Belief in Imitator Claims: the Role of Source Memory

Sharmistha Law, University of Toronto
ABSTRACT - This study investigates the effectiveness of imitator ads. Results suggest that imitator ads confuse subjects, as evidenced by low recognition rates compared to ads seen for the first time. Overall, subjects report a lower belief in imitator ads relative to novel ads. Memory for the source of an imitator ad is a major determinant of its believability. When subjects confuse an imitator ad with a similar ad seen previously, they express greater belief in it. On the other hand, when subjects correctly recognize an ad as an imitator, they are strongly skeptical of it. Marketers need to consider these findings, especially the effect of source memory, when implementing imitator strategies.
[ to cite ]:
Sharmistha Law (1995) ,"Belief in Imitator Claims: the Role of Source Memory", in NA - Advances in Consumer Research Volume 22, eds. Frank R. Kardes and Mita Sujan, Provo, UT : Association for Consumer Research, Pages: 165-170.

Advances in Consumer Research Volume 22, 1995      Pages 165-170

BELIEF IN IMITATOR CLAIMS: THE ROLE OF SOURCE MEMORY

Sharmistha Law, University of Toronto

ABSTRACT -

This study investigates the effectiveness of imitator ads. Results suggest that imitator ads confuse subjects, as evidenced by low recognition rates compared to ads seen for the first time. Overall, subjects report a lower belief in imitator ads relative to novel ads. Memory for the source of an imitator ad is a major determinant of its believability. When subjects confuse an imitator ad with a similar ad seen previously, they express greater belief in it. On the other hand, when subjects correctly recognize an ad as an imitator, they are strongly skeptical of it. Marketers need to consider these findings, especially the effect of source memory, when implementing imitator strategies.

INTRODUCTION

In 1988, the Clearly Canadian Beverage Company introduced a transparent cola, portraying a pure, natural, environmentally conscious image. Its introduction was soon followed by Crystal Pepsi. In 1992, Colgate brought out a clear version of its Softsoap. A few months later, Procter and Gamble launched a clear version of Ivory (The Globe and Mail, 1994). Examples of such "me-too" products abound in today's marketplace. Imitation strategies include similar package designs, similar claims about product performance, or similar brand names. Competitor slogans and tag lines may also be mimicked. Creators of imitator ads attempt to position their product next to a more established brand in the hope that consumers will generalize attributes of the original brand (usually a leading brand) to their product (Foxman, Muehling, and Berger, 1990).

Imitation strategies can have harmful consequences for the consumer. Confused consumers may be at a physical risk when they unintentionally buy products other than the one they intended to buy. For example, when Lever Brothers introduced Sunlight dishwashing detergent, the package's similarity to Minute Maid lemon juice led several consumers (33 adults and 45 children in the state of Maryland alone) to mistakenly ingest Sunlight (Reiling, 1982). More generally, consumers who mistakenly buy me-too brands can suffer negative consequences if such products perform less well than the original brand. Therefore, from the public policy point of view, there is indeed a need for greater understanding of the factors that contribute to consumer confusion. Identification of these factors and the circumstances under which they operate will aid policy makers in reducing the detrimental effects of this marketing strategy.

Some critical questions for both marketers and researchers with respect to consumer confusion are: How do consumers react to imitator ads? Do advertisers benefit from using an imitator strategy? Or, do me-too claims hurt the product? This study addresses these questions by investigating the impact of imitator ads on consumer confusion and consumer belief regarding the claims.

CONSUMER BRAND CONFUSION

Empirical studies demonstrate that consumers become confused when there is a high degree of package similarity (e.g., Foxman, Darrel, Muehling, and Berger, 1990; Loken, Ross, and Hinkle, 1986; Ward, Loken, Ross, and Hasapopoulos, 1986), when copycat brands match the timing of the market leader's advertising (Trout and Ries, 1985 cited in Park and Hahn, 1991). For example, Foxman et al. (1990) showed subjects slides of packages of well known consumer products. Subjects were shown a slide depicting packages of the original brand alongside five, physically similar imitator brands, and were required to identify the original brand. Results indicated that 17% to 27% of the subjects wrongly identified an imitator brand as the original brand.

An explanation for why consumer brand confusions occur is suggested by the phenomenon of category priming and the associated notion of assimilation effects (e.g., Herr, 1989; Herr, Sherman, and Fazio, 1983; Meyers-Levy and Sternthal, 1993). Higgins, Rholes, and Jones (1977) found that, subjects primed with exemplars of a particular category are more likely to use that category in evaluating a subsequently presented target stimulus. When the target stimulus closely resembled the previously presented object, subjects categorized the target as being the object previously presented.

The preceding discussion suggests that prior exposure to marketing claims will affect consumers' judgments of new claims and, consumer brand confusion will occur when a new imitator claim closely resembles the original claim.

CONSUMER BELIEFS AND THE EFFECT OF SOURCE MISATTRIBUTION

While confusion is clearly an important concern for marketers and consumers, little research has been done on the impact of brand confusion on product evaluation. On the one hand, brand confusion may lead to an increased false belief for imitator claims when the imitation goes undetected. On the other hand, consumers may become unduly skeptical or suspicious when a market leader's product claim is repeated by an imitator, thereby resulting in diminished belief in the claim.

Previous research has shown that the mere repetition of statements causes them to be rated as more valid than non-repeated control statements (e.g., Hasher, Goldstein, and Toppino, 1977; Bacon, 1979; Arkes, Boehm, and Xu, 1991; Schwartz, 1982), a phenomenon called the "truth effect." It is argued that subjects' sense of having seen the item before (i.e., feelings of familiarity) leads to heightened truth ratings (e.g., Bacon, 1979; Hawkins and Hoch, 1992). For instance, Hawkins and Hoch found that (a) repetition of marketing claims increased belief in the claims, and (b) statements perceived to be repeated, whether or not they were actually repeated, were given higher validity ratings than statements perceived to be non-repeated.

Law and Hawkins (1994), building on the truth effect research, investigated the role which memory for the source of the claim has on the subjects' belief regarding the claim. They demonstrated that the subjects' belief in a statement is influenced not only by their familiarity with it, but also by whether or not the second presentation is accompanied by a correct recall of the source of the prior experience. Amongst repeated items, those items which were misattributed to a source outside the experiment received the highest truth ratings.

Given the findings of Law and Hawkins (1994) that, source misattribution leads to increased belief in claims, it seems reasonable to expect this effect with respect to imitator claims. That is, the previous exposure to an original claim should result in feelings of familiarity with an imitator claim, and in the absence of correct recognition (that is, if subjects fail to recognize that their initial exposure to the ad had been associated with a different brand), subjects will judge the claim to be more believable than if they are able to accurately identify it as an imitator.

The purpose of this study is threefold: (1) to examine the level of confusion (that is, misrecognition) of imitator ads; (2) to assess the believability of imitator claims, and (3) to investigate the role of source misattribution on subjects' belief about imitator claims.

The preceding discussion leads to the following hypotheses:

H1: Incidence of Confusion. Imitator ads will lead to consumer confusion as reflected in a lower accuracy of recognition relative to other ads seen for the first time.

H2: Belief in Imitator ads. Subjects' beliefs in imitator ads will be different relative to other ads seen for the first time.

H3a: Role of Source Memory. Misattribution of the source of an imitator claim (that is, not realizing that this is an imitator claim) will increase subjects' belief in it.

H3b: Conversely, correctly recognizing an imitation claim will decrease subjects' belief in it.

Since subjects' level of fatigue and/or boredom may influence how they react to and process commercial messages, the relationship between arousal level and truth ratings was also examined in this study. To test our hypotheses, imitator ads were operationalized as ads which contained a plausible claim that had been previously presented with a different brand name.

METHOD

Subjects

Subjects were 35 undergraduate students who were enrolled in an introductory management course and took part for extra marks. They were run in small groups ranging in size from 2 to 7 with each group being randomly assigned to one of four separate booklet conditions (to be discussed shortly). Three subjects were excluded from the study as they failed to comprehend or act upon the instructions. One subject, who was able to guess the hypotheses of the study, was also excluded from the data analysis. Therefore, the hypotheses were tested with a total of 31 subjects.

Design

The study involved a 2 x 2 completely within subjects design. Session (Study vs. Test) and Claim Type (Original vs. Imitator claims) were the two repeated factors.

Stimuli Construction

The experimental stimuli consisted of fictitious marketing claims about brands and products. Two stages of pretesting were conducted to ensure that the brand names were familiar and the claims were plausible to Canadian consumers.

Pretest 1. This pretest was conducted in a sample of 15 subjects to identify familiar brand names. Brand names with an average familiarity of 2.5 or higher on a 5-point scale (1=Very Unfamiliar; 5=Very Familiar) were selected from a list of 112 names spanning 50 different product categories (e.g., shampoo, airlines, car rental companies). The mean familiarity rating for the chosen brands was 3.9.

Pretest 2. The selected brand names from Pretest 1 were then paired with fictitious statements to form plausible product related claims. Some of these statements were from Hawkins and Hoch (1992). The plausibility of these claims were pretested using another 15 subjects. However, as one subject did not complete the questionnaire, the data analysis for the pretest was conducted using 14 subjects. Those statements with a mean truth rating of approximately 3.0 (on a 5-point scale with 1=Definitely False; 5=Definitely True), standard deviation of less than 1, and with a unimodal distributions were included in the experimental stimuli list. The mean plausibility rating and standard deviation of the statements included in the main experiment was 3.07 and 0.82, respectively.

The final test material consisted of 64 fictitious marketing claims. For each claim, two brand names (an original and an imitator) could be used interchangeably. Using these statements, 4 counterbalanced booklet types (A, B, C, D) were created. Each booklet had 32 Study items and 32 Test items. At Study, each subject saw a set of 16 new items (called original) which would be altered slightly and re-presented as imitator items; and a different set of 16 new items (called new1). At Test, each subject saw a set of 16 imitator items (which were ads corresponding to the originals seen at Study) and a different set of 16 items (called new2). The imitator ads were exactly the same claim as the original but with a different brand name.

To ensure the results of the experiment were free of any systematic bias due to the nature of the claims or the particular brand name, each statement occurred an equal number of times as an original, new1, new2, and an imitator. In addition, each brand name was paired an equal number of times with an original, new1, new2, and an imitator product claim. Subsequent analysis of the results showed no significant effect of booklet [F(3,108)=0.78, p>.5] and no significant interaction between booklet and item-type [F(9,108)=1.16, p>.3], confirming that the process of stimulus randomization had had its desired effect.

The Study list consisted of 70 product-related statements which were composed of the following: 16 original items; 16 new1 items; 10 buffer items used to minimize recency and primacy effects; and 28 filler items included to increase the length of list. The Test booklet contained 50 product-related claims: 5 buffer items included to allow subjects to get acquainted with the task; 16 imitator items (corresponding to the 16 original items); 16 new2 items; and 13 filler items. Table 1 describes the template of the Study and Test booklets.

Procedure

The Study consisted of three parts: Study, Filler, and Test. Subjects were told that the purpose was to assess the effectiveness of product claims. During the Study session, subjects were shown 70 claims for 10 seconds each (a time period determined on the basis of pretests) and asked to rate how true of false they thought each statement was (see Dependent Measures section for the scale used). Following this, there was a 20 minute interval during which subjects filled out a questionnaire for another study and completed a mental folding task. This part of the study served as a Filler task to minimize the effects of short term memory. At Test, subjects were told that they would see another set of marketing statements and would be required to respond to two questions. For each statement displayed, subjects were given 10 seconds in which to fill out the two dependent measures (to be discussed shortly). They were also instructed not to leave any questions blank.

Dependent Measures

In the Study session, subjects rated the truth of each item. For each item seen in the Test session, subjects first rated its truth and then indicated the source of their prior exposure to it.

TABLE 1

STUDY AND TEST BOOKLET TEMPLATES

Belief in a Claim. Subjects' truth rating for each presented item was measured on a continuous 16 cm. Likert scaled anchored at "Definitely False" on the left, to "Definitely True" on the right. Belief in a claim was measured as the distance (in cms.) from the left hand side of the scale.

Source Attribution. In the Test session, after rating the truth of an item, subjects were asked to indicate where they had encountered the statement: Before: from TV, magazine, friend, etc. (that is, outside the experiment); First Session (that is, during Study); Before and First Session; or Never.

Incidence of Confusion. Confusion is considered to have occurred if a subject misidentified the source of a statement. In other words, if a subject indicated that an imitator statement had been seen during the Study phase or outside of the experiment or both, then that response would constitute an instance of brand confusion.

Other Measures. Mackay, Cox, Burrows, and Lazzerini's (1978) Arousal scale was adapted to measure subjects' self-reported activation once before the Study session and again after the Test session. Subjects were given a list consisting of the following adjectives: active, drowsy, vigorous, sleepy, aroused, alert, passive, lively, idle, stimulated, somnolent, tired, activated, sluggish, and energetic. They were instructed to respond to the scale as follows: 'definitely feel' (++), 'feel slightly' (+), 'do not understand or cannot decide' (?), and 'definitely do not feel' (-). The score for this scale runs from 0 to 15 with smaller numbers indicating greater fatigue.

RESULTS AND DISCUSSION

Incidence of Confusion

H1 predicted that subjects would commit more errors in accurately recognizing imitator ads compared to new ads. To test this hypothesis, subjects' ability correctly classify imitator claims and new claims as never having seen before, was compared. Confirming H1, subjects' recognition accuracy of imitator items was found to be significantly lower than that of new2 items (mean correct recognition: 51% vs. 71%, p<.001). Indeed, accurate recognition rate was almost at chance (that is, 50%) for the imitator ads, suggesting subjects' confusion with imitator claims.

Belief in Imitator claims

Consistent with H2, subjects rated the validity of imitator ads significantly lower relative to new claims (see Figure 1). For each subject, we calculated a mean response to original, new1, new2, and imitator items. A within-subject ANOVA of these means revealed a significant effect of item type [F(3,90)=15.40, p<.001] such that original and new1 items shown at Study and new2 items shown at Test were evaluated as being more true than imitator claims displayed at Test (Moriginal=9.28; Mnew1=9.54; Mnew2=9.1; Mimitator=7.91). In addition, thanks to the extensive pretesting, no significant difference was found between the truth ratings of original, new1, and new2 claims (p>.05).

These results suggest an overall detrimental effect of imitator ads on consumer belief. It appears that subjects are generally more skeptical of imitator claims compared to novel claims seen for the first time.

Role of Source Misattribution

The principal comparison of interest in this experiment was the role of source misattribution on the truth effect. H3a predicted that copycat marketing claims thought to have been seen before (that is, misattributed) will be believed to be truer than would claims (correctly) identified as being new. A within subject ANOVA post-hoc blocked by source attribution, confirmed this hypothesis. There was a significant main effect of source attribution with the imitators thought to have been seen outside the experiment being rated highest in truth value [F(2,53)=28.83, p<.001].

Hypothesis H3b predicted that a claim correctly recognized as being an imitator and hence attributed as being a new claim, would be less believable compared to a claim thought to be presented earlier. This hypothesis was confirmed. The mean truth rating means for the different source attributions along with a count of responses in each category were, in the order Never, First Session, and Before: 6.89 (n=31, SE=0.30), 8.36 (n=29, SE=0.39), and 10.07 (n=26, SE=0.44). See Figure 2. Post hoc comparisons revealed that these means were all significantly different at p<.01.

FIGURE 1

TRUTH RATINGS FOR THE FOUR ITEM TYPES

It is noteworthy that, when subjects are confused about the source of an imitator claim (that is, they fail to detect it is an imitation and attribute it to an outside source), they rate the statement to be truer. However, when they correctly recognize that the ad is an imitator, they rate it significantly lower. This finding qualifies the results reported in support of H2 as it demonstrates that while consumers are generally skeptical about imitator ads, imitator ads which go undetected actually benefit from the sense of familiarity they induce in the consumer, and hence are rated more true compared to novel claims.

Measure of Fatigue

Truth ratings of new statements were found to be generally lower at Test than at Study (though not significantly so). It may be argued that fatigue or boredom, might be accounting for the differences in truth ratings across the experimental sessions. Two results rule out this interpretation. First, comparison of self-reported levels of arousal measured by the Arousal scale showed no significant difference across Study and Test [t(30)=1.00, SE=0.77, p>.3]. Second, the standard errors for the four item types were, in the order new1, new2, imitators, originals: 0.28, 0.21, 0.27, and 0.25, indicating low and equal variability around the performance means. These findings allow us to reject the possibility that ratings at Study were affected by subject fatigue or boredom.

CONCLUSION

Imitation advertising is a common occurrence in today's advertising campaigns. Results presented here replicate previous research by showing that imitator ads do indeed result in consumer confusion. Furthermore, this research shows that consumers' belief in imitator ads is influenced not only by their feelings of familiarity with the message but also by whether or not the feeling is accompanied by recall of the source of the prior experience.

Particularly noteworthy is the finding that, when subjects misattribute imitator ads to a source outside the experiment, they judge them to be most believable. In contrast, when subjects are able to accurately identify an imitator ad, they tend to rate it significantly lower in validity relative to completely new ads. In sum, these results suggest that the use of "me-too" advertising is a double-edged sword. More specifically, imitation claims presented in an ad does not appear to always benefit from similarities shared with the original claims. It may be that an awareness that a presented claim mimics a previously shown claim functions as a "discounting cue" (Hovland, Lumsdaine, and Sheffied, 1949) which attenuates its believability. As the use of me-too claims increases, the users of these strategies need to be aware of the issues raised in this paper.

How imitator ads impact the image of the original brand has not been addressed in this study. It will be interesting to investigate if imitator advertising affects the believability of the claims made by the original brand. Also, it will be of interest to examine the effectiveness of imitator ads which, unlike the ones used in this study, claim the same benefit as an original brand but do not use identical wording.

Finally, in this study the imitator ads were seen soon after the originals whereas in the real world, days or even weeks may transpire between two such exposures. Viewing the imitator ads soon after exposure to the originals may have resulted in enhanced "discounting" as the originals may have been still fresh in the subjects' memory, leading the subjects to judge the imitators more severely than they normally would. Future studies can partially mitigate this concern by increasing the delay between Study and Test sessions or by collecting think-aloud protocols which would examine subjects' cognitive processing during exposure to the ads.

FIGURE 2

EFFECT OF SOURCE ATTRIBUTION ON THE TRUTH RATING OF IMITATOR CLAIMS

REFERENCES

Arkes, Hal R., Lawrence Boehm, and Gang Xu (1991), "Determinants of Judged Validity," Journal of Experimental Social Psychology, 27, 576-605.

Bacon, Frederick T. (1979), "Credibility of Repeated Statements: Memory for Trivia," Journal of Experimental Psychology: Human Learning and Memory, 5 (3), 241-252.

Burke, Raymond R., Wayne S. DeSarbo, Richard L. Oliver, and Thomas S. Robertson (1988), "Deception by Implication: An Experimental Investigation," Journal of Consumer Research, 14 (March), 483-493.

Foxman, Ellen R., Darrel D. Muehling, and Phil W. Berger (1990), "An Investigation of Factors Contributing to Consumer Brand Confusion," Journal of Consumer Affairs, 24 (1), 170-189.

Hasher, Lynn, David Goldstein, and Thomas Toppino (1977), "Frequency and the Conference of Referential Validity," Journal of Verbal Learning and Verbal Behavior, 16 (1), 107-112.

Hawkins, Scott A. and Stephen J. Hoch (1992), "Low-Involvement Learning: Memory without Evaluation," Journal of Consumer Research, 19 (September), 212-225.

Herr, Paul M. (1989), "Priming Price: Prior Knowledge and context Effects," Journal of Consumer Research, 16 (June), 67-75.

Herr, Paul M., Steven J. Sherman, and Russell H. Fazio (1983), "On the Consequences of Priming: Assimilation and Contrast Effects," Journal of Experimental Social Psychology, 19 (July), 323-340.

Higgins, Tory E., William S. Rholes, and Carl R. Jones (1977), "Category Accessibility and Impression Formation," Journal of Experimental Social Psychology, 13, 141-154.

Hovland, Carl I., Arthur A. Lumsdaine, and Fred D. Sheffield (1949), Experiments in Mass Communication. Princeton: Princeton University Press.

Law, Sharmistha and Scott A. Hawkins (1994), "Advertising Repetition and Consumer Beliefs: The Role of Source Memory," Measuring Advertising Effectiveness. NY: Lawrence Erlbaum (forthcoming).

Loken, Barbara, Ivan Ross, and Ronald L. Hinkle (1986), "Consumer confusion of origin and brand similarity perceptions," Journal of Public Policy and Marketing, 5, 195-211.

Mackay, Colin, Tom Cox, Grenville Burrows, and Tony Lazzerini (1978), "An Inventory for the Measurement of Self-reported Stress and Arousal," British Journal of Social Clinical Psychology, 17, 282-284.

Meyers-Levy, Joan and Brian Sternthal (1993), "A Two-factor Explanation of Assimilation and Contrast Effects," Journal of Marketing and Research, 30 (August), 359-368.

Reiling, Lynn (1982), "Consumer Misuse Mars Sampling for Sunlight Dishwashing Liquid," Marketing News, 3 (September), 1.

Schwartz, Marian (1982), "Repetition and Rated Truth Value of Statements," American Journal of Psychology, 95, 393-407.

The Globe and Mail (1994), "Why Outlook is Murkey for Clear Products," March 16.

Ward, James, Barbara Loken, Ivan Ross, and Tedi Hasapopoulos (1986), "The Influence of Physical Similarity Generalization of Affect and Attribute Perceptions from National Brands to Private Label Brands," in AMA Educator's Proceedings, Series Number 52, Terence A. Shimp et al. (eds.), Chicago, IL: American Marketing Association, 51-56.

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