Logo Passei Direto
Buscar
Material
páginas com resultados encontrados.
páginas com resultados encontrados.
left-side-bubbles-backgroundright-side-bubbles-background

Crie sua conta grátis para liberar esse material. 🤩

Já tem uma conta?

Ao continuar, você aceita os Termos de Uso e Política de Privacidade

left-side-bubbles-backgroundright-side-bubbles-background

Crie sua conta grátis para liberar esse material. 🤩

Já tem uma conta?

Ao continuar, você aceita os Termos de Uso e Política de Privacidade

left-side-bubbles-backgroundright-side-bubbles-background

Crie sua conta grátis para liberar esse material. 🤩

Já tem uma conta?

Ao continuar, você aceita os Termos de Uso e Política de Privacidade

left-side-bubbles-backgroundright-side-bubbles-background

Crie sua conta grátis para liberar esse material. 🤩

Já tem uma conta?

Ao continuar, você aceita os Termos de Uso e Política de Privacidade

left-side-bubbles-backgroundright-side-bubbles-background

Crie sua conta grátis para liberar esse material. 🤩

Já tem uma conta?

Ao continuar, você aceita os Termos de Uso e Política de Privacidade

left-side-bubbles-backgroundright-side-bubbles-background

Crie sua conta grátis para liberar esse material. 🤩

Já tem uma conta?

Ao continuar, você aceita os Termos de Uso e Política de Privacidade

left-side-bubbles-backgroundright-side-bubbles-background

Crie sua conta grátis para liberar esse material. 🤩

Já tem uma conta?

Ao continuar, você aceita os Termos de Uso e Política de Privacidade

left-side-bubbles-backgroundright-side-bubbles-background

Crie sua conta grátis para liberar esse material. 🤩

Já tem uma conta?

Ao continuar, você aceita os Termos de Uso e Política de Privacidade

left-side-bubbles-backgroundright-side-bubbles-background

Crie sua conta grátis para liberar esse material. 🤩

Já tem uma conta?

Ao continuar, você aceita os Termos de Uso e Política de Privacidade

left-side-bubbles-backgroundright-side-bubbles-background

Crie sua conta grátis para liberar esse material. 🤩

Já tem uma conta?

Ao continuar, você aceita os Termos de Uso e Política de Privacidade

Prévia do material em texto

<p>The Role of Attitude Accessibility in the</p><p>Attitude-to-Behavior Process</p><p>RUSSELL H. FAZIO</p><p>MARTHA C. POWELL</p><p>CAROL J.WILLIAMS*</p><p>Attitudes toward a number of products and the accessibility of those attitudes as</p><p>indicated by the latency of response to an attitudinal inquiry were assessed. Sub-</p><p>jects with highly accessible attitudes toward a given product displayed greater atti-</p><p>tude-behavior correspondence than did those with relatively less accessible atti-</p><p>tudes. Furthermore, subjects with less accessible attitudes displayed more sensi-</p><p>tivity to the salience afforded a product by its position In the front row, as opposed</p><p>to the back row, than did subjects with more accessible attitudes. The implications</p><p>of these data for a model of the process by which attitudes guide behavior are</p><p>discussed.</p><p>T he topic of attitudes clearly occupies a central</p><p>position in research on consumer behavior (e.g.,</p><p>Engei and Blackwell 1982; Kassarjian and Kassarjian</p><p>1979). In part, the attention given to the attitude con-</p><p>struct stems from the fact that much advertising can</p><p>be described as social influence attempts aimed at cre-</p><p>ating positive attitudes toward the product. Indeed,</p><p>considerable research has been concerned with un-</p><p>derstanding this persuasion process (e.g., Boyd, Ray,</p><p>and Strong 1972; Lutz 1975; Petty, Cacioppo, and</p><p>Schumann 1983). The assumption underlying such</p><p>persuasive attempts is that the development of posi-</p><p>tive attitudes will produce a corresponding change in</p><p>behavior. As a result, consumer research also has</p><p>been concerned with understanding the relation be-</p><p>tween attitudes and subsequent behavior (e.g., Day</p><p>and Deutscher 1982; Ryan and Bonfield 1975; Smith</p><p>and Swinyard 1983). It is this attitude-behavior rela-</p><p>tion that constitutes the focus of the present article.</p><p>Fazio and his colleagues proposed a model of the</p><p>process by which attitudes guide behavior (Fazio</p><p>1986; Fazio, Powell, and Herr 1983). In brief, the</p><p>model views behavior in any given situation as a func-</p><p>*Russell H. Fazio is Professor of Psychology, and Martha C. Pow-</p><p>ell and Carol J. Williams are Research Assistants, all in the Depart-</p><p>ment of Psychology, Indiana University, Bloomington, IN 47405.</p><p>This article was prepared while the first author was supported by</p><p>the National Institute of Mental Health Research Scientist Devel-</p><p>opment Award MH00452. The research was supported by a grant</p><p>from the Ogilvy Center for Research and Development. The au-</p><p>thors thank Clark Leavitt and Alexander Biel of the Ogilvy Center</p><p>for their helpful consultation while the research was being designed</p><p>and David Sanbonmatsu for his assistance in data collection. Por-</p><p>tions of these data were presented at the 1986 Association for Con-</p><p>sumer Research Annual Conference in Toronto.</p><p>280</p><p>tion of the individual's immediate perceptions of the</p><p>attitude object in the context of the situation in which</p><p>the object is encountered. "Perception" refers to the</p><p>individual's current feelings about, or appraisal of,</p><p>the object as experienced in the immediate situation.</p><p>According to the model, attitudes guide such apprais-</p><p>als of the object, but only if they have been activated</p><p>from memory upon observation of the object. Hence,</p><p>the accessibility of the attitude from memory is postu-</p><p>lated to act as a critical determinant of whether the</p><p>attitude-to-behavior process is initiated. Various as-</p><p>pects of the model have received empirical support</p><p>from both correlational and experimental investiga-</p><p>tions in such contexts as voting behavior (Fazio and</p><p>Williams 1986), the evaluation of evidence related to</p><p>a social policy issue (Houston and Fazio 1989), and</p><p>"free-play" behavior with intellectual puzzles (Fazio</p><p>et al. 1982; see Fazio 1989 for a review). The major</p><p>goal of the present research was to assess the general-</p><p>ity of the model to a consumer behavior domain. The</p><p>research focuses upon the relation between attitudes</p><p>toward specific products and product-selection be-</p><p>havior.</p><p>The hypothesis here is that the status of individuals'</p><p>attitudes along an attitude/non-attitude continuum</p><p>moderates the attitude-behavior relation. Fazio and</p><p>his colleagues (Fazio 1989; Fazio etai. 1986) have de-</p><p>fined this continuum in terms of the strength of the</p><p>association in memory between the object and the in-</p><p>dividual's evaluation of the object. At one end of the</p><p>continuum is the non-attitude; no a priori evaluation</p><p>of the object is available in memory. If asked to ex-</p><p>press an attitude, the individual needs to construct</p><p>one on the spot. Such constructions presumably will</p><p>involve some assessment of attributes of the object</p><p>©JOURNAL OF CONSUMER RESEARCH •Vol. 16»December 1989</p><p>ATTITUDE ACCESSIBILITY 281</p><p>that are recalled from memory and/or apparent in the</p><p>external environment. However, at all other points</p><p>along the continuum, an evaluation is available in</p><p>memory. Moving along the continuum, the strength</p><p>of the association between the evaluation and the ob-</p><p>ject—and hence the accessibility of the attitude—in-</p><p>creases. At the upper end of the continuum is a well-</p><p>learned, strong association—sufficiently strong that</p><p>the evaluation is capable of automatic activation</p><p>from memory upon mere observation or mention of</p><p>the object. Attitude-behavior consistency is expected</p><p>to vary as a function of position along this attitude/</p><p>non-attitude continuum. Individuals who possess</p><p>highly accessible attitudes toward a given product are</p><p>expected to be more attitudinally consistent in their</p><p>product-selection behavior than are individuals</p><p>whose attitudes are relatively less accessible from</p><p>memory.</p><p>As mentioned earlier, evidence consistent with this</p><p>hypothesized moderating role of attitude accessibility</p><p>has been obtained in previous research. Most relevant</p><p>is a correlational field study by Fazio and Williams</p><p>(1986) in which it was found that the relation between</p><p>attitudes toward Reagan and self-reported voting be-</p><p>havior in the 1984 presidential election varied as a</p><p>function of the accessibility of the attitude. Attitude</p><p>accessibility was measured via latency of response to</p><p>the attitudinal inquiry. Individuals who were able to</p><p>indicate their attitude relatively quickly (the high atti-</p><p>tude accessibility group) displayed greater consis-</p><p>tency between those attitudes and subsequent voting</p><p>behavior than did individuals who responded rela-</p><p>tively slowly (the low attitude accessibility group).</p><p>In addition to examining the applicability of̂ the</p><p>model to product-selection behavior, the present re-</p><p>search was intended to address an unavoidable short-</p><p>coming of the voting behavior study—the self-report</p><p>nature of the behavior measure. As in any investiga-</p><p>tion of choices made within the confines of a voting</p><p>booth, Fazio and Williams (1986) were forced to rely</p><p>upon participants' reports of how they had voted. The</p><p>respondents were telephoned within a day or two of</p><p>the election and asked to reveal whether they had</p><p>voted and, if so, for whom. Although we do not dis-</p><p>cern any plausible explanation for how such self-re-</p><p>ports might have been biased by both the attitude ex-</p><p>pressed months earlier and the accessibility of the at-</p><p>titude, the shortcoming of this method of observing</p><p>the attitude-to-behavior relation cannot be denied.</p><p>The present investigation concerned actual overt be-</p><p>havior. Subjects selected items from a set of products</p><p>arranged on a table. These selections, which subjects</p><p>took home as "a token of our appreciation" for their</p><p>having participated in the study, constituted the be-</p><p>havioral measure.</p><p>The present investigation involves, as did the Fazio</p><p>and Williams (1986) study, measurement of attitude</p><p>accessibility via latency of response to an attitudinal</p><p>inquiry. Hence, it is important to review briefly what</p><p>is known about the validity of this measure as an indi-</p><p>cation of the accessibility of attitudes and their posi-</p><p>tion along the attitude/non-attitude continuum.</p><p>First, the latency measure has been found to reflect</p><p>what has been postulated to be the conceptual vari-</p><p>able that determines the chronic accessibility of an</p><p>at-</p><p>titude—namely, the strength of the association be-</p><p>tween the object and the evaluation. A number of ex-</p><p>periments have manipulated the strength of this</p><p>object-evaluation association by having subjects ex-</p><p>press their attitudes repeatedly. This research has in-</p><p>dicated that such repeated expression enhances the</p><p>speed with which individuals respond to later inqui-</p><p>ries concerning their attitudes (Fazio et al. 1982; Pow-</p><p>ell and Fazio 1984).</p><p>Second, and more important, the latency measure</p><p>provides a good approximation of the likelihood that</p><p>the attitude will be activated from memory automati-</p><p>cally upon mere observation of the object. It has been</p><p>demonstrated that attitude objects preselected on the</p><p>basis of an individual's having responded quickly to</p><p>an attitudinal inquiry are more likely to activate the</p><p>attitude automatically upon subsequent presentation</p><p>of the object than are attitudes characterized by rela-</p><p>tively slow latencies of response to an attitudinal in-</p><p>quiry (Fazio et al. 1986). Sanbonmatsu and Fazio</p><p>(1986) observed the same result in a study specifically</p><p>concerned with the automatic activation of attitudes</p><p>toward products. Thus, the latency with which one</p><p>responds to an attitudinal inquiry is sensitive to the</p><p>strength of the object-evaluation association and pro-</p><p>vides an indication of the likelihood that the attitude</p><p>will be activated spontaneously upon one's encoun-</p><p>tering the object.</p><p>Subjects in the present study responded to attitudi-</p><p>nal inquiries concerning a large number of products.</p><p>A subset of these products served as the target attitude</p><p>objects and subsequently were made available as be-</p><p>havioral alternatives. According to the process</p><p>model, these behavioral selections should be a func-</p><p>tion of the individual's perceptions of the object in</p><p>the immediate situation. Is the individual's appraisal</p><p>of the object at that particular moment favorable or</p><p>unfavorable? An attitude that is highly accessible</p><p>from memory and, hence, likely to be activated auto-</p><p>matically upon the individual's observation of the ob-</p><p>ject is apt to result in immediate perceptions that are</p><p>congruent with the attitude. In contrast, when the at-</p><p>titude is not activated from memory, immediate per-</p><p>ceptions are less likely to be influenced by a pre-</p><p>viously constructed evaluation that is available in</p><p>memory. Instead, these immediate appraisals are</p><p>likely to be based upon momentarily salient and po-</p><p>tentially unrepresentative features (memory-based</p><p>and/or stimulus-based) of the object that are not nec-</p><p>essarily evaluated in a manner congruent with the at-</p><p>titude. As a result, greater attitude-behavior consis-</p><p>282 THE JOURNAL OF CONSUMER RESEARCH</p><p>tency is expected when attitudes are highly accessible</p><p>than when attitudes are relatively less accessible from</p><p>memory.</p><p>METHOD</p><p>Subjects</p><p>One hundred one individuals participated in the</p><p>experiment. Sixty-one participated in partial fulfill-</p><p>ment of an introductory psychology course require-</p><p>ment. The remaining subjects responded to an ad</p><p>placed in the local newspaper and participated in re-</p><p>turn for monetary payment.</p><p>Procedure</p><p>Subjects were informed that the experiment con-</p><p>cerned a new way of measuring attitudes via a micro-</p><p>computer. Their initial task involved responding to</p><p>each of 100 attitude objects, all of which were familiar</p><p>and commonly available products. The subjects were</p><p>instructed to press one oftwo keys, labeled "like" and</p><p>"dislike," to indicate their feelings about each object.</p><p>The subjects were told to "respond as quickly and as</p><p>accurately as possible." The presentation was con-</p><p>trolled by an Apple 11+ computer. The order in which</p><p>the objects were presented was randomized for each</p><p>subject. A three-second interval separated each trial.</p><p>The subject's response was recorded, along with the</p><p>latency of the response (from stimulus onset to re-</p><p>sponse) to the nearest millisecond. To familiarize</p><p>subjects with the procedure, subjects performed a se-</p><p>ries of practice trials involving different products than</p><p>those used in the actual experimental list.</p><p>Following the computerized task, subjects com-</p><p>pleted a number of questionnaires, only the first of</p><p>which was relevant to the present purposes. The sub-</p><p>jects provided a scalar attitudinal rating of each of the</p><p>100 objects. These ratings were made on a 1 (ex-</p><p>tremely bad) to 7 (extremely good) scale.</p><p>Behavior Measure</p><p>As their final task, subjects were given the opportu-</p><p>nity to select five products from a set of 10 alterna-</p><p>tives. The 10 products, which had been roughly</p><p>equated for perceived value through informal pretest-</p><p>ing, were a Snickers candy bar, a small bag of Fritos</p><p>corn chips, two boxes of Sun-Maid raisins, a small can</p><p>of Star-Kist tuna, a can of Dr. Pepper, a box of</p><p>Cracker Jacks, a bag of Planters peanuts, two cans of</p><p>V-8 juice, two 5-stick packs of Dentyne gum, and a</p><p>Mounds candy bar. The products were arranged in</p><p>two rows of five on a table and were covered by a ta-</p><p>blecloth.</p><p>At the appropriate time, the experimenter removed</p><p>the tablecloth and told the subject to choose five of</p><p>the items as a "token ofour appreciation" for having</p><p>participated. The experimenter unobtrusively re-</p><p>corded the products selected and the order in which</p><p>they were selected. The selected items were then</p><p>placed in a paper bag for the subject to take home.</p><p>RESULTS</p><p>The mean latency of response across objects and</p><p>subjects was 1.14 seconds. When a subject's dichoto-</p><p>mous response during the computerized task was in-</p><p>consistent with the subsequent scalar rating, it was</p><p>considered an error. An error was defined as a sub-</p><p>ject's responding "like" during the computerized la-</p><p>tency task but rating the product more negatively</p><p>than the neutral point on the questionnaire scale, or</p><p>responding "dislike" during the computer task but</p><p>assigning a rating above the neutral point on the scale.</p><p>Such errors were infrequent. On the average, subjects</p><p>responded inconsistently on 7.3 percent of the 100</p><p>trials. The particular datum from any trial on which</p><p>an error occurred was omitted from the analysis.</p><p>Previous research on attitude accessibility has ob-</p><p>served a relation between attitude accessibility, as</p><p>measured by latency of response to an attitudinal in-</p><p>quiry, and attitude extremity, as indexed by deviation</p><p>of a scalar value from the neutral point of the scale</p><p>(Fazio and Williams 1986; Powell and Fazio 1984).</p><p>The same relation was evident in the present data set.</p><p>Faster response times were associated with more ex-</p><p>treme ratings. The within-subject correlation be-</p><p>tween response latency and attitude extremity</p><p>reached a level of significance for 46 percent of the</p><p>subjects. The mean correlation across subjects was</p><p>-0.18 (/(lOO) = 11.47, p< 0.001). As a result of this</p><p>relation, it was necessary that all tests of the hypothe-</p><p>sis regarding the moderating role of attitude accessi-</p><p>bility be conducted in a manner that did not con-</p><p>found attitude accessibility and attitude extremity.</p><p>Both within-subject and between-subject analyses</p><p>were conducted. In both cases, the analysis that was</p><p>performed examined the impact of accessibility as</p><p>measured by response latency while considering</p><p>equivalent attitude scores.</p><p>Within-Subjects Analysis</p><p>To overcome the naturally existing relation be-</p><p>tween attitude accessibility and extremity, the within-</p><p>subjects analysis focused upon any two or more target</p><p>products for which a given subject had assigned the</p><p>same attitude scalar rating. We restricted ourselves to</p><p>these attitudinal "ties" to ensure that the role of atti-</p><p>tude accessibility was examined in a manner that was</p><p>independent of attitude scores. For each subject, the</p><p>proportion of ties for which the subject's behavior</p><p>concurred with a prediction based upon the subject's</p><p>response latencies was determined. According to the</p><p>ATTITUDE ACCESSIBILITY 283</p><p>hypothesis, subjects with highly accessible attitudes</p><p>should behave more consistently than those with less</p><p>accessible attitudes. Thus, for any ties at positions</p><p>more positive than the neutral point, we predicted</p><p>that the subject would either select the product associ-</p><p>ated with the faster response latency before selecting</p><p>the other product or select the former product and not</p><p>the latter. The prediction was reversed for any ties at</p><p>attitudinal positions more negative than the neutral</p><p>point. Any ties at the neutral point were not included</p><p>in the analysis. Averaged across the 100 subjects with</p><p>analyzable ties, the proportion of instances confirm-</p><p>ing the predictions was 0.59, significantly greater than</p><p>the value of 0.50 expected by chance alone, (/(99)</p><p>= 3.09, ; ;< 0.0025).</p><p>Between-Subjects Analyses</p><p>The between-subjects analyses involved classifying</p><p>subjects into groups of high, moderate, and low atti-</p><p>tude accessibility for each product. As in the Fazio</p><p>and Williams (1986) study, this group assignment was</p><p>performed at each and every level of the attitude scale</p><p>to avoid confounding attitude scores with attitude ac-</p><p>cessibility. For example, the subsample of subjects</p><p>who had assigned a rating of 7 (extremely good) to a</p><p>given target product were trichotomized on the basis</p><p>of their accessibility scores into high, moderate, and</p><p>low groups. The subsamples who had rated the prod-</p><p>uct 6, 5, and so on were similarly trichotomized. This</p><p>classification procedure was performed anew for each</p><p>of the 10 target products. In this way, the attitude dis-</p><p>tributions for any given product were equivalent in</p><p>the high, moderate, and low attitude accessibility</p><p>groups.</p><p>Unlike the Fazio and Williams (1986) investiga-</p><p>tion, however, attitude accessibility was not indexed</p><p>simply by raw response latency. Instead, the accessi-</p><p>bility measure employed for classification purposes</p><p>was the z-score of a given subject's response latency</p><p>for a given target product relative to that subject's</p><p>mean and standard deviation of the latencies for the</p><p>90 filler products. This within-subject z-score serves</p><p>to place a given subject's latency of response to a</p><p>given object within his or her distribution of latencies</p><p>for the filler products (see Fazio forthcoming for a dis-</p><p>cussion of indices of baseline speed of responding).</p><p>This change was necessitated by the fact that inter-</p><p>item correlations among the 10 response latencies</p><p>were substantial (average r = 0.42), whereas inter-</p><p>item correlations in the Fazio and Williams investiga-</p><p>tion were quite minimal (average r = 0.19; see Foot-</p><p>note 1 of Fazio and Williams 1986).' The use of raw</p><p>'Why the two studies differed in this regard is not clear. However,</p><p>one difference in the attitude accessibility measurement procedures</p><p>employed in the two studies is worth noting. The Fazio and Wil-</p><p>liams (1986) investigation involved subjects responding to an au-</p><p>latency in the present case would have led to the con-</p><p>sistent classification of individuals with tendencies to</p><p>respond quickly to any inquiry as high accessibility</p><p>subjects.</p><p>In one of the analyses that was conducted, behavior</p><p>was coded on a 0 to 5 scale, with products that were</p><p>not selected by a given subject being assigned a score</p><p>of 0 and the product that was selected first a score of 5.</p><p>The remaining products were assigned intermediate</p><p>scores indicative of the order of selection. For each</p><p>product, the correlation between attitude and behav-</p><p>ior was computed within the high, moderate, and low</p><p>accessibility groups. These correlations were ana-</p><p>lyzed, following a Fisher's r-to-z transformation, via</p><p>an a priori polynomial contrast examining whether</p><p>they displayed a significant linear trend as a function</p><p>of level of attitude accessibility. This prediction was</p><p>confirmed (F( 1,9) = 8.01, p < 0.02).^ Averaged across</p><p>the 10 products, the mean correlations (following a</p><p>retransformation of the average zs back to rs) were</p><p>0.62, 0.54, and 0.50 for the high, moderate, and low</p><p>attitude accessibility groups, respectively.</p><p>The analysis involved the scoring of behavior as a</p><p>function of order of selection and, thus, assumes that</p><p>subjects chose first their most preferred alternative,</p><p>then their second most preferred item, and so on. The</p><p>data also were analyzed with behavior coded dichoto-</p><p>mously as a function of whether the product was or</p><p>was not selected. This analysis also confirmed the pre-</p><p>diction. The mean correlations were 0.61, 0.59, and</p><p>0.51 in the high, moderate, and low attitude accessi-</p><p>bility groups, respectively, and, as in the previous</p><p>analysis, revealed a linear trend (F(l,9) = 5.09,</p><p>p = 0.051).</p><p>Thus, regardless of how the data are examined, the</p><p>findings converge upon the notion that attitude acces-</p><p>sibility moderated the attitude-behavior relation. The</p><p>more accessible a subject's attitude was toward a</p><p>given product, the more likely it was that product se-</p><p>lection behavior was consistent with that attitude.</p><p>Position EfFects</p><p>As was mentioned earlier, the attitude-to-behavior</p><p>process model views behavior as a function of the in-</p><p>diotaped statement, the end of which contained the electronic</p><p>marker that initiated the timing. In contrast, the present study in-</p><p>volved displaying the name of an object on a computer screen, the</p><p>onset of which initiated the timing. The time necessary to read the</p><p>visual display may have enhanced the amount of covariation ob-</p><p>served. Reading time may be a fairly constant individual difference</p><p>that formed a component of each response latency.</p><p>^Given the small number of units of analysis involved in this sta-</p><p>tistical test, i.e., the 10 products, it seemed desirable to also exam-</p><p>ine the data nonparametrically. The data were analyzed via the</p><p>Page test for ordered alternatives (Page 1963; Seigel and Castellan</p><p>1988), which did reveal a significant linear trend. The same was</p><p>true of all additional linear trends reported here.</p><p>284 THE JOURNAL OF CONSUMER RESEARCH</p><p>dividual's perceptions of the object in the immediate</p><p>situation. A highly accessible attitude is likely to be</p><p>activated from memory automatically and is pre-</p><p>sumed to determine this immediate perception of the</p><p>object. In contrast, an attitude that is relatively low in</p><p>accessibility is less likely to be activated upon obser-</p><p>vation of the object. As a result of low accessibility,</p><p>the immediate perception is likely to be influenced by</p><p>momentarily salient features of the object.</p><p>This hypothesized greater influence of a momen-</p><p>tarily salient dimension for objects associated with a</p><p>relatively inaccessible attitude than for objects associ-</p><p>ated with a highly accessible attitude was apparent in</p><p>a rather surprising fashion in subjects' selection be-</p><p>havior. The 10 products were arranged in two rows of</p><p>five for the subjects' viewing and selection.' On the</p><p>assumption that the products positioned in the front</p><p>row were more salient than those in the back row, we</p><p>suggest that the selection of a given product will be</p><p>more influenced by row status if the attitude is low in</p><p>accessibility than if its accessibility is high. The data</p><p>revealed precisely this pattern. Among the products</p><p>in the front row (Snickers, Mounds, peanuts, Den-</p><p>tyne, and tuna), the lower the accessibility of their at-</p><p>titudes, the greater the likelihood that subjects se-</p><p>lected the product. Indeed, across these products, the</p><p>average proportion of subjects who selected the prod-</p><p>uct was 0.55, 0.59, and 0.59 in the high, moderate,</p><p>and low accessibility groups. For products in the back</p><p>row, the reverse was true. The lower the accessibility</p><p>of the attitude, the less likely subjects were to select</p><p>the product (average proportions of 0.50, 0.46, and</p><p>0.40 for the high, moderate, and low groups, respec-</p><p>tively). The row effects—the difference between the</p><p>mean proportion of subjects who selected a product</p><p>positioned in the front row and the mean proportion</p><p>who selected an alternative that had been positioned</p><p>in the back row—were 0.05, 0.13, and 0.19 for the</p><p>high, moderate, and low attitude accessibility groups,</p><p>respectively. Analysis of these data indicated that the</p><p>row effects constituted a significant linear trend</p><p>) = 5.96,/? < 0.05)."</p><p>^Products were arranged randomly at the beginning of the experi-</p><p>ment and then fixed across</p><p>subjects. As a result, an overall prefer-</p><p>ence for front-row products over back-row products (analogous to</p><p>a main effect) may reflect either the influence of row status and/</p><p>or the products themselves. Obviously, no clear inferences can be</p><p>drawn from data reflecting an overall preference for the products</p><p>in one row over the products in the other row. However, this con-</p><p>founding does not affect our ability to examine the degree to which</p><p>the exhibited preference for front-row products relates to the level</p><p>of attitude accessibility (conceptually analogous to an interaction</p><p>effect).</p><p>•"The actual analysis was performed on scores for each product</p><p>that represented the proportion of subjects who selected the prod-</p><p>uct if it had been in the front row and the proportion who did not</p><p>select it if the product had been in the back row. A greater influence</p><p>of row status would be apparent if these scores increased linearly as</p><p>Thus, the lower the attitude accessibility, the more</p><p>selection behavior was influenced by the relative sa-</p><p>lience afforded a product by its positioning. In con-</p><p>trast, the linear trends presented earlier with respect</p><p>to attitude-behavior correlations suggest that the</p><p>greater the attitude accessibility, the more selection</p><p>behavior was influenced by the attitude.</p><p>DISCUSSION</p><p>The Role of Attitude Accessibility</p><p>The present findings are consistent with the hy-</p><p>pothesis that attitude accessibility exerts a moderat-</p><p>ing role upon the attitude-behavior relation, just as is</p><p>postulated by the process model. The more accessible</p><p>the attitude, the more predictive it was of subsequent</p><p>behavior. Furthermore, this evidence was obtained in</p><p>a situation involving actual behavior as opposed to</p><p>self-reports of behavior. Subjects chose, took home,</p><p>and presumably consumed the products. Thus, the</p><p>data are very consistent with the attitude-to-behavior</p><p>process model.</p><p>When interpreted within the context of the process</p><p>model, the findings suggest that subjects scanned the</p><p>set of available alternatives and arrived at immediate</p><p>perceptions of each (or at least some subset) of the</p><p>products. In the case of objects for which a strong</p><p>evaluative associate existed in memory, this evalua-</p><p>tion was activated automatically from memory upon</p><p>the subject's observation of the object. That is, obser-</p><p>vation of the object automatically activated the atti-</p><p>tude from memory if the chronic accessibility of the</p><p>attitude, as estimated by a relatively fast latency of</p><p>response to an earlier attitudinal inquiry, was high.</p><p>Such activated attitudes strongly determined the sub-</p><p>ject's feelings toward the object in the immediate situ-</p><p>ation. If the object-evaluation association for a given</p><p>product was weak and, hence, the chronic accessibil-</p><p>ity of the attitude was relatively low, then the immedi-</p><p>ate perception was likely to be influenced by momen-</p><p>tarily salient thoughts or features of the object. The</p><p>data suggest that the perception of such objects was</p><p>affected by the greater salience afforded an object po-</p><p>sitioned in the front row as opposed to the back row.</p><p>Other momentarily salient factors that might have</p><p>operated include product attributes that were recalled</p><p>accessibility decreased, which, as reported, they did (mean propor-</p><p>tions of 0.52,0.56, and 0.60 for the high, moderate, and low accessi-</p><p>bility groups). For ease of understanding, the data are presented in</p><p>the text in terms of row effects, i.e., the difference in the proportion</p><p>of subjects who selected front-row versus back-row products.</p><p>Additional analyses on the frequency with which a given product</p><p>was selected revealed that level of attitude accessibility was inde-</p><p>pendent of the specific products that were positioned in each row.</p><p>The data did not display any appreciable deviations from the over-</p><p>all linear trend that was apparent for the front-row (x^(8) = 4.15, p</p><p>> 0.25) or back-row products (x^(8) = 1.34, p > 0.25).</p><p>ATTITUDE ACCESSIBILITY 285</p><p>from memory, how thirsty or hungry the subject was,</p><p>how recently the subject had eaten a particular food</p><p>(e.g., a candy bar), what snack foods the subject re-</p><p>called as having stored at home, and the like. In any</p><p>case, such factors appear to exert a greater influence</p><p>when the individual's attitude toward a given object</p><p>is unlikely to have been activated automatically upon</p><p>observation of the object than when the attitude is</p><p>highly accessible from memory and capable of auto-</p><p>matic activation.</p><p>Product Positioning and the</p><p>Consideration Set</p><p>The differential influence of row status as a func-</p><p>tion of attitude accessibility is clearly evident from</p><p>the data. The precise mechanisms through which this</p><p>influence occurred is less clear. One possibility con-</p><p>cerns the potential influence of salient positioning on</p><p>the likelihood that a given alternative will receive</p><p>much consideration in the selection process. Con-</p><p>sumer behavior researchers have distinguished the</p><p>"evoked" or "consideration" set—those alternatives</p><p>that are actually considered for potential selection—</p><p>from the pool of alternatives that are available (e.g..</p><p>Baker etai. 1986; Howard and Sheth 1969). Although</p><p>the present investigation did not involve the collec-</p><p>tion of any data relevant to the question of which</p><p>items constituted an individual's consideration set,</p><p>the observed influence of both attitude accessibility</p><p>and position may have operated at the level of inclu-</p><p>sion or exclusion from the consideration set.</p><p>For ease of discussion, imagine two individuals,</p><p>one of whom (A) holds highly accessible attitudes to-</p><p>ward some subset of the 10 alternatives and one of</p><p>whom (B) does not have highly accessible attitudes</p><p>toward any of the alternatives. When scanning the</p><p>items, individual A may especially notice, and attend</p><p>to, those products toward which s/he has a highly ac-</p><p>cessible attitude because the affective associate to the</p><p>object will be activated automatically from memory</p><p>upon observation of the object. Indeed, some very re-</p><p>cent research indicates that objects toward which an</p><p>individual holds highly accessible attitudes, be they</p><p>positively or negatively valenced, are more likely to</p><p>be noticed when presented in the visual field than are</p><p>objects toward which the individual holds less acces-</p><p>sible attitudes. Roskos-Ewoldsen and Fazio (1989)</p><p>found support for this hypothesis in two investiga-</p><p>tions, one of which involved the measurement of atti-</p><p>tude accessibility via response latency (as in the pres-</p><p>ent study) and one of which involved the experimen-</p><p>tal manipulation of attitude accessibility.</p><p>Thus, these data suggest that our hypothetical indi-</p><p>vidual A will quickly and easily notice those products</p><p>concerning which accessible attitudes are held even if</p><p>those items are positioned in the relatively disadvan-</p><p>tageous back row (see Alba and Hutchinson 1987 for</p><p>a related discussion of the likelihood of product de-</p><p>tection). Any products within this subset that are neg-</p><p>atively valued will be quickly rejected from inclusion</p><p>in the consideration set. In other words, the percep-</p><p>tion or appraisal of the product in the immediate situ-</p><p>ation, determined as it is by the negative attitude acti-</p><p>vated from memory, will lead to its rejection. In con-</p><p>trast, those products that are both positively valued</p><p>and characterized by high attitude accessibility are</p><p>likely to be included in individual A's considera-</p><p>tion set.</p><p>Construction of the consideration set may proceed</p><p>quite differently for our hypothetical individual B.</p><p>Lacking highly accessible attitudes, this individual's</p><p>attention is unlikely to be attitudinally guided. Those</p><p>products enjoying the relative salience afforded them</p><p>by their position in the front row are more likely to</p><p>be included in the consideration set than are those po-</p><p>sitioned in the back row. Thus, the probabilities that</p><p>individual B will select a front-row product over a</p><p>back-row product may be enhanced. Obviously, the</p><p>role of attitude accessibility and product salience in</p><p>defining the consideration set merits further investi-</p><p>gation.</p><p>The Attitude Accessibility</p><p>Measure</p><p>Our interpretation of the attitude-behavior consis-</p><p>tency data rests on the validity of response latency to</p><p>an attitudinal inquiry as a measure of the chronic ac-</p><p>cessibility of the attitude. As indicated earlier, evi-</p><p>dence regarding such validity is provided by previous</p><p>research (Fazio et al. 1986) demonstrating that this</p><p>measure relates to the likelihood of automatic activa-</p><p>tion of the attitude upon exposure to the object.</p><p>Those objects for which an individual could respond</p><p>relatively quickly to a direct attitudinal query also</p><p>were likely to activate the attitude from memory</p><p>upon their mere presentation. In contrast, objects for</p><p>which response latencies to an inquiry were slow</p><p>showed little evidence of producing automatic attitu-</p><p>dinal evaluation upon their presentation.</p><p>Although we can be confident that the latency mea-</p><p>sure reflects the chronic accessibility of the attitude,</p><p>it must be recognized that the present investigation</p><p>is correlational in nature. Attitude accessibility was</p><p>measured and not manipulated. Thus, whatever di-</p><p>mensions are naturally associated with attitude acces-</p><p>sibility may have contributed to the differences that</p><p>were observed with respect to attitude-behavior con-</p><p>sistency. For example, familiarity may moderate the</p><p>attitude-behavior relation. Because the present corre-</p><p>lational investigation involved the measurement of</p><p>attitude accessibility and because objects with which</p><p>individuals are familiar are likely to be ones toward</p><p>which they have accessible attitudes, the distinct</p><p>moderating influences of attitude accessibility and fa-</p><p>miliarity cannot be ascertained from the present data.</p><p>286 THE JOURNAL OF CONSUMER RESEARCH</p><p>However, two counterpoints should be noted. First,</p><p>experimental work in which attitude accessibility was</p><p>manipulated directly, by varying the number of times</p><p>that subjects expressed their attitudes, has indicated</p><p>a causal impact of attitude accessibility upon both at-</p><p>titude-behavior consistency (Fazio etai. 1982, Exper-</p><p>iment 4) and attitude-judgment consistency (Hous-</p><p>ton and Fazio 1989). Although further experimental</p><p>research is undoubtedly needed to isolate the causal</p><p>influence of attitude accessibility, this finding does</p><p>bolster our confidence that the present results are not</p><p>due solely to the common dependence of attitude ac-</p><p>cessibility and attitude-behavior consistency on some</p><p>third variable, such as familiarity.</p><p>Second, we should note, as did Fazio and Williams</p><p>(1986), what we view as the primary advantage of fo-</p><p>cusing upon attitude accessibility as opposed to some</p><p>other strength-related dimension of attitude, such as</p><p>confidence in one's attitude (e.g., Fazio and Zanna</p><p>1978), the manner of attitude formation (e.g., Fazio</p><p>and Zanna 1981), affective-cognitive consistency</p><p>(e.g., Norman 1975), or amount of information about</p><p>the attitude object (e.g., Davidson et al. 1985). Unlike</p><p>other indicants of attitude strength, the construct of</p><p>attitude accessibility operates at an information-pro-</p><p>cessing level of analysis. As a result, it is much more</p><p>clearly relevant to the issue of the process by which</p><p>attitudes guide behavior. Although other indicants of</p><p>the strength of an attitude may be associated with atti-</p><p>tude accessibility and, hence, with our classification</p><p>of subjects in the present study, the construct of atti-</p><p>tude accessibility has clear implications with respect</p><p>to the basic mechanisms involved in the production</p><p>of attitude-consistent behavior.</p><p>Indeed, it has been postulated that various identi-</p><p>fied moderators of the attitude-behavior relation may</p><p>exert their impact because they reflect the strength of</p><p>the object-evaluation association and, hence, the ac-</p><p>cessibility of the attitude (Fazio 1986). This has been</p><p>demonstrated with respect to one such moderator—</p><p>the manner of attitude formation. Attitudes based</p><p>upon direct behavioral experience with the object</p><p>have been found to be both more predictive of later</p><p>behavior (Fazio and Zanna 1981) and more accessi-</p><p>ble from memory (Fazio et al. 1982, 1983) than atti-</p><p>tudes based upon indirect experience. Similar evi-</p><p>dence has been obtained with respect to the moderat-</p><p>ing variables of personal importance of the attitude</p><p>issue (Krosnick forthcoming) and of individual</p><p>differences in self-monitoring tendencies (Kardes et</p><p>al. 1986). The same may hold true for other variables</p><p>that have been identified as moderators of the atti-</p><p>tude-behavior relation. If so, the construct of attitude</p><p>accessibility and the process model may provide a</p><p>conceptual framework for integrating this set of mod-</p><p>erators and for understanding how and why they in-</p><p>fluence attitude-behavior consistency.</p><p>Affect Referral</p><p>The attitude-behavior process model bears some</p><p>similarity to Wright's (1975) theoretical statements</p><p>regarding what he terms affect referral. Affect referral</p><p>is a decision-making strategy in which the individual</p><p>avoids reviewing any specific attribute information</p><p>concerning the alternatives, but instead relies upon</p><p>previously formed, global affective judgments of the</p><p>alternatives. Although Wright did not detail any</p><p>mechanisms or process by which such affect referral</p><p>occurs, the affect referral strategy that he mentions is</p><p>one that our model obviously endorses. Indeed, the</p><p>model can be viewed as specifying the mechanism un-</p><p>derlying, and the conditions necessary for, such affect</p><p>referral. Position along the attitude/non-attitude</p><p>continuum dictates the extent to which an affect re-</p><p>ferral strategy is possible. A strong object-evaluation</p><p>association in memory and, hence, a highly accessible</p><p>attitude is a prerequisite for such affect referral. In</p><p>such a case, the attitude will be activated from mem-</p><p>ory and influence, if not completely determine, the</p><p>individual's appraisal of the object in the immediate</p><p>situation. The sparser the immediate environment is</p><p>with respect to the availability of new information</p><p>about the attitude object, the less critical the selective</p><p>processing component ofour model becomes. In a sit-</p><p>uation in which no new information is presented (as</p><p>in the present investigation), the selective processing</p><p>amounts to an effortless application of the attitude</p><p>stored in memory as one's appraisal of the object in</p><p>the immediate situation—much as Wright appears to</p><p>mean by his affect referral strategy. With a highly ac-</p><p>cessible attitude, the immediate appraisal is more</p><p>likely to be congruent with the attitude stored in</p><p>memory than it is when a relatively inaccessible atti-</p><p>tude is involved. Active construction of an immediate</p><p>appraisal is not necessary.</p><p>In contrast, the selective processing component as-</p><p>sumes more importance in an environment that is in-</p><p>formation rich. When the situation makes new infor-</p><p>mation about the object available (as in the Fazio and</p><p>Williams voting behavior study or in an ongoing so-</p><p>cial interaction), interpretations of this information</p><p>will be colored by an activated attitude. As a conse-</p><p>quence of what is potentially extensive selective pro-</p><p>cessing in this case, immediate appraisals are more</p><p>likely to be congruent with attitudes among individu-</p><p>als whose attitudes are highly accessible than among</p><p>individuals whose attitudes are not. Thus, the extent</p><p>of selective processing depends upon the degree to</p><p>which the situation provides new information. In ei-</p><p>ther case, the attitude-behavior process model points</p><p>to the importance of correspondence between percep-</p><p>tions of the object in the immediate situation and atti-</p><p>tudes toward the object. Such correspondence is more</p><p>likely when the attitude involves a strong object-eval-</p><p>uation association and is capable of automatic activa-</p><p>ATTITUDE ACCESSIBILITY 287</p><p>tion from memory upon mere observation of the ob-</p><p>ject.</p><p>Final Implications</p><p>The findings from the present investigation, as well</p><p>as our discussion of them, attest to the relevance of</p><p>the attitude-behavior process model to the consumer</p><p>behavior domain. The accessibility of an attitude</p><p>from memory does appear to moderate the extent to</p><p>which that attitude guides product</p><p>selection behav-</p><p>ior. A major implication of this finding concerns the</p><p>role of advertising as a social influence agent intended</p><p>to promote the desired consumer behavior. Appar-</p><p>ently, inducing a positive attitude toward the brand is</p><p>not in and of itself sufficient to have much influence</p><p>upon consumer behavior. The general suggestion</p><p>offered by the model and the present findings is that</p><p>the desired behavior is most likely to occur if the atti-</p><p>tude is highly accessible from memory. Thus, if the</p><p>goal of advertising is to influence purchase behavior,</p><p>then one needs to be concerned not only with the va-</p><p>lence of the resulting attitude but also with its accessi-</p><p>bility from memory (see Herr and Fazio 1988 for a</p><p>more general discussion of this issue). How social in-</p><p>fluence attempts can be designed to accomplish this</p><p>end constitutes an important challenge for the future.</p><p>[Received August 1988. Revised May 1989.]</p><p>REFERENCES</p><p>Alba, Joseph W. and J. Wesley Hutchinson (1987), "Di-</p><p>mensions of Consumer Expertise," Journal of Con-</p><p>sumer Research, 13 (March), 411-454.</p><p>Baker, William, J. Wesley Hutchinson, Danny Moore, and</p><p>Prakash Nedungadi (1986), "Brand Familiarity and</p><p>Advertising: Effects on the Evoked Set and Brand Pref-</p><p>erence," in Advances in Consumer Research, Vol. 13,</p><p>ed. Richard J. Lutz, Provo, UT: Association for Con-</p><p>sumer Research, 637-642.</p><p>Boyd, Harper W., Michael L. Ray, and Edward C. Strong</p><p>(1972), "An Attitudinal Framework for Advertising</p><p>Strategy," Journal of Marketing, 36 (April), 27-33.</p><p>Davidson, Andrew R., Steven Yantis, Marel Norwood, and</p><p>Daniel E. Montano (1985), "Amount of Information</p><p>About the Attitude Object and Attitude-Behavior Con-</p><p>sistency," Journal of Personality and Social Psychol-</p><p>ogy, 49 (November), 1184-1198.</p><p>Day, George S. and Terry Deutscher (1982), "Attitudinal</p><p>Predictions of Choices of Major Appliance Brands,"</p><p>Journal of Marketing Research, 19 (May), 192-198.</p><p>Engel, James F. and Roger D. Blackwell (1982), Consumer</p><p>Behavior, Hinsdale, IL: Dryden.</p><p>Fazio, Russell H. (1986), "How Do Attitudes Guide Behav-</p><p>ior?" in The Handbook of Motivation and Cognition:</p><p>Foundations of Social Behavior, eds. Richard M. Sor-</p><p>rentino and E. Tory Higgins, New York: Guilford,</p><p>204-243.</p><p>(1989), "On the Power and Functionality of Atti-</p><p>tudes: The Role of Attitude Accessibility," in Attitude</p><p>Structure and Function, eds. Anthony R. Pratkanis et</p><p>al., Hillsdale, NJ: Lawrence Eribaum Associates, 153-</p><p>179.</p><p>(forthcoming), "A Practical Guide to the Use of Re-</p><p>sponse Latency in Social Psychological Research," in</p><p>Review of Personality and Social Psychology, Vol. 11,</p><p>eds. Clyde Hendrick and Margaret S. Clark, Beverly</p><p>Hills, CA: Sage.</p><p>, Jeaw-Mei Chen, Elizabeth C. McDonel, and Steven</p><p>J. Sherman (1982), "Attitude Accessibility, Attitude-</p><p>Behavior Consistency, and the Strength of the Object-</p><p>Evaluation Association," Journal of Experimental So-</p><p>cial Psychology, 18 (July), 339-357.</p><p>, Martha C. Powell, and Paul M. Herr (1983), "To-</p><p>ward a Process Model of the Attitude-Behavior Rela-</p><p>tion: Accessing One's Attitude Upon Mere Observa-</p><p>tion of the Attitude Object," Journal of Personality and</p><p>Social Psychology 44 (April), 723-735.</p><p>, David M. Sanbonmatsu, Martha C. Powell, and</p><p>Frank R. Kardes (1986), "On the Automatic Activa-</p><p>tion of Attitudes," Journal of Personality and Social</p><p>Psychology, 50 (February), 229-238.</p><p>and Carol J. Williams (1986), "Attitude Accessibil-</p><p>ity as a Moderator of the Attitude-Perception and Atti-</p><p>tude-Behavior Relations: An Investigation of the 1984</p><p>Presidential Election," Journal of Personality and So-</p><p>cial Psychology. 51 (September), 505-514.</p><p>and Mark P. Zanna (1978), "Attitudinal Qualities</p><p>Relating to the Strength of the Attitude-Behavior Rela-</p><p>tionship," Journal of Experimental Social Psychology,</p><p>14 (4), 398-408.</p><p>and Mark P. Zanna (1981), "Direct Experience and</p><p>Attitude-Behavior Consistency," in Advances in Ex-</p><p>perimental Social Psychology, Vol. 14, ed. Leonard</p><p>Berkowitz, New York: Academic Press, 161-202.</p><p>Herr, Paul M. and Russell H. Fazio (forthcoming), "The</p><p>Attitude-to-Behavior Process: Implications for Con-</p><p>sumer Behavior," in Ad Exposure, Memory, and</p><p>Choice, ed. Andrew A. Mitchell, Hillsdale, NJ: Law-</p><p>rence Eribaum Associates.</p><p>Houston, David A. and Russell H. Fazio (1989), "Biased</p><p>Processing as a Function of Attitude Accessibility:</p><p>Making Objective Judgments Subjectively," Social</p><p>Cognition, 7 (Spring), 51-66.</p><p>Howard, John A. and Jagdish N. Sheth (1969), The Theory</p><p>of Buyer Behavior. New York: John Wiley.</p><p>Kardes, Frank R., David M. Sanbonmatsu, Richard T.</p><p>Voss, and Russell H. Fazio (1986), "Self-Monitoring</p><p>and Attitude Accessibility," Personality and Social</p><p>Psychology Bulletin, 12 (December), 468-474.</p><p>Kassarjian, Harold H. and Waltraud M. Kassarjian (1979),</p><p>"Attitudes Under Low Commitment Conditions," in</p><p>Attitude Research Plays for High Stakes, eds. John C.</p><p>Maloney and Bernard Silverman, Chicago: American</p><p>Marketing Association, 3-15.</p><p>Krosnick, Jon A. (forthcoming), "Attitude Importance and</p><p>Attitude Accessibility," Personality and Social Psy-</p><p>chology Bulletin.</p><p>Lutz, Richard J. (1975), "Changing Brand Attitudes</p><p>Through Modification of Cognitive Structure," Jour-</p><p>nal of Consumer Research, 1 (March), 49-59.</p><p>288 THE JOURNAL OF CONSUMER RESEARCH</p><p>Norman, Ross (1975), "Affective-Cognitive Consistency,</p><p>Attitudes, Conformity, and Behavior," Journal of Per-</p><p>sonality and Social Psychology, 32 (July), 83-91.</p><p>Page, Ellis B. (1963), "Ordered Hypotheses for Multiple</p><p>Treatments: A Significance Test for Linear Ranks,"</p><p>Journal of the American Statistical Association. 58</p><p>(March), 216-230.</p><p>Petty, Richard E., John T. Cacioppo, and David Schumann</p><p>(1983), "Central and Peripheral Routes to Advertising</p><p>Effectiveness: The Moderating Role of Involvement,"</p><p>Journal of Consumer Research, 10 (September), 135-</p><p>146.</p><p>Powell, Martha C. and Russell H. Fazio (1984), "Attitude</p><p>Accessibility as a Function of Repeated Attitudinal Ex-</p><p>pression," Personality and Social Psychology Bulletin,</p><p>10 (March), 139-148.</p><p>Roskos-Ewoldsen, David R. and Russell H. Fazio (1989),</p><p>"Do Affectively-Laden Objects Attract Attention?" pa-</p><p>per presented at the annual meeting of the Midwestern</p><p>Psychological Association, Chicago.</p><p>Ryan, Michael J. and E.H. Bonfield (1975), "The Fishbein</p><p>Extended Model and Consumer Behavior," Journal of</p><p>Consumer Research, 2 (September), 118-136.</p><p>Sanbonmatsu, David M. and Russell H. Fazio (1986), "The</p><p>Automatic Activation of Attitudes Toward Products,"</p><p>paper presented at the annual meeting of the Associa-</p><p>tion for Consumer Research, Toronto, Ontario, Can-</p><p>ada.</p><p>Seigel, Sidney and N. John Castellan (1988), Nonparamet-</p><p>ric Statistics for the Behavioral Sciences, New York:</p><p>McGraw-Hill.</p><p>Smith, Robert E. and William R. Swinyard (1983), "Atti-</p><p>tude-Behavior Consistency: The Impact of Product</p><p>Trial Versus Advertising," Journal of Marketing Re-</p><p>search, 20 (August), 257-267.</p><p>Wright, Peter (1975), "Consumer Choice Strategies: Sim-</p><p>plifying Vs. Optimizing," Journal of Marketing Re-</p><p>search, 12 (February), 60-67.</p>

Mais conteúdos dessa disciplina