Visual Information Processing of Television Commercials: Cognitive Effects

ABSTRACT - Advertisers have utilized various techniques for assessing the effectiveness of television commercials. The present manuscript describes a theory and methodology for examining viewers' chunking of ongoing visual information when viewing commercials. In an exploratory study, viewers indicated the number of chunks of information in television commercials while being exposed to either the visual stimuli or to both the audio and visual stimuli. It was hypothesized that the number of boundaries of information chunks (breakpoints) indicated by observers would affect information processing and, subsequently, output measures of advertising effectiveness (i.e., attitudes & beliefs). Results consistent with this process perspective were found.

Citation:

Thomas D. Jensen and Larry W. Rottmeyer (1986) ,"Visual Information Processing of Television Commercials: Cognitive Effects", in NA - Advances in Consumer Research Volume 13, eds. Richard J. Lutz, Provo, UT : Association for Consumer Research, Pages: 158-163.

Advances in Consumer Research Volume 13, 1986      Pages 158-163

VISUAL INFORMATION PROCESSING OF TELEVISION COMMERCIALS: COGNITIVE EFFECTS

Thomas D. Jensen, University of Arkansas

Larry W. Rottmeyer, University of Arkansas

ABSTRACT -

Advertisers have utilized various techniques for assessing the effectiveness of television commercials. The present manuscript describes a theory and methodology for examining viewers' chunking of ongoing visual information when viewing commercials. In an exploratory study, viewers indicated the number of chunks of information in television commercials while being exposed to either the visual stimuli or to both the audio and visual stimuli. It was hypothesized that the number of boundaries of information chunks (breakpoints) indicated by observers would affect information processing and, subsequently, output measures of advertising effectiveness (i.e., attitudes & beliefs). Results consistent with this process perspective were found.

INTRODUCTION

A variety of research techniques and measures have been proposed and developed for assessing the effectiveness of television commercials. These techniques include memorial measures (e.g., recall, & recognition), physiological measures (e.g., eye movements, tachistoscope, galvanic skin responses, pupil dilations, & electroencephalograms), and cognitive measures (e.g., cognitive rehearsals, free associations, attitudes, beliefs, & intentions) as well as more behaviorally oriented measures (e.g., sales following test markets, inquiries via reader cards, calls to toll-free telephone numbers requesting additional information). Although all of these techniques have advantages and disadvantages (see Engel, Warshaw, & Kinnear 1983), with the exception of the physiological measures, the majority of the measurement techniques do not allow the simultaneous collection of data during consumer viewing of the commercials and, hence, are post hoc. Consequently, these measures fail to acknowledge that the manner in which the stimuli are processed may affect subsequent cognitive measures and behaviors (i.e., purchase). For the physiological measurement techniques, mixed results have been found and, hence, have had little practical implications for the development of commercials (Watson & Gatchel 1979; Hensel 1970).

Previous research in the field of advertising has concentrated on the stimuli inputs and cognitive outputs related to television commercials such as the attention properties of the commercial and recall of these attention points. However, little or no attention has been given to the processing (what & how) involved during the actual viewing of the commercial. When information processing has been addressed, it has almost exclusively been conducted on auditory as opposed to visual stimuli.

Although there is an absence of research addressing visual information processing of television commercials, some research does exist examining the processing of visual stimuli for print advertisements (i.e., pictures vs. text). Generally, these studies have found that pictures are more easily recalled than accompanying text or text containing the same basic information as the pictures, and that print advertisements influence product related attitudes and beliefs, both negatively and positively. In one of the few studies to actually examine the effects on consumers of print advertising from a process perspective, Edell & Staelin (1983) found differences in the processing of text and pictures based upon the relevancy between the two components as well as for subsequent cognitions (e.g., attitudes). Specifically, these authors found differential effects of the sd structure (relevancy of pictures and text) and ad content (objective versus subjective) on the processing of information and subsequent measures of cognitive responses, beliefs, attitudes, and intentions. These dependent measures were found to be hierarchical whereby the activation of one process variable impacted upon subsequent variables.

The present process conceptualization for television commercials parallels that of Edell and Staelin s (1983) conceptualization for the processing of print advertisements with one major exception: the present study examines the manner in which the consumers process the visual information contained within the commercial as opposed to any content differences between commercials per se. In this regard, the present conceptualization suggests that the stimuli are not necessarily the sole key to what material is processed and how the material is processed. Rather, the conceptualization suggests an interactive process of the advertisement, information processing, and cognitive activity. In the case of television advertisements, the processing of audio material should be similar to the processing of text material in print advertisements and, hence, similar to output measures from text material. For visual material, however, the motion inherent in television commercials forces viewers to process a series of "pictures" or "frames"; each "picture" potentially having the property of conveying information to the viewer and interacting with the audio portion as well as with viewer processing variables. This perspective suggests that consumers may segment the visual stimuli into chunks. The chunking of the visual portion of commercials and related variables represents the major thrust of the present study. The number of chunks of visual information that is perceived by a viewer should impact upon various output measures of beliefs and attitudes. In this regard, a theoretical perspective and methodological technique for examining the visual information process are warranted. The feature change hypothesis and breakpoint analYses serve this PurPOSe.

Feature Change Hypothesis & Breakpoint Analyses. Darrin

Newtson and his colleagues (Newtson 1973, 1976; Newtson & Engquist 1976; Newtson, Engquist, & Bois 1977; Newtson, Rinder, Miller, & LaCross 1978) have demonstrated that observers of an action sequence tend to divide the action into meaningful segments for the purpose of extracting information. The number of segments indicated by observers has been shown to be affected by the instructional set provided to the observers (virtually all studies in this area), the predictability of the action sequence (Newtson 1973; Newtson et al. 1978; Wilder 1978a, 1978b; Jensen & Schroeder 1982), the sex-type of the observer (Deaux & Majors 1977), and by the number of actors (Jensen 1983, 1984). The indication of the boundaries for a segment are termed breakpoints. What is important for information processing, however, is not the actual breakpoints but, rather, the change or action between two successive breakpoints. This contention, termed the feature change hypothesis, has received considerable empirical support (e.g., Newtson et al. 1977; Newtson et al. 1978).

The feature change hypothesis postulates that stimuli factors such as movement, rate of movement, changes in the rate of movement, predictability, and changes in the relationship between two objects could alter an individual s segmentation strategy when observing ongoing action. Viewer factors such as attention, the sex-type, personality, and previous exposure to the material or related material may also affect the degree of perceptual analysis of an action sequence. Furthermore, this hypothesis suggests that stimulus and viewer factors may interact in affecting the processing of visual information and subsequent cognitions. The research methodology in most of the studies have had observers view a videotape or movie while indicating "meaningful actions" (breakpoints) by either pressing a button connected to a continuous event recorder, pressing a button connected to a high speed computer, or by indicating tally marks on a piece of paper.

The present study was not designed to be a direct test of the feature change hypothesis but, rather, was designed to test the relationship between the chunking of visual information in commercials and the subsequent impact upon cognitions. However, in testing this relationship, ancillary evidence for the feature change hypothesis may be gleaned. In the present study, it was predicted that individuals viewing television commercials would segment television commercials into chunks when provided with differential instructions utilizing a breakpoint methodology and that the number of breakpoints indicated would be related to subsequent cognitive output variables (e.g., attitudes & beliefs). First, it was predicted that individuals instructed to indicate the more detailed or "smaller" actions in the commercial that were meaningful to them would indicate more breakpoints than individuals instructed to indicate the "larger" actions. More importantly, it was predicted that a positive relationship would be found between the number of breakpoints indicated by individuals and subsequent cognitive output measures (i.e., attitudes & beliefs). Previous research has indicated a positive relationship between the number of breakpoints and attributions (i.e., Newtson 1973; Newtson & Rindner 1979; Wilder 1978a) and affect (i.e., Lassiter & Stone 1983) concerning the actor(s) in stimulus materials.

The above predict ions and previous research findings are based upon the premise that an increase in the number of chunks of visual information results in more positive cognitions. However, it is feasible that with an increase in the number of chunks of information, an individual would simply process more information to confirm pre-viewing cognitions. In other words, individuals who were initially positive toward the brand, commercial, or actors within the commercial could become more positive after more detailed processing while individuals who were initially negative could become more negative. In this regard, following the viewing of a commercial, individuals cognitions could become more extreme or increase in magnitude in either a positive or negative direction. Hence, it was predicted that individuals indicating a finer unit of perceptual analysis would holt more extreme cognitions.

Finally, it was predicted that the presence versus absence of the audio portion of the commercials would affect the number of breakpoint indicated, hence, lending support to the interactive perspective of the stimulus material and viewer visual information processing. No specific predictions were made about the direction of the influence when both audio and visual stimuli of a commercial were provided to viewers as opposed to only the visual stimuli.

METHODOLOGY

Subjects & Procedures

A total of 209 students (96 males & 113 females) enrolled in either a consumer behavior course or a principles of marketing course volunteered to participate in the study for extra course credit. The 2 to 5 subjects who attended each session were instructed that the study addressed the manner in which individuals extracted information from television commercials. The experimenter explained to the subjects that previous research had indicated that people segment action sequences into meaningful segments or units and provided the subjects with a brief example of how an act ion sequence could be segmented into either small or large segments. The subjects were informed that they would be making judgments of meaningful actions in commercials by pressing a handheld metal tally register whenever, in their judgment, one meaningful action ended and another action began.

Approximately one fourth of the subjects were instructed to press the tally register for the smaLlest (fine unitization instructions) meaningful actions, one fourth of the subjects for the largest (gross unitization instructions), and one fourth of the subjects for the most natural (natural unitization instructions). For example, individuals provided with fine unitization instructions were asked to "indicate their judgments of the smallest meaningful behaviors or actions portrayed in the commercial by pressing the tally register." The remaining one fourth of the subjects were not provided with any information or instructions concerning the segmentation of action (no unitization instructions): they were informed that the study dealt with the manner in which individuals processed information from commercials and that following the viewing of some commercials they would be requested to complete a questionnaire concerning their impressions of the commercials.

One half of the subjects were shown the commercials with the audio turned off (visual condition). The remaining subjects were provided with both the visual and audio portions of the commercials (audio/visual conditions). Hence, the factorial combination of instructions (fine, natural, gross, or none) and stimuli (visual or audio/ visual) resulted in eight conditions that were conducted according to a block randomized schedule.

For those individuals provided with unitization instructions, between each of the six commercials the experimenter recorded the number of breakpoints indicated by each subject for that commercial. Following the presentation of the commercials, all of the subjects completed a questionnaire concerning their impressions of the commercials. When the subjects had completed the questionnaire the experimenter collected the booklets and thoroughly debriefed the subjects.

Stimulus Materials: Commercials & Questionnaire

Television commercials were recorded on a 1/2" VCR color television recorder using two major criterion: the commercial must contain ongoing action sequences and the primary emphasis other than brand must be the actor and the relationship between the actor and the brand. Using these criterion, 66 commercials were selected for further consideration, of which six were utilized in the present study based upon the clarity of the picture, length (30 seconds), the advertised product being a relatively frequently purchased nondurable good, and the presence of visual information. In order to reduce any order effects, two randomly ordered presentations were developed with approximately one half of the subjects receiving each of the two different orders of presentation.

The questionnaire requested subjects to indicate their beliefs about the primary actor/actress (13 items), commercial (11 items), and brand (7 items) for each of the six commercials. Subjects were also requested to indicate their attitude (liking) toward the actor/actress, commercial, and brand for each of the six commercials. All belief and attitude measures were collected on seven point semantic differential scales with the positive and negative poles randomly determined both within and between commercials. However, for all of the analyses presented below the scales have been transformed such that 7 represents the positive pole. Finally, a series of demographic items, the frequency of previous purchases of the brands and exposure to the commercials, and future purchase intentions were collected.

RESULTS

The effects of the instructions (fine, natural, gross, or none) and stimuli (visual only or audio/visual) on the various dependent measures were tested utilizing multivariate analysis of variance (MANOVA) across commercials where the main effects of instructions and stimuli as well as the interaction served as predictor variables. The results of these analyses for all dependent variables are reported in Table 1. As a manipulation check of the assignment of individuals to conditions, two analyses were conducted utilizing the individual s self-reported frequency of exposure to the commercials and pre-experimental purchase patterns of the advertised brands. None of the F values approached statistical significance for these dependent variables, verifying the random assignment procedure.

TABLE 1

MANOVAS ACROSS SIX COMMERCIALS: F VALUES AND df

Breakpoints

The analyses conducted on the number of breakpoints indicated for the commercials revealed, as predicted, significant main effects of instructions and stimuli. Collapsing across commercials, individuals given fine unitization instructions indicated more breakpoints (X 9.08) than individuals given natural unitization instructions (X = 7.41) with individuals given gross unitization instructions indicating a number of breakpoints between the two extremes (X = 8.58). Individuals receiving only the visual stimuli indicated more breakpoints (X = 9.32) than individuals receiving both the audio and visual portions of the commercial (X = 7.36). The interaction of these two factors did not approach significance.

Purchase Intentions

Analyses conducted on purchase intentions for the advertised brands revealed no differences between conditions.

Attitudes

Commercial. The F values associated with the effects of instructions and stimuli on attitudes towards the commercials revealed a significant F value for instructions. Individuals provided with fine unitization instructions indicated greater liking for the commercials (X = 4.85) than did individuals who were not provided with any unitization instructions (X = 4.51). Individuals provided with gross or natural unitization instructions indicated a degree of liking for the commercials somewhere in between those individuals in the other two conditions (X = 4.66 and X = 4.62, respectively). The F values for stimuli and the interaction of instructions X stimuli did not approach statistical significance.

Brand. Analyses conducted for participants liking of the brand revealed only a significant effect of stimuli. Individuals receiving both the audio and visual portions of the commercials indicated a greater liking for the brand (X = 4.67) than individuals receiving only the visual stimuli (X = 4.51). The main effect of instructions and the interaction of instructions X stimuli did not approach significance.

Actors. Analyses conducted on individuals' attitudes toward the actors revealed a marginally significant F value, E = .11, for the main effect of instructions. This findings was due to individuals receiving the fine unitization instructions (X = 4.95) or no unitization instructions (X = 4.83) indicating a more positive attitude toward the actors than individuals receiving either gross (X = 4.72) or natural unitization instructions (X = 4.74).

Attitude Extremity

In order to assess whether the attitude extremity was affected by the experimental manipulations as opposed to the direction and strength of the attitude as in previous studies, the scales were collapsed about the midpoint. Hence, attitude extremity measures ranged from O to 3 with O indicating neutral and 3 indicating a more extreme attitude regardless of direction (e.g., favorable versus unfavorable).

Commercial. The analyses conducted on the attitude extremity for the commercial revealed significant main effects for instructions and stimuli as well as a significant instructions X stimuli interaction. Basically, individuals provided with natural unitization instructions held more extreme attitudes (X - 1.81) than individuals provided with no unitization instructions (X - 1.67), gross unitization instructions (X = 1.63) or fine unitization instructions (X = 1.58). Although the MANOVA revealed a main effect of instructions. this finding was due to individuals indicating a more extreme attitude when presented with both the audio and visual stimuli (X = 1.79) as compared to only the visual stimuli (X 1.55) for two of the commercials while for the other four commercials this trend was reversed (visual = 1.73, audio/visual = 1.61), qualifying the overall significance when collapsing across all six commercials (X = 1.67). The significant interaction was due to individuals given fine unitization instructions indicating more extreme attitudes when receiving only the visual stimuli (X = 1.65) as compared to both the audio and visual stimuli (X = 1.51) while for the other instructional conditions no significant difference was found (audio/visual = 1.73 visual (1.68).

Brand. Analyses conducted for attitude extremity in reference to the brands advertised did not reveal any significant differences.

Actors. Although not revealing any significant main effects, this analysis revealed a significant instruction X stimuli interaction. Individuals receiving only the visual stimuli indicated more extreme attitudes when provided with natural unitization instructions (X = 1.68) than when provided with either fine (X = 1.56), gross (X = 1.56), or no (X = 1.53) unitization instructions. For individuals receiving audio and visual stimuli, however, more extreme attitudes were found for natural unitization instructions (X = 1.76) than for gross (X = 1.58) or no (X = 1.50) unitization instructions, which in turn were more extreme than for fine unitization instructions (X =1.37).

Beliefs

In order to examine the participants beliefs, all items used in tapping the beliefs were reassigned such that 7 was the positive pole and 1 represented the negative pole. In the case of items that tit not possess easily discernible positive and negative poles (e.g., traditional versus contemporary), the scales were transformed when necessary such that the poles were labeled consistently (e.g., traditional = 1). The items tapping subjects beliefs were then summed for beliefs about the commercial (11 items), brand (7 items), and actors/ actresses (13 items). Hence, neutral beliefs would be indicated by scores of 44, 28, and 52 for commercial, brand, and actors/actresses, respectively. Collapsing across commercials, Cronbach's alphas were .81 for commercial beliefs, .56 for brand beliefs, and .78 for actor beliefs.

Commercial. A significant effect of instructions for participants beliefs about the commercial was revealed. Individuals provided with unitization instructions tended to have slightly more positive beliefs (fine = 49.76, gross = 48.99, & natural = 49.18) than individuals not receiving any unitization instructions (X = 48.07).

Brand. This analysis did not reveal any significant main effects nor interactions.

Actors. Individuals receiving different instructions held differential beliefs about the actors. As with commercial beliefs, individuals receiving unitization instructions hat more positive beliefs about the actors (fine = 59.51, gross = 59.07, & natural = 58.75) than individuals not receiving any unitization instructions (none = 57.98).

Belief Extremity

As with attitudes, the possibility exists that the experimental manipulations utilized in the present study may have affected the extremeness of beliefs as opposed to or in addition to beliefs per se. In this respect, the instructional manipulations may have caused individuals to attend closer to the details contained within the commercial itself. This increased attention could result in the individuals observing some aspect of the commercial which, in turn, could justify their existing beliefs and, subsequently, increase the magnitude or extremity of those beliefs. In order to test this proposition, individuals belief scores were collapsed about the neutral midpoint for commercial, brand, and actor beliefs. In this regard, a zero belief extremity score represents a neutral or very weakly held belief for the commercial, brand, or actor. Scores of 33, 21, and 39 represent extreme beliefs, either positive or negative, concerning the commercial, brand, and actor. respectively.

Commercial. Analyses conducted for extremity of beliefs about the commercial did not reveal any significant effects.

Brand. The F values for extremity of brand beliefs were not significant.

Actors. Analyses conducted for the extremity of beliefs regarding the actors in the commercials revealed a significant main effect of instructions. Individuals receiving gross (X = 11.36) or natural (X = 11.49) instruct ions indicated held more extreme beliefs when compared to individuals not receiving any unitization instructions (X = 9.78) while individuals receiving fine unitization instructions possessed beliefs somewhere between the two extremes (X = 10.61). No other significant effects were found.

DISCUSSION

In the present study, individuals directed to attend to the finer details of the action in a commercial perceive more meaningful behaviors than individuals provided with alternative instructional sets. T he fact that individuals indicated more breakpoints when provided with "gross" as opposed to "natural" unit instructions is inconsistent with previous studies examining the unitization of ongoing action sequences: most research has found that, although not necessarily statistically significant, individuals indicate more breakpoints when provided with natural as opposed to gross unitization instructions. Two possibilities for this discrepancy exist. First, under "natural" unitization instructions observers may have believed that the semantic context of natural could be interpreted to imply "as if you were watching the commercials at home." In this regard, individuals may have reacted by not paying as close of attention to the commercial as they did under "gross" unitization instructions and, hence, passively processed the incoming visual and auditory information. Second, in a similar semantic vein, individuals may have interpreted the "natural" in the instructions as referring to "common everyday occurrences" rather than "naturally meaningful" actions. Given the uniqueness of the actions contained within the commercials and the propensity for these actions to attract attention, individuals may have viewed the commercials as containing few actions or behaviors that they might engage in or that they might view others exhibiting. Also, some combination of the above two reasons may have been responsible for the difference between gross and natural unitization instructions.

The finding in the present study that individuals indicated more breakpoints when provided with only the visual portion of the commercials as contrasted to both the audio and visual portions of the commercials is, although unique in the present context (e.g., commercials), consistent with previous studies requesting individuals to focus on either the visual or the verbal behaviors of two actors having a conversation (e.g., Strenta & Kleck 1984). At least three explanations for this finding are feasible. First, it is possible that the audio stimuli interfered with the processing of the visual information. The inclusion of the audio portion in the present study may have caused individuals to engage in more "gross" processing of visual information due to the interference/distraction of the auditory stimuli. Second, the audio message may have caused the individual to shift toward a "gross" unitization strategy of the visual information, either actively or passively, in order to capture both sensory inputs. This explanation suggests a change (reduction) in visual information processing with the increase in information. Third, individuals may lack the attention capacity to process both stimuli and may therefore switch attention between the audio and visual, resulting in the appearance of a gross" visual analysis.

All of the above explanations, although not exhaustive, assume a limited capacity input channel. Individuals may have had to allot processing time to the audio stimuli in order to capture both sets of information. In this regard, visual messages may not be adequately processed, not as deeply processed, processed in less detail, and/or processed in larger chunks when presented with audio messages. Hence, one would predict that more complex and/or distracting audio messages would result in poorer visual information processing and have a subsequent impact upon memory for the components (cf., Bither & Wright 1973). Likewise, under conditions of the same audio message, complex visual messages would not be processed to the same extent nor in the same manner as simple visual messages. These explanations are also consistent with the work of Posner and Snyder (1975a, 1975b) who have shown that if attention is attracted to one modality, information from another modality is not processed to the same extent as if it were presented separately and with Edell and Staelin s (1983) findings that the consistency between the messages presented in the pictures and text of print advertisements affects memory and cognitions. Furthermore, these interpretations are also consistent with Jensen and Schroeder s (1982) findings that an increase in the number of actors does not result in a corresponding increase in the number of breakpoints.

The relationship between the visual information processing as operationally defined by the number of breakpoints and subsequent measures of cognitive processes revealed some interesting findings. As expected, individuals cognitions about the commercial were more affected by the extent of visual information processing than individuals cognitions about either the actors or brands. This finding is not surprising when one acknowledges that the extent or magnitude of processing manipulation involved attention properties as well as an interaction of attention and cognitive activities (see Newtson 1980). Individuals may be more prone to cognitive change concerning an object that they have had relatively little experience, such as the commercial, whereas individuals may possess relatively strong attitudes and beliefs regarding the brand. Also, a strongly held cognition would be relatively difficult to alter following a single exposure. For cognitions about the actor, individuals would be unlikely to form attitudes since the observer would not foresee interacting with the actor or making any decisions involving the actor. However, individuals should be able to form beliefs about the actors, assigning them various characteristics or attributes (e.g., talent). These attributes, in turn, should be related to the attention an individual pays to the actors and, hence, as found in the present study, to the extent or magnitude of processing. This finding is consistent with previous studies examining the segmentation of ongoing behavior and subsequent causal and ability attributions (e.g., Newtson 1973; Newtson & Rindner 1979; Wilder 1978a). Also, for both cognitions regarding the commercial and actors, the positive relationship between the extent of processing (i.e., number of breakpoints) and attitudes or beliefs is consistent with earlier studies of affect (e.g., Lassiter & Stone 1983).

An interesting finding concerning cognitions and measures of cognitive extremity was found for attitudes and beliefs about the commercial and beliefs about the actors. Given the random assignment of individuals to conditions and subsequent manipulation checks, the amount of processing affected cognitive change in a positive direction which, in turn, was inversely related to measures of cognitive extremity. Individuals provided with instructions invoking a finer level of analysis were more likely to change their attitudes and beliefs in a favorable direction but, subsequently, appeared less sure or confident of those attitudes and beliefs. The possibility, strongly suggested in the present study, that cognitive change is inversely related to subsequent measures of extremity, confidence, or strength following persuasive communications warrants additional research.

Potential limitations of the present study are similar to those associated with other exploratory studies and the need for more sophisticated attention in future research efforts. First, the results of the present study for either the breakpoints, cognitions, or both measures may be tue to demand characteristics. The plausibility of demand characteristics causing the individuals to respond in some set manner is weakened, however, when one notes the consistency of the present results with previous studies, findings from other studies examining the motion between successive breakpoints as opposed to nonbreakpoints, and similarities in memory for individuals provided with natural as compared to no unitization instructions in recognizing breakpoints and nonbreakpoints (for example, see Newtson et al. 1977). For the measures of cognitions, it also seems unlikely that demand characteristics could account for the more positive actor and commercial related cognitions while not also impacting upon brand related cognitions. Nonetheless, additional studies may need to develop appropriate methodologies to examine the possibility of demand characteristics in spuriously producing the obtained results.

Second, one can always question the generalizability of the findings in a laboratory situation to more "real world" settings. Individuals in the present study knew they were involved in a study examining the "effectiveness" of various commercials. The possibility exists that some of the findings may have been tue to the artificial situation imposed upon the participants.

Third, individuals in the present study were exposed to only a single presentation of each commercial. Given that most individuals had some previous exposure to the commercials, questions arise about the effects of multiple exposures and novelty of the commercials on visual information processing and subsequent cognitions.

Fourth, the present study did not examine the relationship between visual information processing and cognitive responses nor memorial factors (e.g., Edell & Staelin 1983). Studies are justified which examine the simultaneous information processing of television commercials, cognitive responses, and recall as well as recognition. Furthermore, the present study measured cognitions immediately following the viewing of the commercials. Future research should examine the persistence of some of the findings (e.g., attitude change) following differential processing as well as probe for new findings following the passage of time (e.g., changes in confidence).

Fifth, the present study examined commercials for relatively frequently bought products, some of which may not have been especially relevant for all of the participants. Other studies examining visual versus auditory (e.g., Bither & Wright 1973) or pictorial versus textual (Edell & Staelin 1983) have concentrated their efforts on less frequently purchased, higher ticket products (e.g., automobiles). Investigators may need to consider the degree of involvement of the observers for the advertised products.

In conclusion, echoing and elaborating upon the call by Jensen and Schroeder (1982), the ultimate veracity of the segmentation of ongoing action rests in the identification of related psychological processes (e.g., cognitions) and the resulting impact upon behavior. Furthermore, future research should be concerned with the identification of the conditions likely to induce the various levels of processing. For example, Newtson (1973) and Wilder (1978a; 1978b) have shown that the predictability of the action affects the unitization rate. Marketing research efforts designed to identify the characteristics needed in television commercials to invoke fine or gross levels of visual information processing should prove to be valuable both from theoretical and applied perspectives.

REFERENCES

Cohen, D. E., & Ebbesen, E. B. (1979) "Observational Goals and Schema Situation: A Theoretical Framework for Behavior Perception." Journal of Experimental Social Psychology, 15, 305-329.

Deaux, D., & Majors, B. (1977) "Sex-Related Patterns in the Unit of Perception." Personality and Social Psychology Bulletin, 3, 297-300.

Edell, J. A. & Staelin, R. (1983) 'She Information Processing of Pictures in Print Advertisements." Journal of Consumer Research, 10, 45-61.

Engel, J. F., Warshaw, M. R., & Kinnear, T. C. (1983) Promotional Strategy, 5th edition, Richard D. Irwin, Inc.: Homewood Illinois.

Hensel, J. S. (1970) Physiological Measures of Advertising Effectiveness: A Theoretical and Empirical Investigation. (Ph.D. dissertation, Ohio State University).

Jensen, T. D. (1983) Behavior Segmentation Strategies: Single versus Dyadic Actor Situations. Paper presented at the 29th Annual Convention of the Southwestern Psychological Association, San Antonio, Texas, April.

Jensen, T. D. (1984) Behavior Segmentation of Dyadic Actor Situations: Interactive versus Co-acting Sequences. Paper presented at the 30th Annual Convention of the Southwestern Psychological Association, New Orleans, April.

Jensen, T. D. & Schroeder, D. A. (1982) "Behavior Segmentation in a Dyadic Situation." Personality and Social Psychology Bulletin, 8, 264-272.

Lassiter, G. D., & Stone, J. I. (1984) "Affective Consequence of Variation in Behavior Perception: When Liking is in the Level of Analysis." Personality and Social Psychology Bulletin, 10, 253-259.

Massad, C. M., Hubbard, M., & Newtson, D. (1979) "Selective Perception of Events." Journal of Personality and Social Psychology, 15, 513-532.

Newtson, D. (1973) "Attribution and the Unit of Perception of Ongoing Behavior." Journal of Personality and Social Psychology, 28, 23-28.

Newtson, D. (1976) "Foundations of Attribution: The Unit of Perception of Ongoing Behavior." In J. Harvey, W. Ickes, & R. Kidd (Eds.), New Directions in Attribution Research, Erlbaum: Hillsdale, N. J.

Newtson, D. & Engquist, G. (1976) "The Perceptual Organization of Ongoing Behavior." Journal of Experimental Social Psychology, 12, 436-450.

Newtson, D., Engquist, G., & Bois, J. (1977) "The Objective Basis of Behavior Units." Journal of Personality and Social Psychology, 35, 847-862.

Newtson, D. & Rindner, R. J. (1979) "Variation in Behavior Perception and Ability Attribution." Journal of Personality and Social Psychology, 37, 1847-1858.

Newtson, D., Rindner, R., Miller, R., & LaCross, K. (1978) "Effects of Availability of Feature Changes on Behavior Segmentation." Journal of Experimental Social Psychology, 14, 379-388.

Posner, M. I., & Snyder, C. R. R. (1975a) "Attention and Cognitive Control." In R. L. Solso (Et.), Information Processing and Cognition: The Loyola Symposium. Erlbaum Associates: Hillsdale, N. J;

Posner, M. I., & Snyder, C. R. R. (1975b) "Facilitation and Inhibition in the Processing of Signals." In P. M. A. Rabbitt & S. Dornic (Eds.), Attention and Performance V. Academic Press: New York.

Strenta, A. C., & Kleck, R. E. (1984) "Physical Disability and the Perception of Social Interaction: It s Not What You Look At But How You Look At It." Personality and Social Psychology Bulletin, 10, 279-288.

Watson, P. J. & Gatchel, R. J. (1979) "Autonomic Measures of Advertising," Journal of Advertising Research, 19, 15-26.

Wilder, D. (1978a) "Effect of Predictability on Units of Perception and Attribution." Personality and Social Psychology Bulletin, 4, 281-284.

Wilder, D. (1978b) "Predictability of Behaviors, Goals, and Units of Perception." Personality and Social Psychology Bulletin, 4, 604-607.

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