CONSUMER PROCESSING OF TELEVISION COMMERCIALS: A MODEL AND PRELIMINARY TEST
CONSUMER PROCESSING OF TELEVISION COMMERCIALS:
A MODEL AND PRELIMINARY TEST
Kenneth R. Lord, Mercer University and
Sanjay Putrevu, University of Western Australia
A Cognitive Arousal Model is proposed to explain how program, ad, and consumer characteristics affect viewer processing of television commercials. Ad response is modeled as a function of the amount and type of thoughts generated upon exposure to the message. The viewer’s cognitive state, ranging from passivity to arousal to distraction, is affected by the interaction of the consumer’s pre-existing level of involvement in the topic of the ad, program involvement at the time the commercial appears, involvement in the ad per se, and whether the source of the ad involvement is relevant or irrelevant to the message’s central arguments. Various propositions are advanced regarding the viewer’s cognitive state as well as how such a state might affect ad response. The propositions are tested using a pre-existing data set that contains the relevant manipulations. The model finds support in this exploratory test.
Introduction
An area of inquiry that had made its way into the advertising literature at least by the early 1970s (e.g., Kennedy, 1971) is the effect of program involvement on viewer response to television commercials. The ensuing years have seen occasional forays into the domain of program-context effects, resulting in an accumulation of findings that creates an appreciation of the complexity of the impact of program environments on viewer responses to television advertising. In some ways, however, the body of evidence, still rather sparse, creates more questions than it answers. Even the most basic question posed by pioneering research in this area nearly thirty years ago-- whether program involvement affects viewer response to commercials positively or negatively (or randomly or not at all)--still does not have a clear answer, due to findings that are inconsistent and sometimes apparently contradictory.
The objective of this paper is to develop a model that addresses the cognitive-arousal implications of product involvement, program involvement and ad involvement on viewer processing of television commercials. In considering how consumers’ involvement in each of these areas can affect the level and type of commercial processing that is likely to occur, we hope to offer a model that helps advertisers choose appropriate vehicles to promote their products. A reanalysis of a data set from earlier published work in this area (Lord and Burnkrant, 1993) provides preliminary support for some of the propositions emanating from the model.
The following section summarizes the major findings of the program-involvement literature. A model and research propositions are then developed to account for the divergent findings of earlier studies. Tests of the propositions for which relevant data exist in the Lord and Burnkrant (1993) data set are then reported. Finally, a research agenda is proposed.
Program Involvement Literature
The most common prediction of early research in program-involvement effects was that of a negative relationship between program involvement and viewer response to commercials. This was based on an assumption that ongoing processing of a program at the time of a commercial break interferes with a consumer’s ability to attend to the ads. Between 1971 and 1981, three published studies found support for such a hypothesis (Kennedy, 1971; Bryant and Comisky, 1978; Soldow and Principe, 1981; Lord and Burnkrant, 1988a). There was, however, some disagreement between scholars. For example, Krugman (1983) argued that the most favorable environment for a television commercial is a program that engages a high level of viewer interest. The more engaging the program, he reasoned, the greater the likelihood that consumer attention will be engaged and remain focused on the screen at the time a commercial appears on the screen. His results provided support for this prediction, apparently contradicting the findings of the earlier studies.
Findings of subsequent research added to the understanding of program-context effects by adding to the list of variables that appear to affect consumer ad processing, but did not succeed in producing a consistent, unambiguous resolution to the inconsistencies of earlier studies. Lord and Burnkrant (1998a) tested the impact of program involvement across two programs (high and low involvement) for two commercials. While they claimed tentative support for a negative effect of that construct on commercial processing, that finding was significant for only one of the two ads. Murry, Lastovicka and Singh (1992) found a significant effect of program involvement on attitude toward the ad, but did not report its direction. Lord, Lee and Sauer (1994), turning their attention to radio, found an effect in the opposite direction; i.e., program involvement contributed positively to listener processing of commercials. Similarly, Lord and Putrevu (1996), in a test of viewer response to ads in four successive Superbowl broadcasts, found that ad recall increased with program-involvement levels.
Recently, a few scholars have conceptualized commercial response as a nonlinear function of program involvement (Lord, Lee and Sauer, 1994; Tavassoli, Shultz and Fitzsimmons, 1995). According to this view, television viewers are likely to be passive when program involvement is low, cognitively aroused at moderate levels of involvement, and distracted when the program is highly involving. This is based on an assumption that viewers require a modest level of stimulation to engage their attentional resources, but that excessive stimulation results in overload or confusion as to the appropriate target for the focus of attention. Tavassoli et al. (1995) found support for such a relationship.
What the literature makes clear is that program involvement is but one dimension of a complex interplay between program, ad and audience characteristics that may affect viewer response to commercials. Lord and Burnkrant (1988b) developed a model in which the influence of program involvement on commercial processing was expected to vary based on whether or not the ad contained an attention-engaging device. Their rationale was that such a device would interrupt the program-relevant thought likely to occur when program involvement is high, focusing attention on the commercial. According to this model, an adequate level of commercial processing would be most likely to occur under two conditions: (1) when program involvement is low; (2) when an ad in a high-involvement program position begins with an attention-engaging device. In 1993, the same authors published results of a study employing the variables addressed by their earlier model. By that point in time, they had departed from the predictions of the initial model and favored a more complex explanation in which commercial response was viewed as an interaction of program and ad involvement levels and the presence or absence of attention-engaging devices. Similarly, Lord and Putrevu (1996) showed that Superbowl commercial recall was a function of both program and ad involvement.
In summary, prior research provides substantial evidence that program involvement affects viewer response to television commercials, but the nature of that effect is unclear. Evidence has been advanced for both negative and positive linear relationships between program involvement and ad response, for a nonlinear relationship, and for interactions with other variables. These mixed results have arisen from a body of literature that has used widely varying research procedures and has been inconsistent in its choice of dependent variables (e.g., cognitive response, recall, attitude toward the ad, purchase intention).
Model and Propositions
The development of the model and accompanying propositions occurs in four stages: cognitive states of television viewers, antecedents of cognitive arousal, cognitive response activity, and commercial recall.
Cognitive States
An individual’s cognitive state varies as a function of a number of internal and external determinants. A viewer’s level of arousal at the time program action stops and a commercial begins provides the cognitive environment within which the advertisement is to be processed (Lord and Burnkrant, 1988a). Arousal may be low (passive cognitive state), in which case minimal thought occurs. Alternatively, viewers may be in an active cognitive state, characterized by high levels of thought. The level of arousal may become sufficiently intense to interfere with the viewer’s ability to direct cognitive resources at specified stimuli with precision. That is, a viewer confronted with multiple unrelated arousing stimuli may be unable rapidly and efficiently to engage in such cognitive operations as information rehearsal or attentional shifts (cognitive distraction).
Antecedents of Cognitive Arousal
A variety of factors may give rise to the cognitive states identified above. These include: (1) product involvement — the perceived personal relevance of the advertised product or issue to the viewer; (2) program involvement — the extent to which program content engages a viewer’s attention and interest, and hence induces elaboration; (3) ad involvement — the ad’s innate ability to engage the attention and interest of its audience.
Involvement in the advertised product or issue may arise for a variety of reasons: e.g., prior personal experience, the anticipation of future commercial-relevant behavior, or an unsatisfied need or want that has placed the viewer in a state of problem recognition. Whatever its source, recognition that the message relates to a product or issue of personal relevance to the viewer should render relevant associations in memory readily accessible to the viewer, stimulating a high level of cognitive arousal. Such associations are less likely to surface when involvement in the product or issue is low, thereby leaving the viewer in a more passive cognitive state.
A high level of program involvement may be expected to induce a high level of thought that, while it is focused on program action rather than advertising content at the time the commercial begins, leaves the viewer in a heightened state of cognitive arousal. Low program involvement, by contrast, may be expected to stimulate less program-relevant thought, with a stronger likelihood that the advertiser will find the viewer in a cognitively passive state.
An ad may also be high or low in involvement to a given viewer. Involvement may stem from the ad’s direct relevance to the viewer’s needs, wants or interests, or from the advertiser’s use of stimuli that involuntarily draw the viewer’s attention. For instance, a novel and arresting stimulus positioned at the beginning of a commercial (attention-engaging device) may reorient viewer attention and induce some ad-relevant thought. Whether ad involvement arises from the relevance of the message topic or the use of an attention-engaging device, its effect will be to increase the viewer’s overall level of cognitive arousal.
The above arguments are formalized in the following proposition:
P1: Viewers will be in a state of higher cognitive arousal when:
product involvement is high than when it is low.
program involvement is high than when it is low.
ad involvement is high than when it is low.
The absence (or low level) of these three mediators of arousal provides a television-viewing environment conducive to cognitive passivity. The presence (or high level) of any one of the three variables — product, program or ad involvement — may stimulate thought sufficiently to leave the viewer in a cognitively active state as commercial exposure occurs. However, two or more of these conditions, such as high product and high program involvement, may operate in combination with one another. This creates a situation in which multiple factors inducing cognitive activity, each with a somewhat different focus (i.e., directing viewers’ thoughts toward the advertised product or issue, the program, or the ad), compete for limited cognitive resources. The result may be a state of cognitive distraction, a situation in which the demands of the processing task may exceed cognitive capacity:
P2: Antecedents of cognitive arousal (product, program, and ad involvement), occurring in combination, will lead to cognitive distraction.
When this occurs, the source of ad involvement may influence whether viewers’ cognitive state is active or distracted. A viewer whose high product involvement encourages commercial-relevant thought, at the same time that high program involvement encourages the elaboration of program content, is in a state of cognitive distraction. If ad involvement is stimulated by the use of an attention-engaging device that is not directly relevant to the message’s central claims, the device itself may become a third stimulus competing for interpretation in an already crowded cognitive environment. This could strengthen the probability of distraction. Alternatively, ad involvement that is based on the message’s relevance to viewer needs and motivations or the use of a claim-relevant attention-engaging device, may serve to simplify rather than to complicate the viewer’s information-processing task. This could serve to reduce the likelihood of cognitive distraction. These predictions are contained in the following proposition:
P3a. A relevant source enhances message processing in a distraction-prone environment by simplifying the information-processing task.
P3b. An irrelevant source complicates message processing in a distraction-prone environment by serving as an additional stimulus.
Cognitive Response Activity
The cognitive state of a television viewer appears to influence viewer’s commercial-relevant thoughts (cognitive-response activity). As any advertiser recognizes, a consumer’s primary purposes in television viewing are entertainment and relaxation, not learning about products by carefully processing commercial information. Thus, Howard (1977) notes that viewers whose program involvement is low passively watch commercials as part of the entertainment. Within the framework of the proposed model, low program involvement, if not attenuated by other arousal factors, would leave the viewer in a passive cognitive state. Howard’s theorizing would seem to predict that low program elaboration would lead to relatively few commercial-relevant cognitive responses. Alternatively, television viewers in a cognitively active (but not distracted) state, induced either through high product, program, or ad involvement, should demonstrate relatively higher levels of cognitive-response activity. As cognitive distraction occurs, however, commercial-relevant cognitive-response levels are expected to drop below those observed in cognitively active states. Therefore, a combination of the antecedents of cognitive arousal identified earlier, attracting viewer thought in a variety of directions, renders viewers less able to generate commercial-relevant thoughts. The following proposition captures these expectations:
P4: The extent of commercial-relevant cognitive response will depend upon the viewer’s level of cognitive arousal, with viewers in a state of cognitive arousal generating more commercial-relevant cognitive responses than those in a state of cognitive passivity or distraction.
Additionally, the nature of the thoughts generated may vary as a function of the viewer’s cognitive state. Cognitive responses may derive directly from message content or from information stored in memory that is activated through the elaboration process upon program/commercial exposure. For purposes of this discussion, those that clearly stem from the message itself are called "message thoughts," while responses departing sufficiently from message content to indicate the tapping of other stored information are referred to as "own thoughts." The relatively few commercial-relevant cognitive responses emerging in passive and distracted states are likely to reflect a stronger tendency toward message thoughts. Cognitively aroused viewers, on the other hand, will be more likely to engage in the depth and type of elaboration that favors own-thought production. This is because own thoughts result from the association of stimulus details with pertinent information, images or impressions stored in long-term memory. Message thoughts, however, may arise from the relatively simpler activity of cognitive rehearsal of message content. Simple cognitive operations can be executed by viewers in a relatively passive state, but a heightened level of arousal appears to promote more complex processing:
P5: The nature of commercial-relevant cognitive response will depend upon the viewer’s level of cognitive arousal, with those in a state of cognitive arousal producing more "own thoughts," and those in a state of cognitive passivity or distraction generating more "message thoughts".
Recall
The types of thoughts generated in varying cognitive states have the potential to mediate the types of message elements recalled. The generation of message thoughts implies that viewers have engaged to some extent in the rehearsal of message content. Alternatively, own thoughts represent the retrieval of personal experiences, associations, and images, and may well interfere with the rehearsal of an ad’s specific informational details. As central message elements face such interference from the own thoughts of cognitively active viewers, peripheral cues (e.g., spokesperson, music, colors) may be more likely to intrude upon these reflections and enter long-term memory:
P6: Recall of central message arguments will be higher among viewers whose cognitive responses are comprised primarily of message thoughts, while recall of peripheral cues will be greater for those generating primarily own thoughts.
The model and propositions are depicted graphically in Figure 1.
Figure 1: Cognitive Arousal Model
Empirical Evidence
To provide some preliminary evidence in support of the above propositions, we use the data set described as Experiment 1 in Lord and Burnkrant (1993). The design includes manipulations of program involvement (high and low) and attention-engaging devices (present and absent) for two commercials about products/issues. One is for a product shown to be low in involvement for the sample (Mobil 1 motor oil), while the other is about an issue the sample finds to be highly involving (drinking and driving). The attention-engaging device employed in the device-present condition of the commercial for the high-involvement issue was relevant to the message’s central message arguments. That used for the ad for the low-involvement issue, however, was an irrelevant device (a sequence of loud beeps designed to attract attention to the screen, but not necessarily to prime the viewer to think about the topic of the ensuing message). For purposes of the present investigation, ad involvement is considered to be low for the low-involvement product without an attention-engaging device, high but "irrelevant" for the low-involvement product with the irrelevant device, and both high and "relevant" for the high-involvement product with the relevant attention-engaging device. The sample included 264 student subjects. Dependent variables included measures of the cognitive response and recall constructs addressed in the above propositions.
We have divided cognitive responses into message- and own-thought categories and analyzed recall results based on the central/peripheral breakdown required by the propositions. Two judges blind to the experimental condition coded cognitive-response protocols, categorizing each listed thought as representing "message" or "own" thought. The correlation between the two judges with respect to the number of thoughts categorized as "message" (r = .90) and "own" (r = .82) indicates adequate inter-judge reliability. The mean of the two judges’ tallies of the number of responses in each category served as the criterion measure for subsequent analysis. A comparable approach was used in assigning responses to a free-recall measure (subjects were asked to write down everything they could remember about the test ad) to central and peripheral categories, again demonstrating adequate reliability between judges (r = .90 and .88, respectively). All correlations were significant (p < .01).
The predictions that low program involvement would be associated with a low level of cognitive arousal (P1), and that cognitive passivity would induce a low level of cognitive-response production (P4), enjoy directional but not significant support. For both ads, subjects generated fewer commercial-relevant thoughts when program involvement was low (Mobil 1 ad, mean 4.76; drinking-and-driving ad, mean 5.89) than when it was high (4.99 and 6.47 for Mobil 1 and drinking/driving, respectively; both p > .10). Significant interactions between program and ad involvement for both ads (Mobil 1: F = 12.50, p < .01; drinking/driving: F = 4.55, p < .05) show a pattern of results more supportive of the model’s predictions for cognitive arousal and thought production (P1, P3, P4). Relatively high levels of commercial-relevant thought emerged under conditions expected to lead to a state of cognitive arousal (but not distraction) — e.g., high program involvement/low product and ad involvement (mean 5.49) or high product involvement and high/relevant ad involvement (mean 6.96). The distraction phenomenon appears to have resulted as predicted (P2, P3, P4), with subjects exposed to an advertisement in a state of high program involvement generating significantly more commercial-relevant thoughts in the absence of an irrelevant attention-engaging device than in its presence (means of 5.49 and 4.49, respectively, p < .05).
An analysis of the types of thoughts generated, conducted as a test of P5, reveals that own thoughts dominated in all conditions. However, the greatest elicitation of message thoughts arose under conditions of cognitive passivity (all types of involvement low — mean .90) and distraction (high program involvement, high irrelevant ad involvement — mean .81). Means for other conditions were less than .65 (relevant contrasts significant at p < .01 for drinking and driving, only at p < .10 for Mobil 1). Program involvement/ad involvement interactions were significant for both ads (Mobil 1: F = 12.02, p < .01; drinking/driving: F = 3.58, p < .01). As predicted, conditions expected to be associated with a high level of cognitive arousal evoked higher levels of own-thought production than those characterized by passivity or distraction. Thus, viewers of the low-involvement Mobil 1 ad generated the highest number of own thoughts (mean 4.80) when program involvement was high. The number of own thoughts was greatest among those viewing the high-involvement drinking-and-driving commercial when program involvement was high and the relevant attention-engaging device directed attentional focus to the topic of the ad (mean 6.45). The pattern of results for message and own thoughts is therefore consistent with the model.
Recall results provided some support for the model’s expectation that recall of central message arguments will be greater at lower levels of own-thought production, with peripheral-cue recall prevailing when a high level of cognitive arousal causes viewers to engage in extensive elaboration. Viewers exposed to the Mobil 1 ad in the low-involvement program position (presumably in a cognitive passive state) recalled more central message arguments than did those seeing the same ad in the high program-involvement condition (means of 7.17 and 6.27, respectively; F = 4.85, p < .05). A significant interaction (F = 4.09, p < .05) reveals that the attention-engaging device that facilitated own-thought production among drinking-and-driving ad viewers when program involvement was high detracted from their recall of central message arguments (means of 7.19 and 5.79 when the device was absent and present, respectively, p < .05). Correspondingly, the presence of the device resulted in the recall of more peripheral cues (means of 3.29 and 2.50 when the device was present and absent, respectively; F = 4.04, p < .05).
Discussion
The results provide a plausible explanation of the way in which program, ad and audience characteristics interact to affect viewer response to commercial messages. However, it falls to future research, using a design, sample and stimuli better equipped to comprehensively test the model, to fully establish its validity, determine its generalizability, and identify its boundary conditions. Specifically, the model should be tested using a full-factorial design incorporating validated manipulations of all constructs. Testing should also employ a wider array of types of programs, ads and attention-engaging devices, and a sample that is representative of viewing audiences.
The model makes a contribution from both a conceptual and a strategic point of view. Conceptually, the model’s use of cognitive arousal as an explanatory variable allows a more comprehensive explanation of the impact of program environments on commercial processing and a basis for explaining some discrepant results in prior studies. Viewing cognitive activity as spanning a continuum from passivity to arousal to distraction is consistent with the nonlinear explanations of program-involvement effects mentioned earlier. But the model goes further to identify determinants of the viewer’s cognitive state and the way in which they interact to affect cognitive response to and recall of advertisements. The model suggests at least two possible explanations for the inconsistent results in this area. First, the way in which program involvement affects recall will depend on whether product/issue involvement and ad involvement exist at a level likely to enhance arousal and attentional focus or to create confusion and distraction. Since these other variables, seldom controlled in prior research, are likely to vary from situation to situation, different results for program involvement should be expected. Second, higher-order responses (e.g., attitude and purchase intention) are likely to depend on lower-order cognitive responses. Thus, an effective prediction or explanation of such common dependent variables as recall and attitude requires an understanding of the level and type of cognitive response generated upon ad exposure.
A common approach to the study of advertising effectiveness that illustrates the importance of this broader view of commercial response involves the use of recall as the sole dependent variable. A natural assumption would be that ads inducing low recall of central message claims are ineffective. The model and results presented in this study imply, however, that low recall of message arguments does not always reflect a low level of commercial processing. When that diminished recall is a result of the viewer’s "own thoughts" (commercial-relevant elaborations), what is remembered may be of greater value in inducing the desired behavior than the ability to restate specific message claims from memory. Without attention to the states of cognitive activity and their antecedents and consequences, such a possibility could be overlooked.
Strategically, the model suggests a need for advertisers carefully to weigh their objectives with respect to amount and type of message processing they desire of the viewing audience. The selection of an appropriate advertising strategy requires an initial assessment of whether extensive personal elaboration of the message topic — potentially at the expense of high recall of central message arguments — or extensive recall of specific commercial details is most likely to satisfy the advertiser’s objectives.
If consumer learning is a major objective of an advertising campaign, a commercial should be placed in an environment conducive to the recall of central message arguments. When product/issue involvement within the target audience is low, placement of commercials in a program position characterized by low involvement appears to provide such a setting. If, on the other hand, high program involvement is considered likely, opening the message with a dramatic attention-engaging device may provide sufficient distraction to suppress the personal associations (own thoughts) that may otherwise reduce the processing of specific message details. Facilitating learning for viewers high in product/issue involvement may also require the suppression of own thoughts via an attention-engaging stimulus.
Many advertisements are low in complexity, conveying only one or two significant points of information. Simple brand recognition, rather than extensive learning, may be a sufficient outcome for such ads. In such a scenario, the viewer’s elaborative thoughts may be of greater value than extensive central recall. Under such circumstances, an active but undistracted cognitive state may be desirable. If initial product involvement is low, high program involvement may serve to induce the desired state of cognitive activity. Alternatively, an attention-engaging device may perform the same function if program involvement is low. To the extent possible, however, it would be desirable to avoid the combination of multiple antecedents of cognitive arousal, since they may interact to distract viewers from commercial-relevant thought. When product or issue involvement is high, a low-involvement program position appears to be desirable. If the program environment is likely to elicit high levels of viewer elaboration, a device that immediately, dramatically and unambiguously calls attention to the product or issue in which the viewer is already highly involved may alleviate the program’s distraction potential.
References
Bryant, J., & Comisky, P. W. (1978). The effect of positioning a message within differentially cognitively involving portions of a television segment on recall of the message. Human Communication Research, 5 (Fall), 63-75.
Howard, J. A. (1977). Consumer behavior: Application of theory. New York: McGraw-Hill.
Kennedy, J. R. (1971). How program environment affects TV commercials. Journal of Advertising Research, 11 (February), 33-38.
Krugman, H. E. (1983). Television program interest and commercial interruption: Are commercials on interesting programs less effective? Journal of Advertising Research, 23 (February/March), 21-23.
Lord, K. R., & Burnkrant, R. E. (1988). Television program elaboration effects on commercial processing. In M. J. Houston, (Ed.). Advances in consumer research, 15 (pp. 213-218). Provo, UT: Association for Consumer Research.
Lord, K. R., & Burnkrant, R. E. (1988). Commercial processing efficiency: A program elaboration model. In G. Frazier, C. Ingene, D. Aaker, A. Ghosh, T. Kinnear, S. Levy, R. Staelin & J. Summers (Eds.). 1988 AMA educators’ proceedings: Efficiency and effectiveness in marketing (pp. 112-116). Chicago: American Marketing Association.
Lord, K. R., & Burnkrant, R. E. (1993). Attention versus distraction: The interactive effect of program involvement and attentional devices on commercial processing. Journal of Advertising, 22 (1), 47-60.
Lord, K. R., Lee, M. S., & Sauer, P. L. (1994). Program context antecedents of attitude toward radio commercials. Journal of the Academy of Marketing Science, 22 (1), 3-15.
Lord, K. R., & Putrevu, S. (1996). Superbowl ad recall: Program and ad involvement effects. In P. M. Herr and J. Kim (Eds.) Proceedings of the society for consumer psychology (pp. 1-7). Boulder, CO: University of Colorado Graduate School of Business.
Murry, J. P. Jr., Lastovicka, J. L., & Singh, S N. (1992). Feeling and liking responses to television programs: An examination of two explanations for media-context effects. Journal of Consumer Research, 18 (4), 441-451.
Soldow, G., & Principe, V. (1981). Response to commercials as a function of program context. Journal of Advertising Research, 21 (April), 59-65.
Tavassoli, N. T., Schultz, C. J. II, & Fitzsimmons, G. J. (1995). Program involvement: Are moderate levels best for ad memory and attitude toward the ad? Journal of Advertising Research, 35 (September/October), 61-72.