A Study of Selected Issues in Vividness Reserach: the Role of Attention and Elaboration Enhancing Cues

Craig A. Kelley, California State University
ABSTRACT - Research on the vividness effect has been controversial. The purpose of this study is to address some of the methodological and theoretical issues that exist in previous research. Methodologically, a multiple-item vividness scale was developed to serve as a manipulation check and a within-subjects design was used to manipulate differential attention. Theoretically, an additional test of the availability-valence hypothesis was conducted. The results indicated that a vividness effect does exist.
[ to cite ]:
Craig A. Kelley (1989) ,"A Study of Selected Issues in Vividness Reserach: the Role of Attention and Elaboration Enhancing Cues", in NA - Advances in Consumer Research Volume 16, eds. Thomas K. Srull, Provo, UT : Association for Consumer Research, Pages: 574-580.

Advances in Consumer Research Volume 16, 1989      Pages 574-580

A STUDY OF SELECTED ISSUES IN VIVIDNESS RESERACH: THE ROLE OF ATTENTION AND ELABORATION ENHANCING CUES

Craig A. Kelley, California State University

ABSTRACT -

Research on the vividness effect has been controversial. The purpose of this study is to address some of the methodological and theoretical issues that exist in previous research. Methodologically, a multiple-item vividness scale was developed to serve as a manipulation check and a within-subjects design was used to manipulate differential attention. Theoretically, an additional test of the availability-valence hypothesis was conducted. The results indicated that a vividness effect does exist.

It has been argued repeatedly in the literature that a communication message is more memorable and persuasive if vivid stimuli are included in the message (e.g., Dickson 1982; Kisielius and Sternthal 1984, 1986; Nisbett and Ross 1980; Taylor and Thompson 1982). However, the premise that vivid stimuli have a greater impact on memory than pallid stimuli has not always been supported by empirical research. Kisielius and Sternthal (1986) and Taylor and Thompson (1982) have advanced several explanations for the lack of many studies' success in demonstrating the vividness effect. These reasons have both methodological and theoretical roots.

Taylor and Thompson (1982) argued that two major methodological problems may be contributing reasons why the vividness effect has not been found in many studies. The first problem is that many studies lack a manipulation check. The second problem involves the way that vivid and pallid stimuli are presented to the subject. The subject may process messages with vivid stimuli only if the stimuli compete with pallid stimuli for the subject's limited memory resources.

From a theoretical perspective, Kisielius and Sternthal (1986) suggested that the presence or absence of significant research results could be explained by the availability-valence hypothesis. According to this hypothesis, the availability and favorableness of information stored in memory determines the subsequent decision. The reason messages containing vivid stimuli are more memorable and persuasive is that they cause the individual to cognitively elaborate on the message information. In addition, the favorableness of the available information determines the direction of the decision.

The purpose of the present study is to report the results of a series of experiments which addressed the methodological and theoretical issues mentioned above. Specifically, a reliable multiple-item vividness scale will be developed to serve as a manipulation check and differential attention will be manipulated with a within-subjects design. In addition, the availability-valence hypothesis of the vividness effect will be tested.

METHODOLOGICAL ISSUES IN VIVIDNESS RESEARCH

The absence of a reliable measure of the vividness construct is a common problem found in many investigations of the vividness effect (Taylor and Thompson 1982). Without a reliable measure of vividness there is no way to determine if experiments, which have reportedly manipulated messages containing vivid and pallid stimuli, have actually done so (Kisielius and Sternthal 1984; Reyes, Thompson, and Bower 1980). A study by Dickson (1982) did contain a manipulation check. However, the scale used in this study consisted of a single-item, seven-point Likert-type scale which does not allow for an assessment of internal reliability.

Another methodological problem encountered in research of the vividness effect is the almost exclusive use of between-subjects experimental designs. Taylor and Thompson (1982) argue that a vividness effect may only occur if there is competition among stimuli for the subject's attention. This would entail the use of a within-subjects experimental design. In a between-subjects design, the subjects would process the stimuli, whether vivid or pallid, that were presented to them and no difference in elaboration of the information would be found.

A study by Reyes, Thompson, and Bower (1980) used a within-subjects design by pairing nine vivid (or pallid) prosecution arguments with nine pallid (or vivid) defense arguments to determine the relative influence of each set of arguments had on the subjects' judgments of guilt or innocence. The vivid arguments did not have an immediate persuasive effect, but did have a delayed effect. However, a manipulation check was not included in the study and no attempt to control for the equivalency of the information contained in the sets of vivid and pallid prosecution and defense arguments. Nevertheless, the study did illustrate that the vividness effect was found when differential attention was manipulated in a within-subjects design.

THEORETICAL ISSUES IN VIVIDNESS RESEARCH

The encoding and the availability-valence hypothesis have been advanced to explain why messages that contain vivid stimuli are more persuasive and memorable when compared to messages that contain pallid stimuli. Each of these will be discussed in turn and then the differences reconciled.

Encoding Explanation

The encoding explanation posits that the amount of information that is encoded determines whether it will be readily available in memory when decisions are made (Tversky and Kahneman 1973). In essence, an individual has a selection bias when message information is encoded. Vivid stimuli are more available in memory and have a greater influence on decisions because they are attended to and encoded in greater numbers than pallid stimuli (Nisbett and Ross 1980; Taylor and Fiske 1978).

There has been considerable speculation as to why more vivid stimuli might be encoded. Forgas (1983) suggested that vivid stimuli are encoded because they are novel. Paivio (1971) argued that vivid stimuli are more imaginable, and therefore, selected for encoding in both verbal and image form. Finally, vivid stimuli may be selected for encoding because they are emotionally involving (Nisbett and Ross 1980).

Little evidence exists to support the encoding explanation. Dickson (1982) primed his subjects by asking them to imagine the consequences of a sudden failure of their refrigerator and then asked them to read a case history containing either concrete or statistical information. The vivid, concretely-written case history did not have a greater effect on the primed subjects' judgments, as expected. Fiske et al. (1979) instructed their subjects to first imagine one of three character perspectives and then to read a story involving the three characters. They found that the subjects' judgments were not influenced by imagining the character's perspective. Manis et al. (1980) and Purdy and Luepritz (1982) did not find support for the hypothesis that decisions are influenced by encoding vivid stimuli in verbal and image form. Finally, Borgida and Nisbett (1977) found that emotional interactions between students had a stronger impact on an evaluation of a college psychology class than the exchange of pallid, statistical data.

Availability-valence Hypothesis

Kisielius and Sternthal (1986) suggested the encoding explanation did not adequately explain the effect of vivid stimuli on decisions. Alternatively, they argued that the mixed empirical results of previous studies of the vividness effect could be explained with the availability-valence hypothesis. According to this hypothesis, the information that is most available in memory will be used to make decisions. Subsequent decisions depend on how favorable the stored information is perceived by the individual.

Cognitive elaboration is thought to be an important factor in determining which information will become available in memory. The prediction behind the vividness effect is that vivid stimuli cause an individual to engage in a greater amount of cognitive elaboration. When a decision is made, the vivid stimuli would be more available because more pathways would be involved in the retrieval of information.

Once activated, how the available information affects a decision depends on the favorableness of the information. Kisielius and Sternthal (1986) suggested that vivid stimuli contained in a communication will enhance persuasion if they activate associations that are similar in meaning. Pallid stimuli are believed to inhibit elaboration, and therefore, if the communication is similar in its meaning, the pallid information will not be as persuasive. Conversely, vivid stimuli may cause a communication to be less persuasive if they activate an individual's idiosyncratic associations which tend to be less favorable towards the position being advocated in the communication. Pallid stimuli would be more persuasive because they limit idiosyncratic associations and the individual relies on the communication, which is designed to persuade, to make a decision.

Little research has tested the availability-valence hypothesis. When Reyes et al., (1980) found a vividness effect after a forty-eight hour delay, it could have been due to the time it takes for the individual to elaborate on the information. Edell and Staelin (1983) did not find a vividness effect when subjects were allowed to elaborate on vivid, pictorial advertisements versus pallid, verbal advertisements. However, Kisielius and Sternthal (1984) did find a vividness effect when they manipulated cognitive elaboration and the favorableness of the message information.

The two theoretical explanations for the vividness effect can be reconciled. Before information is processed it first must attract the attention of the individual processor, Once attention is focused on stimuli (whether it is vivid or pallid), the individual elaborates the stimuli. Therefore, vivid stimuli contain two properties that make them more memorable. Attentional cues direct the processor's attention away from the other information contained in a message and elaborative cues that facilitate cognitive elaboration such that the vivid stimuli are more available for making decisions.

THE STUDY

The present study addressed the major issues highlighted previously. The first experiment was designed to determine if vivid stimuli attract more attention than pallid stimuli. If attention is directed toward vivid stimuli in a communication, the next question is whether the inclusion of attentional stimuli increases the comprehension of a message over verbal vivid stimuli by themselves. Finally, the last experiment sought to determine if vivid stimuli are elaborated to a greater extent than pallid stimuli.

EXPERIMENT I

The purposes of Experiment I were to develop an internally reliable measure of the vividness construct and to determine if vivid stimuli attract more attention. The context in which the vivid and pallid stimuli were manipulated was a warning message embedded in a product label. An incongruency between current beliefs about the product class and the warning was necessary to prevent the subjects from responding to the dependent variable measures with previously stored knowledge. It was also important that a neutral experimental product be selected to prevent the gender of the subjects from influencing their responses (Kisielius and Sternthal 1984). Extensive pretesting indicated that a fluoride dental rinse was viewed by the subject population as safe and that a product warning cautioning against swallowing such a rinse was inconsistent with this belief.

Vividness Scale

Extensive pretesing was then conducted to develop a multiple-item scale to serve as a manipulation check. The end result of the pretesting was an eight-item vividness scale. The items were seven point Likert-type measures where each of the seven points were anchored. For example, the end anchors of one item were very confusing and very clear. Another item end anchors were not very descriptive and very descriptive. An overall vividness score was calculated by summing the item values. The range of possible scores was 8 (pallid) to 56 (vivid). A confirmatory factor analysis indicated the eight-item scale was unidimensional.

Stimulus Materials

After the multiple-item vividness scale had been developed, the vivid and pallid warning messages for the experimental product were developed. Before the warnings were used in the experiment, three expert judges used the legal criteria of an adequate warning to determine whether the two warnings were equivalent in meaning. The warning messages were equivalent if they: (1) described the nature of the risk, (2) identified the magnitude of the risk, and (3) described how to avoid the risk. The judges were asked to respond to a short questionnaire which was designed to interpret the warning messages relative to the above criteria. The judges only disagreed on one question and it was decided by a majority vote (Brucks, Mitchell and Staelin 1984).

Color, size of print, and location of the warning message on a label have all been used by manufacturers to attract attention to a warning. The inclusion of a symbol or a picture in the warning is also used to increase the amount of attention paid to a warning. In addition, a contrasting border was used in previous research as a means of enhancing the attentional qualities of a warning (Funkhouser 1986). From these examples, a warning symbol (i.e., a circle with a slash across it, superimposed on the word swallow) and a red contrasting border drawn around the warning message were selected lo operationalize the vivid warning message. The verbal portion of the warning message was written in concrete language. The pallid warning message did not include the attractive stimuli and the verbal portion of the warning was written in abstract language. The label contained a fictitious brand name, the warning and other product claims. Approximately 25% of the label was devoted to the warning.

Sample

The sample consisted of 25 volunteer subjects selected from the student population enrolled in marketing courses at a western university. The subjects participated in the experiment individually. Approximately equal numbers of men and women participated in the experiment.

Procedure

Two levels of vividness were manipulated in a within-subjects design. The subject entered a room where two slide projects were aimed at each of two different projection screens which faced each other. The subject was seated between the two screens.

A different slide was flashed on each of the screens simultaneously for a period of ten seconds. The first set of slides served to orient the subject to the experimental task. One slide featured a drawing of a tree and the other displayed the word "tree." The subject then responded to the following question: "If something caught you attention, please describe what it was." The responses to the orientation slides were not included in the analysis.

The second set of slides contained the experimental product labels with the vivid or pallid warning message embedded in them. The information on the labels was the same except for the warning. The slides were shown for ten seconds and then both screens were darkened. Pretesting indicated that ten seconds was enough time to read only a portion of the two product labels. Since time was a limiting factor, the portion of the labels which was read was presumed to be the most attractive part of the label.

The subject responded to the same questions as mentioned above. Again, the subject was shown slides of the experimental product labels, this time for one minute. This allowed the subject to read a larger portion of the two warning messages. Then the following questions were asked: 'Which slide could you select to read?" and "Why would you select this slide over the other one?"

Once these questions were answered, the slides containing the experimental product labels were shown again, one at a time, and the manipulation check for vividness was administered. The product label that was viewed first was alternated from subject to subject. The subject was then debriefed. After the subject left the room, the slides in the right projector were exchanged with those in the left projector. This was done so reduce a confounding effect that might result if the subjects had the tendency to lean toward the right or left side. Then the next subject entered the room to start the experiment. The experimental sessions lasted approximately ten minutes each.

Results

The scores of each item of the eight-item vividness scale were summed to get a total vividness score. A dependent t test (t = 12.11; df = 24; p < 0.001) indicated that the vivid warning was significantly more vivid than the pallid warning. A Cronbach coefficient alpha of .94 indicated that the eight-item measure was internally reliable.

All of the subjects stated that something in the two slides caught their attention. In responding to the first open-ended question, 24 of the subjects indicated that it was one or more of the stimuli embedded in the vivid warning message that caught their attention. All 25 subjects responded that they would select to read the label with the vivid warning message if they had to make a choice.

The second open-ended question attempted to determine why the label with the vivid warning was selected over the label with the pallid warning. In general, the subjects indicated that the vivid warning was more attractive and clearer than its pallid counterpart.

There are two major implications of the results of Experiment I. First, a reliable multiple-item measure that could be used as a manipulation check of the vividness construct was developed. This multiple-item scale distinguishes the present study from previous published research on the vividness effect since manipulation checks were lacking in many of these studies. Second, the results supported Taylor and Thompson's (1982) argument that the vividness effect might occur if the vivid stimuli have a chance to compete with pallid stimuli for the subjects' attention.

EXPERIMENT II

The purpose of Experiment II was to determine whether the inclusion of the attentional stimuli (i.e., the red border and warning symbol) communicated more information than the verbal message by itself. If there was no additional information communicated, a test of whether vivid stimuli are elaborated to a greater extent than pallid stimuli could be conducted.

Stimulus Materials

Two levels of border (present and absent), two levels of symbol (present and absent), and two levels of verbal message (concrete and abstract) were manipulated in a 2 x 2 x 2 factorial design. The dependent variable, comprehension, was defined as what the warning message said (Funkhouser 1984). Comprehension was measured by questions with "yes, "no," and "do not remember" responses. The "do not remember" response was included to control for a response bias that might occur if the subject was forced to judge the accuracy of any particular question (Ford and Yalch 1982).

Ten comprehension questions were asked. Six of the ten questions were related to the warning message. Of the six questions, there were three "yes' and three "no" correct answers. This was done to control for a response bias that might occur if unequal numbers of "yes" and "no" questions are asked (Ford and Yalch 1982). The other four questions were related to information not contained in the warning and were included to keep the subjects from guessing the purpose of the experiment.

Sample

One hundred sixty-four students enrolled in marketing classes at a western university constituted the sample. The subjects were run in small groups and were randomly assigned to the treatments.

Procedure

The subjects were given 45 seconds to read the label. Forty-five seconds was the average amount of time the subjects in the pretest need to read the label. After time was called, the subjects turned the page and responded to the 10 comprehension questions. The subjects were instructed not to turn back to the label when answering the questions. Next, the experimental materials were collected and the subjects debriefed.

The six comprehension questions were categorized as "correct," "incorrect," and "do not remember." The number of correct responses was adjusted for guessing by subtracting the number of incorrect responses from the number of correct responses. This difference was then converted to a proportion by dividing by the total possible correct responses. The result was to transform the raw data to a continuum of proportions ranging from -1 to +1 (Sax 1980).

Results

Table 1 shows the ANOVA results. There were no significant differences-between the treatments (P = 1.690; p < 0.115). Therefore, the two warning treatments used in Experiment I were selected to explore the final issues of interest in the study.

EXPERIMENT III

The purpose of Experiment II was to determine if vivid stimuli are elaborated to a greater extend than pallid stimuli, and if so, whether elaboration makes a difference in comprehending the warning-message.

Stimulus Materials

The same labels as used in Experiment I were used. In addition, the same ten comprehension questions used in Experiment II were again used in Experiment m. Three neutral nonexperimental product labels were developed and included in the experiment to disguise the purpose of the experiment. The labels contained a pallid warning message and were similar in design as the experimental labels. Fictitious brand names were developed and the warning was placed in the same place on the label. The nature of the claims that were made on the labels were similar, but the wording was changed to reflect the different products. Four product labels were used in the experiment because too few labels would make the vivid warning too conspicuous while too many labels would cause the subjects to suffer fatigue with the experimental task. Ten comprehension questions were also developed for each of these nonexperimental labels. However, the answers to these questions were not included in the analysis.

Subjects

Sixty students enrolled in marketing classes at a western university volunteered to participate in the experiment. Thirty subjects were randomly assigned to each treatment. To increase the amount of experimental control, the subjects participated in the experiment in groups of 5 or less.

Procedure

The three nonexperimental product labels and either the vivid or pallid experimental product label were randomly placed in a booklet to control for order effects. The subjects were given 45 seconds to read the first label. After time was called, the subjects turned the page and were given another 45 seconds to read the second label. This process continued until the fourth label had been read. Then the booklet containing the labels was collected and a second booklet containing the measures of comprehension of the experimental and nonexperimental labels was distributed. After the subjects had completed their responses, they were instructed to return two days later.

TABLE 1

EXPERIMENT II ANOVA RESULTS

The second session began by having the subjects respond to the same ten comprehension questions related to the experimental product as asked in the first session. These questions were followed by Marks's (1973) Visual Vividness Imagery Questionnaire. This measure was included to serve as a covariate to control for differences in the subjects' ability to imagine (Childers and Houston 1984) and has been judged to be reliable and valid (White, Sheehan, and Ashton 1977). Next, the subjects were exposed to the same experimental product label that they had in the first session and then were asked to complete the eight-item vividness scale developed for Experiment I. The manipulation of vividness had to be taken at the end of the second session to prevent the subjects from guessing the purpose of the study. Finally, the subjects were debriefed.

Results

The vividness scale responses were summed and an independent t-test confirmed that the vivid experimental warning and pallid experimental warning were significantly different (t = 7.69; df = 58; p < 0.001). The Cronbach coefficient value for the scale was .83. Therefore, the vividness manipulation was successful.

The comprehension questions related to the experimental product in each session were categorized as "correct," "incorrect," and "do not remember." The individual subject's scores were then adjusted for guessing and converted to a proportion as was done in Experiment II. A repeated measures ANCOVA was then performed and the results are presented in Table 2.

A significant vividness and time main effect were found. These results indicate that a vividness effect occurred both in the immediate and delayed situation. A vividness effect in the immediate case indicates that vivid stimuli that are incongruent with existing knowledge is processed and elaborated to a greater extent than pallid stimuli. In the delayed situation, because the vivid stimuli had caused more elaboration in the first session, the vivid stimuli were then more available in memory for making decisions. Thus, both results are consistent with the availability- valence hypothesis.

DISCUSSION

The series of experiments conducted in this study address many issues of vividness research. From a methodological perspective, a vividness effect was found when differential attention was manipulated. This suggests that attentional cues may be a necessary factor in making information vivid. In addition, the experiments included a multiple-item scale as a manipulation check.

Experiment II determined that attentional cues used to make a message vivid are not sufficient to cause a difference in cognitive elaboration. A verbal component to the warning may be necessary to stimulate cognitive elaboration and possibly yield greater comprehension. Experiment m demonstrated that elaboration is stimulated by vivid stimuli embedded in verbal aspects of a message. Therefore, vivid stimuli appear to require attentional and elaboration (i.e., verbal) cues to make a difference in making decisions.

These experiments did not manipulate the favorableness of the warning message. As Kisielius and Sternthal (1986) state, the favorableness of the information may increase, decrease, or have no effect on making decisions. Future experiments should also focus on the confounding effect that may have been introduced by using attention and verbal cues to create the vivid and pallid experimental manipulations. The inclusion of both types of cues may have altered the processing of the experimental warning message, thus preventing stronger conclusions from being drawn concerning the availability-valence hypothesis.

TABLE 2

EXPERIMENT III ANCOVA RESULTS

REFERENCES

Borgida, Eugene and Richard E. Nisbett (1977), "The Differential Impact of Abstract vs Concrete Information on Decisions," Journal of Applied Social Psychology, 7, 258-271.

Brucks, Michael, Andrew Mitchell, and Richard Staelin (1984), 'The Effect of Nutritional Information Disclosure in Advertising: An Information Processing Approach," Journal of Public Policy and Marketing, 3, 1-25.

Childers, Terry and Michael Houston (1984), "Conditions for a Picture Superiority Effect on Consumer Memory," Journal of Consumer Research, 11 (September), 643-654.

Dickson, Peter R. (1982), "The Impact of Encoding Case and Statistical Information on Consumer Judgments," Journal of Consumer Research, 8 (March), 398-406.

Edell, Julie and Richard Staelin (1983), "The Information Processing of Pictures in Print Advertisements," Journal of Consumer Research, 10 (June), 45-61.

Fiske, Susan, Shelly Taylor, Nancy Etcoff, and Julie Laufer (1979), "Imaging, Empathy and Causal Attributions," Journal of Experimental Social Psychology, 15, 356-372.

Ford, Gary T. and Richard Yalch (1982), "Viewer Miscomprehension of Televised Communication A Comment," Journal of Marketing, 46 (Fall), 37-31.

Forgas, Joseph (1983), "The Effects of Prototypicality and Cultural Salience on Perceptions of People," Journal of Research in Personality 17 (June), 153-183.

Funkhouser, G. Ray (1984), "An Empirical Study of Consumer's Sensitivity to the Wording of Affirmative Disclosure Messages," Journal of Public Policy and Marketing, 3, 26-37.

Kisielius, Jolita and Brian Sternthal (1986), "Examining the Vividness Controversy: An Availability-Valence Interpretation," Journal of Consumer Research, 12 (March), 418-431.

Kisielius, Jolita and Brian Sternthal (1984), "Detecting and Explaining Vividness Effects in Attitudinal Judgments," Journal of Marketing Research, 21 (February), 54-64.

Manis, M., J. Dovalina, N. Avis, and S. Cardoze (1980), "Base Rates Can Affect Individual Predictions," Journal of Personality and Social Psychology, 38, 231-248.

Marks, David (1973), "Visual Imagery Differences in the Recall of Pictures," British Journal of Psychology, 64, 11-24.

Nisbett, Richard and L. Ross (1980), Human Inference: Strategies and Shortcomings of Social Judgment, Englewood Cliffs, NJ: Prentice Hall.

Paivio, Allan (1971), Imagery and Verbal Processes, New York: Holt, Reinhart and Winston.

Purdy, Jesse and Roy Luepuitz (1982), "Immediate and Long-term Retention for Pictorial and Verbal Stimuli," Perceptual and Motor Skills, 55 (December), 1079-1082.

Reyes, Robert, William Thompson, and Gordon Bower (1980), "Judgmental Biases Resulting from Differing Availability Arguments," Journal of Personality and Social Psychology, 39 (July), 2-12.

Sax, George (1980), Principles of Educational and Psychological Measurement and Evaluation, Belmont, CA: Wadsworth Publishing.

Taylor, Shelly E. and Susan T. Fiske (1978), "Salience, Attention, and Attribution: Top of the Head Pherlomena," in Advances in Experimental Social Psychology, vol. 11, ed. by L. Berkowitz, New York: Academic Press.

Taylor, Shelly E. and Susan Thompson (1982), "Stalking the Elusive Vividness Effect," Psychological Review, 89 (March), 155-181.

Tversky, A. and David Kahneman (1973), "Availability: A Heuristic for Judging Frequency and Probability," Cognitive Psychology, 5, 207-232.

White, K., P. Sheehan, and R. Ashton (1977), "Imagery Assessment: A Survey of Self-Report Measures," Journal of Mental Imagery, 1, 145-170.

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