DOI: 10.2501/SOO21849906060065 March 2006 49 Reconsidering Recall and Emotion in Advertising ABHILASHA MEHTA Gallup & Robinson, Inc. [email protected] gallup-robinson.com SCOTT C. PURVIS Gallup & Robinson, Inc. [email protected] Recall, one of the key metrics in advertising testing has been criticized over the years as favoring rational advertising over emotional advertising. An analysis and reconsideration of the available evidence show that emotional advertising is not penalized by recall, and that emotiona l content in well-executed commercials can actually boost recall. Strong empirical ev idence shows that recall, when used in combination with other measures, is a valid measure of advertising effectiveness and, as the analysis here illustrates, does not miss the emotion in advertising that builds brands. INTRODUCTION Recall is one of the several major measures used in advertising effectiveness testing today, in addi- tion to others such as persuasion and advertising liking. However, despite a strong base of empiri- cal validation, recall has been among the most criticized of the measures. And while many of these criticisms have long since been resolved, doubts about the measure linger from the days when recall was used by many as the solitary indicator of advertising effectiveness. Among the more important of the historical criticisms of recall was that it favors more “ratio- nal” commercials over more “emotional” ones. During the late 1970s and early 1980s, several researchers suggested and re ported that the recall of rational commercials wa s, on average, higher than the recall of emotional ones. This viewpoint subsided in later years as other research and the reanalysis of the early st udies showed no inherent disadvantage. Additionally, several important val- idation studies in the past tow decades have de- livered strong independent empirical evidence of the role of recall in identifying commercials that produced higher in-market business results. Re- cently though, Unilever along with one of its research partners, Ameritest, has resurrected the issue and concluded “recall misses the emotion in Advertising that builds brands,” using new data to bring into question once again the value of recall when measuring emotion based advertising (Kastenholz and Young, 2003). This article recaps the state of knowledge on the important subject of recall and emotion in advertising and helps show more clearly the value of recall in current advertising research. Although the days of recall as the sole measure of adver- tising effectiveness have long since passed, the analyses here show it is an important evaluative tool for understanding the effectiveness of both types of advertising, emotional or rational. BACKGROUND Advertisers have long believed that advertising must arouse some emotion to be effective. This affective response is important for two main rea- sons. First, the key to bran RECONSIDERING RECALL AND EMOTION IN ADVERTISING Advertisers have long believed that advertising must arouse some emotion to be effective. to be effective, there is little agreement among advertising researchers about how exactly emotion works to influence the overall impact of advertising, or even how emotional response in advertising can be measured or evaluated. As debate about how to measure advertising effectiveness continues, the issue of recall, one of the leading measures of advertising intrusive- ness, and its connection with emotion is sometimes at the center of the debate. Or more precisely, some critics of recall even question whether there is an interaction between the two at all. What is the rela- tionship between recall and emotion in advertising? It is clear that the answer to this question is important in better under- standing how best to test advertising for its effectiveness. SALES VALIDITY OF RECALL The issue of the validity of recall needs to be briefly summarized before recall's re- lationship with emotion is discussed. Any true measure of advertising effectiveness must show validity in predicting future in-market performance. All major copy testers have their own empirical support demonstrating the validity of their mea- sures. While they often place different emphasis on their measures (particularly among recall, persuasion, and advertising liking), that recall has value in evaluating advertising effectiveness is nearly univer- sally accepted, with a variety of supplier and independent studies demonstrating its sales validity (e.g., see Dubow, 1994; also IRI's "How Advertising Works" study by Lubetkin, 1991 and Lodish et al., 1995, although some authors have minimized the findings of this study after removing some of the data points/so-called "outli- ers," by no means does the research show "... no evidence of a relationship be- tween related recall scores and sales ef- fects ..." as Kastenholz and Young, 2003 concluded). Additionally, the Advertising Research Foundation Copy Validity Re- search Project (ARF CRVP; Haley and Baldinger, 1991) found recall to be a valid measure of advertising effectiveness, sec- ond only to advertising liking. CONCEPTUAL FRAMEWORK: EMOTION, ATTENTIONING, AND MEMORY To understand how emotion works in ad- vertising or how it interacts with the re- call measure in advertising testing, a basic knowledge of the memory process is use- ful. Memory is a critical part of consumer behavior and of how advertising influ- ences consumer behavior. Consumers usu- ally do not make brand purchase choices at the time of advertising exposure; rather, it is the memory of the advertising mes- sages that influence consumers. Recall's importance stems from the fact that recall measures some aspect of this memory of the advertising. New advances in our understanding of how the brain functions have helped clar- ify how consumers respond to the deluge of media stimuli around them, and how memory is built. The process of Attention- ing is said to govern what stimuli should be given attention, with memory traces being formed or strengthened based on the lengtli and deptii of attention given to a particular stimulus. The longer and deeper the attention, the stronger the memory traces. As a result, when conscious learn- ing is the goal, focused attention is given to the material at issue, and the attention is kept on it as long and as much as RECONSIDERING RECALL AND EMOTION IN ADVERTISING further reinforced by repeated exposure cess, are related, as well as how emo- tally derived from the attempts to de- to the stimulus even if the stimulus does tional advertising is not penalized by recall. scribe television commercials in words" not receive conscious attention—^because In fact, under some situations or when (p, 51), of repeated attentioning—and can be ex- influenced by moderating variables, highly This conclusion, however, does not rec- pected to create some measurable mem- emotional advertising actually enhances oncile with other findings. Our research ory traces of the stimulus even without recall, has shown that there is a moderately high, conscious learning. Recall should be able positive correlation between recall (simi- to tap into this memory as well (this is RECALL AND ATTENTION lar to ASI's and described by respondent memory, regardless as how it was ere- Does recall measure attention? It is gen- in words) and attention obtained from ated), and the emotional response in the erally accepted that the two have a recognition of the advertising via reexpo- initial attentioning process should thus moderately strong, positive correlation. sure to the actual advertising, and not via influence recall, Kastenholz and Young (2003), however, a description of the advertising in words Recall measurement requires verbal report a very low correlation between ASI (see Table 2), The results do show that proof of advertising exposure. The tradi- recall and Ameritest attention, which is Ameritest and MB attention are different tional criticism against recall vis-a-vis based on asking what commercials are from ASI and Gallup & Robinson (G&R) emotion that "feeling" advertisements, ad- found interesting, and a low negative cor- attention, but the reason offered (ASI's vertisements with high emotional content relation between ASI recall and Millward verbal cue) does not seem to hold up in that are expected to evoke emotions, will Brown (MB) attention, which is based on light of G&R nonverbal-based results. The be penalized by recall compared to "think- the claimed active enjoyment of commer- better question may be whether "interest- ing" advertisements, is based on this re- cial, but a stronger positive correlation ing" as in Ameritest attention and "active quirement of verbal proof. It is argued between ASI recall and ASI attention, enjoyment" in MB attention are really that because the emotive content is impor- which is based on recognition of the ad- about "breakthrough"? The concepts of tant in feeling advertising, respondents vertising via a verbal description (see "interest" and "active enjoyment" seem would have difficulty verbalizing their Table 1), These results are then used to closer to a positive reaction after attention memory of these types of commercials, argue that ASI recall and MB/Ameritest or breakthrough: something could be no- This view was reinforced by the earlier attention cannot be both measuring "break- ticed and remembered even if it is not brain theories that believed that the two through" power, and the strong positive necessarily found to be "interesting" or hemispheres of the brain functioned indi- relationship between ASI recall and ASI "enjoyable," Stapel (1994) reports that vidually, and that the left-brain functions attention is explained as ",,, probably included verbal and cognitive issues while because both ASI measures are fundamen- the right-brain functions included nonver- bal image and picture memory functions and storage, TABLE 2 Brain theorists today, though, do not agree TABLE 1 Correlations between with the two hemisphere/left-right brain Correlations between "Breakthrough" Measure division. In fact, there is only one memory "BreakthrOUgh" MeaSUreS _ ^ o ii of the advertising that includes all ele- ^^ Q'^ff^.^^^ Pretesting ments: the visuals, music, words, experi- ences, etc. Further, the memory trace is SyStemS and ReCall ^ Recall distributed throughout the brain, raisins ,_, _ „ '^^' Recall RECONSIDERING RECALL AND EMOTION IN ADVERTISING The concepts of "interest" and "active enjoyment" seem closer to a positive reaction after attention TABLE 3 Correlations between Liking and Recall or breaktiirough: something couid be noticed and source .-,... . -1 * J ,1. •- ipsos-ASI, as reported by remembered even if it is not necessarily found to be Kaste.hoi, 3,d v^ung (2003) 39 * "interesting" or "enjoyable." AS': wa'ter and pubitsky (1994) .34* G&R (2004) .46** "Significant at 99"/<, CL 'Significance level not reported among respondents who neither liked an runs counter to the idea that a merely advertisement nor found it interesting, a entertaining commercial can be the most substantial proportion recalled the adver- sales effective commercial" (p, 52), Greene's tisement (30 percent) or recognized the research suggested that liking may have advertisement (59 percent). Among those little to do with the traditional concept of who found the advertisement interesting, entertainment and that viewers seem to substantially more recalled the advertise- respond to the question about liking more ment (66 percent) or recognized it (87 in terms of its communications, RECONSIDERING RECALL AND EMOTION IN ADVERTISING ads test—representing hundreds of differ- TABLE 5 ent clients and agencies,,, confirms that p^^g,, ^gveis for Different Types of Commerclals recall scores are related to both interest in the message and involvement with the Highly Highly creative execution," Rational Emotionai Correlation analysis from our own find- AV Commercials Commercials ings show recall to be significantly and Synchronization (%) (%) positively correlated to a number of pos- ^ -^ '^ Above average 12 16 itive diagnostics and negatively/not re- ^^^^^ ^ ^^e) (319) (117) lated to unfavorable ones (see Table 4). Average 10 10 RECALL AND EiVIOTiONAL ADVERTISiNG (.?^.^^r,,^?,?? ,*.^?,?? ':^331 Researchers have directly studied the re- Rpinw avpra^p 8 R lationship between recall and different (Base = 527) f99) (428) types of advertising. As mentioned ear- Her, researchers (Krugman, 1977) hypoth- Total 11 10 esized that because recall was a verbal/ left-brain activity and television advertising was largely a right-brained function, re- call for television advertising would be