BRAIN WAVES, PICTURE SORTS ® AND BRANDING MOMENTS
This paper describes a method to identify potential branding moments in television commercials. It involves the convergence between two fundamentally different nonverbal moment-by-moment measurement techniques. The first is a picture-sorting technique. The second is brain-wave measurement. At the intersection of the two can be found special moments where the personalized meaning of the esthetic experience of an advertisement appears to flow into the positioning concept of a brand.
The mind of the consumer can be thought of as being continuously engaged in the process of defining the self and orienting it with respect to the outside world. A brand’s image is constructed in relationship to the consumer’s concept of self. A brand’s positioning is determined with reference to the marketing universe of competitors. But, it is the dramatic tension between the inner-directed process of brand image building and the outer-directed process of positioning a product in the marketplace that energizes the yin-yang of advertising creative as it is received by the mind of the consumer.
While many advertising professionals appear to use the terms “brand image” (Ogilvy, 1983) and “brand positioning” (Reis and Trout, 1981) interchangeably, or think of one as a subset of the other, our experimental data and current models of memory and information theory instead support the notion that they are, in fact, two quite different dimensions of advertising communication—dimensions that can be measured separately.
Advertising researchers who use moment-by-moment approaches to studying television commercials quickly learn that the “80/20 rule” of marketing applies to advertising executions as well as it does to other marketing phenomenon. A small percentage of “peak” moments in a TV commercial do much of the “work” in driving traditional advertising performance measures. (To use day-after recall as an example, see Young and Robinson, 1989.) The results reported below lead us to hypothesize the following: that within effective, brand-building television commercials there are special branding moments that occur at the boundaries between the inner-directed and outer-directed information content of an ad. Branding moments in effective television commercials forge a link between brand image and product positioning. As such, these moments may be critical for understanding the long term contribution a television commercial will make to brand equity.
This paper describes a method to identify potential branding moments in television commercials. It involves the convergence between two fundamentally different non-verbal moment-by-moment measurement techniques. The first is a picture-sorting technique. The second is brain wave measurement. At the intersection of the two can be found special moments where the personalized meaning of the esthetic experience of an ad appears to flow into the positioning concept of a brand.
The use of electroencephalogram (EEG) measurements to describe the electrical activity of the brain while a viewer watches television was first described by Krugman (1971). Subsequently, a number of researchers followed up on this work, including Olson and Ray (1983), Rothschild (1988), and Alwitt (1989). A particular focus of attention in the past has been on hemispheric variations in how the right brain processes information differently than the left brain. This is taken from Diamond and Beaumont, (1973), which was reviewed by Krugman (1980). Differences in brain activity as a function of the type of wave measured, for example, alpha waves versus beta waves, have been commented on by Olson (1989). Over a decade ago Rothschild et. al., (1989) suggested the need for research with more precise brain-wave measurements correlating to individual scenes or claims in the commercial.
The Ameritest Picture Sorts ™ method for measuring consumer recall of individual images in a commercial was described by Young and Robinson (1987). Later they showed the relationship between the Flow of Attention™ generated by this method and the recall (1989) and persuasiveness measures (1992) produced by Research Systems Corporation, a leading copytesting system. Young and Robinson (1991) also used their method to show that the semantic and esthetic information content of advertising, and how it is processed by consumers, varies as a function of the life stage of a brand. From an information theory standpoint, new product advertising is fundamentally different than advertising for established brands.
More recently, Braithwaite and Swindells (1995) and Heath (1999) took the research on brain activity associated with the processing of advertising a step further to a discussion of how differences in processing might affect how brands are constructed in the mind.
The advertising used in this study consisted of eight television commercials, two from each of four packaged goods brands in the personal-care product category. Each pair consisted of one strong and one weak commercial for the brand, defined in terms of statistically significant differences for either the attention-getting power of the ad (in a clutter reel environment), the motivation power of the ad (weighted purchase intent), or both. Thus, the sample of ads, while restricted to one product category, was taken as representing a meaningful range of television advertising experience in terms of important measures of advertising performance.
In the Ameritest commercial pre-test system, each ad was tested monadically among a sample of 125 consumers drawn from female heads of household aged 18 to 54, recruited by mall intercept, with demographic quotas balanced to census. For the eight ads in this study, therefore, 1000 interviews were collected. Each ad was shown in a controlled viewing situation with a personal interview lasting approximately twenty-five minutes. The interview consisted of open-ended and closed-ended questions, with the Picture Sorts data being collected during the last five minutes of the interview. In this diagnostic portion of the interview, respondents were given a shuffled deck of photo images (typically from 10 to 30 photo images) taken from the commercial itself. The number of pictures is determined by the visual complexity of the ad and not by a mechanical rule. The procedure is to take as many pictures as necessary to describe the visual information content of the ad.
The deck of pictures forms a natural vocabulary, based on the commercial iconography itself, for the respondents to use in reconstructing the visual experience they had of the advertising. To construct the Flow of Attention™ respondents are asked to sort images into two piles: the images they remember seeing in the ad and the ones they do not. The results of this sort are displayed in the “graphical user interface” shown in Exhibit 1. The pictures are shown in the order they actually appear in the ad and the height of each picture, measured from the top of the picture, represents the percentage of respondents who remember seeing that image in the commercial. On average, respondents remember approximately two out of three images in a typical commercial (See Young and Robinson, 1989).
These same eight ads were re-tested with a brain-waves measurement system among a sample of 100 women with demographic characteristics similar to the respondents used in the original Ameritests. A proprietary, non-invasive “dry” headset was used to measure the electrical activity of the brain while respondents watched these television commercials. The EEG activity is measured across four commonly accepted frequency bands: delta (4 Hz), theta (4-8 Hz), alpha (8-12 Hz) and beta (10-30 Hz.). Since past research has shown that beta waves are directly related to alertness, focus and task engagement, while alpha and theta are inversely related, these measurements were combined into a single measure, the Engagement Index™, (EI), which is calculated EI= beta ÷ (alpha + theta). (See Pope et al., (1995); Freeman et al., (1999).) Measurements are taken every 0.2 seconds
Comparing Two Non-Verbal Measures
To compare the response functions produced by these two non-verbal measurement systems, it was necessary to take a subset of the brain wave measurements that coincided with the images that were used in the Ameritest Picture Sort. When a particular image was on the television screen for a period of time that extended across several brain-wave measurements, an average of the brain-wave measurements was used. The synchronized measures of the two time series resulted in a total of 201 moments for analysis across the eight commercials.
In addition, the two response functions were normalized by indexing them to the average level for the respective measures within each ad, with the average set equal to 100 for both the Flow of Attention and the Engagement Index.
The synchronized graphs of the two response functions for the eight commercials laid end to end (for illustration purposes only) are shown in Exhibit 2, with the different colored areas highlighting the peaks and valleys of the Picture Sorts function and the brain waves Engagement Index function. Visual inspection reveals that the two response functions are quite different. Indeed, statistical analysis reveals that the two functions are, in fact, totally uncorrelated. A regression model built to predict the Picture Sorts response function with data from the brain wave response function produces an r-squared value of only .009.
Several initial explanations for the failure to find a correlation between the two data sets were explored and rejected.
One possible explanation for the lack of correlation is that there might be some kind of time delay between the two measures. However, use of lagged variables in the regression modeling of the time series data did not improve the fit between the two. This result makes sense given that there is no significant reaction-time bias in either measurement approach. The brain wave measurement is made in real time while the respondent watches the commercial and, unlike other moment-by-moment approaches, does not suffer from the confounding variable of respondent brain-hand reaction time. The Picture Sorts data, an after-the-fact self-reported reconstruction of the advertising experience rather than a mechanical measurement, also contain no reaction-time variable that could time-shift the response function.
A second explanation for the lack of correlation is that the Picture Sorts data only represent how viewers process the visual information content of the commercial while the brain wave measurements represent the total information processing of the commercial—visuals plus copy plus music. While this would lead us to expect only a partial correlation between the two measurement approaches at best, this does not explain the total lack of correlation between the Picture Sorts and brain-waves measurements.
A third explanation of why the two response functions are so different can be based on the fact that brain wave measurements are made while the respondent is watching the commercial whereas Picture Sorts measurements are made up to twenty minutes after viewing the ad. Consequently, the Picture Sorts, in some sense, represents information that has been more deeply processed than that measured by the brain wave EEG. Pursuing this line of reasoning, we conjectured that the response function of the Picture Sorts, representing information that has been filtered through the mind for a period of time, might be less volatile than the brain wave function. In other words, there might be a sharpening or focusing effect as a result of the additional processing resulting in fewer peaks and valleys in the Picture Sorts “wave” function than in the brain waves. To explore this hypothesis, we examined the volatility of the two data sets.
We define a “peak” moment in either the brain wave or the Picture Sorts data sets as one that scores 115 or higher on the index—i.e., at least fifteen percent above the commercial average. Similarly, a “valley” is a moment that scores 85 or lower on the index—i.e., at least fifteen percent lower than the average. Table 1 shows the results of this analysis.
The two response functions are equally volatile—they contain the same number of peaks and valleys. Slightly more than one out of four of the moments in the ads are peak moments according to this definition and slightly fewer than one out of four are valleys. But, importantly, fewer than ten percent of the moments in these eight commercials are simultaneously peak moments for both brain wave arousal and Picture Sorts memorability.
A sensitivity analysis of the cut-off points used to define them (e.g., 120 instead of 115) confirmed that peak moments for brain waves and Picture Sorts are different moments. These two non-verbal measurement approaches appear to be measuring fundamentally different things in the experience of television commercials, and both approaches appear to be important.
A content analysis of these moments, as shown in the Dove deodorant example in Exhibit 3, suggests that both types of moments are essential to the effectiveness of the advertising. In this example, peak Picture Sort moments include two scenes when a woman and man hug each other on-screen. The hug image is a key visual for the Dove brand because (a) it says that with this deodorant you can get close to another person (the social benefit of using the deodorant category) and (b) the hug dramatizes the brand-differentiating idea that Dove is the one deodorant that is soft on your skin. In contrast, brain wave activity peaks more often in the second half of the ad on product visuals, when the new product is introduced and on moments where strategic information is conveyed, such as the fact that this deodorant “contains 1/4 moisturizer.”
Semantic vs Esthetic Information
Information theorists such as Moles (1968) suggest the existence of two types of information that are present in all messages. The first type of information is semantic information; it is logical, structured, and translatable into a foreign language or from one channel of communication to another. Importantly, from a behaviorist conception, it serves to prepare actions in the world—e.g., purchase behavior. From an advertising standpoint this may be thought of as the strategic message content in an advertisement.
The second type of information is esthetic information. It is specific to the channel that transmits it and is profoundly changed by being transferred from one channel to another. This is the information in a picture than cannot be translated into words; it is the information in a piece of music that cannot be described with verbatim playback; it is the poetry of language. Indeed, one of the main reasons that Picture Sorts were developed was to create visual “vocabularies” to probe the esthetic component of advertising. Esthetic information, which pertains to emotions, might be thought of philosophically as personal information which shapes internal states of mind. From an advertising standpoint, it is the execution that makes the communication a creative act.
Visual inspection of the peak brain wave and Picture Sorts moments across the eight ads suggests that the differences in what these two non-verbal approaches are measuring could be related to the dominant type of information present in each moment. To explore this hypothesis, two independent market research data coders (who had not seen the measurement data) were employed to code the information content in these moments.
The type of information present in each of the two hundred commercial moments was coded as follows. The moment was classified as semantic if the visual or associated audio contained strategic information such as a product claim or support point—e.g., contains 1/4 moisturizer—or a package shot, product demo, etc. Otherwise, by default, the moment was classified as esthetic. Examples of the results of the coding are shown in Exhibit 4. In general, the two coders agreed on classification over ninety percent of the time and discussion between them was used to resolve the few differences in classification that remained.
Across all two hundred moments in these ads, eighty percent of the visuals were classified as having primarily esthetic content, while twenty percent contained semantic information (See Table 2). In contrast, only about sixty percent of the audio contained only esthetic content—dialogue—while roughly forty percent contained semantic content. In other words, semantic information was being delivered disproportionately by the audio channel.
Looking at the brain wave analysis, it is apparent that peak moments of brain arousal are relatively more likely (p.05) to be identified with semantic information content. In fact, over half the peak brain wave moments contain semantic information. In contrast, the visuals in these commercials are dominated by esthetic content and the peak moments mirror this dominance. However, looking at the audio content synchronized with the peak Picture Sorts moments, it is apparent that peak moments are relatively less likely to contain semantic content (p.05) and are more likely to be moments of pure esthetic content.
While these differences represent tendencies rather than dichotomies, it can be said that brain wave arousal is associated with the thinking part of the communication—i.e., where information is being learned—while the peak Picture Sorts moments tend to be associated with the feeling part of the ad where some emotion or sensation is being experienced.
This interpretation is consistent with several previous research studies. Zielske (1982) showed that day-after-recall tests penalize emotional advertising—i.e., recall tests under-represent the impact of emotional, esthetic content. Young and Robinson (1989) showed that recall test scores are driven by focusing peak moments in an ad on semantic information content. Moreover, Young and Robinson (l991) also demonstrated that a principle characteristic of new product commercials, as opposed to established brand advertising, is that new product ads are more likely to contain (and viewers are more likely to process) semantic information. In other words, the job of a new product commercial is to “teach” the consumers about the new product—why it exists and how it fits into their world. Consistent with this, the relationship between recall tests, which posit a learning model of advertising effectiveness, and in-market measures of effectiveness such as brand awareness or brand salience is much more clearly established in the literature for new product advertising than it is for established brand advertising. (See Lodish et. al., 1995)
Semantic vs Episodic Memory
A theory of how advertising works in the mind is inescapably linked to a theory of memory. There exists a well-established literature about memory that parallels the construction begun above. E. Tulving (1983, 2000) , a leading figure in modern memory research , describes current theories about the existence of two fundamentally different memory systems that operate in the mind. (There is also a third system, procedural memory, which is not germane to the discussion here but may be important for retail research.) Both systems are important for understanding the mental processes that allow advertising to work and the two dovetail neatly with the two types of information being discussed. The first is the episodic memory system and the second is the semantic memory system. Table 3 summarizes key differences between the two memory systems.
The most important difference is the frame of reference given for the two systems. For the episodic system, the frame of reference is the self—the internal “I” of the consumer. For the semantic system the frame of reference is the universe “outside”—from a marketing standpoint, the competition. When they think of memory, most people tend to think of it in the first, personalized sense, as described by William James (1890):
“Memory requires more than mere dating of a fact in the past. It must be dated in my past. In other words, I must think that I directly experienced its occurrence. It must have that ‘warmth and intimacy’ which were so often spoken of in the chapter on the Self, as characterizing all experiences ‘appropriated” by the thinker as his own.”
In contrast, memorization of objective or impersonal “factual information”—e.g., the multiplication tables, is input to the semantic memory system.
According to Tulving (1983), the source of information for the episodic system is sensation. For example, the tactile sensation evoked by watching a woman touch her hair frequently occurs as a peak Picture Sorts moment in many hair care commercials and would be an instance of the kind of sensory “experience” that is input to the episodic system. On the other hand, for information to be stored in semantic memory it needs to be in the form of concepts that “must be understood and comprehended, that is, related to existing knowledge.” Tulving (1983).
The basic units of information for the semantic system are facts or selling propositions. In contrast, the basic units of information for the episodic system are events or episodes. Tulving’s description of an event could easily include many television commercials:
“An event is something that occurs in a particular situation. It has always a beginning and an end in time, and the interval between the two temporal boundaries is filled with some activity, frequently but not always by one or more ‘actors’. Events recorded in the episodic system always involve the rememberer, either as one of the actors or as an observer of the event.” (1983)
Because the episodic system is organized by time, dramatic narrative structure or story-telling is more important for this system. From a visual narrative standpoint, the peak moments in a Flow of Attention graph are typically the dramatic highlights of a commercial storyline—much like the best remembered scenes of a favorite movie.
While language plays a more important role in the semantic system, the episodic system is fed by experience. Indeed, it could be argued that the purpose of showing scenes of a product or products being consumed or used in television commercials is to create a form of virtual consumption experience for the brand. When fed into the episodic memory system, these might become indistinguishable from real experiences of the individual—much like the phenomenon of “false memories” of childhood that some psychologists have reportedly been able to create. In our database, these moments of virtual consumption also frequently occur at peak Picture Sorts moments in food and beverage advertising.
Generally, it is thought that the episodic system is more context-dependent than the semantic system in which knowledge is de-contextualized. Moreover, affect is considered more important for the episodic system and less important for the semantic system. From an advertising research perspective, these differences may have important implications for understanding media. Media program environment may play more of a role in influencing how advertising works with the episodic system than it does with the semantic system.
The retrieval questions used to access the two systems are also quite different. Information from the semantic system is accessed with questions such as “What brand of deodorant contains 1/4 moisturizer?” Information from the episodic system may require more projective-type probing of time and place, employing questions of “When?” and “Where?” as might be used in a qualitative focus group setting to evoke brand imagery. Indeed, the concept of place plays a central role in methods of training memory and is an important, if not very well understood, concept for branding, as in “Marlboro Country.”
Finally, for advertisers, one of the most significant observations about the differences between the two memory systems commented on by Tulving and echoing the words of James is that “Remembered past events somehow belong (italics mine) to the rememberer.” Indeed, one way to define a “brand” is that it is a product which is, in some emotional sense, owned by the consumer. Whereas using advertising to convey information to the semantic system is important for informing the consumer about the difference between the advertised product and competing alternatives, it appears that the episodic system is central to the mental process by which a product is turned into a brand which is essential in creating brand loyalty.
One of the challenges facing students of advertising is understanding the processes by which brands grow in the mind of the consumer. A useful metaphor for our model is to think of the growth ring of a tree.
It is a general operating principle for scientists working with the emerging mathematical science of complexity theory that the most interesting things—indeed, life itself—happen at the edges of complex dynamic systems, at the boundary between order and randomness. A leading figure in this area, Stuart Kaufman (2000) has described a construct called the “adjacent possible”—a future one-step removed from the present—for illuminating the growth and evolution for biological systems.
In the realm of psychology, Mihalyi Czikszentmihalyi (1990) has described the “flow” state as a mechanism for explaining the growth of the self. Simply put, the flow state occurs at the edge of human experience when you are pushing yourself to perform your personal best, balancing the limits of your skill set against the difficulty of a challenging task you have set yourself. According to Czikszentmihalyi, each time you move toward the edge to enter the flow state of mind for a time, you add a new growth ring for the self.
In the realm of social science, Goffman (1963) has described the high drama which frequently takes place when one crosses social boundaries, such as in a movement from a “backstage” to a “frontstage” social situation. An insightful student of media, Joshua Meyrowitz (1985) used Goffman’s theoretical framework to describe how the movement of perceptual boundaries caused by new media has had major impact on human behavior.
As a continuation of this line of thought, it is useful to consider how the concept of boundaries can be used to understand how advertising can grow a brand. It is an old business rule learned by experienced creatives that in selling a client a new advertising campaign, one should bring three creative concepts. The first concept is the one that is “too safe,” the one the client asked for; the second concept is “too risky,” the far-out execution the creative would really like to do but the client will be too risk-adverse to buy; and in ‘Goldilocks’ fashion, the third concept is “just right,” a fresh idea that is close enough to where the brand has been to be saleable. The logic of this third idea is that it is “edgy” enough to bring the brand a new idea or feeling, but “close enough” in to still make a connection with the roots of the brand. In short, to add another experiential “growth ring” to a brand, one must find ideas close to the boundaries between what the brand owns in the mind of the consumer and what the brand does not own. Growth occurs at the balance point between the past and future, with a focus not on being but on becoming.
If we zoom in on this concept further, we need to ask ourselves what content characterizes this third category of creating concepts—ads that can add “growth rings” to brands? Our hypothesis is shown in Exhibit 5. We suggest that there are special branding moments, which occur in effective, brand-building advertising. Branding moments occur at the boundary between outer-directed semantic information, which defines a brands positioning in the world, and inner-directed esthetic information, where experience created by the advertising is attached to the self. Branding moments are where the mind meets the marketplace.
Branding moments in an ad are those special, ownable moments when a brands positioning is expressed in a fresh new way. Branding moments capture both the essence (positioning) and drama of the brand—hence, brand images have simultaneous semantic and esthetic content (image).
In order to develop a theory of branding moments it is first necessary to have a way of identifying them. The two non-verbal, moment-by-moment techniques described here provide us with a method for doing so. We suggest that branding moments can be identified from the convergence of these two measurement approaches. As we have seen, advertising moments that generate a peak for both the Picture Sort™ and the brain wave Engagement Index represent less than ten percent of the total content of television commercials—and, in fact, such moments do not appear to occur in all commercials. We suggest that branding moments can be found at those intersections. Three examples will be used to illustrate the point. (See Exhibit 6).
The first example comes from the introductory advertising for ThermaSilk shampoo. The reason-for-being for this new product was that, by using a revolutionary formula, this shampoo would actually transform something that is normally bad for your hair—the damaging heat from blow dryers—into something that is actually good for the health of your hair. From pre-testing measures, this television commercial was very attention getting and motivating, and was subsequently very successful in launching this new product. But only one moment from the ad, the one shown in Exhibit 6, generated a peak in both the Picture Sorts™ and brain wave measurements. It’s a visual from the “thermal world” showing a woman using a blow dryer on her hair, with the copy “a blow dryer makes your hair look and feel healthier.” This moment, more than any other in the ad, summed up the strategic selling proposition of the brand. Moreover, the use of unusual “thermalized” imagery was the key esthetic element of the campaign. In fact, when the campaign awareness was tracked in-market at the end of the first year, using a brand-blinded, recognition-based approach to measurement, the thermalized imagery was determined to be an important cue for consumers to use in correctly identifying ThermaSilk as the brand being advertised.
The second example is from Suave® shampoo advertising, another campaign with proven market sales effectiveness. Suave is a value brand with a strategy of convincing consumers that, in terms of the product benefits that are really important to them, all products are pretty much the same so that there is no need to pay a higher price for premium brands. The executional concept is to show two women with great-looking hair, one of who uses Suave and one who uses a premium brand, and have the consumer audience engage in the “interactive game” of figuring out “who uses Suave?”. The point is,” if you can’t tell the difference, why pay more?”. As you can see, this particular visual, a side-by-side shot of two women with great looking hair that you almost want to touch, and the challenge copy combined execution and strategy into a single gestalt. A branding moment? This is the only moment in the commercial where the Picture Sort and brain waves peaked at the same time.
The last example comes from a revised version of the Dove® deodorant execution shown earlier. Based on diagnostic findings from research on the original ad, the execution was revised to provide a better set-up at the beginning of the ad: the Dove parent brand was identified more clearly up front and the importance of having soft skin even under your arms was established before introducing the new product. With these changes, an interesting thing happened. The hug shot, which followed these changes, was again a Picture Sorts peak but was also now a brain wave peak. For the reasons given above this was clearly a key visual in the ad but apparently, with the new set-up, the significance of the raised arms cueing the deodorant category and linking to the strategic benefit also came into focus for the audience. Importantly, the “Dove hug” has the look and feel of a long-term “ownable” advertising equity, i.e. a branding moment.
There are two important points to be made from the last example. First, there is no copy synchronized with this visual, only music—so our concept of branding moments is more than moments in a commercial where strategic copy is synchronized with arresting visuals. Second, branding moments do not stand alone or apart from the other elements in the commercial. The Dove example illustrates this point.
Presently, branding moments represent merely a hypothetical construct that emerges out of the empirical finding that Picture Sorts and brain waves appear to be measuring two very different types of information processing by the mind of a consumer engaged in watching television commercials. One type of processing, of esthetic content, is measured by the Picture Sorts, while the other typed of processing, of semantic content, is measured in particular by EEG response to brain activity. These two types of processing appear to be related to two fundamental marketing processes—building a brand image and positioning a brand in the marketplace.
In only a very small percentage of moments, both measures of information processing peak simultaneously—a fact suggesting that these moments are somehow special. Proving the existence of branding moments might, therefore, provide us with a theoretical mechanism for understanding the long-term, brand-building effects of advertising.
Alwitt, Linda. EEG Activity Reflects the Content of Commercials. Psychological Measures of Advertising Effects: Theory. Hillsdale, NJ: Erlbaum Assoc. 1989: 201-220.
Braithwaite, Alan and Swindells, Alan. “Qualitative Advertising Research.” Proceedings of the Market Research Society Conference, (1995).
Czikszentmihalyi, Mihalyi, Flow: The Psychology of Optimal Human Experience, Harper Collins, NY, NY (1990).
Diamond, S.J. and Beaumont, J. G. “Differences in Vigilance Performance of the Right and Left Hemispheres,” Cortex, 9. (1973): 259-265.
Freeman, Frederick et. al. “Evaluation of An Adaptive Automation Using Three EEG Indices with a Visual Tracking Task,” Biological Psychology, 50. (1998): 61-76.
Heath, Robert. “Just Popping Down to the Shops for a Packet of Image Statements: A New Theory of How Consumers Perceive Brands.” International Journal of Advertising, (Vol) 41, (No) 2, (1999): 153-169.
Moles, Abraham. Information Theory and Esthetic Perception. Joel Cohen Trans. Urbana, IL. University of Illinois Press. (1968).
Ogilvy, David. Ogilvy and Advertising. Crown Publishers. NY, NY. (1983).
Olson, Jerry. “Using Brain Wave Measures to Assess Advertising Effects.” Marketing Science Institute. (1983) Report No. 83-108.
Olson, Jerry C. and William J. Ray. “Exploring the Usefulness of Brain Waves as Measures of Advertising Response.” Marketing Science Institute. (1989) Report No. 89-116 Cambridge, MA.
Pope et. al “Biocybernetic System Evaluates indices of operator Engagement in Automated Task,” Biological Psychology, 40. (1995): 187-195.
Reis, Al and Jack Trout. Positioning: The Battle for Your Mind. NY, NY. Warner Books. (1981).
Rothchild, Michael et. al. “Brain Activity of the Processing of Television Commercials.” Communication Research, (Vol) 13, (No) 2, (1988).
Rothschild, M., & Hyun, Y.J. (1990). “Predicting Memory for Components of TV Commercials from EEG.” Journal of Consumer Research, (Vol) 16, (1989): 472-478.
Shannon, C.E. and W. Weaver. The Mathematical Theory of Communication. Urbana, IL. University of Illinois Press. (1949).
Tulving, E. Elements of Episodic Memory. Oxford, England. Oxford University Press. (1983).
Young, Charles E. and Robinson, Michael. “Guideline: Tracking the Commercial Viewer’s Wandering Attention.” Journal of Advertising Research (Vol) 27, (No) 3, (1987): 15-22.
Young, Charles E. and Robinson, Michael. “The Visual Experience of New and Established Product Commercials.” Advances in Consumer Research. Howland, R. and Solomon, M. (Vol) 18, (1991): 545-551.
Young, Charles E. and Robinson, Michael. “Visual Connectedness and Persuasion.” Journal of Advertising Research, (Vol) 32, (No) 2, (1992): 51-59.
Zeilske, Herbert. “Does Day-After –Recall Penalize ‘Feeling’ Ads?” Journal of Advertising Research, (Vol) 22, (1982).