journal of consumer research, inc. - miami

Journal of Consumer Research, Inc.

Preconscious Processing Effects: The Independence of Attitude Formation and ConsciousThoughtAuthor(s): Chris JaniszewskiReviewed work(s):Source: Journal of Consumer Research, Vol. 15, No. 2 (Sep., 1988), pp. 199-209Published by: The University of Chicago PressStable URL: http://www.jstor.org/stable/2489525 .

Accessed: 26/12/2012 12:49

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

The University of Chicago Press and Journal of Consumer Research, Inc. are collaborating with JSTOR todigitize, preserve and extend access to Journal of Consumer Research.

http://www.jstor.org

This content downloaded on Wed, 26 Dec 2012 12:49:48 PMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=ucpress

http://www.jstor.org/stable/2489525?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


Preconscious Processing Effects: The

Independence of Attitude Formation

and Conscious Thought

CHRIS JANISZEWSKI*

Two experiments investigate the formation of attitudes toward unattended stimuli. In Experiment 1, a presentation format that encourages processing at a preconscious level demonstrates that attitude formation can occur independently of conscious consideration. Alternative theoretical explanations are offered to account for the purported independence of conscious thought and preference formation, and Experiment 2 is a test of these alternatives. The results of Experiment 2 suggest that consumers use differential hemispheric strategies for task performance to form preconsciously based attitudes. A post hoc analysis is conducted to ad- vance more explicit claims about the operational nature of the underlying preconscious processes.

J ndividuals often approach consumption decisions in an uninvolved or passive fashion (Kassarjian

1981; Krugman 1965; Obermiller 1985; Olshavsky and Granbois 1979; Peter and Olson 1987). Interest in the passive processing positioh has stimulated investigation of learning mechanisms that presume little or no conscious cognitive effort, most notably, the mechanisms implied by AAd formation (Edell and Burke 1984; MacKenzie, Lutz, and Belch 1986; Mitchell 1986; Mitchell and Olson 1981; Shimp 1981), mere exposure effects (Obermiller 1985; Saw- yer 1975; Zajonc 1968), and classical conditioning (Bierley, McSweeney, and Vannieuwkerk 1985; Gorn 1982; Gresham and Shimp 1985). This article investigates the preference formation parameters of AAd, mere exposure, and classical conditioning that may operate independently of consciousness, but does not investigate the relationship between conscious and subconscious parameters of preference formation (see MacKenzie et al. 1986; Mitchell 1986; Mitchell and Olson 1981; Obermiller 1985).

Empirical investigations claiming to demonstrate attitude formation without the participation of conscious thought are methodologically and theoreti- cally controversial (Allen and Madden 1985; Dawson

1973; Dixon 1981; Page 1974). Methodologically, many current experimental procedures may be simply inadequate for assessing the relationship between conscious processing and attitude formation. Most empirical investigations use a posttest only design with a delayed measure of conscious processing or awareness (e.g., Bierley et al. 1985; Moreland and Za- jonc 1979; Wilson 1979). The delay in the measure of conscious processing requires individuals to "remember" their "awareness," thus rendering any comparison between treatment group awareness levels and chance awareness levels suspect (Page 1974). This paradox has led Merikle and Cheesman to con- clude that "despite numerous studies directed at es- tablishing whether or not unconscious perceptual processes provide sufficient information to influence higher level decision processes, there are still no generally accepted answers" (1987, p. 298).

Theoretical objections to the claim that lasting behavioral changes resulting from the presentation of information not consciously considered (Holender 1986) are grounded in the assumption that preattentive processes consist of extensive, temporary analy- ses that can encourage a shift in attention, but that are insufficient to create lasting behavioral changes (Greenwald and Leavitt 1984). Yet, evidence suggests that preattentive processing can influence subsequent verbal reports (Fowler et al. 1981; Marcel 1983; Mer- ikle and Cheesman 1987). If the response to a stimulus can be influenced by previous preconscious processing of the stimulus, or related stimuli, then con-

*Chris Janiszewski is an Assistant Professor of Marketing, Col- lege of Business Administration, University of Florida, Gainesville, FL 3261 1. The author would like to acknowledge the helpful com- ments of Joseph Alba, Chris Allen, Alan Sawyer, and three anony- mous reviewers on earlier versions of this article.

199 ? JOURNAL OF CONSUMER RESEARCH o Vol. 15 0 September 1988



200 THE JOURNAL OF CONSUMER RESEARCH

sumer researchers are currently ignoring a potentially influential dimension of consumer decision making.

Given the methodological difficulties associated with demonstrating that conscious processes have not occurred, two experiments were designed to demonstrate the lack of a relationship between conscious thought and the antecedents to attitude formation rather than the lack of conscious thought during attitude formation. Preference is a conscious experience (see Lazarus 1982), but the inputs used to form the preference can be subconsciously influenced. Dem- onstrations that preference altering manipulations cannot be causally related to conscious processes are used to support the contention that preconscious processing can result in meaningful changes in behavior. Subsequent to the empirical demonstrations, a post hoc analysis is conducted to differentiate between competing theoretical accounts of the preconscious processes that may be responsible for attitude formation.

DEMONSTRATING THE INDEPENDENCE OF CONSCIOUS

THOUGHT AND ATTITUDE FORMATION

One method of assessing whether conscious thought is causally related to preference formation might involve experiments with sets of manipulations in which varied levels of preference are formed, and a comparison between experimental conditions suggests consciousness is not responsible for the effect. Suppose a group of individuals randomly is assigned to read one of two versions of a newspaper. In the first version of the newspaper, an ad is placed to the left of an article. In the second version of the newspaper, the same ad is placed to the right of the same article. In each case, a second ad fills the open space on the opposite side of the article. Except for alternating the placement of these two ads, everything else in the newspaper is held constant.

After reading the article, subjects are asked to eval- uate the ad of interest and to provide an indication of whether they recognize the ad as having been in the newspaper. If the groups differ in their evaluation of the ad, but do not differ in their levels of recognition, then the results might suggest preferences had been formed independently of conscious processing. The constant levels of recognition, an indicator of conscious processing (see Holender 1986; Moreland and Zajonc 1979; Seamon, Brody, and Kauff 1983; Wil- son 1979; Zajonc 1980), across the two groups would imply no relationship between conscious thought and attitude formation.

CONTRALATERAL CONDUCTION, HEMISPHERIC PROCESSING

STYLES, AND PREATTENTIVE PROCESSING

Contralateral Conduction The proposed effects of placement on ad evaluation

depend on the neurological connection between the sensory receptors of the visual system and the brain hemispheres. The human visual system is organized so stimuli located within the individual's field of foveal vision, roughly 1.5 degrees to the left or right of the current point of focus (Berlucchi 1972), are sent to both hemispheres for processing. However, stimuli placed to the left of the current field of focus are initially sent to the right hemisphere for processing, and stimuli placed to the right of the current field of focus are initially sent to the left hemisphere for processing (Curtis 1968, pp. 766-767). Referred to as contralateral conduction, the instantiation of peripherally placed information in the opposing hemisphere is commonly accepted as a biological fact that forms the basis for almost all hemispheric investigations related to the processing of visual stimuli.'

Hemispheric Processing Styles The possibility that placement influences evalua-

tion of a peripherally placed stimulus stems from the hypothesis that the hemispheres have different processing styles (see Allen 1983; Alwitt 1981; Dimond 1972; Hansen 1981). Processing style refers to the procedure or process each hemisphere uses when attempting to perform a task.

The processing style of the right hemisphere can be described as holistic, because it uses a template matching procedure when attempting to perform any task. This procedure invokes templates or prototypes to give meaning to incoming information and applies alternative templates to elaborate upon incoming or stored information. Right hemisphere compatible tasks may have organizational requirements more congruous with the procedural abilities of the right hemisphere relative to the left hemisphere (see Allen 1983). Examples of right hemisphere compatible (RHC) tasks include the processing of music (Gordan 1983) and visuospatial information (Beaumont 1982) and the formation of inferences and conclusions (Brownell, Potter, and Bihrle 1986). Note that performance of each type of RHC task relies on the ability of the right hemisphere to simultaneously integrate multiple pieces of information.

'Extensive reviews of visual field placement effects can be found in Beaton (1985), Beaumont (1982), and Young (1982).



PRECONSCIOUS PROCESSING 201

The processing style of the left hemisphere is best described as unit-integrative. The left hemisphere recognizes well-learned individual units, then serially integrates or combines them into some meaningful whole. Left hemisphere compatible tasks may have organizational requirements more congruous with the procedural abilities of the left hemisphere relative to the right hemisphere. Examples of left hemisphere compatible (LHC) tasks include counting (Davidoff 1982), processing unfamiliar words (Searleman 1983) and single sentences (Brownell et al. 1986), phonetic and syntactical processing (Zaidel 1978), and expres- sion via language (Allen 1983; Searleman 1983). Note that performance of LHC tasks relies on the ability of the left hemisphere to store and combine a set of serially presented events or stimuli.

Preattentive Processing Preattentive processing refers to the simultaneous,

preconscious monitoring of all sensory channels for events that will require a shift in attention (see Green- wald and Leavitt 1984). Placement effects predicted to result from contralateral organization of the visual system and the unique processing styles of the two hemispheres are assumed to be a function of the preattentive processing of the peripherally placed information (Beaton 1985; Beaumont 1982; Dixon 1981; Young 1982). Information in the unattended sensory channels must be processed to a level that will allow the individual to decide whether the other events in his or her environment warrant a shift in attention (Greenwald and Leavitt 1984; Kahneman 1973). This preattentive or preconscious processing is hypothesized to be responsible for the placement effect predicted in the newspaper example.

Summary The existence of a preattentive processing mecha-

nism that surveys the environment prior to conscious consideration suggests that the contralateral organization of an individual's visual system could be used to encourage processing at different sites simply by manipulating the placement of information within the environment. Information residing within the individual's peripheral field of vision presents the potential for differentiated preattentive processing at the hemispheric level, as each hemisphere will attempt to use its unique procedural style to identify the incoming information. Preferences should be formed to the extent that the processes used in perception and conscious identification are similar to the processes used during preconscious identification. If hemispheric elaborations precede conscious perception, then preattentive processing of the stimulus should result in greater preference for the stimulus.

In addition, recognition measures on the peripherally placed stimuli should provide insight as to the extent attention has shifted, and conscious consideration has occurred, within each placement condition.2

EXPERIMENT 1 Experiment 1 investigates the manipulation sug-

gested in the newspaper example presented earlier. The ad of interest was a pictorial ad. Placing the pictorial ad to the left of the article should result in it being sent to the right hemisphere for processing, whereas placing the ad to the right of the article should result in it being sent to the left hemisphere for processing. When later evaluated, the pictorial ad should receive a more favorable evaluation when placed to the left of the editorial, because at the time of conscious evaluation the right hemisphere is used to holistically per- ceive the picture ad. This conscious perception should be easier because of similar preconscious processing in the right hemisphere.

METHOD

Stimuli Purely pictorial stimuli have been used as RHC

tasks in a variety of investigations (for a discussion, see Beaton 1985, pp. 68-72; Davidoff 1982; Young and Ratcliff 1983). The stimulus of interest was a pictorial ad for a furrier that depicted a woman wearing a hooded fur coat. The ad contained no text and had appeared in the Thursday edition of a metropolitan newspaper about three months prior to the running of the experiment. The stimulus ad was cut from a much larger ad. The ad's age and its method of con- struction ensured its likely novelty for experimental purposes.

The filler ad placed opposite the pictorial ad was selected to function as a blocking task. Blocking tasks discourage the interhemispheric transfer of information related to the studied activity (e.g., processing the pictorial ad) by engaging the hemisphere not of interest to the experimenter. Because the filler ad would be used to block in both experimental conditions, an attempt was made to equate the blocking potential of the ad. Thus, a Sulka Clothier ad composed of 112 words in no more than one-quarter inch type was cho- sen as the filler ad. The small typesize should result in degraded information being sent to the nonstudied hemisphere and, across conditions, equivalent levels

2Procedurally, it is crucial that the ad is viewed peripherally a majority of the time. If individuals shifted their attention and consciously considered the ad, the ad would be instantiated in both hemispheres and no difference in evaluation would occur. The effect depends on the advantage gained (lost) from instantiation in a compatible (incompatible) hemisphere.




of processing activity. For these reasons, the evaluation of the blocking ad was of no experimental interest.

Subjects Forty-three graduate students recruited from two

Marketing Management classes were assigned randomly to treatment groups. Subjects were paid cash for their participation in the experiment. Subjects were run in groups ranging in size from six to 10 over a two-day period.

Procedure Subjects were told they would read a series of arti-

cles, ostensibly because the experimenter was inter- ested in people's biases in the processing of certain words. Subjects were informed that the articles had been placed in a mock newspaper to encourage lei- surely reading and that such reading facilitated the manifestation of the biasing phenomenon. Subjects were told they could search for the key words within the text, but that the task would be quite difficult. The instructions generally increased interest in, and attention to, the articles, thus reducing the likelihood subjects would attend to the peripherally placed ad.

Subjects were instructed to read four articles marked with orange dots. The instructions stated subjects were to close the newspaper and await further instructions when they had finished reading. The newspapers were then distributed to the subjects.

Instrument The instrument looked similar to a student newspa-

per. Fourteen articles, 16 ads, two photos, and a cross- word puzzle were spread over 10 pages. The only in- consistency was the newspaper was made by stapling together a set of five 11 X 17 inch photocopied sheets rather than by assembling and folding a set of 22 X 11 inch press-printed sheets.

The four marked articles were special interest edi- torials selected from past issues of a student newspaper. The first article discussed student hostels and was placed on page two, a left-side page. It had no ads flanking it and was used to reinforce the experimental guise and thereby reduced subjects' interest in ac- tively processing the experimental ad flanking the second article. The second article discussed a Chinese bookstore and was positioned on the extreme left of page five, a right-side page. An ad was placed to its right, on the same page, and to its left, on the adjacent page. The ads were adjacent to the last eight of 20 paragraphs. The remaining area on the two pages was filled with two additional articles and six additional ads.

The placement of the pictorial ad to the left or the right of the article constituted the experimental manipulation. The pictorial ad occupied a visual space ranging from approximately 2.5 degrees to 17.5 degrees when the subject viewed the editorial margin closest to the ad and approximately 16 degrees to 29 degrees when the subject viewed the editorial margin furthest from the ad.3 Thus, the pictorial ad was out- side of foveal vision when subjects read the editorial.4 This manipulation did not ensure subjects would only attend to the editorial, as no physical or methodological constraints limited the subjects' attention. Subjects were free to shift their attention to the ad at all times. Support for the claim that errant attention to the ad was noninfluential would exist if there were equivalent recognition levels among experimental conditions and a lack of difference in ad evaluations between those subjects that recognized the ad and those that did not.

The third and fourth articles, on pages seven and eight respectively, discussed a hitchhiking professor and a local Spanish radio station. The reading of these articles acted as a filler task.

These procedures were designed to maximize subjects' motivation to engage in active processing of the articles and, at the same time, to provide considerable exposure to the peripheral stimulus of interest. Lead- ing subjects to believe that the experiment involved the reading of the text resulted in approximately a three-minute peripheral exposure to the picture ad. The three-minute exposure is significantly longer than the millisecond exposures often used in laterality studies.

Dependent Measure Booklet After all subjects had finished the reading tasks, the

newspapers were collected and a dependent measure booklet was distributed. The first page of the dependent measure booklet contained instructions that took approximately one minute to read. Following the instructions were 14 ads: the critical picture ad, the opposing ad, six filler ads, and six ads that had been placed elsewhere in the newspaper. Two filler ads were placed first and the remaining ads followed in a random order that was constant for all subjects. The picture ad was the ninth ad evaluated.

3Although acuity concerns often result in the restriction of peripheral presentations from two to six degrees from the point of fixation, researchers who have violated this methodological prac- tice have not had difficulty producing visual field effects (see Beau- mont 1982 for discussion).

4Again, the reader is encouraged to refer to any introductory biology text for a discussion of the contralateral properties of the visual system. The general acceptability of the paradigm for manipulating the location of instantiation is evidenced by its prevalent use in the Journal of Experimental Psychology during the last 10 years.




TABLE 1

EVALUATION OF PICTURE AD

Placement of pictorial stimulus Contextual recognition Left visual field Right visual field

n=19 n=20 No 33.3 (8.1) 28.3 (7.9)

n=2 n=2 Yes 32.5 (7.8) 25.5 (9.2)

n=21 n=22 Total 33.2 (7.9) 28.0 (7.8)

The dependent measures of interest were the subjects' evaluations of the picture ad, recognition of the ad, and familiarity with the ad. Subjects viewed the ads and indicated their initial impressions on five, nine-point, bipolar adjective items: Unappealing/Ap- pealing, Unattractive/Attractive, Bad/Good, Unlike- able/Likable, Unpleasant/Pleasant. After subjects had evaluated all ads, subjects were asked to indicate their gender, handedness, and whether they had read any of the articles prior to that day. Finally, the recognition and familiarity measures were administered. Subjects were asked to go back and view the ads a second time and indicate whether they recognized the ad as having been in the mock newspaper (Yes/No) and whether they had seen the ad prior to that day (Yes/No).

Results

The five items measuring the subjects' evaluative responses to the picture ad were found to be reliable (a = 0.93) and unidimensional (ML Confirmatory analysis; Goodness-of-Fit Index = 0.9; Root Mean Square residual = 0.03). The five items were summed for each subject and used as an indicator of a subject's evaluation of the ad. Bartlett-Box homogeneity test was not significant. A plot of the residuals supported the assumption of the independence of errors.

First, a t-test was performed to determine whether placement influenced preference for the pictorial ad. The pictorial ad was more preferred when placed to the left of the article: Left visual field X = 33.19, Right visual field X = 28.0, t(41) = 2.16, p < 0.05, c2 = 0.088). There was no influence of handedness, gender, or past exposure to the article.

Table 1 shows the evaluations of the picture ad cat- egorized by level of contextual recognition. A difference of proportions test was performed to assess whether the levels of recognition were equivalent between cells. A z-test was not significant (z = 0.05, 1 - B = 0.86). In addition, tests for a positive influence

of recognition on ad evaluations (F = 0.17) or a differential effect of recognition on ad evaluations by treatment condition (F = 0.06) were not significant.

DISCUSSION The results generally support the claim that prefer-

ences can be generated independently of conscious thought. The equivalent levels of recognition across experimental conditions suggest that the degree of conscious processing of the ad was equivalent in the two experimental conditions. If conscious processing of the ad was equivalent across the two conditions, then conscious processing could not be a causal agent responsible for the differences in preference for the pictorial ad. This does not imply that preferences are not experienced as a result of conscious consideration of some internal analysis, but it does imply that consciousness was not causally related to the change in internal states that was responsible for the significant differences in preference for the pictorial ad.

It is important to note that failure of the manipulation to restrict processing of the pictorial ad to a preattentive or preconscious level would probably have resulted in a null effect, because the differences in preference relied on the instantiation of the ad in a single hemisphere. Had subjects engaged in foveal viewing of the pictorial ad, both hemispheres would have had equal access to the information, resulting in no placement effect on ad evaluation. The low levels of recognition also support a conclusion of preconscious processing effects. Recognition levels for the pictorial ad (9 percent) were not significantly different from the recognition levels for the filler ads that were absent from all newspapers (14 percent), suggesting recognition was at a chance level. Chance recognition has often been used to support claims of noncon- scious processing (Wilson 1979; Zajonc 1980).

Given that the support for preconscious preference formation rests on the assumption that the placement effect is hemispherically based, alternative explanations for the placement effect must be identified and explored. The placement effect could have resulted from a perceptual bias too subtle to be detected by the recognition measure. For example, each time the eyes shifted left to read the next line, they may have drifted a little too far left and attended to the pictorial ad (Carpenter 1977). Because perceptual drift may result in conscious processing bursts only milliseconds in length, the recognition level would not be higher in the left placement condition. Hence, the difference in preference scores could be attributed to the perceptual drift bias associated with conscious processing rather than the hemispheric advantage associated with preconscious processing.

A second alternative explanation for the results of Experiment 1 is derived from the hemispheric litera-




ture. Researchers have hypothesized that low involvement processing is a right hemisphere task (Hansen 1981; Krugman 1971), resulting in a left visual field presentation bias for any type of stimulus. Therefore, the results of Experiment 1 could be attributed to the physiological fact that stimuli placed in the left visual field have access to a low involvement processing mechanism located in the right hemisphere. Experiment 2 was designed to test if either perceptual drift and/or a specialized low involvement processor could be supported as rival explanations for the results of Experiment 1.

EXPERIMENT 2 Experiment 2 used the contrasting predictions of

hemispheric theory and the proposed alternative explanations to create an empirical test of the three po- sitions. The hemispheric advantage explanation hypothesizes that the differences in evaluation in Exper- iment 1 were the result of different processing styles at the hemispheric level. The right hemisphere is hypothesized to have an advantage for processing information with holistic orientations and the left hemisphere is hypothesized to have an advantage for processing information with a serial orientation. Therefore, adding a second set of conditions in which the pictorial stimulus used in Experiment 1 is re- placed with a verbal stimulus should allow the left hemisphere to enjoy a relative advantage and thus re- verse the direction of the placement effect.

In contrast, the low involvement processing explanation advanced by Hansen (1981) and Krugman (197 1) would predict that varying the stimulus should not affect preference formation. Regardless of the stimulus type, stimuli in the left visual field should have access to the low involvement processor in the right hemisphere, resulting in a left visual field placement effect.

To address the alternative hypothesis of perceptual drift, the focal task was altered from reading an article to searching for 14 hidden objects in a sketch of a Thanksgiving feast. Because an object search does not necessarily proceed in a left to right fashion, perceptual drift should be a less viable alternative hypothesis. Therefore, if perceptual drift was responsible for the placement effect in Experiment 1, it should pro- duce a null effect for the manipulations in Experi- ment 2.

METHOD

Stimuli The stimuli of interest were a pictorial ad for Air

France Airlines and a verbal ad for a cola soft drink. The pictorial stimulus was cut from an ad for Air

France that initially contained verbal and pictorial information. The experimenter constructed the verbal ad; the ad read in bold block letters "COKE IS IT," one word per line. The blocking stimulus, placed in the visual field opposite the target ads, was a verbal ad for an area nightclub. The method of ad construc- tion made it likely all ads were novel for experimental purposes.

Subjects Ninety-six undergraduate students from two uni-

versities were assigned randomly to treatment groups. Subjects were given extra credit for their participation in the experiment and were run in groups ranging in sizes from 15 to 23. A covariate analysis showed no influence of subject pool on stimuli evaluations and no interaction of subject pool with treatment vari- ables. Therefore, the groups of subjects were com- bined for the analysis.

Procedure The procedure was similar to the procedure in Ex-

periment 1 with three exceptions. First, the attended task was changed to a picture search task with an associated change in instructions. Second, subjects performed only one task, not four as in Experiment 1. Third, subjects were asked to cease engaging in the search task after two minutes and 30 seconds had elapsed. The search task had been calibrated to take about three minutes so the cutoff equated exposure times among treatments and discouraged processing of the peripherally placed stimuli.

Design Experiment 2 used a 2 X 2 between-subjects facto-

rial design in which type of peripherally placed stimulus (pictorial or verbal) and placement of the peripheral stimulus (left visual field or right visual field) were manipulated. The dependent measures were identical to those used in Experiment 1.

Predictions The rival hypotheses presented in the discussion

section for Experiment 1 predict three patterns of effects. The perceptual drift hypothesis predicts a null effect of the treatments because the attended task has been changed to a task that does not require left to right visual movements. The more random search pattern associated with locating an object hidden within a sketch should remove or at least attenuate any perceptual drift effect.

The specialized low involvement processor hypothesis (Hansen 1981; Krugman 1971) predicts a main




TABLE 2

EVALUATION OF PERIPHERALLY PLACED ADS

Placement of stimulus

Stimulus type Left visual field Right visual field

n=24 n=25 Pictorial 32.7 (7.8) 29.3 (8.5)

n=24 n=23 Verbal 21.5 (7.4) 26.3 (7.7)

effect of placement, but no effect of the type of peripherally placed stimulus. Noninvolved processing is hypothesized to occur in the right hemisphere, so a processing advantage should occur only when either stimulus is placed to the left of the attended task.

The hemispheric advantage hypothesis predicts an interaction between placement of the peripheral stimulus and type of the peripheral stimulus. The pictorial ad should be more preferred when placed to the left of the picture search task, and the verbal ad should be more preferred when placed to the right of the task.

RESULTS

The five items measuring the subjects' evaluative responses to the picture were reliable (a = 0.91) and unidimensional (ML Confirmatory analysis; Good- ness-of-Fit Index = 0.88; Root Mean Square residual = 0.033). The five items were summed for each subject and used as an indicator of a subject's evaluation of the ad. Bartlett-Box homogeneity test was not significant. A plot of the residuals supported the assumption of the independence of errors. Table 2 shows means and standard deviations.

A test for an interaction between the type of stimulus and the placement of the stimulus was statistically significant (F(1,92) = 6.52, p < 0.05, co2 = 0.047). A test for a main effect of placement was not significant (F(1,92) = 0.20). A test for a main effect of stimulus type was significant (F(1,92) = 19.18, p < 0.05, co2 = 0.15). Simple main effects tests of placement for each stimulus were significant for the verbal stimulus (F(1,92) = 4.39, p < 0.05, W2 = 0.03) and not significant for the pictorial stimulus (F(1,92) = 2.27). Handedness and gender did not exert influence.

Tests were performed to determine whether conscious processing could have been responsible for the observed interaction effect. Recognition levels for the pictorial ad were three of 24 in the LVF condition and five of 25 in the RVF condition. Recognition levels for the verbal ad were seven of 24 in the LVF condition and six of 23 in the RVF condition. Difference of proportion tests comparing recognition levels be-

tween conditions were not significant (Pictorial ad: z = 0.72; Verbal ad: z = 0.22). In addition, tests for a positive influence of recognition on ad evaluations (F = 0.06) or a differential effect of recognition on ad evaluations by placement (F = 0.13), stimulus type (F = 1.04), or a combination of the two (F = 0.12) were not significant.

DISCUSSION The results of Experiment 2 are concordant with

the hemispheric processing theory explanation. There was a significant interaction between placement and type of stimulus on evaluation of the stimulus. When the ads were placed so they could be instantiated in the hemisphere with a compatible processing style, they were evaluated more favorably than if they were initially represented within the incompatible hemisphere. A placement by stimulus type interaction effect is inconsistent with the alternative explanations of perceptual drift and a specialized low involvement processor.

Despite the encouraging results, caution must be exercised when embracing the hemispheric explanation of the placement by stimulus type interaction. The failure to replicate the placement effect on the pictorial stimulus used in Experiment 2 (see simple main effect test) suggests the possibility of problems with the robustness of the phenomenon or the con- ceptualization of the underlying process. The Air France ad used in Experiment 2 was more complex than the pictorial stimulus used in Experiment 1, suggesting the complexity of the peripherally placed stimulus may limit the extent to which hemispheric instantiation can influence preference formation. In addition, the finding of a placement effect on evaluation of a pictorial ad when a verbal attended task is used (Experiment 1) and a placement effect on a verbal ad, but not a pictorial ad, when a visual attended task is used (Experiment 2) suggests that attended stimuli may influence the processing of unattended stimuli. Therefore, an equally plausible explanation for the results is that the attended task interfered with the processing of the peripherally placed ad when the ad was instantiated in the hemisphere responsible for the processing of the attended task. This interference hypothesis is in direct contrast to the facilitation hypothesis advocated throughout the article and, therefore, is a serious threat to the veracity of the hemispheric advantage conclusion. Had ad evaluations from a no exposure control group been included in either design, the control group mean could have been used to differentiate between these competing hypotheses. A control group mean equivalent to the ad evaluation means in incompatible conditions would be supportive of a facilitation effect, but a control group mean equivalent to the ad evaluation




means in the compatible conditions would have been supportive of an interference effect.

The experimental results also illustrate preference formation can depend on previous processing (placement effect) and current, conscious processes (stimulus type effect). Varying the peripheral placement of an ad varied the hemisphere of instantiation and influenced the subsequent evaluation of the ad. The low levels of recognition suggest the hemispherically based antecedent to an evaluative judgment can be formed preconsciously. The failure of recognition to influence evaluation of the ad across conditions, or differentially within conditions, is further evidence for preconscious processes.5

Yet, it must be recognized that the stimulus type main effect dwarfs the preconscious processing effect (W2 of 0.15 as opposed to the placement W2 of 0.047). The preference for the pictorial ad over the verbal ad is testimony to the strength of the associations accessed at the time of evaluation. Therefore, the placement effect is probably the result of the initial repre- sentation priming a procedure instrumental in the later evaluative process, not the result of unconscious positive and negative associations. Currently, two popular viewpoints are consistent with this proposal.

Perceptual fluency (see Jacoby 1983) is one theoretical option. Perceptual fluency refers to the feeling of familiarity one experiences when encountering an ap- parently novel stimulus. Perceptual fluency, subsequent familiarity, and the associated enhancement of preference judgments are thought to accrue from the activated representations of stimuli that are independent of any elaboration or contextual reference. Ex- posure allows for the stimulus to be represented, hence, liked, but the length or acuity of exposure pre- vents the elaboration required to recognize the stimulus at a later time. As in Experiments I and 2, perceptual fluency effects are often independent of recognition (e.g., Johnston, Dark, and Jacoby 1985; Seamon et al. 1983; Seamon, Marsh, and Brody 1984).

A second explanation that might account for the hemispheric advantage effect is the "direct path" hypothesis advanced by Zajonc and his colleagues (Moreland and Zajonc 1979; Wilson 1979; Zajonc 1980). Zajonc hypothesizes that the direct path from the retina to the hypothalamus may allow for preference formation without the participation of cognition. The neural instantiation of the stimulus is responsible for later reports of preference. Similar to the perceptual fluency position, the "direct path" hy-

pothesis predicts an independence of preference and recognition. In contrast to the perceptual fluency position, the "direct path" argument does not propose an intermediary familiarity judgment.

The subjects' familiarity ratings of the target ads used in Experiments 1 and 2 may provide some insight into the validity of the alternative explanations for the hemispheric advantage effect. Recall that subjects were asked if they had seen the ads prior to the experimental session. Because the ads were novel, the "prior exposure" measure is conceptually similar to the Old/New measure of familiarity often used in preference formation investigations (e.g., Jacoby 1983; Johnson et al. 1985; Obermiller 1985; Seamon et al. 1983; Seamon et al. 1984; Wilson 1979). Famil- iarity was not significantly correlated with preference ratings for the target ads: -0.06 (Experiment 1) and 0.00 (Experiment 2) for the pictorial ads and 0.20 (Experiment 2) for the verbal ad. These correlations are inconsistent with the familiarity-based explanation advocated in the perceptual fluency literature.

Supporting a theoretical position with null effects is unwise, and, therefore, no conclusions can be drawn concerning the "direct path" explanation for preconscious affective responses. At best, the post hoc analysis of the familiarity ratings suggests the underlying process responsible for the differences in preference are inconsistent with the perceptual fluency position. Further study may isolate the preconscious processes responsible for the differences in preference that result from stimulus characteristics and placement manipulations.

IMPLICATIONS AND FUTURE RESEARCH ISSUES

The results have implications for the areas of AAd and classical conditioning research. Previous researchers have suggested that AAd may be formed as a result of mood transfer (Lutz 1985) and execution likability (Batra and Ray 1985). However, the afore- mentioned results suggest a third dimension, the organization of information within the ad, may be involved in the formation of attitude toward the ad. For example, Ellis and Miller (1981) suggest that some organizational formats (e.g., verbal information on the right side of the ad and pictorial information on the left side of the ad) are preferred over others (e.g., verbal information on the left side of the ad and pictorial information on the right side of the ad). The organizational effect may accrue from the preconscious processing of one type of information while the individual is attending to a second type of information. Con- sistent with other proposed antecedents to AAd formation, the organization of the stimuli within the

5A placement compatible with the processing styles of the hemispheres (i.e., pictorial stimulus in the LVF or a verbal stimulus in the RVF) could possibly motivate short shifts in attention that are insufficient for recognition. An eye scanning apparatus or a more sensitive recognition measure could have tested the veracity of this rival hypothesis.




ad would generally not be a consciously identifiable influence of preference formation.

The findings also have implications for classical conditioning research. Currently, classical conditioning researchers are debating the role of awareness in conditioned responses. Researchers adopting a cognitive perspective (see Allen and Madden 1985; Dawson 1973; Page 1974) contend that contingency awareness may be necessary for the development of a conditioned response, but researchers adopting a be- havioristic perspective (see Gorn 1982; McSweeney and Bierley 1984) contend the investigation of internal states is inconsistent with the behaviorist meta- physic. However, cognitivists claim that attitude change, the dependent measure behaviorists often use to investigate conditioning within the consumption context, is an event often associated with cognitive interpretation. Experiments 1 and 2 suggest some antecedents to affective responses may be formed independently of conscious processes. If this is the case, one can imagine opportunities to create conditioned affective responses independent of conscious thought. For example, one may be able to vary an individual's attitude toward a brand by varying the dis- tance between affectively loaded descriptors and a brand name (e.g., either descriptors directly follow the brand name or are separated from the brand name by a series of words). Demonstrating that attitudes may be formed by varying sentence structure would suggest the conditioning procedure is much more powerful than currently hypothesized.

The results should also encourage investigations into the possibility that preconscious processes can create meaningful changes in behavior. Recent research has shown that subconscious processes are involved in preference judgments, free associations, story and fantasy productions, the learning of gram- matical rules, lexical decisions, word definition, and perceptual identification (see Holender 1986; Schacter 1987). Demonstrations that some components of preference formation and learning may occur with limited conscious monitoring should encourage consumer researchers to attempt to identify and explain the automatic processing that may paral- lel conscious processes.

The hemispheric principles used to create the empirical demonstrations of preconscious preference formation represent a possible approach for investi- gating and predicting behavioral changes that might result from preconscious processing of stimuli. Al- though the hemispheric principles used to hypothe- size the placement effect were limited to intrahemi- spheric processes, extensive literature addresses interhemispheric processes (see Allen 1983) that could be adapted to study the processing of simultaneously presented stimuli (e.g., picture-verbal, picture-picture, and verbal-music ads). For example, hemi-

spheric theory could be used to investigate the influence of peripherally presented material on the processing of attended material (i.e., how recall, recognition, comprehension, and elaboration of a message are influenced by other simultaneously presented stimuli). The ability to predict and demonstrate such effects would not only intensify interest in studying the influence of the message environment on attended processing, but could also establish links between levels of processing that are often assumed not to exist (Greenwald and Leavitt 1984).

CONCLUSION

The possibility that preferences may be formed independently of conscious consideration suggests that consumer attitudes may be influenced without consumers being able to identify the antecedents or the processes responsible for that influence. If forming attitudes independently of conscious consideration is possible, insight could be gained through a thorough understanding of the parameters of the phenomenon. The evidence presented suggests that differences in the ability of each hemisphere to represent a stimulus can be linked to preference formation. The usefulness of hemispheric theory for hypothesizing preconsciously based attitude formation effects suggests that it may be a valuable framework for future investigations into the influence of nonattended cues on subsequent behavior.

[Received November 1987. Revised March 1988.1

REFERENCES Allen, Max (1983), "Models of Hemispheric Specializa-

tion," Psychological Bulletin, 93 (1), 72-104. Allen, Chris T. and Thomas J. Madden (1985), "A Closer

Look at Classical Conditioning," Journal of Consumer Research, 12 (December), 301-315.

Alwitt, Linda F. (198 1), "Two Neural Mechanisms Related to Modes of Selective Attention," Journal of Experi- mental Psychology: Human Perception and Perfor- mance, 7 (April), 324-332.

Batra, Rajeev and Michael L. Ray (1985), "How AdvertIs- ing Works at Contact," in Psychological Processes and Advertising Effects: Theory, Research, and Applica- tions, eds. Linda F. Alwitt and Andrew A. Mitchell, Hillsdale, NJ: Lawrence Erlbaum Associates, 13-43.

Beaton, Alan (1985), Left Side, Right Side, New Haven, CT: Yale University Press.

Beaumont, J. Graham (1982), "Studies With Verbal Stim- uli," in Divided Visual Field Studies of Cerebral Orga- nization, ed. J. Graham Beaumont, London: Academic Press, 58-86.

Berlucchi, G. (1972), "Anatomical and Physiological As- pects of Visual Functions of Corpus Callosum," Brain Research, 37 (January), 371-392.




Bierley, Calvin, Frances K. McSweeney, and Renee Van- nieuwkerk (1985), "Classical Conditioning of Prefer- ences for Stimuli," Journal of Consumer Research, 12 (December), 316-323.

Brownell, Hiram H., Heather H. Potter, and Amy M. Bihrle (1986), "Inference Deficits in Right Brain-Damaged Patients," Brain and Language, 27 (March), 310-321.

Carpenter, Roger H.S. (1977), Movement of the Eyes, Lon- don: Pion.

Curtis, Helena (1968), Biology, New York: Worth. Davidoff, Jules (1982), "Studies With Non-verbal Stimuli,"

in Divided Visual Field Studies of Cerebral Organiza- tion, ed. J. Graham Beaumont, London: Academic Press, 30-57.

Dawson, Michael (1973), "Can Classical Conditioning Oc- cur Without Contingency Learning? A Review and Evaluation of the Evidence," Psychophysiology, 10 (January), 82-86.

Dimond, Stuart J. (1972), The Double Brain, New York: Halstead.

Dixon, Norman F. (1981), Preconscious Processing, New York: John Wiley.

Edell, Julie A. and Marion C. Burke (1984), "The Moderat- ing Effect of Attitude Toward an Ad on Ad Effective- ness Under Different Processing Conditions," in Ad- vances in Consumer Research, Vol. 11, ed. Thomas C. Kinnear, Provo, UT: Association for Consumer Re- search, 644-649.

Ellis, Andrew W. and Diane Miller (1981), "Left and Wrong in Adverts: Neuropsychological Correlates of Aesthetic Preference," British Journal of Psychology, 72 (May), 225-229.

Fowler, Carol A., George Wolford, Ronald Slade, and Louis Tassinary (1981), "Lexical Access With and Without Awareness," Journal of Experimental Psychology: General, 110 (September), 341-362.

Gordan, Harold W. (1983), "Music and the Right Hemi- sphere," in Functions of the Right Cerebral Hemi- sphere, ed. Andrew W. Young, London: Academic Press, 65-86.

Gorn, Gerald J. (1982), "The Effects of Music in Advertis- ing on Choice Behavior: A Classical Conditioning Ap- proach," Journal of Marketing, 46 (Winter), 94- 101.

Greenwald, Anthony G. and Clark Leavitt (1984), "Audi- ence Involvement in Advertising: Four Levels," Jour- nal of ConsumerResearch, 11 (June), 581-592.

Gresham, Larry G. and Terence A. Shimp (1985), "Attitude Toward the Advertisement and Brand Attitudes: A Classical Conditioning Perspective," Journal ofAdver- tising, 14 (1), 10-17.

Hansen, Flemming (1981), "Hemispherical Lateralization: Implications for Understanding Consumer Behavior," Journal of Consumer Research, 8 (June), 23-36.

Holender, Daniel (1986), "Semantic Activation Without Conscious Identification in Dichotic Listening, Para- foveal Vision, and Visual Masking: A Survey and Appraisal," The Behavioral and Brain Sciences, 9 (March), 1-66.

Jacoby, Larry L. (1983), "Perceptual Enhancement: Persis- tent Effects of an Experience," Journal ofExperimental Psychology: Learning, Memory, and Cognition," 9 (March), 21-38.

Johnston, William A., Veronica Dark, and Larry L. Jacoby (1985), "Perceptual Fluency and Recognition Judg- ments," Journal of Experimental Psychology: Learn- ing, Memory, and Cognition, " 11 (March), 3-11.

Kahneman, Daniel (1973), Attention and Effort, Engle- wood Cliffs, NJ: Prentice-Hall.

Kassarjian, Harold H. (1981) "Low Involvement: A Second Look," in Advances in Consumer Research, Vol. 8, ed. Kent B. Monroe, Ann Arbor, MI: Association for Con- sumer Research, 31-34.

Krugman, Herbert E. (1965), "The Impact of Television Advertising: Learning Without Involvement," Public Opinion Quarterly, 29 (Fall), 349-365.

(1971), "Brain Wave Measures of Media Involve- ment," Journal ofAdvertisingResearch, 11 (February), 3-19.

Lazarus, Richard S. (1982), "Thoughts on the Relation Be- tween Emotion and Cognition," American Psycholo- gist, 37 (September), 1019-1024.

Lutz, Richard J. (1985), "Affective and Cognitive Anteced- ents of Attitude Toward the Ad: A Conceptual Frame- work," in Psychological Processes and Advertising Effects: Theory, Research, andApplications, eds. Linda F. Alwitt and Andrew A. Mitchell, Hillsdale, NJ: Law- rence Erlbaum Associates, 45-64.

MacKenzie, Scott B., Richard J. Lutz, and George E. Belch (1986), "The Role of Attitude Toward the Ad as a Me- diator of Advertising Effectiveness: A Test of Compet- ing Explanations," Journal of Marketing Research, 23 (May), 130-143.

Marcel, Anthony J. (1983), "Conscious and Unconscious Perception: Experiments on Visual Masking and Word Recognition," Cognitive Psychology, 15 (April), 197- 237.

McSweeney, Frances K. and Calvin Bierley (1984), "Recent Developments in Classical Conditioning," Journal of Consumer Research, 11 (September), 619-631.

Merikle, Philip M. and Jim Cheesman (1987) "Current Sta- tus of Research on Subliminal Perception," in Ad- vances in Consumer Research, Vol. 14, eds. Melanie Wallendorf and Paul Anderson, Provo, UT: Associa- tion for Consumer Research, 298-302.

Mitchell, Andrew A. (1986), "The Effect of Verbal and Vi- sual Components of Advertisements on Brand Atti- tudes and Attitude Toward the Advertisement," Jour- nal of Consumer Research, 13 (June), 12-24.

and Jerry C. Olson (1981), "Are Product Attribute Beliefs the Only Media of Advertising Effects on Brand Attitudes?" Journal of Marketing Research, 18 (Au- gust), 318-332.

Moreland, Richard L. and Robert B. Zajonc (1979), "Expo- sure Effects May Not Depend on Stimulus Recogni- tion," Journal of Personality and Social Psychology, 37 (June), 1085-1089.

Obermiller, Carl (1985), "Varieties of Mere Exposure: The Effects of Processing Style and Repetition on Affective Response," Journal of Consumer Research, 12 (June), 17-30.

Olshavsky, Richard W. and Donald H. Granbois (1979), "Consumer Decision Making-Fact or Fiction?" Jour- nal of Consumer Research, 6 (September), 93- 100.




Page, Monte M. (1974), "Demand Characteristics and the Classical Conditioning of Attitudes Experiment," Journal of Personality and Social Psychology, 30 (4), 468-476.

Peter, J. Paul and Jerry C. Olson (1987), Consumer Behav- ior, Homewood IL: Erwin and Associates.

Sawyer, Alan G. (1975), "Demand Artifacts in Laboratory Experiments in Consumer Research," Journal of Con- sumer Research, 1 (March), 20-30.

Schacter, Daniel L. (1987), "Implicit Memory: History and Current Status," Journal of Experimental Psychology: Learning, Memory, and Cognition, 13 (September), 501-518.

Seamon, John G., Nathan Brody, and David Kauff(1983), "Affective Discrimination of Stimuli That Are Not Recognized: Effects of Shadowing, Masking, and Cere- bral Laterality," Journal of Experimental Psychology: Learning, Memory, and Cognition, 9 (September), 544-555.

, Richard L. Marsh, and Nathan Brody (1984), "Crit- ical Importance of Exposure Duration for Affective Discrimination of Stimuli That Are Not Recognized," Journal of Experimental Psychology: Learning, Mem- ory, and Cognition, 1O (September), 465-469.

Searleman, Alan A. (1983), "Language Capabilities of the Right Hemisphere," in Functions of the Right Cerebral

Hemisphere, ed. E. Percemen, New York: Academic Press, 87-111.

Shimp, Terence A. (1981), "Attitude Toward the Ad as a Mediator of Consumer Brand Choice," Journal ofAd- vertising, 10 (2), 9-15.

Wilson, William R. (1979), "Feeling More Than We Can Know: Exposure Effects Without Learning," Journal of PersonalityandSocialPsychology, 37 (June), 811-821.

Young, Andrew W. (1982), "Methodological Theoretical Bases," in Divided Visual Field Studies of Cerebral Or- ganization, ed. J. Graham Beaumont, London: Aca- demic Press, 11-29.

and Graham Ratcliff( 1983), "Visuospatial Abilities of the Right Hemisphere" in Functions of the Right Ce- rebral Hemisphere, ed. Andrew W. Young, London: Academic Press, 1-32.

Zaidel, Edward (1978), "Lexical Organization in the Right Hemisphere," in Cerebral Correlates of Conscious Ex- perience, eds. Pierre A. Buser and Arlette Rougeul- Buser, New York: North-Holland, 177-197.

Zajonc, Robert B. (1968), "Attitudinal Effects of Mere Ex- posure," Journal of Personality and Social Psychology Monograph Supplement, 9 (June), 1-27.

(1980), "Feeling and Thinking: Preferences Need no Inference," American Psychologist, 35 (February), 151-175.



journal of consumer research, inc. - miami

Documents