Special Session Summary Influence Attempts Beyond People’S Awareness

C. Miguel Brendl, INSEAD
Amitava Chattopadhyay, INSEAD
[ to cite ]:
C. Miguel Brendl and Amitava Chattopadhyay (2003) ,"Special Session Summary Influence Attempts Beyond People’S Awareness", in NA - Advances in Consumer Research Volume 30, eds. Punam Anand Keller and Dennis W. Rook, Valdosta, GA : Association for Consumer Research, Pages: 151-153.

Advances in Consumer Research Volume 30, 2003     Pages 151-153



C. Miguel Brendl, INSEAD

Amitava Chattopadhyay, INSEAD

The session consisted of three papers followed by a discussion led by Professor Robert S. Wyer, Jr. The three papers built on the session theme as follows: "Are brands containing name letters preferred?" builds on the phenomenon that people like their first initials more than other letters. We present evidence that when choice objects contain first initials, the positive affect associated with these initials can transfer to the choice object. We show that choices can be influenced by this name letter effect without decision makers being aware of this influence. This renders the influence via name letter difficult to control for a perceiver.

In "Subliminal priming and choice" we present evidence that under certain conditions subliminal priming affects choices. Here it is very clear that perceivers are not aware of the subliminal prime and hence of this source of influence on their choice. Thus, the influence of a subliminal prime is completely out of the control of a perceiver.

"Implicit and explicit evaluations: A declaration of independence" introduces a response time procedure that can measure the uncontrollable evaluative component of an overall evaluation. The measure allows us to investigate the attributes of uncontrollable evaluative responses. Specifically, we ask whether such uncontrollable evaluative responses are or are not independent of explicit attitudes; that is, attitudes that have controllable components. While research on implicit attitudes currently assumes that these two types of attitudes are independent, we find evidence that they are reasonably correlated and hence in some relation of dependency. We discuss the implications of this finding for understanding the processes underlying phenomena such as the name letter effect or subliminal priming.

We present below abstracts of each of the three papers.



C. Miguel Brendl, INSEAD

Amitava Chattopadhyay, INSEAD

Brett W. Pelham, SUNY, Buffalo

Mauricio Carvallo, SUNY Buffalo

Evan T. Pritchard, University of Winnipeg

Research on name-letter preference, on the mere ownership effect, and on self-regulation and self-enhancement all converge to suggest that most people possess positive unconscious associations about themselves. For this reason, people prefer things that are associated with the self (e.g., the letters in one’s name, the numbers in one’s birthday) to things that are not. We refer to such preferences as implicit egotism. This talk will summarize a recent program of research that examines whether making brand names similar to consumers’ names increases the likelihood that consumers will choose the brand. In essence we ask whether the affect that is derived, for instance, from a name letter, spills over to the brand.

One straightforward prediction is that people will prefer and be more likely to choose products or services whose names prominently feature the letters in their own first or last names. We tested this hypothesis in two studies. In a first study we asked American subjects to rate and rank order a set of candy bars that were widely available in the US market. Color images of the candy bars were shown on a single screen projected in front of the classroom. Subjects filled out their ranking and rating of the candy bars on an identical sheet in black and white, in front of them. The results showed that subjects’ preference rankings and evaluations of name letter matching brands were higher than those of non-name letter matching brands. In a second experimental study, subjects, run in yoked pairs, were given a choice between two teas. The teas were marked with brand names composed of the first three alphabets of the subjects’ names and a random three-letter ending. A yoked pair of subjects saw the same pair of names, but for each subject one brand name had a three-letter overlap with their name and the other did not. Subjects were allowed to taste the two teas and asked to make a choice of which tea they would like as a gift to take home. They were given a 30-gram bag of the tea brand they chose to take home with them. The results showed that subjects were significantly more likely to choose the tea branded with the name letter overlap to that branded without this overlap. Further, these subjects were unaware of the influence of the name letter on their decision. Even when the subjects were probed whether the name of the tea could have affected their decision, most subjects dismissed this possibility. The results are the same when the analyses are conducted only with the latter group of subjects.

While these studies show that implicit egotism can affect choices, they do not speak to the potential influences of implicit egotism in real life decision. A number of studies using archival, real life data show that important real life decisions are also influenced by the name letter effect. For example, a series of archival studies showed that people are disproportionately likely to live in cities or states whose names resemble their own first or last names. An additional archival study extended this finding to birthday number preferences. People were disproportionately likely to live in cities whose names began with their birthday numbers (e.g. Two Harbors, MN). A second series of studies suggested that people disproportionately choose careers whose labels resemble their names (e.g., people named Dennis or Denise are over represented among dentists, people named Cheryl or Charlotte are disproportionately likely to work in businesses or sell products whose names begin with the letter C (e.g., cookies, candles, cakes, etc.). A third series of archival studies employed exhaustive sampling procedures to show that people were more likely to marry other people whose last names began with the same letters as their own. This name-letter matching effect held within specific ethnic groups (e.g., Latinos, Korean Americans), indicating that the effect cannot be attributed to ethnic matching. Along similar lines, people contributed more money to the 2000 Presidential election campaign funds of candidates if their last names began with the same letters as the last names of these two candidates.

Taken together, the combination of experiments and archival studies, spanning a wide range of choices that are both important life decisions and more mundane consumer choices, provides substantial evidence that choices are influenced by the name letter effect. This evidence further suggests that unconscious influences of product features on choices go far beyond physical cues (e.g., color) and atmospherics. Brands should be more preferred, the better they establish an association to consumer’s selves, even if this association is established along features that consumers would regard as completely irrelevant, if they were aware of their influence. They suggest that the positive associations people have about themselves have an unconscious influence on people’s daily judgments and decisions. This perspective stands in sharp contrast to many models of rational choice and attests to the importance of understanding implicit social cognition and implicit self-esteem.



Steven J. Spencer, University of Waterloo

Erin J. Strahan, University of Waterloo

Mark P. Zanna, University of Waterloo

Ironically, many people have both hopes and fears that subliminal persuasion might work. For example, millions buy subliminal self-help tapes to lose weight or to increase their self-esteem. Such tapes are appealing because they promise a low-effort way to improve your life. Just press 'play’ on the tape recorder, go to sleep, and let your subconscious do all the work. On the other hand, to many people, subliminal persuasion is frightening. Imagine you are at the movies and, just before the feature is about to start, you are overcome with an urge to drink a Coke. You run to the concession stand, buy a Coke, and drink it. A couple of weeks later, you read in the newspaper that the owner of the theatre is being sued for flashing the subliminal message, "Drink Coke," on the screen throughout a series of ads that precedes the movie. How would you feel? Well, if you’re like me, the thought that you might have been influenced by a subliminal message would make you feel angryCand, if not afraid, at least nervous.

Although there remains a widespread belief in the general public that subliminal persuasion can and does work, over the past several years psychological research has strongly suggested that it doesn=t and can=t. For example, Elliot Aronson and Anthony Pratkanis reviewed more than 150 articles from the mass media and over 200 academic papers on subliminal processes and found no evidence that subliminal messages influence behaviour. While we do not question this 'received wisdom,’ we do think there is another way that subliminal information might affect persuasion. Put simply, we think that it might be possible to use subliminal priming techniques to improve or enhance the effectiveness of conventional, supraliminal persuasive appeals. That is, we think it might be possible to use subliminal priming (which the literature suggests is possible) to activate a concept (or to create a psychological state) that a persuasive communicator could take advantage of.

To examine this idea, we present two experiments: In the first experiment, in a 2 x 2 between subjects design, we first manipulated participants’ level of thirst and then subliminally presented them with either thirst-related words or control words. While the manipulations had no effect on participants’ self-reported, conscious ratings of thirst, there was a significant interactive effect of the two factors on how much of the drink provided in the taste test was consumed by the participants. In the thirsty/thirst prime condition, participants drank on average 203 ml of the drink compared to 162ml (not thirsty/thirst-related prime), 148 ml (not thirsty/neutral prime), and 136 ml (thirsty/neutral prime), in the other three conditions.

In a second, follow up experiment, thirsty participants were subliminally presented with either thirst-related words or control wordsCas in the first study, this as a between-participant factorCafter which they viewed advertisements for two new sports beverages, this was a within-participant factor. One advertisement was designed to convey the message that one of the two fictitious brands is the best thirst-quenching sports drink on the market. The advertisement for the other fictitious brand was designed to convey the message that it was the best electrolyte-replacing beverage on the market. Our main hypothesis was that participants who were subliminally presented with thirst-related words would prefer the brand positioned as thirst quenching. In the absence of the thirst prime, participants would be equally likely to choose either brand. Both the preference and the choice data were consistent with this prediction. Thus, experiment two replicates and extends the findings of the first experiment.

In conclusion, this research demonstrates that, under certain conditions, subliminal priming techniques can enhance persuasion. It has not escaped our attention that these findings could be exploited and, possibly, used for immoral purposes. Certainly, this possibility gives us pause. However, upon reflection, we believe that one way to control such techniques is to fully understand how subliminal priming works. For example, one important (and encouraging) finding from our research is that subliminal priming appears to only (or, perhaps, primarily) have an effect when the subliminal prime taps into people’s acute or chronic motivational states. Thus, although our findings do have the potential to be exploited, they also suggest, that as we learn more about subliminal priming, we may have less to fear than we might have thought. In any case, we believe that the wide dissemination of scientific knowledge on this topic is ultimately the best protection against exploitation.



C. Miguel Brendl, INSEAD

Arthur B. Markman, University of Texas at Austin

Claude Messner, University of Heidelberg

[Our apologies to Larry Jacoby for modifying the title from his chapter: Jacoby, Larry L., Andrew P. Yonelinas, and Janine M. Jennings (1997), "The relation between conscious and unconscious (automatic) influences: A declaration of independence," in Scientific approaches to consciousness, Jonathan d. Cohen and Jonathan W. Schooler, Eds. Mahwah, New Jersey.]

In the previous two presentations perceivers were not aware of the source of an attitude, neither for the fact that their preference for a certain tea stemmed partly from name letter liking, nor that their preference for a drink was exacerbated by a subliminal prime. Recent research in social psychology considers the possibility that perceivers might not even be aware of the attitude itself. Hence, this type of attitude is called an implicit attitude. This assumption stems from findings that correlations between the Implicit Association Test (IAT) and self-report measures of attitudes are often very low (e.g., Greenwald, McGhee, and Schwartz 1998; Karpinksi and Hilton 2001), which some investigators interpret as evidence in favor of the assumption that explicit and implicit attitudes are independent, even rooted in independent processing systems. We hypothesize that the lack of correlations between implicit and explicit evaluations is due to measurement error. We have collected data showing that when this error is reduced, correlations emerge. We believe that attitudes measured by response time methods are uncontrollable evaluative responses and are in that respect automatic. However, in our view there is currently no evidence to show that these attitudes are unconscious, nor that they are part of an independent processing system.

We introduce the Evaluative Movement Assessment (EMA), a response time measure that we have developed capable of tapping spontaneous evaluations. The respondent's task is to evaluate repeatedly a "target word" (e.g., the brand name Coke) or a series of different stimuli (e.g., different pictures representing one brand) that appear on a computer screen one by one. Next to each target word/picture the respondent's first name appears on the screen. For example, a respondent named Arthur would see in one trial:


Respondents are instructed to indicate whether a word is positive or negative by moving it on the monitor with the help of a joystick either toward their name or away from their name. In Study 1 respondents were faster to move positive words (e.g., flower) toward than away from their first name, and vice versa for negative words (e.g., Hitler). In Study 2 in one set of trials EMA forces respondents to move, for example, "Coke" as quickly as possible away from their name, while in another set of trials it forces the same respondent to move it toward their name. Although respondents consciously try to respond as fast as possible in each set of trials, respondents with a positive uncontrollable evaluation of Coke execute toward movements faster than away movements. Response time differences between toward and away movements are interpreted as attitude scores (EMA-scores) that reflect an uncontrollable component of an evaluative response.

The mentioned zero correlations of "implicit" and explicit attitudes are based on using the Implicit Association Test (Greenwald, McGhee, and Schwartz 1998), which always measures preferences of one attitude object to the other. In Study 2 we measured attitudes toward 5 attitude objects using both EMA and rating scales. When analyzing attitudes toward just two objects we also show zero correlations between EMA-scores and ratings. However, when correlating EMA-scores with ratings for all 5 attitude objects we show sizeable correlations.

We conclude that attitudes measured via response times tap uncontrollable evaluative responses. Current evidence does not support an independent-systems view. To the contrary, a dependence is observable. We suspect that these uncontrollable responses contribute to explicit attitudes. While it is conceivable that there are independent attitude systems that typically work in synchrony, we currently lack evidence in support of that hypothesis.