Conditioning and Learning Process

By Peter, J.P., Olson, J.C.

Edited by Paul Ducham


A number of consumer research studies have investigated classical conditioning. The authors of one program of research state, “There are enough demonstrations in our literature to accept the fact of classical conditioning of consumers’ attitudes toward consumption objects.” Others have argued that classical conditioning may be most useful in marketing in low-involvement situations:

 Consumer Involvement is low when the products have only minor quality differences from one another. . . . This is especially the case in saturated markets with mature products. It is exactly in these markets that product differentiation by means of emotional conditioning is the preferred strategy of influencing consumers.

 Because most products are mature and many markets are saturated, classical conditioning is likely to be a useful strategy for low-involvement purchases. However, classical conditioning can also be useful for high-involvement situations, such as the purchase of athletic shoes by teenagers. Marketers of brands such as Nike, Adidas, and Fila seem well aware that the presence of superstar athletes in their commercials performing exciting slam dunks and fast breaks can condition positive affect to their products and lead to increased sales. Automobile companies seem well aware that the presence of attractive models, exciting locations, and popular background music can influence purchase through classical conditioning.


The use of classical conditioning as a marketing tool has several implications. First, classical conditioning directs attention to the presentation of stimuli that, because of previous conditioning, elicit affect in consumers. In some cases, these feelings are likely to increase the probability of certain behaviors and/or decrease the probability of other behaviors. Second, in many cases, marketers may find it useful to condition responses to stimuli. By repeatedly pairing Tiger Woods, an exciting golfer and sports personality, with Tag Heuer watches and Nike golf clothes, these products may generate greater excitement and increased purchases. Exhibit 9.2 summarizes a number of marketing tactics consistent with classical conditioning principles.

Operant Conditioning

 Operant conditioning is the process of altering the probability of a behavior being emitted by changing the consequences of the behavior. It differs from classical conditioning in at least two important ways. First, whereas classical conditioning is concerned with involuntary responses, operant conditioning deals with behaviors that are usually assumed to be under the conscious control of the individual. By conscious control, behaviorists mean under the control of the skeletal nervous system, which governs the “striped” muscles; they are not stating that behaviors are under the control of cognitions. Second, although classically conditioned behaviors are elicited by stimuli that occur before the response, operant behaviors are emitted because of consequences that occur after the behavior.

 In any given situation, at any given time, there is a certain probability that an individual will emit a particular behavior. If all of the possible behaviors are arranged in descending order of probability of occurrence, the result is a response hierarchy . Operant conditioning has occurred when the probability that an individual will emit a behavior is altered by changing the events or consequences that follow the behavior.

 Some events or consequences increase the frequency with which a given behavior is likely to be repeated. For example, if a reward, such as a cash rebate, is given at the time of purchase, it may increase the probability that a shopper will purchase in the same store in the future. In this case, because the reward increases the probability of the behavior being repeated, it is called positive reinforcement . Positive reinforcement is likely the most common type of consequence marketers use to influence consumer behavior. In general, the greater the amount of the reward and the sooner it is received after the behavior, the more likely the behavior will be reinforced and the consumer will perform similar behaviors in the future. For example, a $1 coupon for Tropicana orange juice would likely increase the probability of purchase of juice and lead to future purchases of this product than would a 50-cent coupon. Similarly, if the coupon is redeemable at the time of purchase, it will likely be more effective than a mail-in coupon for which the consumer has to wait for the reward.

 The frequency of consumer behavior can also be increased by removing aversive stimuli. This is called negative reinforcement . For example, if a consumer purchases a product to get a salesperson to quit pressuring him or her, the consumer may be negatively reinforced. That is, by performing the behavior of purchasing, the aversive stimuli (the actions of the pushy salesperson) are removed. In the future, when confronted with pushy salespeople, operant conditioning would predict that the consumer will be more likely to purchase again.

 Sometimes operant techniques are used to decrease the probability of a response. If the environment is arranged so that a particular response results in neutral consequences, over a period of time that response will diminish in frequency. This process is referred to as extinction . For example, at one time the A&P grocery chain was the largest retailer in the world. However, one mistake it made was to overstock its own brands (which had higher profit margins) and understock nationally branded merchandise. Consumers who were loyal to a number of nationally branded products often could not obtain them at an A&P store. Eventually many consumers quit shopping at A&P, partially because they could not obtain their favorite brands. Thus, A&P inadvertently used extinction on its own customers.

 If a response is followed by a noxious or aversive event, the frequency of the response is also likely to decrease. The term punishment is usually used to describe this process. 6 For example, suppose you went to a clothing store and the salespeople were rude to you. Wouldn’t this decrease the chances that you would go back there? Punishment is often confused with negative reinforcement, but they are distinctly different concepts. Exhibit 9.3 presents a summary of the four methods of operant conditioning.

 There are a number of other important ideas about operant conditioning. We discuss three—reinforcement schedules, shaping, and discriminative stimuli—that have major implications for designing marketing strategies to influence consumers’ behavior.

Exhibit 9.2


A number of different reinforcement schedules can be employed. For example, it is possible to arrange conditions so a positive reinforcer is administered after every desired behavior. This is called a continuous reinforcement schedule . Marketers usually try to keep the quality of their products and services constant so that they will be continuously reinforcing with every purchase, but this is difficult to do. For example, frequent product recalls for automobiles indicate a failure to maintain product quality. Services such as airlines may be unable to control contingencies such as bad weather; overbooked, canceled, and late flights; and unfriendly employees, which can make flights nonreinforcing. Because sporting events may be boring or the home team may get beaten, they may not be continuously reinforcing for some consumers.

 Conditions can also be arranged so that every second, third, or tenth time the behavior is performed, it is reinforced. This is called a fixed ratio schedule . Similarly, it is possible to have a reinforcer follow a desired behavior on an average of, say, one-half, one-third, or one-fourth the time the behavior occurs, but not necessarily every second, third, or fourth time. This is called a variable ratio schedule . The various state lotteries are examples of prizes awarded on variable ratio schedules.

 The variable ratio schedules are of particular interest because they produce high rates of behavior that are reasonably resistant to extinction. Gambling devices are good examples. Slot machines are very effective in producing high rates of response, even under conditions that often result in substantial financial losses. This property of the ratio schedule is particularly important for marketers because it suggests that a great deal of desired behavior can be developed and maintained with relatively small, infrequent rewards. Deslauriers and Everett found that by giving a free token for riding a bus on a variable ratio schedule, the same amount of bus riding could be obtained as when rewards were given on a continuous schedule. Thus, for approximately one-third the cost of the continuous schedule, the same amount of behavior was sustained.

 Numerous other examples of the use of variable ratio schedules can be found in marketing practices. In addition to state lotteries, common examples include sweepstakes, contests, and door prizes, in which individuals must behave in a certain way to be eligible for a prize. Consumer Insight 9.1 discusses the use of variable ratio schedules for selling Pepsi and Mountain Dew products.

 Consumer Insight 9.1

Using Variable Ratio Schedules to Increase Pepsi Purchases

PepsiCo ran an “Unlock the Great Taste and Win” sweepstakes. The grand prizes in the contest were two Lamborghini sports cars with an estimated retail value of $215,000 each. Other prizes included Kawasaki jet skis, compact vending machines, vacations, and sterling silver key chains.

 Although consumers could receive two game chances without purchases by writing the company, most game chances were distributed through purchase of Pepsi and Mountain Dew products. Here’s how it worked. With purchases of multipacks—12-, 20-, or 24-can packages— consumers had a chance to receive a free, inexpensive key chain. Behind the key chain package was notification of any major prize won. However, only one out of two multipacks contained the key chain and chance of winning. Thus, on average, consumers would have to purchase two multipacks to get a chance at the major prizes. PepsiCo used a variable ratio schedule to allocate prize chances to increase the probability that consumers would purchase several multipacks.

 Some bottle caps on 2-liter, 3-liter, and 16-ounce nonreturnable bottles also contained chances to win, but no key chain. Also, the odds of winning were better when consumers bought the more expensive multipacks. For example, the odds of winning the grand prizes from a multipack purchase were 1 in 18,444,000, whereas the odds of winning the grand prizes from a bottle purchase were 1 in 113,118,597. In addition, only by purchasing multipacks could a consumer win the sterling silver key chains valued at $50; the bottle purchases allowed winning only a brass key chain valued at $10. All prizes were awarded on variable ratio schedules.

 Overall, variable ratio schedules were used to allocate the chances to win prizes as well as the prizes themselves. By offering the chances to win and the inexpensive key chains on a variable ratio schedule, PepsiCo increased the probability of consumers making more than one purchase. Also, the cost of the key chains was only half what it would have been if every multipack contained one. Offering major prizes on a variable ratio schedule is likely the only way expensive products can be used as reinforcers for purchase of inexpensive products and still be profitable. To learn more about Pepsi, visit its Web site at


Another operant conditioning concept that has important implications for marketing and consumer behavior is shaping . Shaping is important because, given consumers’ existing response hierarchies, the probability that they will make a particular desired response may be very small. In general, shaping involves a process of arranging conditions that change the probabilities of certain behaviors not as ends in themselves but to increase the probabilities of other behaviors. Usually shaping involves the positive reinforcement of successive approximations of the desired behavior or of behaviors that must be performed before the desired response can be emitted.

 Many firms employ marketing activities that are roughly analogous to shaping. For example, loss leaders and other special deals are used to reward individuals for coming to a store. Once customers are in the store, the probability that they will make other desired responses (such as purchasing full-priced items) is much greater than when they are not in the store. Carnivals held in shopping centers or auto dealer parking lots may be viewed as attempts to shape behavior because consumers are more likely to come in and purchase when they are already in the parking lot than when they are at home. Similarly, free trial periods may be employed to increase the likelihood that the user will have contact with the product so that he or she can experience the product’s reinforcing properties. Real estate companies that offer free trips to look over resort property are employing a shaping tactic, as are casinos that offer free trips to gamblers. In both cases, moving people to the place of purchase (or place of gambling) increases the probability that these behaviors will be performed.

 Shaping is not confined to a one-step process; it can be used to influence several stages in a purchase sequence. For example, suppose a car dealer wants to shape an automobile purchase. Free coffee and doughnuts are offered to anyone who comes to the dealership. Five dollars in cash is offered to any licensed driver who test-drives a car. A $500 rebate is offered to anyone who purchases a car. This example demonstrates not only how operant principles can be used in a multistep process but also how they can be used in a high-involvement purchase situation.


It is important to distinguish between the reinforcement and discriminative functions played by stimuli in the operant model. So far in this section, we have focused on the reinforcing function. However, the mere presence or absence of certain stimuli can serve to change the probabilities of behavior. These are called discriminative stimuli .

 Discriminative stimuli are often said to “set the occasion” for behaviors. This means discriminative stimuli can be presented before a behavior and can influence whether the behavior occurs. In fact, discriminative stimuli allow operant conditioners to account for the effects or antecedents to behavior on changing behavior. For example, suppose Pizza Hut runs an ad that offers a free quart of Pepsi with every large pizza purchased. This offer may increase the probability of purchasing a large pizza from Pizza Hut. However, the offer itself is not a reinforcer since it is offered before the behavior. Rather, the offer is a discriminative stimulus.

 Many marketing stimuli are discriminative. Store signs (“50 percent off sale”) and store logos (Wal-Mart’s sign, Target’s Bullseye) or distinctive brand marks (the Nike swoosh, the Oakley O, the Polo insignia) are examples of discriminative stimuli. Previous experiences have perhaps taught consumers that purchase behavior will be rewarded when the distinctive symbol is present and will not be rewarded when the symbol is absent. For example, many consumers purchase Ralph Lauren shirts, jackets, and shorts that have the embroidered polo player symbol on them and avoid other Ralph Lauren apparel that does not have this symbol. A number of competitors have tried to copy the polo player symbol because of its power as a discriminative stimulus. Clearly, much of marketing strategy involves developing discriminative stimuli that increase certain behaviors.


Many marketing strategies and tactics are consistent with operant conditioning principles. If these are carefully designed, they can be quite effective in influencing consumer behavior. Many marketing tactics involve giving rewards after a purchase to increase its probability in the future. These rewards include rebates, contest tickets, bonuses, prizes, in-package coupons, and courteous thanks from salespeople. Although most strategies involve keeping product and service quality on a continuous reinforcement schedule, other types of rewards can be offered on a partial reinforcement schedule. Shaping is used to develop earlier behaviors in a purchase sequence to increase the chances of later behaviors. Finally, many store and brand symbols and logos have become discriminative stimuli for some consumers. Exhibit 9.4 summarizes a number of these tactics.

Vicarious Learning

Vicarious learning refers to processes by which people change their behaviors because they observed the actions of other people and the consequences that occurred. In general, people tend to imitate the behavior of others when they see it leads to positive consequences and to avoid performing the behavior of others when they see that negative consequences occur.

 Vicarious learning is also called modeling. Overt modeling involves consumers actually observing the model, such as seeing a salesperson demonstrating a Hoover vacuum cleaner in a store (live modeling) or seeing a commercial that depicts this behavior (symbolic modeling).

 Exhibit 9.5 shows the vicarious learning process. Many advertisements and TV commercials show models buying and using products and receiving positive consequences for doing so. Advil commercials have shown people suffering from arthritis pain but smiling and enjoying activities after they take the product. Toothpaste and deodorant commercials frequently show people being accepted and admired after using particular brands of these products.

Exhibit 9.4

Exhibit 9.5


There are three major uses of vicarious learning or modeling in marketing strategy. First, modeling can be used to help observers acquire one or more new response patterns that previously did not exist in their behavioral repertoires. Second, modeling can be used to decrease or inhibit un desired behaviors. Third, modeling can lead to response facilitation, whereby the behavior of others “serves merely as discriminative stimuli for the observer in facilitating the occurrence of previously learned responses.”

 Developing New Responses .  Modeling can be used to develop new responses that previously were not in the consumer’s behavioral repertoire. Consider the videocassette machines used in a variety of department and other stores to demonstrate use of a product. Sears has long used this method to demonstrate the appropriate and safe use of its chain saws. The appropriate uses of Berkeley fishing equipment and Olt duck calls are also demonstrated in this way. New behaviors are also frequently modeled in TV commercials. For example, insurance is traditionally purchased from an agent either at the agent’s office or in the consumer’s home, not in retail stores. Sears used a modeling strategy when it began in-store sales of Allstate insurance. Basically, the TV commercial shows a family coming to the Sears store and dropping off its old insurance policy for comparisons with Allstate rates. After a pleasant shopping trip, the family returns and is told that Allstate can provide a better deal, thus modeling the positive consequences of the new behavior. Similarly, Arm & Hammer baking soda ads showed new uses of the product as a carpet and refrigerator freshener and portrayed the models being complimented on the freshness of their homes. WD-40 lubricant ads also model new uses of the product.

 Inhibiting Undesired Responses.  Modeling can also be used to decrease the probability of undesired behaviors. Because of the ethical and practical problems involved in using punishment to influence consumer behavior, we have given little attention to ways of reducing the frequency of undesired responses. Such problems are far less prevalent when aversive consequences are administered to models rather than to actual consumers, however. Thus, vicarious learning may be one of the few approaches that can be used to reduce the frequency of unwanted elements in the behavioral repertoire of a potential or present customer.

 It is well known from the modeling literature that, under appropriate conditions, observers who see a model experience aversive outcomes following a particular act will reduce their tendency to exhibit that behavior. Similarly, vicarious learning can employ extinction to reduce the frequency of behavior.

 Consider the following examples. Hefty bags have been advertised on TV using a modeling approach. Various family members are shown taking out the trash in “bargain bags.” Of course, the bargain bag breaks and garbage is spewed all over the driveway. This is a very annoying experience! The frustrated family member is then told about Hefty bags, uses them successfully, and is socially reinforced for doing so. Head and Shoulders shampoo commercials have shown people initially being found attractive by members of the opposite sex but then being rejected when the models scratch their heads, indicating they may have dandruff. Following the use of the advertised product, the model is shown being happily greeted by an attractive member of the opposite sex.

 A common use of this type of modeling is in public service advertising. Many behaviors considered socially undesirable can be modeled and shown to have aversive consequences. These behaviors include littering, smoking, driving drunk, using drugs, overeating, wasting energy, and polluting. One commercial, for example, showed a drunken driver being caught, taken to court, and given a considerable fine and jail sentence for his behavior.

 Response Facilitation.  In addition to developing new behaviors and inhibiting undesired ones, modeling can be used to facilitate the occurrence of desired behaviors that are currently in the consumer’s repertoire. Modeling has been used extensively in advertising not only to illustrate the uses of a product but also to show what types of people use it and in what settings. Because many of these uses involve behaviors consumers already perform, the model’s function is merely to facilitate these responses by depicting positive consequences for using the product appropriately. For example, Nyquil ads show adult cold sufferers using the product before going to bed and then sleeping comfortably. This technique also appears frequently in advertising for high-status products. Such ads do not demonstrate any new behaviors but show the positive consequences of using the product. A series of Lowenbrau ads stressing the use of this beer for very special occasions is a good example.

 It is also possible to influence emotional behavior through a vicarious learning approach. Bandura noted that many emotional behaviors can be acquired through observations of others as well as through direct classical conditioning:

 Vicarious emotional conditioning results from observing others experience positive or negative emotional effects in conjunction with particular stimulus events. Both direct and vicarious conditioning processes are governed by the same basic principles of associative learning, but they differ in the force of the emotional arousal. In the direct prototype, the learner himself is the recipient of pain- or pleasure-producing stimulation, whereas in vicarious forms somebody else experiences the reinforcing stimulation and his affective expressions, in turn, serve as the arousal stimuli for the observer.

 To the degree that positive emotions toward a product are desired, vicarious emotional conditioning may also be useful for the design of effective advertisements.


There is no question that watching a model perform a behavior often increases the likelihood that the observer will also perform the behavior. It is well established in the psychological literature that in many situations, modeling is effective in changing behavior, as illustrated in Consumer Insight 9.2. However, certain factors have been found to increase the likelihood that vicarious learning will occur. These factors can be divided into three groups: (1) model and modeled behavior characteristics, (2) observer characteristics, and (3) characteristics of modeled consequences.

 Model and Modeled Behavior Characteristics.  Several personal characteristics of observed models influence the probability that an observer will imitate the modeled behavior. Models who are found to be attractive may be sought out, whereas less attractive models may be ignored. Models who are perceived to be credible and successful exert greater influence than those who are not. In addition, high-status and competent models are more influential in determining modeling success.

 Observers are also influenced by the manner in which the modeled behavior is performed. If the sequence of the modeled behavior is detailed very carefully and vividly, modeling effects tend to increase. The rate of learning also depends on the salience and complexity of the modeled behaviors. Interestingly, models who display a bit of apprehension and difficulty and yet complete the task are more effective than models displaying no struggle or difficulty. A reason for this was suggested by Manz and Sims:

 It appears that an observer can identify more with a model who struggles and overcomes the difficulties of a threatening task than a model who apparently has no problem. A model who is seen as possessing substantially greater abilities may not be considered a reasonable reference point for the observer. However, experts who display little difficulty in completing a task (e.g., professional athletes) may serve as ideals to be emulated in nonthreatening situations.

 Another factor that influences the effectiveness of models is the perceived similarity of the model to the observer. This finding supports the common practices of using models similar to people in the target market in commercials and attempting to increase similarities between customers and salespeople when hiring and assigning sales personnel. Many advertisers take advantage of these characteristics in developing commercials. These characteristics may also influence whether modeling aids in the diffusion of new products, an issue discussed in Consumer Insight 9.3.

 Characteristics of Observers.  Any number of individual-difference variables in observers can be expected to mediate successful modeling. For example, individual differences in cognitive processing as well as in physical ability to perform a modeled behavior may affect the process. Bandura suggests that in many cases observers who are dependent, lack confidence and self-esteem, and have been frequently rewarded for imitative behavior are especially prone to adopt the behavior of successful models. 14 However, perceptive and confident people readily emulate idealized models who demonstrate highly useful behaviors.

 Perhaps most important is the value the observer places on the consequences of the modeled behavior. For example, if consumers value the social approval obtained by a model in the Grecian Formula (hair coloring) commercial, they are more likely to purchase and use the product.

 Characteristics of Modeled Consequences.  Just as operant conditioning places importance on the consequences of behavior, so does vicarious learning. Of course, in vicarious learning, the observer does not experience the consequences directly. Thus, a major advantage of vicarious learning for consumers is that they can learn effective purchase and use behavior while avoiding negative consequences.

 Research has demonstrated that positively reinforcing a model’s behavior is a key factor in facilitating vicarious learning. In terms of consumer behavior, much fruitful research could be done on identifying appropriate reinforcers for various types of products. Currently, however, little is known about what types of positive consequences would be most effective to model. Similarly, for modeling applications that seek to decrease undesired behaviors, the most effective types of negative consequences to model in commercials are unknown. Although modeling has been demonstrated to be useful in deterring smoking, reducing drinking, reducing uncooperative behavior of children, and reducing energy consumption, many other areas of consumer behavior are unexplored.

 Consumer Insight 9.2

Do Professional Models Make Women Feel Bad about Their Appearance?

Many cues in our culture may communicate to women and young girls that being thin and fit is a prerequisite to being considered attractive in our society. Professional models, such as Kate Moss, Heidi Klum, and Giselle Bündchen, and many TV and movie stars who appear in commercials and ads and grace the covers of magazines usually appear to be tall, thin, and fit. In fact, many modeling agencies will not hire fashion models unless they are at least 5’7” tall. Pictures of the models’ bodies are often computer generated and computer enhanced to appear thinner and more shapely. Even Barbie dolls may be a problem; if Barbie were life-size, she would be 5’9” tall and have measurements of 36-18-33.

 Critics argue that these images have negative effects on many women and girls. Some suggest that eating disorders and self-esteem problems have resulted; 90 percent of the 8 million Americans with severe eating disorders are women. One study of 803 women found that, in 1985, 30 percent were dissatisfied with their appearance, whereas in 1995, 48 percent were dissatisfied. Increasingly, women believe there are two standards to serve: thinness and looking fit, according to the study’s author. Overall, 46 percent were dissatisfied with their weight; 40 percent with their muscle tone; 47 percent with their hips, buttocks, thighs, and legs; 51 percent with their waists and stomachs; and 25 percent with their chests, shoulders, and arms.

 Do you think models used in ads and commercials have negative effects on women’s body images? Do you think this causes problems like eating disorders and low self-esteem? Is it appropriate to use computer-enhanced models to sell fashions, exercise clothing and equipment, and cosmetic surgery?

 Consumer Insight 9.3

Diffusion of Innovations: A Modeling Process?

 Modeling plays a prime role in spreading new ideas, products, and social practices within a society or from one society to another. Successful diffusion of innovations follows a common pattern: (1) New products and behaviors are introduced by prominent examples, (2) the product/behavior is adopted at a rapidly accelerating rate, and (3) adoption then either stabilizes or declines, depending on the product/behavior’s functional role. The general pattern of diffusion is similar, but the mode of transmission, the speed and extent of adoption, and the life span of innovations vary for different products and forms of behavior.

 Modeling affects adoption of innovations in several different ways. It instructs people in new styles of behavior through social, pictorial, or verbal displays. Some observers are initially reluctant to buy new products or embark on new undertakings that involve risks until they see the advantages gained by earlier adopters. Modeled benefits accelerate diffusion by weakening the restraints of more cautious, later adopters. As acceptance spreads, the new gains further social support. Models not only exemplify and legitimize innovations, they also serve as advocates for products by encouraging others to adopt them.


Vicarious learning or modeling has many implications for marketing strategies designed to influence consumer behavior. First, modeling can be helpful in developing Information Contact behaviors. For example, TV commercials could show consumers how to contact a company’s Web page to get more information about products. The commercials could also show consumers how to order products from the Internet or by phone. Second, modeling can be used to increase store contact and Product Contact behaviors by demonstrating how consumers can get to a store or mall or find products. Commercials could also show consumers enjoying the shopping experience at the store and enthusiastically looking over products. Third, modeling can be used to influence Funds Access and transactions, such as the commercials for Master Debit Cards that showed a consumer using one to complete a transaction in time to make a flight, while another consumer who was writing a check missed the flight. Fourth, modeling can help to influence consumption by demonstrating how a product can be used safely and effectively. Infomercials for fishing tackle like the Banjo Minnow show how to rig it and how it catches many species of fish. Multitask woodworking equipment infomercials show how to set the machine up for different jobs and the excellent results it obtains. Modeling can also be used to affect disposition by showing consumers safe ways to dispose of hazardous products like motor oil or paint. Finally, modeling can affect communication by showing ads in which consumers tell others about how good a product is and encouraging them to buy it.

 In sum, advertisements and commercials commonly use modeling to influence consumer behavior. By carefully analyzing the salient characteristics of the models and modeled behaviors, the target consumers, and the consequences depicted in ads, marketers can increase advertising effectiveness. Exhibit 9.6 summarizes some uses of modeling to influence consumer behavior.

Exhibit 9.6