Experiment 2. Results supported the view that participants who formed implementation intentions were less likely to miss taking a pill every day compared to controls. Evidence suggested that implementation intentions were effective because they allowed participants to pass control of behaviour to the environmental cues contained in the implementation intention. Implications of the study and some suggestions for future research are outlined. Copyright © 1999 John Wiley & Sons, Ltd.

The theory of planned behavior, organizational behavior and human decision processes

Article

Jan 1991

Icek Ajzen

Habit versus planned behaviour: A field experiment

Article

Mar 1998

A field experiment investigated the prediction and change in repeated behaviour in the domain of travel mode choices. Car use during seven days was predicted from habit strength (measured by self-reported frequency of past behaviour, as well as by a more covert measure based on personal scripts incorporating the behaviour), and antecedents of behaviour as conceptualized in the theory of planned behaviour (attitude, subjective norm, perceived behavioural control and behavioural intention). Both habit measures predicted behaviour in addition to intention and perceived control. Significant habit x intention interactions indicated that intentions were only significantly related to behaviour when habit was weak, whereas no intention-behaviour relation existed when habit was strong. During the seven-day registration of behaviour, half of the respondents were asked to think about the circumstances under which the behaviour was executed. Compared to control participants, the behaviour of experimental participants was more strongly related to their previously expressed intentions. However, the habit-behaviour relation was unaffected. The results demonstrate that, although external incentives may increase the enactment of intentions, habits set boundary conditions for the applicability of the theory of planned behaviour.

Extending the theory of planned behavior: A review and avenues for further research

Article

Jan 1998

This paper describes and reviews the theory of planned behavior (TPB). The focus is on evidence supporting the further extension of the TPB in various ways. Empirical and theoretical evidence to support the addition of 6 variables to the TPB is reviewed: belief salience measures, past behaviodhabit, perceived behavioral control (PBC) vs. self-efficacy, moral norms, self-identity, and affective beliefs. In each case there appears to be growing empirical evidence to support their addition to the TPB and some understanding of the processes by which they may be related to other TPB variables, intentions , and behavior. Two avenues for expansion of the TPB are presented. First, the possibility of incorporating the TPB into a dual-process model of attitude-behavior relationships is reviewed. Second, the expansion of the TPB to include consideration of the volitional processes determining how goal intentions may lead to goal achievement is discussed. The theory of planned behavior (TPB) is a widely applied expectancy-value model of attitude-behavior relationships which has met with some degree of success in predicting a variety of behaviors present paper examines avenues for development of this theory as a way of furthering our understanding of the relationship between attitudes and behavior. This is achieved in two ways: a review of the evidence supporting the addition of six different variables to the TPB, and a review of two avenues for expanding this theory. Six additional variables are reviewed: belief salience, past behaviodhabit, perceived behavioral control versus self-efficacy, moral norms, self-identity, and affective beliefs. Two avenues for model expansion are considered: multiple processes by which attitudes influence 'Correspondence concerning this article should be addressed to Mark Conner, School of Psychology , University of Leeds, Leeds LS2 9JT. United Kingdom.

The four horsemen of automaticity: Awareness, intention, efficiency and control in social perception and cognition

Article

J.A. Bargh

The unbearable automaticity of being

Article

Jul 1999
AM PSYCHOL

What was noted by E. J. Langer (1978) remains true today; that much of contemporary psychological research is based on the assumption that people are consciously and systematically processing incoming information in order to construe and interpret their world and to plan and engage in courses of action. As did E. J. Langer, the authors question this assumption. First, they review evidence that the ability to exercise such conscious, intentional control is actually quite limited, so that most of moment-to-moment psychological life must occur through nonconscious means if it is to occur at all. The authors then describe the different possible mechanisms that produce automatic, environmental control over these various phenomena and review evidence establishing both the existence of these mechanisms as well as their consequences for judgments, emotions, and behavior. Three major forms of automatic self-regulation are identified: an automatic effect of perception on action, automatic goal pursuit, and a continual automatic evaluation of one's experience. From the accumulating evidence, the authors conclude that these various nonconscious mental systems perform the lion's share of the self-regulatory burden, beneficently keeping the individual grounded in his or her current environment.

Controlled and automatic human information processing: I. Detection, search and attention

Article

Jan 1977

Verbal Attitudes and Overt Behavior: A Study of Racial Discrimination

Article

Mar 1965

Lawrence S. Linn

The present study in attempting to measure the relationship between racial attitudes and overt behavior asked Ss to pose for a photograph with a Negro of the opposite sex. Discrepancies between verbal attitudes and subsequent overt behavior involving those attitudes was found in 59 percent of the cases. The relationship between attitude (prejudice) and behavior (discrimination) is seen to be a function of the level of social involvement with the attitude object as well as the amount of prior experience with it. One implication of the study is that statements or predictions of racial behavior based on attitude measurements have little reliability unless first validated empirically.

Multiple Regression: Testing and Interpreting Interactions

Article

Jan 1994
J OPER RES SOC

Residual Effects of Past on Later Behavior: Habituation and Reasoned Action Perspectives

Abstract

Recommended publications

Residual Effects of Past on Later Behavior: Habituation and Reasoned Action Perspectives

Choice of Travel Mode in the Theory of Planned Behavior: The Roles of Past Behavior, Habit, and Reas...

The theory of planned behaviour: Reactions and reflections

Perceived Behavioral Control, Self‐Efficacy, Locus of Control, and the Theory of Planned Behavior