The Use of an Information Processing Approach to Understand Advertising Effects

Andrew A. Mitchell, Carnegie-Mellon University
ABSTRACT - An information processing approach for examining individual level advertising effects is discussed. This approach focuses on the two critical factors that affect the type of processing that occurs during exposure to an advertisement--attention levels and the processing strategy used by an individual. Some preliminary results are presented from a study that examines the effect of these two factors on brand attitudes and the amount of product information that can be retrieved from long term memory.
[ to cite ]:
Andrew A. Mitchell (1980) ,"The Use of an Information Processing Approach to Understand Advertising Effects", in NA - Advances in Consumer Research Volume 07, eds. Jerry C. Olson, Ann Abor, MI : Association for Consumer Research, Pages: 171-177.

Advances in Consumer Research Volume 7, 1980     Pages 171-177

THE USE OF AN INFORMATION PROCESSING APPROACH TO UNDERSTAND ADVERTISING EFFECTS

Andrew A. Mitchell, Carnegie-Mellon University

[This research was partially supported by NSF Grant DAR 76-81806 from the National Science Foundation.]

ABSTRACT -

An information processing approach for examining individual level advertising effects is discussed. This approach focuses on the two critical factors that affect the type of processing that occurs during exposure to an advertisement--attention levels and the processing strategy used by an individual. Some preliminary results are presented from a study that examines the effect of these two factors on brand attitudes and the amount of product information that can be retrieved from long term memory.

INTRODUCTION

Two fundamentally different approaches have frequently been used in consumer research. Implicit in these research approaches is a definition of the role that consumer research should play within the general context of behavioral research. Under the first approach, the primary role of consumer research is to test theories or mini-theories developed in one of the primary behavioral disciplines (e.g., psychology) within a consumer behavior context. Examples of this approach include the testing of attitude models and attribution theory in a consumer setting.

An alternative approach is to assume that some of the questions to be answered and the phenomena to be understood in consumer behavior are different from those studied in the primary behavioral disciplines. In these situations it may be necessary for consumer researchers to initiate a series of studies directed at answering these questions and understanding these phenomena.

In examining advertising effects, the latter approach would seem to be appropriate. Although social psychologists have provided a useful starting point in their research on communication effects, there are a number of important aspects of advertising that have not been examined. One of these is the effect of visual or pictorial communications. A large portion of advertising today is almost entirely visual. In these advertisements very little verbal information is provided and what little is provided often seems to be almost an afterthought. Exactly how do these predominately pictorial advertisements affect consumer behavior? Are the resulting cognitive processes similar to those used in processing verbal information. Do consumers convert the visual information into verbal information and then use this verbal information to form an evaluation of the advertised brand? Is this the only way that visual information affects brand attitudes? Most of the previous research on communication effects has used verbal stimuli and aside from some initial research in the area (e.g., Lutz and Lutz, 1978; Mitchell and Olson, 1977, 1979; Rossiter and Percy, 1980), we do not have a clear understanding of the effect of visual communications.

Another important aspect of advertising that is not well understood is the effect of "involvement" on the processing of information from advertisements. Many researchers have hypothesized that much of advertising is viewed under conditions of "low involvement", however, there appears to be little understanding of what this means in terms of psychological processes (Mitchell, 1979), to say nothing of what advertising effects may be under these conditions.

In this paper, I would like to discuss our current research on advertising and our attempts to understand the effects of involvement and pictorial advertising. In the next section, I will briefly discuss our overall conceptual approach to understanding these effects which uses an information processing framework. Next, I will discuss some preliminary results from an experiment which examines the effect of "involvement" and pictorial advertising. Finally, I will discuss some implications of this research in terms of developing an overall understanding of advertising effects.

THEORETICAL FRAMEWORK

In using an information processing approach to understand advertising effects, our primary focus is on the cognitive process initiated by exposure to an advertisement and the critical points where the individual may exert control over these processes. These processes are typically thought of as a series of stages that information must pass through in order to be stored in long term memory. These stages include sensory analysis, the formation of message beliefs (i.e., comprehension), the evaluation of these beliefs based on prior information stored in long term memory, the formation and change of product beliefs and the formation and change of brand evaluations or attitudes (Figure 1). Under this view information from the environment activates portions of long term memory which is then used to analyze, interpret and evaluate the information in the advertisement. These processes may be automatic or controlled (Shiffren and Schneider,1977). Controlled processes require attention while automatic processes do not. In most cases, the sensory analysis stage will be automatic, while the latter stages will be controlled.

FIGURE 1

STAGES IN THE INFORMATION ACQUISITION PROCESS

Three important points should be noted about this perspective or model. First, it is generally agreed that individuals do not actually process information in discrete stages and that most information acquisition does not occur from a pure data driven process as described by the model in Figure 1 (Norman and Bobrow, 1975). Instead individuals seam to use both a conceptual and a data driven process (Norman, 1979) with information being processed at different stages simultaneously (McClelland, 1979). Second, this model describes only verbal or semantic processing. It has previously been suggested that two different channels exist for acquiring information, a verbal channel and a visual channel, and that these two channels are independent (Brooks, 1973; Nelson and Brooks, 1973). In addition, Zajonc (1979) has suggested that there may also be a separate channel for affect. Third, the critical resource of this system is attentional capacity. Individuals can only attend to a limited number of stimuli at one time.

Individuals exert control over the information acquisition process by determining how much attention they will devote to the advertisement and the strategy that is used in processing the information (Mitchell, Russo and Gardner, 1979; Mitchell, 1980). Both of these factors will affect what information from the advertisement will be stored in long term memory and how it will be organized. For instance, if little or no attention is directed at an advertisement, probably little will be recalled about the advertisement or the advertised brand. In addition, current research indicates that the amount of attention directed at a communication will affect the amount and type of evaluative processing that occurs during exposure to an advertisement (i.e., Osterhouse and Brock, 1974).

Numerous studies indicate that the processing strategy individuals use to process information will affect the recall of that information. A number of different frameworks, which have been called "processing-permanence" frameworks (Mitchell, 1979), for explaining this relationship have recently been proposed. Among these are "depth of processing" (Craik and Lockhart, 1972; Craik and Tulving, 1975), "dual coding" (Paivio, 1971) and "sensory-semantic coding" (Nelson, Reed and McEnvoy, 1977: Nelson, 1979). The "depth of processing" framework, for instance, suggests that both the "level" to which information is analyzed (e.g., sensory vs. semantic levels) and the amount of elaborative processing will affect recall. More specifically, this framework suggests that both semantic processing and the amount of elaboration enhance recall. On the other hand, the dual coding framework suggests that the number of different types of codes (i.e., visual vs. verbal) generated during processing enhance recall.

The relationship between processing strategy and the recall of the processed information has also been demonstrated in the social cognition literature. Hamilton, Katz and Leirer (1979, for instance, have demonstrated that subjects are able to recall more information from a list of sentence predicates when they are asked to form an impression of the person described by the sentences as opposed to when they are simply instructed to remember as many sentences as possible.2 In summary, although the exact relationship between processing strategy and the ability to recall a stimulus is not clearly understood, such a relationship appears to exist (Craik, 1979).

In discussing the effect of strategy in processing information from advertisements, we have made the distinction between a brand and a nonhrand evaluation processing strategy (Mitchell, Russo and Gardner, 1979). In executing a brand evaluation strategy, the individual processes information from the advertisement with the goal of evaluating the advertised brand. In executing a nonbrand evaluation strategy, the individual attends to the advertisement for other reasons such as enjoying the advertisement for its entertainment value. These two factors, attention and strategy, then, become the basis for examining advertising effects (Figure 2).

FIGURE 2

FACTORS AFFECTING THE PROCESSING OF INFORMATION FROM ADVERTISEMENTS

Involvement

In a previous paper, I defined involvement as the amount of arousal or interest evoked by a particular stimulus (Mitchell, 1979). It is important to note that this definition involves two dimensions, intensity and direction. In discussing involvement in the context of an advertisement, intensity refers to the amount of attention devoted to the advertisement and direction refers to the processing strategy.

Under this conceptualization, processing information from an advertisement under conditions of "high involvement'' would require high attention levels and a brand evaluation strategy. Conditions of "low involvement", however, may occur with both low attention and a brand evaluation strategy and high attention and a nonbrand evaluation strategy. This distinction is important since, as we will show, these two conditions may result in differences in the amount of brand information retrieved from long term memory and brand attitudes.

In the next section the preliminary results from an experiment which examines the effect of attention and processing strategy will be discussed.

METHOD

Manipulations: Subjects examined eight different advertisements for eight hypothetical products under four different experimental conditions and a control condition. Under the first experimental condition, subjects were instructed to evaluate the advertised brand and decide whether or not they would be interested in purchasing it. In the second experimental condition, subjects were asked a series of questions about the advertisement. The questions required the subject to examine and read each advertisement. Examples of these questions are "Can you see two ski poles in the picture?" and "How many times does the word 'you' appear in the copy?"

The third and fourth conditions were similar to the first in that subjects were instructed to evaluate the advertisement brand and decide whether or not they would he interested in purchasing it, however, while they were doing this they were also given a distractor task. In the third condition this task involved watching two lights that were randomly flashing on and off approximately 25 times per minute. The subjects were instructed to raise their right hand whenever one of the lights flashed. This condition was designed to examine processing under reduced attention levels. The distractor task in the fourth condition required subjects to count backwards by threes from a three digit number (e.g.,317) while examining the advertisement. This task was designed to simulate exposure to an advertisement with little or no semantic processing. Finally, in the control condition, subjects were not exposed to any of the advertisements for the hypothetical products.

Stimuli

The advertisements differed as to whether they were primarily pictorial or verbal. Some of the pictorial advertisements used a picture to convey information about the advertised brand (e.g., a ski scene for toothpaste), while others did not (e.g., picture of sailboats for a candy bar). The verbal advertisements differed as to whether basic information about the product was presented (e.g., type of functions on a calculator) or evaluative information about how well the brand performed (e.g., how it feels to wear a pair of shoes). The advertisements were designed by an artist and used either an actual photograph or a professional drawing for the pictorial element. In all cases the pictures used in the pictorial advertisements were designed to evoke positive feelings.

Subjects

The subjects were undergraduate students at Carnegie-Mellon University. There were 15 subjects in each condition and they were paid for their participation in the experiment.

Products

The products used in this study were products that most undergraduates had purchased or used. They included a cola, candy bar, toothpaste and a pocket calculator. Each product was given a fictitious brand name. A pretest indicated that the brand names that were selected (e.g. Catan Cola) conveyed little if any product information.

Procedure

In the four experimental conditions, subjects examined the advertisements under the conditions described above. After examining the advertisements they were given a filler task that lasted approximately five minutes to clear short term memory. Next, their response times to a series of questions were measured using a tachistoscope and then they were given another filler task. After this, their attitudes toward the different brands were measured using a short questionnaire, and then, elicitation procedures were used to determine what information the subjects had stored in memory about each brand (Olson and Muderrisoglu, 1979). Finally, they were given a longer questionnaire and then debriefed. The entire procedure lasted approximately an hour.

In the control condition, subjects were given the brand names of the eight products and were asked to evaluate them using only this information A modified version of the long questionnaire used in the four experimental conditions was used for this condition.

Only the data on the brand evaluations and the amount of product information retrieved from long term memory will be discussed in this paper.

Measures

Attitudes were measured on three seven point bipolar scales anchored by the adjectives: good-bad, pleasant-unpleasant, like very much-dislike very much. The mean of these three scales is used as a measure of attitude toward the brand.

RESULTS

Recall of Brand Information

The average amount of information recalled per brand by condition is shown in Figure 3. As can be seen, approximately the same amount of brand information is recalled in conditions I and III -- between 1.4 and 1.5 pieces per brand, respectively. Condition II and IV also have approximately the same amount of information recalled --between .9 and .6 pieces per brand, respectively. The difference between conditions I and III and the difference between II and IV are not significant at the 0.05 level, however, the differences between all other pairs of conditions are significant. [A 0.05 level was used to determine significance of all the results discussed in this study.]

FIGURE 3

AVERAGE AMOUNT OF BRAND INFORMATION RECALLED PER BRAND BY CONDITION

These results indicate that the different types of processing that occurred under the different conditions resulted in differences in the amount of brand information that subjects were able to retrieve from long term memory. In addition, it should be noted that there were significant differences in the amount of information retrieved from long term memory in conditions II and III, the two "low involvement" conditions, and the strong distraction condition (condition IV) and the nonbrand strategy condition (condition II) resulted in the retrieval of similar amounts of brand information.

Brand Attitudes

The overall mean attitude score for each condition are shown in Figure 4. The overall differences between conditions are slight and they are not significant. There were, however, significant differences between conditions for individual brands. Space limitations do not allow us to discuss these for each brand, but we will discuss the differences by condition for three brands (Figure 5). The advertisement for Brand D was primarily pictorial and it provided little information about the brand. For this advertisement, significant differences were found between conditions I and II, conditions II and III, conditions III and IV and conditions I and IV.

FIGURE 4

MEAN ATTITUDE BY CONDITION

The advertisem2nt for Brand B was again pictorial, however, this time the picture conveyed information about the brand. In this case, there are significant differences between conditions I and III and between conditions I and IV.

Finally, the advertisement for Brand H was primarily a verbal advertisement with the information presented at an evaluative level. Here there are significant differences between only conditions I and II.

These results indicate that the different types of processing that occurred in the different conditions also resulted in the formation of different brand attitudes. There were not, however, consistent patterns in the direction o~ the different brand attitudes between conditions. In order to understand these differences, future research should be directed at understanding the mediators of attitude formation in each condition. Finally, it should be noted that differences in attitudes were found between the two "low involvement" conditions. In all, differences in attitude were found between all conditions, except conditions II and IV.

Exposure Effects

Finally, the effect of exposure without semantic processing is presented in Table 1. Here the attitudes of each brand in condition IV are compared to the attitudes of a control group that did not see any advertisements for the hypothetical products. Significant differences between conditions were found for four of the eight products. Three of these products were featured in pictorial advertisements and for these products, the attitudes formed under condition IV were higher than those in the control. Notice that this was also true for the fourth brand presented in a pictorial advertisement, however, the difference was not significant.

FIGURE 5

ATTITUDE TOWARD THE BRAND BY CONDITION

The only product in a verbal advertisement where there was a significant difference between condition IV and the control, resulted in a higher attitude in the control. Since the attitude for this brand was considerably higher than all of the other brands in the control group, this difference seemed to occur because the subjects in the control group tended to evaluate this brand higher than the other brands. [This tendency appeared to be caused by the brand name (Taylor shoe). Although this brand name did not seem to convey any information about the brand in the pretest, apparently it did among subjects in the control group.]

TABLE 1

EFFECT OF EXPOSURE ON ATTITUDES

These results suggest that exposure to an advertisement with little or no semantic processing may result in the formation of attitudes that are different from those formed without exposure to an advertisement. This occurred primarily with pictorial advertisements and it seemed to occur whether or not the picture conveyed any information about the advertised brand.

DISCUSSION

These results have a two general implications. First, they indicate that the type of processing that occurs during exposure to an advertisement has a strong affect on the amount of product information retrieved from long term memory and on the formation of brand attitudes. Second, there does not appear to be any consistent direction in the differences between conditions. Consequently, in order to predict the effect of am advertisement under a particular condition, attention must be directed at understanding the mediators of each condition.

The results also have a number of implications concerning pictorial advertisements and involvement.

Pictorial Advertisements

The results of this study, along with the results of a previous study (Mitchell and Olson, 1979), suggest that individuals convert the visual information to verbal information in processing information from advertisements. They also suggest, however, that this may not be the only way that pictorial advertisements affect brand evaluations or attitudes. In this study, it was found that exposure to advertisements with little verbal processing can produce attitudinal differences if the advertisements are primarily pictorial.

Two different process interpretations of these results are possible. First, subjects may have acquired the visual information, stored it in memory, retrieved it after the distractor task, converted it into verbal information and then formed attitudes. This interpretation is consistent with the notion that individuals have two different channels for acquiring information --a verbal and a nonverbal channel. A second interpretation is that visual stimuli, if they evoke positive feelings, may directly affect attitudes through classical conditioning mechanisms or what we call "affect transfer''. This latter effect was identified in a previous study where it was found that the affective feelings evoked by the picture in an advertisement explained a reliable amount of the variance in brand attitudes after controlling for the effect of brand beliefs with the Fishbein attitude model (Mitchell and Olson, 1979). In addition, this study indicated that a measure of the affective feelings about the advertisement, which we call "attitude toward the advertisement" seem to capture the direct conditioning effect on attitudes.

At this point, it is difficult to determine which of these two process interpretations are correct. It may be that both are. For instance, the subjects' formed more positive attitudes about brands in condition IV when the advertisements used pictures to convey information than when they didn't. This result is consistent with the first process interpretation. However, when the pictorial advertisement was designed to convey little information about the advertised brand, more favorable attitudes were formed in condition IV compared to the control group. This result is consistent with the second process interpretation.

Involvement

The use of our framework suggests that there are two different causes of "low involvement" processes. One is due to attention deficits and the other is due to the use of nonbrand evaluation strategies. Not only are these two processes different conceptually, but they also may result in the formation of different attitudes and the retrieval of differing amounts of information about the advertised brand from long term memory.

The results presented here also suggest that the conditions of a strong distractor task (condition IV) and a nonbrand strategy (condition II) may be very similar if not the same. Conceptually, in each situation, the message is not comprehended as information about the brand, so message beliefs are not formed. In addition, there were no significant differences between these two conditions as to the amount of information retrieved from long term memory or the attitudes that were formed.

Consequently, these results suggest that the two dimensions of attention and strategy may lead to three different models of semantic information acquisition that may occur with advertising (Figure 6). In the first model, information from the advertisement is actively processed and evaluated. This has typically been termed the "high involvement" information acquisition model. In the second model, information from the advertisement is actively processed, but it is not critically evaluated. This occurs when less than full attention is devoted to the advertisement. In the third model, the message is not comprehended in terms of information about the brand and message beliefs are not formed. This may be due to either severe limitations as to the amount of attention devoted to the advertisement or to the use of a nonbrand evaluation strategy.

FIGURE 6

ALTERNATIVE MODELS OF THE SEMANTIC INFORMATION ACQUISITION PROCESS

These three models are what might be termed alternative "pure" processing models. They are presented as a means of understanding alternative modes of processing and their effects. Under natural conditions, individuals will probably use some combination of these models in acquiring information from advertisements.

Research Approach

The research approach discussed here to examine advertising effects differs from previous approaches in four important ways. First, the information acquisition process has been used to identify the critical factors which will affect this process. Then, these factors are experimentally manipulated to understand the resulting effects on brand evaluations and the amount of product information retrieved from long term memory. This differs from most of the previous research in the area which has used either input-output studies which are generally high on external validity, hut provide little information about the cognitive processes that cause the effects, or studies which present advertisements to subjects and then attempt to measure mediators of the resulting process. In general, I believe that these latter studies force subjects into a "high involvement" mode of processing unless the purpose of the experiment is sufficiently disguised.

Second, a finer-grained view is taken of the information acquisition process and the effects of this process. For instance, rather than simply using paper and pencil measures of beliefs, elicitation procedures are used to actually determine what information individuals are able to retrieve from long term memory after processing information from an advertisement These more precise measures should provide a better indication of the effects of this mediator of attitude formation.

Third, this research recognizes that information may be acquired through channels other than a semantic/verbal channel. Previous research (e.g., the cognitive response and cognitive structure approaches) has concentrated almost entirely on the semantic/verbal channel.

Finally, in using this research approach external validity has been deliberately sacrificed to obtain an individual level understanding of advertising effects. This has been done to obtain a theoretical understanding of how differences in attention levels and processing strategies affect the evaluations of the advertised brands, the content and organization of brand information in memory and the mediators of these effects. Once we have understood these effects at the individual level, we can move backward and study the factors which will affect attention levels and processing strategies. A partial list and discussion of these factors can be formed in Mitchell (1980).

The sacrificing of external validity seems to be almost a necessary condition if we are to obtain a fine-grained understanding of consumer behavior from an information processing perspective. In most cases, the measures that are required to obtain this understanding place a severe strain on external validity.

SUMMARY

In this paper, we have taken an information processing approach to understanding advertising effects at the individual level. In using this approach, we have focused on the cognitive processes initiated by exposure to an advertisement and the factors that may affect these processes -- attention and processing strategy.

This framework and the preliminary results of an experiment discussed in the paper provide important insights into the effects of involvement and visual stimuli. It is suggested that there are actually two different "low involvement" processes and that these processes may result in the formation of different attitudes and the retrieval of differing amounts of brand information from long term memory.

In addition, the results of the experiment discussed in this paper suggested that exposure to pictorial advertisements with little verbal processing may affect attitude formation when the pictures in the advertisements evoked positive feelings. One explanation for these results is that the affective feelings evoked by the picture is being transferred to the brand by classical conditioning principles. In a previous study we found that these effects, which we call "affect transfer'', were captured by the construct "attitude toward the advertisement."

REFERENCES

Brooks, L. R. (1968), "Spatial and Verbal Components of the Act of Recall," Canadian Journal of Psychology, 22, 349-350.

Craik, F. I. M. (1979), "Human Memory," Annual Review of Psychology, 60, 63-102.

Craik, F. I. M. and Lockhart, R. S. (1972), "Levels of Processing: A Framework for Memory Research," Journal of Verbal Learning and Verbal Behavior, 11, 671-684.

Craik, F. I. M. and Tulving, E. (1975), "Depth of Processing and the Retention of Words in Episodic Memory," Journal of Experimental Psychology: General, 104, 268-294.

Hamilton, D. L.; L. B. Katz and V. D. Leirer (1980), "Organizational Processes in Impression Formation," in R. Hastie, et. al. Person Memory: The Cognitive Basis of Social Perception, Hillsdale, NJ: Lawrence Erlbaum Associates (forthcoming).

Lutz, K. A. and Lutz, R. J. (1978), "Imagery-Eliciting Strategies: Review and Implications for Research," in H. Keith Hunt, ed., Advances in Consumer Research, Vol. 5. Ann Arbor: Association for Consumer Research, 611-620.

McClelland, 3. L. (1979), "On the Time Relations of Mental Processes: An Examination of Systems of Processes in Cascade," Psychological Review, 86, 1979, 287-330.

Mitchell, A. A. (1979), "The Relationship Between Cognitive Processes and Cognitive Structure," Paper presented at the American Psychological Association Meeting in New York City.

Mitchell, A. A. (1979), "Involvement: A Potentially Important Mediator of Consumer Behavior," in W. Wilkie, ed., Advances in Consumer Research, Vol. 6, Ann Arbor: Association for Consumer Research, 191-196.

Mitchell, A. A. (1980), "Cognitive Processes Initiated by Exposure to Advertising," in R. Harris, ed., Information Processing Research in Advertising, Hillsdale, NJ: Lawrence Erlbaum Associates (forthcoming).

Mitchell, A. A. and J. C. Olson (1977), "Cognitive Effects of Advertising Repetition," in W. D. Perreault, Jr. (ed.) Advances in Consumer Research, Vol. 4, Atlanta: Association for Consumer Research, 213-220.

Mitchell, A. A. and J. C. Olson (1979), "Are Product Beliefs the Only Mediator of Attitude Formation and Change? An Examination of the Theoretical Logic of the Fishbein Model," Working Paper, Graduate School of Industrial Administration, Carnegie-Mellon University, Pittsburgh, PA.

Mitchell, A. A.; Russo, J. E. and Gardner, M. (1979), "Strategy-Induced Low Involvement Perception of Advertising,'' Working Paper, Graduate School of Industrial Administration, Carnegie-Mellon University, Pittsburgh, PA.

Nelson, D. L. (1979), "Remembering Pictures and Words: Appearance Significance and Name." In L. S. Cermak and F. I. M. Craik, eds., Levels of Processing in Human Memory, Hillsdale, NJ: Lawrence Erlbaum Associates, 45-76.

Nelson, D. L. and D. H. Brooks (1973), "Independence of Phonetic and Imaginal Features," Journal of Experimental Psychology, 97, 1-7.

Nelson, D. L.; Reed, V. S. and McEvoy, C. L. (1979), "Learning to Order Pictures and Words: A Model of Sensory and Semantic Encoding." Journal of Experimental Psychology: Human Learning and Memory, 5, 292-314.

Norman, D A (1979), "Perception, Memory and Mental Processes," in L. M. Nilsson (ed.) Perspectives on Memory Research, Hillsdale, NJ: Lawrence Erlbaum Associates, 121-144.

Norman, D. A. and D. G. Bobrow (1975), "On Data-Limited and Resource-Limited Processing," Cognitive Psychology, 7, 44-64.

Olson, J. C. and A. Muderrisoglu (1979), "The Reliability of Responses Obtained by Free Elicitation: Implications for Measuring Attribute Salience and Memory Structure," in W. Wilkie, ed. Advances in Consumer Research, Vol. 6. Ann Arbor: Association for Consumer Research, 269-275.

Osterhouse, R. A. and T. C. Brock (1970), "Distraction Increases Yielding to Propaganda by Inhibiting Counterarguing," Journal of Personality and Social Psychology, 15, 344-358.

Paivio, A. (1971), Imagery and Verbal Processes, New York: Holt, Rinehart and Winston.

Rossiter, J. and L. Percy (1980), "Visual Communication in Advertising," in R. Harris (ed.), Information Processing Research in Advertising, Hillsdale, NJ: Lawrence Erlbaum Associates (forthcoming).

Shiffren, R. M. and W. Schneider (1977), "Controlled and Automatic Human Information Processing: II, Perceptual Learning, Automatic Attending, and a General Theory," Psychological Review, 84, 127-190.

Staats, A. W. and C. K. Staats (1967), "Attitudes Established by Classical Conditioning," in M Fishbein, ed. Readings in Attitude Theory and Measurement, New York: John Wiley & Sons, 377-381

Zajonc, R. B. (1979), "Feeling and Thinking: Preference Need No Inferences," Distinguished Scientific Contribution Address. Presented at the American Psychological Association Meeting in New York.

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