The effects of brand choice on product trial evaluations.

ABSTRACT

Past research suggests that consumers who have greater perceived control over a situation are more likely to evaluate the situation in a favorable manner. In the current study, this theoretical prediction is extended to the domain of product trial and brand choice. Specifically, the results of the study suggest that consumers who experience a product trial resulting from a brand choice process will find the trial to be more useful and valid, and they will feel more expert at judging the brand in the trial compared to those consumers who did not freely choose the trial brand but were simply given a sample of the target brand. There is also some indication that consumers experiencing free choice will judge the brand more favorably in the trial Implications for theory and practice are discussed.

Keywords: Consumer Behavior, Brand Choice; Perceived Control; Product Trial

1. INTRODUCTION

Significant streams of research in both marketing and psychology suggest that direct brand experience (hands-on product trial) tends to create strongly held consumer brand beliefs and attitudes compared to indirect experience, such as reading an ad (e.g., Fazio and Zanna, 1978; Olson and Dover, 1979; Smith, 1993; Wright and Lynch, 1995). Thus product trial is a very powerful source of brand information and one that marketers would benefit from understanding and using to its full advantage. Marketers have several options available to induce consumer trial of their products, including in-home sampling, in-store sampling, discounted trial purchase, and many others. However, in virtually all of the academic studies dealing with trial, consumers are simply provided with a sample of a particular brand and then are asked to answer questions about its performance in the trial. Such situations are obviously not reflective of many of the common ways in which consumers may have trial experiences in the marketplace. For example, in the case of in-store sampling, consumers are able to choose whether or not they want to sample a particular brand, and this choice process would be even more salient in the case of a trial purchase, where funds must be voluntarily outlaid to acquire the product. A major goal of the study reported here is to empirically investigate how consumers' reactions to a trial of a freely chosen brand may differ from their reactions to a "forced" trial such as those used in virtually all past trial research. Significant differences in consumers' processing of these two types of trials could suggest that past studies may over- or under-represent the true power of direct brand experience, as we will discuss below.

Interestingly, research in psychology (e.g., Deci and Ryan, 1985; Averill, 1973) and consumer behavior (e.g., Hui and Bateson, 1991) suggests that people evaluate an experience differently depending on the feeling of control they have with respect to the situation. Specifically, this research has shown that individuals having greater control over an experience were more likely to evaluate that experience favorably. In a product trial context, this indicates that consumers may perceive a particular product trial to be more informative if it was freely chosen. Thus, in laboratory experiments where subjects are not able to exercise free choice in electing to try a product, the effects of direct experience on brand evaluations may be understated, and the brand evaluations themselves may also be understated. Therefore, more naturalistic consumer-controlled product trials may have an even more powerful effect on consumer brand evaluations than previous product trial research suggests.

The present study attempts to investigate this previously unresearched issue by determining (empirically) if perceived control in the context of a trial experience results in more favorable brand and trial evaluations. Specifically, we seek to determine how consumers will evaluate a trial experience when they do (not) have control in choosing the brand they try. In the study reported here, consumers are either: (1) given a sample of the brand to try (i.e., a no-choice situation), or (2) asked to choose between two brands for trial (i.e., free choice situation). We report the effects of this manipulation on brand evaluations and trial cognitions and discuss the theoretical and managerial implications of the findings in detail.

2. LITERATURE REVIEW

2.1 Product Trial

Product trial, which provides consumers with direct (i.e., hands-on) brand experience, has been shown to be superior to advertising in its ability to create strong and confidently held brand beliefs and attitudes. This comes at least partly from the fact that product experience is self-generated data, as opposed to other-generated information such as advertising. Self-generated data is almost always perceived to be more credible and persuasive than other-generated information. As Hoch (2002) phrases it, "Personal experience, unlike information generated by third parties, has that fresh, unvarnished realism that draws us in. Education is pallid, while experience is vivid." However, as also pointed out by Hoch and Deighton (1989) and Hoch (2002), consumer learning from product trial is not infallible. Many product trials are ambiguous, and consumers do not necessarily process trial information in an unbiased manner. For instance, many studies have shown that a pre-trial ad tends to positively bias post-trial brand attitudes (e.g., Kempf and Smith, 1998; Smith, 1993; Olson and Dover, 1979).

Research by Kempf and Smith (1998) empirically demonstrates that not all trials provide equal amounts of perceived information value to consumers. In that article, the authors identify three different types of trial cognitions consumers may produce regarding the informativeness of the trial experience. The most important of these trial cognitions is perceived trial diagnosticity (i.e., the perceived usefulness of the trial for evaluating the brand), which is a function, in part, of the nature of the product's attributes. While most products possess a mixture of experiential and non-experiential attributes, a trial (by definition) only yields direct information about the experiential attributes. Moreover, since many products tend to be dominated by either experiential or non-experiential attributes, it follows that trial will not be equally diagnostic for all products, and these differences in diagnosticity may be noted by consumers.

In addition to product-related perceived trial diagnosticity, Kempf and Smith (1998) identified two other categories of trial cognitions a consumer may generate that will affect a trial's perceived diagnosticity: perceived expertise (the degree to which the consumer feels confident in his/her ability to process the trial information) and perceived validity of the trial (perceptions that a particular trial experience is (not) representative of the product's true performance). These two constructs, in turn, positively influence consumers' overall appraisal of perceived trial diagnosticity. Kempf and Smith (1998) showed empirical evidence that consumers' perceptions of a trial's diagnosticity positively affected post-trial belief strength and brand attitude.

2.2 Perceived Control

Schutz (1967) suggests that a greater feeling of control aids individuals in having satisfactory interactions with their environment. Based on this notion, Hui and Bateson (1991) proposed, tested, and found empirical support for the idea that consumers would rate a service encounter more favorably when they perceived that they had more control over the situation. Control, defined as the need to demonstrate one's competence, superiority, and mastery over the environment, is described by Hui and Bateson (1991) as a fundamental driving human force. This finding appears to be consistent with several studies in psychology showing perceived control to have a positive impact on the psychological state of an individual (Szpiler and Epstein, 1976), task performance (Burger, 1987), and physiological well being (Langer and Rodin, 1976). One explanation for these findings is that individuals simply derive greater satisfaction from situations when they believe they are able to exert some influence (Bateson, 1985).

3. CONCEPTUAL FRAMEWORK AND HYPOTHESES

Hui and Bateson (1991) propose that consumer choice (i.e., the opportunity for a consumer to choose their environment) is a key dimension of perceived control. This is based on Averill's (1973) contention that decisional control is directly related to the amount of perceived control an individual has over a situation. Thus, in the current context of product trial, when consumers are able to choose the brand they try, they should evaluate it more positively than if they were simply randomly assigned a brand. This is also consistent with Hui and Bateson's (1991) finding that service encounters were viewed more positively when consumers had more perceived control over the encounter.

H1: The overall evaluation of a brand (as indicated by expectancy value--[Summation][b.sub.i][e.sub.i]) will be higher for consumers who freely chose the brand they experienced in the trial.

Bateson (1985) and Deci and Ryan (1985) suggest that perceptions of greater control may lead individuals to put forth a greater effort to optimize their decisions and, similarly, to extend a greater effort to complete those tasks. Applying this to the context of information reception, individuals who choose the means via which they receive information should put forth a greater effort to interpret that information. Extending this thinking, individuals who are able to choose the brand they wish to try should be more involved in the trial experience and should process it more deeply and effortfully than those who were simply assigned a trial brand. This deeper processing, in turn, may lead to perceptions that the trial is more diagnostic or useful. Further, individuals feeling greater decisional control or choice over the trial may view the trial to be more valid and may view themselves as being more expert at judging the trial, simply because of the greater effort and attention they devoted to the trial experience. This prediction is supported by Tafarodi, Milne, and Smith (1999), who found that individuals rated their own performance on a task more highly when they possessed greater perceived choice over the task's parameters.

H2: Trial diagnosticity will be higher for consumers who have the opportunity to choose the brand for the product trial, compared to consumers who did not have free choice of brands.

H3: Perceived expertise will be higher for consumers who have the opportunity to choose the brand for the product trial, compared to consumers who did not have free choice of brands.

H4: Perceived trial validity will be higher for consumers who have the opportunity to choose the brand for the product trial, compared to consumers who did not have free choice of brands.

4. METHOD

To assess the effects of choice on consumers' perceptions of trial diagnosticity, expertise, and trial validity, an experiment was conducted. Brand choice, the independent variable, was manipulated at two levels. In one group, subjects were given the opportunity to choose between two advertised brands. In the other, subjects were simply assigned a brand they would be evaluating in the subsequent trial (no brand choice involved). The sample used for the study consisted of 102 undergraduate business students who received class credit for their participation. To increase motivation and involvement with the study tasks, subjects were also entered into a lottery for a $100 first prize and two $50 second prizes. Students were considered an appropriate sample given that they are a major target segment for the stimulus product used in this study--a cola soft drink.

4.1 Experimental Stimuli

A fictitious cola soft drink named Cola-Lite was used for the experiment. Pretest subjects similar to those used in the main study perceived this brand name as appealing and appropriate for the described product type. The product was positioned as a "hybrid" soft drink, trading off the advantages/disadvantages of diet and regular soft drinks (since the time of our experiment, both Coca-Cola[R] and Pepsi[R] have introduced this type of soft drink). The product was described as being sweetened using a fifty-fifty mix of real sugar and artificial sweeteners. It therefore had half the calories of regular soft drinks, but reduced the diet aftertaste of most diet soft drinks. In the experiment, the stimulus product subjects actually consumed consisted of a fifty-fifty mix of Pepsi[R] and Diet Pepsi[R] (and thus, was essentially Pepsi Edge[R]).

In order to add realism to the choice task, all subjects in the experiment were exposed to advertising information about two brands of cola prior to the choice and product trial. The goal of the ad information was to provide choice subjects with a basis on which to meaningfully choose a particular brand for trial. The ad used for the target product in the study (Cola-Lite) was attribute-oriented and contained claims relating to each of the six attributes that pre-test subjects indicated were salient to this product category (described later). The ad was a full-page black and white print ad consisting of a headline designed to gain attention and interest, attribute claims, and a picture of the can with the fictitious name, Cola-Lite.

Because the study called for subjects to make a choice between brands (or be assigned one of two brands, depending on the experimental condition), a second fictitious soft drink was needed. This distracter brand was called Star Cola. It was positioned similarly to Cola-Lite, but the ad gave much less detail about the product. The ad only contained the words "Great Taste," "Low Calories," and "Also Caffeine-Free," along with a picture of the can.

4.2 Procedure

Subjects were randomly assigned to one of the two conditions, choice or non-choice. Separate experimental sessions were conducted in groups of about 25 subjects, with the choice and non-choice groups in separate rooms. In both the choice and non-choice conditions, subjects were given an ad packet consisting of the target ad (for Cola-Lite), the Star Cola ad, plus one filler ad. To increase involvement, subjects were told that both soft drinks represented new brands that would be introduced in the area in the near future. They were given two minutes to review the ad packet, after which they completed the section of the questionnaire containing ad response measures. These included measures of advertising message involvement, attitude toward the ad, pretrial brand belief strength, belief confidence, attribute evaluations, and brand attitude. To avoid biasing subjects' brand choice or post-trial evaluations, these measures were collected for both Cola-Lite and Star Cola.

After completing the pre-trial measures, subjects in the choice group were told they could select one of the two cola brands (for which they had seen ads) to try. They were asked to prominently mark their choice of brands on the questionnaire and then bring the questionnaire up with them to the front of the room to show the administrator, who then gave them the appropriate (chosen) sample. The non-choice subjects were simply told they would be receiving samples of Cola-Lite in the product trial. Separate trays of small cups with cardboard signs showing a color picture of either the Cola-Lite or Star Cola cans contained the samples.

To maintain the validity of the dependent variable comparisons between the choice and non-choice groups, all subjects who chose the Star Cola (the non-target brand) to taste were eliminated from the sample, and their responses were not included in the final data. The hope was that most of the free-choice subjects would choose Cola-Lite, to maximize the usable sample size. In fact, the choice subjects were fairly evenly split between choosing Star Cola and Cola-Lite. While this resulted in losing the data from a significant number of subjects, it also indicates that the choice process was meaningful, in that both brands were reasonably appealing.

Next, subjects were given a small sample ([congruent to] 1.5 oz) of the soft drink they chose (or were assigned, depending on the condition) to taste at their own pace. After tasting the soft drink, subjects completed a thought-listing task in which they were instructed to record any and all thoughts that occurred to them while tasting the cola. They then responded to the remaining post-trial measures, which included trial diagnosticity, trial validity, perceived expertise, attribute belief strength ([bi.sub.i]), and attribute evaluations ([e.sub.i]). Subjects also completed measures of product class involvement and certain demographic variables (age and sex).

4.3 Measures

Salient attributes for soft drinks were obtained in a pretest (n=70) using the free-elicitation task recommended by Fishbein and Ajzen (1975). Pretest subjects (who were similar to the subjects used in the final study) were asked to list the attributes that would be important to them when deciding to purchase a particular brand of soft drink. The most commonly listed attributes (in order of highest to lowest frequency) were: taste (mentioned by 92% of subjects), caffeine level (27%), sweetness (27%), calories (20%), and absence of a diet aftertaste (6%). These five attributes were featured in the Cola-Lite ad.

Consistent with past advertising and product trial studies (e.g., Kempf and Smith, 1998; Marks and Kamins, 1988; Smith, 1993; Smith and Swinyard, 1983; 1988), subjects' cognitive structure or expectancy value from the brand ([summation][b.sub.i][e.sub.i]) was measured using Fishbein and Ajzen's (1975) multi-attribute model. This construct includes brand belief strength ([b.sub.i]) for each attribute and a measure of the value consumers place on each attribute ([e.sub.i]). Brand belief strength was measured by asking subjects, "How likely do you believe it is that (the product) has Attribute X?" for each attribute. A seven-point scale was used for these questions, with endpoints labeled "Very Unlikely" and "Very Likely" (with the midpoint labeled as "Even Chance"). The attribute evaluations were measured by asking subjects how they would evaluate a brand that possessed each of the salient attributes (in individual questions for each attribute). Responses were recorded on a bipolar scale with endpoints labeled "Extremely bad" (-3) and "Extremely good" (+3).

Similar to Kempf and Smith (1998), perceived trial diagnosticity was assessed by asking the respondents: "Overall, how helpful did you feel this taste test was in allowing you to carefully evaluate the new soft drink?" Responses were recorded on a 1-9 scale with the endpoints labeled "Not helpful at all" and "Extremely helpful," and the midpoint was labeled "Somewhat helpful."

Subjects were asked to rate their perceived expertise at judging the product in the trial with the following item: "How confident do you feel in your ability to judge the soft drink in this taste test?" The endpoints of the 1-9 scale were labeled "Not confident at all" and "Extremely confident," and the midpoint was labeled "Moderately confident" (Kempf and Smith, 1998).

Trial validity was measured using a two-item scale (r=.69). The first item was "Do you feel that this trial experience was a valid test of the brand?" The 1-9 response scale was anchored by "Completely Invalid" and "Completely Valid," and the midpoint was labeled "Somewhat Valid." The second item was: "Did you feel that this taste test was a fair test of the brand?" The response was scaled 1-9, with endpoints labeled "Extremely Unfair" and "Extremely Fair," and midpoint labeled "Moderately Fair."

5. RESULTS

5.1 Manipulation Check

To determine whether our brand choice manipulation was successful at instilling significant differences in the level of perceived control between the two experimental groups, a three-item measure of perceived choice was collected from all subjects. These questions asked for the level of agreement/disagreement (1-9 scale) with the following statements: "I feel I was given a free choice as to which of the two soft drinks I got to taste," "I selected the brand I tried of my own free will," and "I felt like I had full 'say so' about which of the two advertised colas I got to taste." This scale had a coefficient alpha of .98, indicating high internal reliability.

The results of this manipulation showed highly significant differences between the choice/no choice groups (choice group mean = 8.59, no choice group mean = 2.37, [t.sub.100]=18.16, p<.0001), indicating a successful manipulation.

5.2 Hypotheses Tests

Hypotheses 1-4 were tested by comparisons of means between the choice and non-choice groups. To control for experiment-wise Type II error due to the intercorrelations among the dependent variables, a MANOVA was performed. The results indicate that the presence or absence of free choice significantly affected the levels of the dependent variables as a group, which provides the opportunity for the follow-up univariate analyses described below (Wilks' Lambda=.87, [F.sub.4,96]=3.67, p<.008).

Hypothesis 1 states that the overall expectancy value ([summation][b.sub.i][e.sub.i]) from the brand will be higher for the choice group than for the non-choice group. This hypothesis was supported. The mean [SIGMA][b.sub.i][e.sub.i] for the choice group was 53.57, and 44.68 for the non-choice group. This difference was statistically significant ([t.sub.100]=1.98, p<.03, one-tailed).

Hypothesis 2 states that consumers who are given a free choice of which brand to try will rate the trial as more diagnostic than those consumers who did not have the opportunity to choose the trial brand. This hypothesis was supported. The mean perceived trial diagnosticity rating for the choice group was 7.63 (on a 9-pt. Scale), compared to 6.57 for the non-choice group ([t.sub.100]=3.71, p<.0002, one-tailed).

According to Hypothesis 3, the choice group was predicted to perceive themselves to be more expert than the non-choice group. This hypothesis was also supported. The mean perceived expertise rating for the choice group was 7.20, compared to 6.38 for the non-choice group ([t.sub.100]=2.29, p<.01, one-tailed).

To test Hypothesis 4, the mean perceived trial validity ratings were compared between the choice and non-choice groups. The choice group's mean validity rating was 6.85, compared to 5.97 for the non-choice, supporting H3 ([t.sub.100]=2.64, p<.005, one-tailed).

5.3 Additional Analyses

Kempf and Smith (1998) proposed that perceived trial diagnosticity is influenced by perceived expertise and perceived trial validity, and trial diagnosticity, in turn, has a direct positive effect on cognitive structure ([sigma][b.sub.i][e.sub.i]). To test these relationships in the current study, two regression analyses were performed. Results showed that both perceived expertise and perceived validity significantly affected overall perceived diagnosticity. The standardized regression coefficient for perceived expertise was .48 ([t.sub.99] =5.4, p<.0001), and for perceived validity the standardized coefficient was .29 ([t.sub.99]=3.2, p<.002). The [R.sup.2] value for this equation was .46. Cognitive structure was then regressed on perceived trial diagnosticity. Again, the results supported Kempf and Smith's (1998) hypotheses, showing a significant positive relationship ([beta]=.23, [t.sub.100]=2.35, p<.01, [R.sup.2]=.05).

6. DISCUSSION

6.1 Implications for Theory

The results of this experiment provide support for the general notion that consumers who believe they have some level of perceived control or choice over the product trial they receive will evaluate that trial as being more diagnostic and more valid, and they will believe themselves to be more expert in evaluating the brand than those who did not feel they had significant control. Further, these results provide some evidence that consumers who had the opportunity to choose which brand they would try will evaluate that brand more favorably than those not having a choice. This study represents an important extension of theory proposed and tested by Hui and Bateson (1991) regarding the effects of choice on brand evaluations. Not only have we shown that the perceived control effects found in that study also hold in a non-service product setting, but we have also shown that perceived control or choice enhances consumers' evaluations of the information source itself (in addition to the brand evaluations), causing them to perceive the trial information as being more useful and valid than would otherwise be the case.

The results of this experiment also provide replication for a portion of the trial processing model presented by Kempf and Smith (1998). Specifically, we have shown that consumers' perceptions of trial diagnosticity, trial validity, and perceived expertise positively influence their post-trial brand evaluations (i.e., [summation][b.sub.i][e.sub.i]).

6.2 Implications for Practice

Given that product trial has been shown to be such a powerful source of brand evaluations and that trials vary significantly in information value, marketing practitioners should be interested in learning how to maximize the positive effects of product trial on brand attitudes. A marketer of a new product has several choices in how to expose potential customers to product trial. One common method of achieving trial is through the use of free samples, which is an effective but often expensive way of administering product trial to large numbers of consumers. Sampling or product demonstration (depending on the type and cost of the product) can be performed in-store, at public gatherings, or in-home. A second way to promote first-time product trial among consumers is to use trial-inducement programs that help to remove barriers to trial purchase. These typically include low introductory prices or coupons with accompanying advertising to create awareness of the offer. If both of these trial programs are feasible, which method should be used by the marketer? Does it make a difference?

A major point of difference between these two trial scenarios is the concept of consumer choice. In a sampling situation, subjects receive the product with no cost, and no a priori choice process is involved. Consider two distinct sampling scenarios. In one situation, a consumer comes to the end of an aisle in

\\ the supermarket and is asked, "Would you like to try a free sample of Brand X cookies?" In a different setting, a consumer opens his/her Sunday newspaper and a small sample of a new hand lotion or laundry detergent is included with the coupons. In both of these situations, there is no forethought, product evaluation or choice process preceding the trial, other than the decision to accept or reject the free sample. Contrast these scenarios with a typical trial purchase, whereby a consumer has seen ads for a new product or brand, or has heard about it from an acquaintance. The consumer then sees the brand in the supermarket and decides to "take a chance" on it, based on expectations or weakly held beliefs created by the ad or word-of-mouth. In this case, the consumer clearly has some, albeit possibly very weak, expectations about the product. Curiosity about the brand and weak brand expectations may also have been generated on the spot in the store based simply on the packaging, resulting in a trial purchase. A choice process has taken place in all these scenarios. Furthermore, the consumer has paid money to acquire the brand for the purpose of trial. This is a significantly different context than the free sample situation described earlier.

Although further research is needed before managerial recommendations can be confidently made, results of the present study at least tentatively suggest that when consumers choose a brand and then subsequently experience a trial of the chosen brand, they will perceive the trial information to be more useful and will evaluate it in a more positive manner. Understanding these differences has important implications for the marketer developing a promotional plan aimed at stimulating trial of a new product.

7. LIMITATIONS

As with most experiments, the present study's results should be interpreted with caution, as it has several limitations (most of which can be addressed with additional experiments). For example, it would be presumptuous to make definitive conclusions from a single study involving one trial for a particular brand representing a single product class. More confident recommendations would be possible after our results are shown to generalize to other product classes and settings. Further, our study was conducted via an experiment with student subjects in a laboratory setting. Future research should attempt to determine if similar results would be obtained from other market segments in more naturalistic marketplace settings. One difficulty we faced, which is inherent in choice research, is the challenge in creating a meaningful brand choice situation in a laboratory setting. A subject's decision not to try a brand results in no data point for that subject. To create a meaningful choice yet still maintain a significant amount of data points, we used a forced choice between two brands to create a feeling of perceived control. In the marketplace, choice is very complex: a consumer may choose to buy a certain brand specifically for the purpose of seeing if they like it, and in other situations they experience a trial because of a free sample. It is difficult to envision a laboratory experiment that exactly mirrors these two trial situations. Despite these limitations, however, the present study yields several interesting insights regarding the effects of perceived control on consumers' processing of trial information and post-trial brand evaluations that should be investigated in future research.

REFERENCES

Averill, James R., "Personal Control over Aversive Stimuli and Its Relationship to Stress," Psychological Bulletin, 80 (4), 1973, 286-303

Bateson, John E. G., "Perceived Control and the Service Encounter," in The Service Encounter: Managing Employee/Customer Interaction in the Service Business, John A. Czpiel et al. (eds.), Lexington, MA, 1985, 67-82.

Burger, Jerry M., "Increased Performance with Increased Perceived Control: A Self-Presentation Interpretation," Journal of Experimental Social Psychology, 23 (4), 1987, 350-360.

Deci, Edward L. and Ryan, Richard M., Intrinsic Motivation and Self-Determination in Human Behavior, Plenum, New York, 1985.

Fazio, Russell H. and Zanna, Mark P., "On the Predictive Validity of Attitudes: The Roles of Direct Experience and Confidence," Journal of Personality, 46 (June), 1978, 228-243.

Fishbein, Martin and Ajzen, Icek, Belief, Attitude, Intention and Behavior: An Introduction to Theory and Research, Addison-Wesley, Reading MA, 1975.

Hoch, Stephen J., "Product Experience is Seductive," Journal of Consumer Research, 29 (December), 2002, 448-454.

Hoch, Stephen J. and Deighton, John, "Managing What Consumers Learn from Experience," Journal of Marketing, 53 (April), 1989, 1-20

Hui, Michael K. and Bateson, John E. G., "Perceived Control and the Effects of Crowding and Consumer Choice on the Service Experience," Journal of Consumer Research, 18 (September), 1991, 174-184

Kempf, DeAnna S. and Smith, Robert E., "Consumer Processing of Product Trial and the Influence of Prior Advertising: A Structural Modeling Approach," Journal of Marketing Research, 35 (August), 1998, 325-338

Langer, Ellen J. and Rodin, Judith, "The Effects of Choice and Enhanced Personal Responsibility for the Aged: A Field Experiment and Institutional Setting," Journal of Personality and Social Psychology, 34 (2), 1976, 191-198

Marks, Lawrence J. and Kamins, Michael A., "The Use of Product Sampling and Advertising: Effects of Sequence of Exposure and Degree of Advertising Claim Exaggeration on Consumers' Belief Strength, Belief Confidence, and Attitudes," Journal of Marketing Research, 25 (August), 1988, 266-281

Olson, Jerry C. and Dover, Philip A., "Disconfirmation of Consumer Expectations Through Product Trial, Journal of Applied Psychology, 64 (April), 1979, 179-89

Schutz, William C., The Interpersonal Underworld, Science & Behavior, Palo Alto, CA, 1967

Smith, Robert E., "Integrating Information from Advertising and Trial: Processes and Effects on Consumer Response to Product Information," Journal of Marketing Research, 30 (May), 1993, 204-219

Smith, Robert E. and Swinyard, William R., "Cognitive Response to Advertising and Trial: Belief Strength, Belief Confidence and Product Curiosity," Journal of Advertising, 17 (3), 1988, 3-14

Smith, Robert E. and Swinyard, William R., "Attitude-Behavior Consistency: The Impact of Product Trial Versus Advertising," Journal of Marketing Research, 20 (August), 1983, 257-267.

Szpiler Jack A. and Epstein, Seymour, "Availability of an Avoidance Response as Related to Automatic Arousal," Journal of Abnormal Psychology, 85 (1), 1976, 73-82

Tafarodi, Romin W., Milne, Alan B. and Smith, Alyson J., "The Confidence of Choice: Evidence for an Augmentation Effect on Self-Perceived Performance," Personality and Social Psychology Bulletin, 25 (11), 1999, 1405-1416

Wright Alice Ann and Lynch, John G., "Communication Effects of Advertising versus Direct Experience when Both Search and Experience Attributes are Present," Journal of Consumer Research, 21 (March), 1995, 708-718

DeAnna S. Kempf, Middle Tennessee State University, Murfreesboro, Tennessee, USA

Russell N. Laczniak, Iowa State University, Ames, Iowa, USA

Dr. DeAnna S. Kempf earned her Ph.D. in Marketing at Indiana University, Bloomington in 1995. She is currently an assistant professor of marketing at Middle Tennessee State University.

Dr. Russell N. Laczniak earned his Ph.D. in Marketing at the University of Nebraska, Lincoln in 1987. He is a professor of marketing and former department chair at Iowa State University. He is currently the editor of the Journal of Advertising.