How an Attention-Getting Device Can Affect Quick Choice Among Similar Alternatives

Robert M. Schindler, University of Chicago
Michael Berbaum, Brandeis University
Donna R. Weinzimer, Yeshiva University
ABSTRACT - Marketers have often been successful at using displays, unusual packaging, and other attention-getting devices to influence consumers when they sake a quick choice from among a set of similar alternatives. An experiment involving a series of lotteries provides evidence that at least some of this influence is due to consumers' information-processing limitations. Although this component of the influence occurs only as the number of similar alternatives exceeds seven, it should result from any device which draws the consumer's attention to a particular alternative.
[ to cite ]:
Robert M. Schindler, Michael Berbaum, and Donna R. Weinzimer (1987) ,"How an Attention-Getting Device Can Affect Quick Choice Among Similar Alternatives", in NA - Advances in Consumer Research Volume 14, eds. Melanie Wallendorf and Paul Anderson, Provo, UT : Association for Consumer Research, Pages: 505-509.

Advances in Consumer Research Volume 14, 1987      Pages 505-509

HOW AN ATTENTION-GETTING DEVICE CAN AFFECT QUICK CHOICE AMONG SIMILAR ALTERNATIVES

Robert M. Schindler, University of Chicago

Michael Berbaum, Brandeis University

Donna R. Weinzimer, Yeshiva University

ABSTRACT -

Marketers have often been successful at using displays, unusual packaging, and other attention-getting devices to influence consumers when they sake a quick choice from among a set of similar alternatives. An experiment involving a series of lotteries provides evidence that at least some of this influence is due to consumers' information-processing limitations. Although this component of the influence occurs only as the number of similar alternatives exceeds seven, it should result from any device which draws the consumer's attention to a particular alternative.

INTRODUCTION

Consumers are often faced with a situation where they must make a quick choice from a crowded market category where they have little basis for deciding among the many alternatives. The epitome of such a situation occurs when a consumer is looking through the Yellow Pages for, say, a furniture store or an insurance agent. Usually, there will be a very large number of alternatives and the consumer may have little basis for deciding which store or agency to call first. Other examples of such choice situations occur when a consumer chooses among numerous brands of, say, paper towels on a supermarket shelf, or when a driver picks a place to eat from the many restaurants lining the highway.

A common strategy marketers use to influence such situations is to do something to make their alternative catch the consumer's attention, or otherwise stand out perceptually from the other alternatives. For example, a roadside dairy restaurant might display a lifesize model of a cow, or an Italian food restaurant could have the "leaning tower of Pizza" in front of its entrance. In a supermarket, a brand of soap pads might use a package of iridescent orange, or a brand of paper towels might arrange special signage or an end-aisle display. And in the Yellow Pages, an insurance agency may invest in a half-page ad where most of the space is devoted to displaying the agency name and phone number in large print.

Research indicates that such marketing techniques are effective in influencing consumer choice. mere have been reports both in the trade press and the academic journals of the effectiveness of supermarket displays (Chevalier 1975; Curhan 1974; Progressive Grocer 1971) and of signs and other attention-getting materials at point-of-purchase (McKinnon, Kelly, and Robison 1981; Progressive Grocer 1977; Woodside and Waddle 1975). In an investigation of the consumer's responses to Yellow Pages advertising, Feldman and Halterman (1963) found that when consumers were asked to select the first druggist or plumber they would call from an unfamiliar phone book, they were 23 times more likely to choose a phone number displayed in an advertisement than one in an ordinary column listing. Further, the existence of this salience effect has also been demonstrated in a laboratory situation. Using a simple lottery task, Schindler and Berbaum (1983) found that subjects were more likely to choose a perceptually salient alternative, even though the payoffs of the choice were determined by a demonstrably random process.

If it is indeed the case that many common marketing techniques work by making one store name or package stand out perceptually from the array of competitors, then understanding the psychological mechanisms by which this perceptual salience can influence choice could lead to more effective use of these marketing techniques. As a first step toward understanding these mechanisms, the factors which could cause this salience effect can be grouped into two general categories.

The first category of factors includes those where the situational meaning which is ascribed to a salient alternative mediates the effect of salience on choice. For example, McKinnon, Kelly, and Robison (1981) hypothesize effects of point-of-purchase signage which are due to the price or benefit messages which are explicitly communicated by the signs. And it is also possible, as Chevalier (1975) suggests, that end-aisle displays may increase sales simply because consumers interpret them as indicating a special bargain. Or, concerning the Schindler and Berbaum (1983) experiment, the subjects may have interpreted salience (incorrectly) to be an indication of an alternative with a greater likelihood of payoff. In any particular marketing situation, consumers nay or may not give salience a meaning which will lead to selection of the perceptually salient alternative. But if such situational meanings are the only factors responsible for the effects of salience on choice, then the decision whether to try to make an alternative stand out from the others should be based entirely on an estimate of the meaning the relevant consumers are likely to ascribe to a salient alternative in that particular situation.

On the other hand, a second category of factors which could cause more frequent choice of salient alternatives includes those which involve the consumer's information processing limitations. SuCh limitations cause consumers to consciously perceive only a small portion of the information which reaches their senses (McGuire 1976, pp. 305-6). In particular, Miller (1956) reviewed evidence suggesting that these limits prevent a person from keeping in mind more than about seven items at any one time. This information processing limitation implies that consumers who are trying to choose among a large number of alternatives will not be able to keep the alternatives simultaneously in mind so they can be co par-d unless they simplify the situation by limiting consideration to an "evoked set" of approximately or fewer alternatives (Howard and sheth 1969, pp. 9899). Parkinson and Reilly (1979) proposed that selection into this evoked set occurs by consumers con id ring the alternatives one at a time, presumably until the evoked set becomes filled. If there are important differences between the alternatives, then these differences will be used to reject alternatives from the evoked set. But if the consumer regards the alternatives as similar, then there will be little basis for rejecting an alternative from the evoked set, and the evoked set will tend to fill up with the first seven or so alternatives perceived. Thus, if an attention-getting device causes one of these otherwise similar alternatives to be thrust into consciousness ahead of the others, it will have a very great likelihood of being included in the consumer's evoked set. This greater likelihood of being in the evoked set would, of course, result in the salient alternative having a greater than chance likelihood of being the final choice.

The present study was designed to determine whether situational meanings are sufficient to account for the effect of salience on choice or if a component concerning human information processing limitations also plays a role. Since such information processing limitations are thought to be relatively constant across people and situations, this component of salience effects would occur no matter how consumers interpret the salience. Of course, since consumers usually will have an interpretation of or a feeling about the attention-getting device, it is possible that this interpretation may be so negative that it outweighs the effects of processing-limitation factors.

The study involved measuring the effect of salience on choice using the lottery technique of Schindler and Berbaum (1983). This technique was designed to simplify the situational context by using numbers as alternatives and by using a simple darkening or lightening of a number as the attention-getting device. In addition, an explicitly random process (a "bingo drum") was used to select the winning numbers, so as to make it clear that the salient number did not have a greater chance than any other number of being the winning number.

The three experimental conditions differed only in the number of alternatives involved in each choice. In the first condition, one alternative stood out in a field of four alternatives; in the second condition, one alternative stood out in a field of 25; and in the third condition, one alternative stood out in a field of 100 alternatives. since the meaning that the subjects of the study ascribe to a salient alternative should not be affected by the number of alternatives, the size of the salience effects found should not differ between the three conditions if only the situational meaning factors are involved. However, if the effect of salience on choice is based on a high likelihood that the salient alternative will be included in the evoked set, then there should be different effects of salience in each of the three conditions. There should be no effect of salience in the four-alternative choices since this number is well within short-term memory capacity and thus all four alternatives can be considered. In the 25-alternative choices, salience should increase the probability of an alternative being chosen from 1/25 to around 1/7, under the assumption that the subjects' idiosyncratic strategies for choosing from the evoked set will result in each evoked set member having an equal chance of being the final choice. An even larger bias toward the salient alternative would be expected in the 100-alternative choices since the high likelihood of the salient alternative being included in the evoked set would give it an approximately 1/7 chance of being chosen while its probability of being chosen without the effects of salience would be only 1/100. The finding of such differences in the size of the salience effects between the three experimental conditions would provide evidence that information processing limitations play a role in causing these salience effects.

METHOD

Subjects

Eighty-four university undergraduates served as subjects in this study in return for credit toward a course requirement. The students were run in groups of approximately 15. In each group, small prizes were awarded to the winners of each round.

Materials

Each subject received a booklet containing three sets of 25 pages, each set separated by a green sheet of paper. A cover sheet displayed the lottery game's name, "Mark-It.

On each of the sheets in a set was printed a matrix of square boxes with each box containing a number. One number on each page was made salient by being solid black in a field of outlined numbers of the same size and typeface. The previous finding (Schindler & Berbaum 19 3) that matrices with a single black number in a field of outlined numbers and matrices with a single outlined number in a field of black numbers produced similar results made it unnecessary to use both types of matrix in this study. of the three 25-page sets in a booklet, one contained 2x2 matrices the four-alternative condition), one contained 5x5 matrices (the 25-alternative condition), and one contained 10x10 matrices (the 100-alternative condition). An example of each size matrix can be seen in the Figure.

FIGURE

EXAMPLES OF (A) THE 4-ALTERNATIVE MATRIX, (B) THE 25-ALTERNATIVE MATRIX, AND (C) THE 100-ALTERNATIVE MATRIX

For each size matrix a different method was used to determine the salient numbers. In the set of four-alternative matrices, a randomly chosen one of the four numbers was salient on seven of the 25 pages in the set, while each of the other three numbers was the salient one on six of the pages. In the set of 25-alternative matrices, each number was the salient one on one of the 25 pages in the set. In the set of 100-alternative matrices, 25 numbers were chosen randomly (without replacement) to serve as the salient ones.

The order of the pages was randomized separately for each of the matrix sets. Also, the order of the three sets in a booklet was randomized separately for each group of subjects.

Procedure

The subjects were told that "Mark-It" was a game in which prizes would be awarded by random drawing, and then the rules of the game were carefully explained (See Appendix). When the experimenter told the subjects to begin, each subject began selecting one number on each page by drawing a line directly through that number. The experimenter paced the subjects orally, allowing them about five seconds to choose a number and turn the page. This was done to encourage the subjects to make the kind of quick decision they might make, say, when choosing a brand of paper towels at the supermarket rather than to try to develop some overly elaborate strategy for this particular game situation.

After completing a set of 25 sheets, the subjects turned back to the first page in that set. The experimenter then rolled a wire bingo drum filled with numbered balls and randomly selected one number that appeared in the preceding matrices. The experimenter called out that number and any subject who had marked the called number on the first sheet in the set received a point. Then the subjects turned to the second page and the procedure was repeated. This continued until two subjects had received enough points to win and receive a prize (for the four-alternative set, four points w re needed to win; in the 25-alternative sets, two points were needed; in the 100-alternative sets only one point was needed). This entire sequence was then completed for the other two sets in the booklet. After the entire booklet was marked, the subjects were asked to complete a questionnaire containing open-ended questions about the strategies they had used in the game.

RESULTS

The number of salient alternatives chosen by a subject was computed separately for each of the three 25-page sets. The mean number of salient alternatives chosen and the number of salient alternatives which would be chosen due to chance alone are displayed in the Table. In the four-alternative choices the average number of times a subject chose the salient alternative did not differ significantly from the number of times any one alternative would be chosen by chance alone (t(83) = 1.54. D > .10). However, in the 25-alternative choices, salience increased the chances of an alternative being chosen from 1/25 to about 1/10 (t(83) = 3.13, p < .005). In the 100-alternative choices, salience increased an alternative's chances of being chosen from 1/100 to about 1/10 (t(83) - 3.52, p < .001). the salience effect for the 100-alternative choices was greater than that for the 25-alternative choices because while the chance probability of choosing any one number in the 100-alternative matrices was one-fourth that of the 25-alternative matrices, the average number of times the salient alternative was chosen in the two conditions did not differ significantly (t(83) - 0.33, n.s.).

TABLE

MEAN NUMBER OF SALIENT ALTERNATIVES CHOSEN AND THE NUMBER OF SALIENT CHOICES WHICH WOULD BE EXPECTED BY CHANCE ALONE FOR EACH OF THE THREE CHOICE SITUATIONS.

A tally of responses to the open-ended questionnaire items indicated that while 63% of the subjects used the words "random," "haphazard," or why chanceS to describe how they made their choices, 308 did report strategies which involved at least sometimes choosing the darkened square. (e.g., "sometimes I picked the darkened number and sometimes I didn't"). However, these strategies were not based on the belief that the darkened squares had a greater than chance likelihood of being chosen. When asked directly about the relative chances of each number being chosen, not one subject reported believing that the darkened numbers had any greater likelihood of being the winning number.

When asked directly what they felt about the darkened number, only eight subjects (108) wrote that they felt that the experimenter had expected them to mark that number. and of the eight, six reported that their choices were guided by strategies which had nothing to do with these expectations.

DISCUSSION

The pattern of results in the experiment supports the view that information processing limitations play a role in causing salience effects. When the choice was among four alternatives, all four could be considered and thus there was no consequence of salience causing an alternative to be considered first. However, when the choice was among 25 or 100 alternatives, most of the alternatives could not be considered. In these choices, the ability of salience to thrust an alternative into consciousness insured consideration of the salient alternatives and thus gave the salient alternative greater-than-chance probability of being chosen.

It could be argued, however, that there are other explanations which could account for the pattern of results. One possibility is that the subjects adopted the strategy of picking at least some salient numbers in each matrix set, just in case they are wore likely to be the winning numbers. The failure of direct post-experimental questioning to turn up even one subject who reported believing that salient alternatives were more likely to win makes this possibility quite unlikely. Another possibility is that the salience effects in the 25- and 100- alternative matrices are due to experimental de and effects (e.g., Sawyer 1975). again, the results of the direct questioning render this alternative unlikely. Further, it is unclear why a desire to conform to the experimenter's expectations would cause a bias toward the salient alternative in the 25- and 100-alternative matrices but not in the 4-alternative matrix .

The finding that the probability of the salient alternative being chosen was about the sane in the 100-alternative matrix as it was in the 25-alternative matrix is also consistent with the view that the subjects were bumping up against their processing limitations in those matrices. the finding that the salient alternative was chosen an average of once every ten choices in those matrices suggests an evoked set size of around ten alternatives. This is larger than Miller's (1956) estimate of the span of short-term memory and is larger than evoked set sizes found in previous research (e.g., Jarvis and Wilcox 1973; Ostlund 1973). However, Cavanagh (1972) found that the number of items which could be simultaneously maintained in short-term memory increased as the size and complexity of the items decreased. And, from examining the pattern of findings in evoked set research, Horton (1983, p. 76) concluded that the size of the evoked set increases as the complexity of the comparisons between the ideas decreases. Thus, the large evoked set size indicated by the present results fits very well with the quite mini al comparisons the subjects are likely to make between alternative numbers in a lottery task.

The support in this study for a processing-limitation component of salience effects suggests to the marketing practitioner that,

(1) displays and other attention-getting devices are especially appropriate in situations where consumers- must choose among a large number of poorly differentiated alternatives, and,

(2) an attention-getting device can have a substantial effect on sales even if it fails to have any relevant meaning to the consumer.

Furthermore, the present viewpoint raises the possibility that the processing-limitation component of the salience effect can occur also as a result of making a brand salient long before choice time. An item which stands out-in an array is more likely to be recalled than any one of the nonsalient items in the array (the "Von Restorff Effect"; e.g. Wallace 1965; Rundus 1971). If an especially attention-getting advertisement makes an item (e.g. a product type, brand, store, etc.) stand out from among its competitors, then the item will be easier to recall at the time when a choice must be made. This increased long-term memory accessibility will cause the item to be one of the first to come to mind, therefore making it virtually certain to be one of the small number of alternatives considered. Thus, whether the salience of an item acts directly or acts by increasing the accessibility of an item's memory trace, at least one component of its effect on sales may be its ability to capture a share of the limited processing capacity of short-term memory.

APPENDIX

INSTRUCTIONS TO SUBJECTS

We'd like you to play a game called Mark- It. This game involves giving out some prizes based on a random drawing. It is very simple to play.

When I tell you to begin, turn to the first page [show a page]. Choose one of the numbers on the page and mark that number by drawing a line directly through it. Choose any number you want, but please choose only one per page. ten, when I tell you to turn the page, go to the next page and repeat this procedure. we will continue like this through the pages of the booklet until we come to the green page. ten we will stop and see who's won the round.

Here's how we'll determine who wins a round: we'll turn back to the first page and turn this drum to pick a number at random. mere will be as many balls in the drum as there are numbers on the page and the selection of a number will be totally random. If the number we pick is the same as the number that you marked, then you get one point. we will then turn to the next page and will again pick a number at random. If this number is the same as the number you marked on the second page, then you get a point. we will continue this procedure until two of you have received enough points to win the round. (we will tell you how many points are necessary to win before we begin a round.) We will play three rounds of Mark-It today, so we will be giving out at least six prizes.

Please note that this drum allows us to draw the numbers totally at random [demonstrate the bingo drum. If you would like to be assured that this apparatus is indeed giving us a fair drawing, you are welcome to come up and examine the drum and the numbered balls inside.

Any questions?

REFERENCES

Cavanagh, J. Patrick (1972), "Relation Between the Immediate Memory span and the Memory Search Rate," Psychological Review, 79 (6), 525-530.

Chevalier, Michel (1975), "Increases in Sales Due to In-Store Displays," Journal of Marketing Research 12 (November), 426-431.

Curhan, Ronald C. (1974), "The Effects of Merchandising and Temporary Promotional Activities on the Sales of Fresh Fruits and Vegetables in Supermarkets," Journal of Marketing Research, 11 (August), 286-294.

Feldman, Sidney P. and Jean C. Halterman (1963), "Consumer Use of the Yellow Pages in Kansas," unpublished manuscript, Division of Business and Economics, Indiana University Northwest, Gary, IN 46408.

Horton, Raymond L. (1984), Buyer Behavior: A Decision- Making Approach, Columbus, OH: Charles E. Merrill Publishing Co.

Howard, John A. and Jagdish N. Sheth (1969), The Theory of Buyer Behavior, New York: John Wiley & Sons.

Jarvis, Lance P. and James B. Wilcox (1973), "Evoked Set Size - Some Theoretical Foundations and Empirical Evidence," in Combined Proceedings Fall Conference of the American Marketing Association, No. 35, ed. Thomas V. Greer, Chicago, IL: American Marketing Association, 236-240.

McGuire, William J. (1976), "Some Internal Psychological Factors Influencing Consumer Choice," Journal of Consumer Research, 2 (March), 302-319.

McKinnon, Gary P., J. Patrick Kelly, and E. Doyle Robinson (1981), "Sales Effects of Point-of-Purchase In-Store Signing," Journal of Retailing, 57 (Summer), 49.

Miller, George A. (1956), 'The Magical Number Seven, Plus or Minus Two: Some Limits on Our Capacity for Processing Information,' Psychological Review, 63, 81.

Mitchell, Andrew A. (1978), "An Information Processing View of Consumer Behavior,' in Research Frontiers in Marketing: Dialogues and Directions, ed. S. C. Jain, Chicago, IL.: American Marketing Association, 188-197.

Ostlund, Lynan E. (1973), 'Evoked Set Size: So e empirical Results,' in Combined Proceedings Fall Conference of the American Marketing Association, No. 35, ed. Thomas V. Greer, Chicago, IL: an Marketing Association. 226-230.

Parkinson, Thomas L. and Michael Reilly (1979), "An Information Processing Approach to evoked Set Formation," in Advances in Consumer Research, Vol. 6, ed. William L. Wilkie, Ann Arbor, MI: Association for Consumer Research 227-231.

Progressive Grocer (1971), "How In-Store Merchandising Can Boost Sales," (October), 94-97.

Progressive Grocer {1977), "How to Turn P-O-P Into Sales Dollars," (June), 86-89.

Rundus, Dewey (1971), "Analysis of Rehearsal Processes in Free Recall," Journal of Experimental Psychology, 89 (l), 63-77.

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

Schindler, Robert M. and Michael Berbaum (1983), "Influence of Salience on Choice," in Advances in Consumer Research, Vol. 10, eds. Richard P. Bagozzi and Alice M. Tybout, Ann Arbor, XI: Association for Consumer Research, 416-418.

Wallace, William P. (1965), "Review of the Historical, Empirical, and Theoretical Status of the Von Restorff Phenomenon," Psychological Bulletin, 63 (6), 410-424.

Woodside, Arch G. and Gerald Waddle (1975), "Sales Effects of In-Store Advertising," Journal of Advertising Research, 15 (June), 29-33.

----------------------------------------