Effects of interactive pictures and ethnicity on recall of brand names
McKelvie, Stuart J
When people desire a product or service, they often have a wide range of alternatives from which to choose. One goal of advertising is to provoke consumers to think of particular brand or company names when the need arises. Because there is a delay between exposure to an advertisement and the purchasing decision, the effect of the message will depend on the consumer’s memory (Keller, 1987). This process may be viewed as a paired-associate learning task in which recall of one item (the brand or company name) is prompted by another (the name of the product or service). For example, for Pilgrim Fences, it is hoped that the consumer will recall “Pilgrim” when considering fences (Esser, Die, Seholm, & Pebley, 1986).
According to Meyers-Levy (1989), brand names perform three functions for consumers and advertisers: They identify products so that people can differentiate the various alternatives, they influence consumer attitudes towards the product, particularly their perception and evaluation of it and, when the brand names are already familiar, they can be used to increase acceptance of a new offering (the brand extension effect
Meyers-Levy, Louie, & Curren, 1994
). In all three cases, it is important to establish what dimensions make a brand name memorable (Meyers-Levy, 1989).
Research on memory for brand names has shown that encoding processes affect recall(Keller, 1987; Saegert & Young, 1983). Particular attention has been given to the effect of pictures on memory for advertisements (Childers & Houston, 1986; Edell & Staelin, 1983; Houston, Childers, & Heckler, 1987).
EFFECTS OF PICTURES ON RECALL
Print advertisers often use pictures to illustrate the paired brand and product names. This practice is consistent with dual-coding memory theory (Paivio, 1971), according to which recall is enhanced if information is coded in two systems (verbal, imagery) rather than one (usually verbal). The easiest way to portray the names is to include a picture in which each component is displayed separately (e.g., a pilgrim in one part of the picture and a fence in another). Although Paivio (1971, pp. 251, 350) suggests that these pictures would increase prompted recall, research with concrete nouns and with simulated advertising material has produced mixed results. When separate pictures accompany words, performance has been higher (Biron & McKelvie, 1986; Epstein, Rock, & Zuckerman, 1960; Paivio & Yarmey, 1986), lower (MacGregor & McKelvie, 1990; Wollen, Weber, & Lowry, 1972), and not different (Esser et al., 1986; McKelvie, Daprat, Monfette, & Cooper, 1993; Morelli, 1970; Wollen & Lowry, 1971) from performance with words alone.
In contrast, there is consistent evidence that recall is enhanced by pictures in which the two components are combined (e.g., the pilgrim is painting the fence). When such “interactive pictures” (Lutz & Lutz, 1977) accompany words, recall is better than with separate pictures or with names alone (Childers & Houston, 1984; Epstein et al., 1960; Esser et al., 1986; Kerst, 1976; Lutz & Lutz, 1977; McKelvie, Cooper, & Monfette, 1992; McKelvie et al., 1993; McKelvie, Sano, & Stout, 1994; Senter & Hoffman, 1976; Wollen & Lowry, 1971; Wollen et al., 1972). These results may be accounted for as follows by Paivio’s (1971) imagery mediation hypothesis. Illustrated names are coded both verbally and visually. When a name alone (e.g., fences) is given during the memory test, it arouses a simple visual image corresponding to its illustration (an image of a fence). In turn, that image arouses a compound image corresponding to the interactive picture (an image of a pilgrim painting a fence). Finally, the second simple component of this image (pilgrim) permits retrieval of the verbal response (“pilgrim”). Following others (McKelvie et al., 1993), we propose that this explanation entails two requirements: (a) the two illustrations should be clearly connected (so that the first simple image will arouse the compound image which in turn arouses the image of the second component), and (b) the two individual illustrations should clearly represent the names that they are designed to portray (so that the names given during the test will arouse their simple images and the second simple images will permit retrieval of the response names).
The first requirement is supported by the greater effect of interactive than of separate pictures on recall. Evidence for the second requirement is provided by Wollen and Lowry’s (1971) finding that interactive pictures did not enhance recall when the illustrations were associated arbitrarily with words pairs. The additive effect of both requirements was demonstrated when the illustrations for brand and product names were separate and did not directly represent the names (MacGregor & McKelvie, 1990). Recall with pictures and words was poorer than with words alone.
Although print advertisers have been advised to use interactive pictures (Alesandrini, 1983; MacInnis & Price, 1987), they rarely do so, preferring instead to keep the illustrations separate or to illustrate only the product or the brand (MacGregor & McKelvie, 1990; McKelvie et al., 1992). One reason may be that many brand or company names are meaningless–e.g., Rona Hardware, store in Lennoxville, Quebec, or abstract–e.g., Johnson Solicitors–(McKelvie et al., 1992), and difficult to illustrate (Alesandrini, 1983; Houston et al., 1987; Lutz & Lutz, 1977), so that the second imagery mediation requirement is hard to satisfy.
Although many potential brand names cannot be directly portrayed, it is important to find out if the positive effect of interactive pictures on recall generalizes to this situation. One purpose of the present experiments was to examine the effect of interactive pictures with illustrations that did not directly represent the brand and product names. Because advertisers are often faced with such names (Alesandrini, 1983; Houston et al., 1987; Lutz Lutz, 1977), this provides a more realistic test of the effectiveness of these pictures than does past research, where the words have usually been concrete nouns or brand and product names that were easy to illustrate (e.g., Lion Matches
Esser et al., 1986
, King Chair
Lutz & Lutz, 1977
, Fox Window
McKelvie et al., 1994
). The present drawings, which were richer in detail than the sketches in previous studies, were also more representative of general print advertising. For example, although Lutz and Lutz (1977) used stimuli from the Yellow Pages, these materials are simpler than the illustrations in newspapers and magazines.
In formulating a hypothesis for the effect of interactive pictures that did not directly represent brand and product names, we considered imagery mediation theory and past research. It was argued above that the higher recall with interactive pictures than with names alone was due to the satisfaction of both imagery mediation requirements. However, with illustrations that did not directly represent the names, the second requirement was only weakly met, which implies that the effect of interactive pictures would disappear or be weaker than with illustrations that clearly represented the names. Because Wollen and Lowry (1971) found no positive effect of interactive pictures on recall when they were related arbitrarily with word pairs and because the present illustrations had some relationship to their names, albeit indirect, it is likely that the effect of interactive pictures would be positive, but weaker than before.
Three studies have examined the effect of interactive pictures on recall when the second imagery mediation requirement was not completely satisfied. Paired-associate recall was higher with than without interactive pictures for word pairs that were both higher and lower in rated imagery (McKelvie et al., 1992, 1993; Wollen t Lowry, 1971). Recall was also enhanced by interactive pictures for people who were unfamiliar with some words because they were learning English (McKelvie et al., 1992). These results suggest that interactive pictures are effective under two conditions (lower name imagery, lower name familiarity) that apply to many brand names that advertisers might wish to use. Every study with simulated advertising material has shown that interactive pictures have a positive effect on recall of brand names, with a mean standardized effect size (d, Cohen, 1977) of 1.20 (McKelvie et al., 1992). This exceeds Cohen’s (1977, p. 26) standard of d = 0.80 for a “large” effect. Of particular interest here, experiments with conditions in which illustrations were both more and less directly related to the names showed a similar size of effect of interactive pictures on recall. This result does not match our theoretical expectation that the effect of interactive pictures would be weaker when illustrations were linked indirectly to the names. One reason for the positive effect of interactive pictures in this condition may be that people expended extra effort to understand the connections. There is evidence that processing with mild discrepancy and incongruity is more extensive than with complete congruity, and that it enhances recall if the incongruity is resolved (Houston et al., 1987; Meyers-Levy et al., 1994). Thus, on the basis of past work with lower imagery and less familiar nouns, the following research hypothesis was formulated for the present interactive pictures:
H sub 1 : Prompted brand name recall will be higher with interactive pictures than with names alone and the effect size will be close to d = 1.20.
To provide another test of the first imagery mediation requirement that illustrations be combined, and because advertisers often employ separate pictures, we included them in a third experimental condition. On the basis of the results of previous comparisons, a second hypotheses was formulated:
H sub 2 : Recall will be higher with interactive pictures than with separate pictures.
Finally, previous studies of the effect of separate pictures compared to names alone have produced mixed results, but because they had a negative effect on recall when they were indirectly related to names and were realistic representatives of print advertising (MacGregor & McKelvie, 1990), we predicted:
H sub 3 : Recall will be lower with separate pictures than with names alone.
As noted above, this prediction is consistent with the assumption that the two imagery mediation requirements have additive effects. Recall should be poorest in this condition because the illustrations were not connected and because their link to the names was weak.
EFFECTS OF ETHNICITY IN ADVERTISING
Pictures may enhance cued brand name recall via imagery mediation but, like brand names, they are also used to influence consumer attitudes towards the product (Childers & Houston, 1986; Edell & Staelin, 1983). For example, it was found that a picture of a fluffy kitten increased the perceived softness of tissue (Mitchell & Olson, 1981). However, varying the content of pictures and brand names may also enhance memory for the association between brand and product names by providing additional encoding cues that capture consumer interest. In particular, illustrations are likely to capture attention because people look at pictures before text in advertisements (Houston et al., 1987). If brand illustrations receive additional processing and they are connected to product illustrations, recall of the brand-product association may be improved.
One theme that has been extensively employed in brand names and in pictures is ethnicity, where national and racial groups are often portrayed in a stereotyped manner (MacGregor & McKelvie, 1990). Such brand names and pictures may be used to imply that the product possesses characteristics associated with the ethnic group, and may also enhance recall of the brand and product names by encoding elaboration. For example, the brand name “Viking,” which is attached to many appliances in Canada, suggests that the product is rugged. Scot Foto in Laval, Quebec (sales promotional flyer, 1994) may indicate that the products are a bargain. In an advertisement for the International Flower Bulb Centre of Holland, the clean, innocent image of a Dutch child in national dress may enhance the perceived freshness of the flowers (Good Housekeeping, 1992). On the other hand, conjuring up a less savory image, the brand name “Bandito” for a video chain is illustrated by a grinning unshaven picture of a Mexican Bandit (MacGregor & McKelvie, 1990). This brand name and image may indicate that the consumer will be “captured” by the product or, more ominously, that the store specializes in violent films.
One problem with ethnic brand names and pictures in advertising is that groups have objected to their stereotyped portrayals (African-Americans
DeMont, 1991; MacGregor, 1992
Meyers, 1990; Strauss, 1992
). For example, although the Hornell Brewing Company may have labelled a beer “Crazy Horse” to imply that the drink had strong effects, the Oglala Sioux were outraged (Teinowitz, 1992; The Burlington Free Press, 1992). They objected to the implied negative stereotype of the drunken Indian and they felt that the name of a revered person had been dishonoured. Such hostile reactions may increase as the population becomes more ethnically mixed. For example, in Canada, by the year 2001, 17.7% of the population will consist of visible minorities and will control at least 20% of the Gross Domestic Product (Samuels, 1992). In the United States, by 2010, slightly more than 33% of children will be black, Hispanic, or Asian (Schwartz & Exter, 1989).
Of course, as the Dutch example demonstrates, ethnic brand names and images in advertising are not all pejorative. Indeed, recognizing the multicultural nature of the society, the Government of Canada has adopted a policy of representing visible and ethnic minorities in all federal advertising and communications (Minister of Supply and Services, 1988). Furthermore, surveys indicate that people believe that multiculturalism is enriching (Feschuk, 1993; Multiculturalism and Citizenship Canada, 1991), and that visible minorities should be more represented in advertising (Canadian Advertising Foundation, 1992a, 1992b). In addition, executives believe that advertisements showing minorities are more effective than those that feature whites only (Canadian Advertising Foundation, 1992b). Consequently, advertisers are adjusting their market strategies to reflect ethnic diversity (Feschuck, 1993).
Although ethnic brand names and pictures in advertising may provoke both positive and negative reactions, they might convey messages about the products and arouse consumer interest. But do they have any effect on brand name recall? In the only study of this question, it was found that, when subjects were exposed to both ethnic and nonethnic material recall for both kinds of brand names was similar either when they were shown alone or accompanied by separate pictures of brands and products (MacGregor & McKelvie, 1990). On this basis, we formulated the following research hypothesis:
H sub 4 : Cued recall will not differ between ethnic and nonethnic brand names when they are presented alone or accompanied by separate pictures.
Although the two kinds of brand names in themselves were equally memorable, the ethnic illustrations may have aroused additional interest, thereby leading to more elaborate processing. However, because the pictures were separate, this processing may not have affected memory for the association between brand and product. Because the present experiment also included interactive pictures that were expected to enhance recall over names alone (H sub 1 ), it was proposed that ethnicity in these pictures would provide a cue that would also enhance processing of the product or service illustration and name:
H sub 5 : Recall will be higher for ethnic than nonethnic brand names when they are accompanied by interactive pictures.
Our indirect relationship between illustrations and names was particularly realistic for ethnic brand names, which are usually difficult to portray. For example, the Swedish furniture company IKEA presents the name in yellow letters on a blue background, hinting at the ethnic group via the colours of the Swedish flag. The present experiments also retained another realistic feature that has been standard practice in previous work with simulated advertising material: The brand and product names were not connected through pre-established associations. Some current brand names in the marketplace have such associations with the product (e.g., Bandito Tortilla Chips has a Mexican connection, Marketing, 1991), which is likely to enhance memory for the paired names via shared elements (Kanungo, 1968; McKelvie et al., 1994), but many do not: Aussie Hair Fixative (People, 1993), Cherokee Clothes (Women’s Day, 1992), Jacques Cartier Pizza (MacGregor & McKelvie, 1990), Nordic Toyota (The Burlington Free Press, 1993), Mac’s Milk (network of stores in Ontario), Mohawk Carpets (The Record, 1986), Old Dutch Potato Chips (Native Network News, 1989), Samurai Camera (Campus Canada, 1989), and Scot Foto (Laval, Quebec, sales promotion flyer, 1994).
In summary, the purpose of Experiment 1 was to investigate the effect of pictures (mode of presentation) and of ethnicity (content) on recall of brand names given product names. Although the within-subjects design is more sensitive to the independent variable than the between-subjects design because it partials out the effect of individual differences, it was only employed for content. Mode of presentation was varied between subjects because it has been found that people exposed to both interactive and separate pictures can be cued to integrate mentally the separate illustrations from their experience with the interactive ones (McKelvie et al., 1992).
Subjects. Participants were 97 volunteer English-speaking undergraduates, mostly from English-and French-Canadian backgrounds. They were allocated randomly to the three mode conditions, with the constraint that the number of men and women was similar. The sample sizes were 16 women and 15 men (Verbal), 16 women and 16 men (Verbal Separate Picture) and 16 women and 18 men (Verbal Interactive Picture).
Materials. To ensure that any differences in recall between ethnic an nonethnic names and pictures was due to ethnicity, materials were matched on variables that influence recall. For names, the variables were imagery, familiarity, pair association and list categorization, which all have positive effects (Deese, 1959; Paivio, 1971; Paivio & Begg, 1981). List categorization was controlled by having three kinds of name: varied (no category), ethnic (ethnic category), and occupational (occupational category). The other variables were dealt with by selecting experimental material as follows.
Initially, 140 names (112 brands, 28 products) were rated for ease of imagery arousal (7-point scale) and for familiarity point scale) by two separate groups of university students (n = 84, 79 respectively). The potential brand names were related to ethnic groups (e.g., Bandito for Mexican) or to occupations (e.g., Beret for Artist), or represented varied categories (e.g., Lava for Volcano). On the basis of the mean and standard deviation of the two ratings for each word, 42 brand-product name pairs were formed by matching the 14 ethnic, 14 occupational, and 14 varied brand names on their imagery and familiarity ratings (frequency distribution control, Christensen 1991, p. 242). Similarly, the product names in each condition were matched on both variables. Natural associations between the brand and product names were avoided.
For each pair, an artist created a separate picture consisting of two black-and-white line drawings, one for each member of the pair, and an interactive picture in which the two drawings were combined. To enhance realism, they were modelled closely on real examples of ethnic advertisements. The artist was given a description of the content of each drawing to ensure that it was not directly representative of its name (e.g., the brand name Cactus was represented by a drawing of a stereotypic Mexican with sombrero and bullet belt; the product spectacles was represented by an eyechart; see Figure 1). (Figure 1 omitted) In addition, he maintained consistency across the three content categories on variables considered important in picture memory (Sharp & Markham, 1992; Snodgrass & Vanderwart, 1980): line complexity, size, whether the size of the brand or product was altered in the combined pictures, and bizarreness in the combined pictures. The ethnic pictures in each set were stereotypic but, to reflect the different reactions to ethnicity in advertising, connotations varied from positive (e.g., Danish Viking helmet indicating strength) to negative (e.g., Chinese mandarin wearing a pigtail) to neutral (e.g., an Arabian pyramid).
Three sets of 42 photographic slides were assembled: name pairs alone, name pairs plus separate pictures, and name pairs plus interactive pictures. In the first case, the names were printed clearly in capital letters in black ink and appeared in the middle of the slide. The brand name always appeared to the left of the product name. For both sets of pictures, the names were placed above the drawings in the centre of the slide. In the separate pictures, the brand drawing was placed on the left of the product drawing.
For words alone, 24 students rated the degree of natural association between the names in each pair (4-point scale). For words plus separate pictures, another 25 students rated visual complexity for each drawing (7-point scale), degree of association between each name and its corresponding drawing (7-point scale), and degree of association between the names themselves (4-point scale). For words plus interactive pictures, a third group of 32 students rated visual complexity of the total picture (7-point scale), ease with which each part of the picture could be identified (7-point scale), the degree to which the two components were perceived as interacting together (7-point scale), and how bizarre, strange or unusual they perceived the whole picture (7-point scale). Degree of interactiveness was rated because it should be separately controlled from bizarreness (Kroll, Schepeler, & Angin, 1986). Finally, to control for the ease with which the ethnic and occupational categories could be identified, a fourth group of 2A students classified each of the brand names into the ethnic, occupational, or varied categories.
From the mean scores for these ratings, we assembled a set of stimulus pairs (Set A, see Table 1) that was matched on all variables across the 8 ethnic, 8 occupational, and 8 varied pairs. (Table 1 omitted) To check matching, analyses of variance were conducted on the ratings with the stimulus as the unit of analysis. For all variables except category identification of brand names, there was no significant effect of content, showing that matching was successful across the ethnic, occupational and varied categories. However, the varied category was easier to identify than the ethnic category, which was in turn easier to identify than the occupational category. It was also found that product names were easier to image and were more familiar than brand names, and that the rated association between drawings and names was higher for products than for brands (M = 5.2 and 4.1 out of 7, respectively). These effects have been found before (McKelvie et al., 1993), and reflect the realistic fact that product names are more concrete than brand names. The rating of 4.1 also shows that we successfully created stimuli in which the illustrations represented the brand names indirectly. On the basis of these results, Set A was used in Experiment 1. Figure 1 shows examples of stimuli in the separate and interactive picture conditions.
Ratings for the remaining 18 stimuli (6 ethnic, 6 occupational, 6 varied; Set B, see Table 1) were also analyzed. Product names were successfully matched across content categories for imagery and for familiarity, but ethnic names were less familiar than the occupational, or varied brand names and, as in Set A, were easier to identify as ethnic than were occupational names as occupational. However, separate occupational product drawings were rated as less closely linked to their names and as more complex than ethnic or varied drawings. As in Set A, product names were easier to image and were more familiar than brand names, and the rated association between drawings and names was higher for products than for brands (M = 5.3 and 4.2 out of 7, respectively).
Procedure. Groups of about 10 were tested in a lab oratory. The procedure was very similar to that inaugurated by Lutz and Lutz (1977) and employed by other investigators (e.g., Esser et al., 1986; MacGregor & McKelvie, 1990; McKelvie et al., 1993). Participants were informed that the purpose of the study was to investigate memory for content often found in newspaper and magazine advertising. They were told that they would view slides on which there was information about a brand or company name and about a product or service, and that they should inspect each slide carefully for a subsequent memory test. However, the precise nature of this test was not described. Each of the slides in the verbal (names alone), verbal separate picture (names plus separate drawings), and verbal interactive picture (names plus interactive drawings), conditions was shown for ten seconds in a single presentation order. After a five-minute break with casual conversation and administration of test instructions, subjects completed a cued recall test in which they were given the product names in one of two random orders, both of which were different from the presentation order. In the space to the left of the product names, they wrote as many of the corresponding the brand names that they could remember.
Following the practice of previous investigators, answers were scored as correct if they were associated with the appropriate product name and if they contained no more than two spelling errors. Table 2 shows the mean scores in each condition. (Table 2 omitted)
A 3 x 2 x 3 (Mode x Gender of Subject x Content) mixed model analysis of variance, with repeated measures on the last factor, gave a significant effect of mode, F(2, 91) = 7.13, p
Effects of pictures. It was predicted that recall would be higher with interactive pictures than with names alone (H sub 1 ) or with separate pictures (H sub 2 ), and that recall with names alone would be higher than with separate pictures (H sub 3 . The confirmation of H sub 2 is consistent with previous research and demonstrates the importance of the first imagery mediation requirement that the illustrations for each name should be connected. The results also confirmed H sub 3 and replicated a previous finding that recall with realistic separate illustrations that were indirectly related to the names was poorer than with names alone (MacGregor & McKelvie, 1990). This suggests that the two imagery mediation requirements have additive effects.
However, contrary to H sub 1 , recall with interactive pictures did not exceed that with names alone, and the effect size of d (.04) was not close to the anticipated 1.20. This novel result conflicts with previous studies in which recall with interactive pictures that were indirectly related to their corresponding names was higher than with names alone (McKelvie et al., 1992, 1993; Wollen & Lowry, 1971). Because the lack of effect of interactive pictures compared to names alone was unexpected from past research, this comparison was reinvestigated in Experiment 2 using stimulus Set B. Like Set A, illustrations did not directly represent names, and they were equated across ethnic, occupational and varied content on the interactive picture variables.
On the basis of Experiment 1, the following prediction was formulated:
H sub 6 : Recall for names with interactive pictures will not differ from that with names alone.
Effects of ethnicity. It was predicted that performance would be similar for ethnic and nonethnic names from various categories when the names were shown alone or accompanied by separate pictures (H sub 4 ), but that recall would be better for ethnic than nonenthnic names with interactive pictures (H sub 5 ). The second novel finding of Experiment 1 was that, for names alone, names plus separate pictures, and names plus interactive pictures, recall was best with varied names followed by ethnic names followed by occupational names. Because previous work with the first two conditions showed no effect of ethnicity on performance (MacGregor & McKelvie, 1990), the results confirm that ethnicity does enhance recall. In addition, although it was expected that ethnicity would have a positive effect with interactive pictures, this prediction was based on the assumption that these pictures would themselves enhance recall. Because the effect of interactive pictures was examined in Experiment 2, the effect of ethnicity was also reinvestigated.
Occupational recall was even poorer than ethnic recall. Although this difference may have occurred because the occupational brand names were more difficult to classify than the ethnic brand names, the pre-experimental judgments were made by subjects who were explicitly informed not only that the names were categorized, but also what the categories were. It is possible that experimental subjects did not spontaneously observe the categories or differentially identify them. To study this matter in Experiment 2, an additional interactive picture condition ended with a post-experimental questionnaire to tap subject’s knowledge of the stimulus categories.
This group was also tested to evaluate the possibility that, with the incomplete memory instructions that did not specify precisely what would be tested later, subjects shown pictures may have attended more to the drawings than to the names (MacGregor McKelvie, 1990). Not only do people usually inspect pictures before text (Houston et al., 1987), but they are likely to dwell more on the pictures when they are interesting and are not directly related to their names (Edell & Staelin, 1983), two conditions that were met here. If our subjects were distracted by the pictures, they may not have linked the drawings to the names. This issue was addressed in the second interactive picture condition in Experiment 2 by explicitly informing subjects that memory for names would be tested, and by telling them to examine all the information provided.
Although the stimuli in Set B were equated on the extraneous variables pertaining to the interactive pictures, the brand names were not perfectly matched for familiarity, where ratings were lower for ethnic than for occupational or varied names. If this factor was important, the reduction in ethnic recall compared to varied recall might be greater in Experiment 2 than in Experiment 1, and might not exceed occupational recall, as it had in Experiment 1. Consequently, the following hypothesis was formulated:
H sub 7 : With or without interactive pictures, recall will be better for varied names than for ethnic or occupational names. The effect size for the difference between varied and ethnic names will exceed d = 0.45.
Subjects. The 92 student participants were allocated randomly to the three mode conditions, with gender controlled: Verbal (16 women, 16 men), Verbal Interactive Picture (16 women, 16 men), and Verbal Interactive Picture Informed (16 women, 12 men).
Materials and Procedure. Stimulus Set B was used, but the procedure for the verbal and verbal interactive picture conditions was the same as in Experiment 1. However, in the verbal interactive picture informed condition, subjects were also told explicitly that memory for the brand and product names would be tested but that they should carefully inspect all the information on each slide. Although they were informed of the name memory test, subjects were not told that it would be cued recall.
When this test was complete, these subjects wrote answers to a series of orally-presented questions about the brand names: (i) Did you notice any particular categories among the brand names (yes or no)? If yes, name them; (ii) Indicate yes or no to show whether you think these categories were present: birds, occupations, vegetables, ethnic, animals, or fruits; (iii) Given that two of these categories were present, pick them.
A 3 x 2 x 3 ( Mode x Gender of Subject x Content) mixed model analysis of variance, with repeated measured on the last factor, was conducted on the correct recall scores. Only the effect of content was significant, F(2, 172)= 11.43, p
For the post-experimental questionnaire, 10 subjects indicated that they thought there were categories of brand names and 19 subjects thought that there were not. A nonsignificant single-variable chi-square test showed that this distribution did not differ from uniformity. Moreover, for those who answered in the affirmative, only four named the ethnic and two named the occupational categories. Responses to the second and third questions were distributed unevenly across categories, X sup 2 (5) = 43.33, 51.36, respectively, p
Effect of pictures. Replicating the results of Experiment 1 with another set of materials, recall in in Experiment 2 did not differ significantly between names alone and names with interactive pictures, confirming H sub 6 . Moreover, performance did not improve when subjects were told that names would be tested and were encouraged to look at all parts of the slide. This indicates that the lack of effect of interactive pictures cannot be attributed to selective attention towards the drawings.
Together, the results of these two experiments show that, when illustrations were complex and realistic and were not directly related to the names that they represented, recall was poorer with separate pictures than with names alone and was not different with interactive pictures than with names alone. These findings are consistent with the proposal that imagery mediation requires that the illustrations for brand and product names should be combined and that the illustrations should be related to their names. When neither requirement was met (separate pictures), recall was poorer than with names alone, and when one requirement was met (interactive pictures), recall improved but was not better than with names alone.
Of course, the present lack of effect of interactive pictures is inconsistent with previous studies that demonstrated a positive effect with indirect relationships between illustrations and names (McKelvie et al., 1992, 1993; Wollen Lowry, 1971). Further experimental work will be required to resolve this inconsistency, but the materials differed in three ways. Firstly, the present illustrations were more complex and realistic than previous ones, which were highly schematic. Secondly, and perhaps due to the sophistication of the drawings, ease of identification of parts in the interactive pictures was rated as 4 (moderate) out of 7 in both experiments, which is probably lower than in past work. Thirdly, most of the present names were concrete and the indirect relationships between pictures and names were created by illustrating an association to the name or a part of the object denoted by the name. In previous investigations, names were less concrete or were less familiar to some subjects. It is possible that the present relationships between illustrations and names were more indirect than in previous work. If so, the extra cognitive effort devoted to examining these relationships may have been wasted because the connections between illustrations and names were not perceived. Meyers-Levy et al. (1994) observe that the extensive processing prompted by indirect relationships will not produce a positive effect unless the incongruity is resolved. Perhaps more frequent exposures or a longer exposure time than ten seconds would permit this condition to be met.
Apart from supporting the two theoretical requirements of imagery mediation, the major implication of the present picture findings is that interactive pictures do not always enhance recall of brand names. Although advice that advertisers should use interactive pictures usually contains the caveat “whenever possible” (Alesandrini, 1983; MacInnis & Price, 1987), the limiting conditions are not laid out. Our results suggest that they include realistic illustrations that are related indirectly to their names. If brand names are difficult to depict, advertisers may be advised to avoid wasting resources on expensive sophisticated illustrations.
Effects of ethnicity. In Experiment 1, where brand name familiarity was equated, recall was highest for varied stimuli, followed by ethnic and then by occupational stimuli. However, in Experiment 2, where familiarity was higher for the varied and occupational than for the ethnic brand names, it was predicted that recall would only be higher for varied than for ethnic or occupational stimuli (H sub 7 ). This hypothesis was confirmed and, although it was not possible to conduct an inferential test, the effect size for the difference between varied and ethnic material was greater in Experiment 2 than in Experiment 1 (d = 0.57, 0.45, respectively).
Although this demonstrates that familiarity is a factor in recall, the consistently poorer performance for the categorized than uncategorized names shows that the extraneous variable of classification did not enhance recall via cueing. Furthermore, the results of the questionnaire given after Experiment 2 indicate that subjects did not notice that some of the names were classified. Indeed, when they were provided with category information, they identified the ethnic and occupational categories more often than the alternatives, but did not differentiate between them as did the pre-experimental subjects who judged each brand name individually.
Assuming that subjects in Experiment 1 were also unaware of the stimulus classifications, the important finding in both experiments is that ethnicity lowered recall. Together with the generally nonsignificant effect of ethnicity in previous work (MacGregor & McKelvie, 1990), this result casts doubt on the usefulness of ethnicity as a technique to increase memorability of brand names.
On the other hand, ethnicity may have an effect with people to whom it is important. Pictures can be more persuasive if they are high rather than low on personal relevance (Miniard et al., 1991). In particular, Pedic (1990) found that advertisements that contained references to Australian nationalism were only more persuasive than non-nationalistic advertisements for subjects who felt nationalistic. Similarly, ethnic content may enhance recall for people who belong to the group portrayed. This suggests that companies should become aware of the types of names and pictures that they should and should not use when advertising their goods and services in heterogeneous markets. Prejudice might also be a factor in the effect of ethnicity on recall, with increased performance for ethnocentric people on out-groups. Of course, this would open the technique to further criticism on ethical grounds.
Limitations of the present experiments. A number of precautions were taken to minimize the effect of extraneous variables, particularly with reference to content. This was important for theory testing, which is best conducted under highly controlled laboratory conditions (Mook, 1983). However, the present research was also conducted to obtain information that would be of practical use to advertisers. This requires that the experiments should be externally valid.
Some features of the materials (lack of initial association between brand and products names, indirect association between illustrations and names, complexity of drawings, stereotypic depictions) were designed to represent aspects of real print advertising. Moreover, the incomplete memory instructions that have been standard practice with other investigators of brand name recall were used because consumers do not process advertisements in anticipation of a particular kind of memory test. However, we did inform our subjects that memory would be tested. A more realistic instruction would have been to simply inspect the materials, perhaps with a cover task involving evaluation of some kind. In addition, print advertisements do not usually display illustrations with brand and product names simply printed above them in the left/right order, respectively. Furthermore, we only presented each pair for ten seconds. Consumers may only glance at most print advertisements. However, they may also be exposed to them more than once. Finally, although researchers have emphasized the importance of understanding how brand names may be made memorable (e.g., Meyers-Levy, 1989), recall is only one part of the process of selecting a product or service. No attempt was made here to examine whether the ethnic content had any effect on product evaluation.
The major findings of the present experiments are that cued recall of brand names was depressed by separate pictures and by ethnic content, and was not enhanced by interactive pictures. The picture findings support the proposal that recall will only be improved by imagery mediation if the illustrations for brand and product names are connected and are seen to represent the corresponding brand and product names. Because many brand names are difficult to depict, these findings show that interactive pictures may have more limited practical use than previously realized. The negative effect of ethnicity adds further to the concerns of those who object to some types of ethnic brand names and stereotypic illustrations in advertisings.
Alesandrini, K. L. (1983). Strategies that influence memory for advertising communications. In R. J. Harris (Ed.), Information processing research in advertising (pp. 66-82). Hillsdale, NJ: Lawrence Erlbaum Associates.
Biron, J., & McKelvie, S. J. (1986). Effects of exposure time, separate, and interactive imagery on recall of advertisements. Canadian Journal of Behaviourial Science, 18, 159-166.
The Burlington Free Press (Burlington, VT). (May 20, 1992, p. 7a).
The Burlington Free Press (Burlington, VT). (June 24, 1993, p. 11).
Campus Canada: The University and College Magazine (April/ May, 1989), p. 2.
Canadian Advertising Foundation. (October, 1992a). Visible minorities in advertising: National consumer survey. Toronto, ON: Author.
Canadian Advertising Foundation. (1992b). Visible minorities in adverting: CEO survey results. Toronto, ON: Author.
Childers, T. L., & Houston, M. J. (1984). Conditions for a picture superiority effect on consumer memory. Journal of Consumer Research, 11, 643-654.
Christensen, L. B. (1991). Experimental methodology (5th ed.). Boston, MA: Allyn and Bacon.
Cohen, J. (1977). Statistical power analysis for the behavioral sciences (rev. ed.). New York: Academic Press.
Deese, J. (1959). Influence of inter-item associative strength upon immediate recall. Psychological Reports, 5, 305-312.
DeMont, J. (1991, August 19). Putting rage to work. Maclean’s, pp. 14-15.
Edell, J. A., & Staelin, R. (1983). The information pressing of pictures in print advertisements. Journal of Consumer Research, 10, 45-61.
Epstein, W., Rock, I., & Zuckerman, C. B. (1960). Meaning and familiarity in associative learning. Psychological Monographs, 74 (4, Whole No. 491).
Esser, J. K., Die, A. H., Seholm, K. J., & Pebley, S. R. (1986). Memory for brand-product information in pictorial advertising. Journal of Psychology, 110, 363-373.
Feschuk, S. (1993, January 7). Survey shows ads need racial diversity. The Globe and Mail (Toronto, ON), p. B7.
Good Housekeeping. (1992, October).
Houston, M. J., Childers, T. L., & Heckler, S. E. (1997). Picture-word consistency and the elaborative processing of advertisements. Journal of Marketing Research, XXIV, 359-369.
Keller, K. L. (1987). Memory factors in advertising: The effect of advertising retrieval cues on brand evaluations. Journal of Consumer Research, 14, 316-333.
Kerst, S. M. (1976). Interactive visual imagery and memory search for words and pictures. Memory & Cognition, 4, 573-580.
Kroll, N. E., Schepeler, E. M., & Angin, K. T. (1986). Bizarre imagery: The misremembered mnemonic. Journal of Experimental Psychology: Learning, Memory, and Cognition, 12, 42-54.
Lutz, K. A., & Lutz, R.J. (1977). Effects of interactive imagery on learning: Applications to advertising. Journal of Applied Psychology, 62, 493-498.
MacGregor, R. M. (1992, November). The coloring of the race issue. Paper presented at the New England Popular Culture Association Conference, Boston, MA.
MacGregor, R. M., & McKelvie, S. J. (1990). Effects of ethnic imagery on recall of brand names. Canadian Journal of Administrative Sciences, 7 (3), 1-9.
MacInnis, D. H., & Price, L. L. (1987). The role of imagery in information processing: Review and extensions. Journal of Consumer Research, 13, 473-491.
Marketing. (1991, October 28, p. 8).
Marketing. (1991, December 16, p. 3).
McKelvie, S. J., Cooper, D., Monfette, P. (1992). To what extent do interactive pictures promote recall? Perceptual and Motor Skills, 75, 627-638.
McKelvie, S. J., Daprat, C., Monfette, P., & Cooper, D. (1993). Effects of pictorial interaction on recall of brand names and lower imagery nouns. Australian Journal of Psychology, 45, 9-15.
McKelvie, S. J., Sano, E., & Stout, D. (1994). Effect of colored separate and interactive pictures on cued recall. Journal of General Psychology, 12, 209-222.
Meyers, J. (1990, April 2). JWT anti-Japan ad is a bomb. Advertising Age, p. 4.
Meyers-Levy, J. (1989). The influence of a brand name’s association set size and word frequency on brand memory. Journal of Consumer Research, 16, 197-207.
Meyers-Levy, J., Louie, T., & Curren, M. T. (1994). How does congruity of brand names affect evaluations of brand names affect evaluations of brand name extensions? Journal of Allied Psychology, 79, 46-53.
Miniard, P. W., Bhatla, S., Lord, K. R., Dickson, P. R., & Unnava, H. R. (1990). Picture-based persuasion processes and the moderating role of involvement. Journal of Consumer Research, 18, 92-107.
Ministry of Supply and Services. (1988). Visible minorities in government communications: A matter of balance. Ottawa, ON: Author.
Mitchell, A. A., & Olson, J. C. (1981). Are product attribute beliefs the only mediator of advertising effects on brand attitudes? Journal of Marketing Research, 18, 318-332.
Mook, D. G. (1983). In defence of external invalidity. American Psychologist, 38, 379-387.
Morelli, G. (1970). Pictures and competing pictures as mediators in paired-associate learning. Perceptual and Motor Skills, 30, 729-730.
Multiculturalism and Citizenship Canada. (August, 1991). Multiculturalism and Canadians: Attitude study 1991: National survey report. Hull, PQ: Author.
Native Network News. (September, 1989, p. 2).
Paivio, A. (1971). Imagery and verbal processes. New York: Holt, Rinehart & Winston.
Paivio, A., & Begg, I. (1981). Psychology of language. Englewood Cliffs, NJ: Prentice-Hall.
Paivio, A., Yarmey, A. D. (1966). Pictures versus words as stimuli in paired-associate learning. Psychonomic Science, 5, 235-236.
Pedic, E (1990). Persuasiveness of nationalistic advertisements. Journal of Applied Social Psychology, 10, 724-738.
People. (1993, January 11).
The Record. (Sherbrooke, QC). (May 15, 1986).
Saegert, J., & Young, R. K. (1983). Levels of processing and memory for advertisements. In L. Percy & A. G. Woodside (Eds.), Advertising and consumer psychology (pp. 440-463). Lexington, MA: Lexington Books.
Samuels, J. (1992). Visible minorities in Canada: A projection. Report commissioned by the Toronto Race Relations Advisory Council on Advertising, Toronto, ON.
Schwartz, J., & Exeter, T. (1989). All our children. American Demographics, May, 34-37.
Scot Foto (Laval, QC). Sales promotional flyer (1994).
Senter, R. J., & Hoffman, R. R. (1976). Bizarreness as a nonessential variable in mnemonic imagery: A confirmation. Bulletin of the Psychonomic Society, 7, 163-164.
Sharp, L., & Markham, R. (1992). The effect of the distinctiveness of bizarre imagery on immediate and delayed recall. Journal of Mental Imagery, 16, 211-220.
Snodgrass, J. G., & Vanderwart, M. A. (1980). A standardized set of 260 pictures: Norms for name agreement, image agreement, familiarity, and visual complexity. Journal of Experimental Psychology: Human Learning and Memory, 6, 174-215.
Strauss, M. (1992, July 2). Hershey responds to race complaint. The Globe and Mail (Toronto, ON), p. B4.
Teinowitz, I. (1992, April 6). Crazy Horse brew incenses Sioux. Adverting Age, p. 38.
Wollen, K. A., & Lowry, D. H. (1971). Effects of imagery on paired-associate learning. Journal of Verbal Learning and Verbal Behavior, 10, 276-284.
Wollen, K. A., Weber, A., & Lowry, D.H. (1972). Bizarreness versus interaction of mental imageas determinants of learning. Cognitive Psychology, 3, 518-523.
Woman’s Day. (1992, September 1).
The authors share equal responsibility for the paper, and wish to thank the Bishop’s University Research Committee for financial support, Stephan Gaines for composing the drawings, Chris Turner for making slides, and Catherine Cloutier and Barbara McLellan for asssistance in collating and analyzing the data.
Address correspondence to either Stuart McKelvie, Department of Psychology or Robert MacGregor, Division of Business Administration, Bishop’s University, Lennoxville, QC, Canada, J1M 1Z7.
Copyright Administrative Sciences Association of Canada Mar 1996
Provided by ProQuest Information and Learning Company. All rights Reserved