Classical Conditioning Effects in Product/Character Pairings Presented to Children

N. Carole Macklin, University of Cincinnati
ABSTRACT - An experiment was designed to test for effects of classical conditioning in children's product preferences. Eighty-four children participated in one of six conditions. The color of a pencil (orange or yellow) was presented with a picture of Smurf in one of three temporal arrangements. The preschoolers saw the pictures of the pencil and Smurf together (simultaneous condition), sequenced separately (forward condition) or unpaired (control). Results indicated no behavioral choice impact. However, a measure indicating cognitive activity resulted in statistical differences among the conditions. Children in the forward conditions were less likely to trade the pencil selected for a sticker. Implications to the classical conditioning framework and to children's advertising are discussed.
[ to cite ]:
N. Carole Macklin (1986) ,"Classical Conditioning Effects in Product/Character Pairings Presented to Children", in NA - Advances in Consumer Research Volume 13, eds. Richard J. Lutz, Provo, UT : Association for Consumer Research, Pages: 198-203.

Advances in Consumer Research Volume 13, 1986      Pages 198-203

CLASSICAL CONDITIONING EFFECTS IN PRODUCT/CHARACTER PAIRINGS PRESENTED TO CHILDREN

N. Carole Macklin, University of Cincinnati

ABSTRACT -

An experiment was designed to test for effects of classical conditioning in children's product preferences. Eighty-four children participated in one of six conditions. The color of a pencil (orange or yellow) was presented with a picture of Smurf in one of three temporal arrangements. The preschoolers saw the pictures of the pencil and Smurf together (simultaneous condition), sequenced separately (forward condition) or unpaired (control). Results indicated no behavioral choice impact. However, a measure indicating cognitive activity resulted in statistical differences among the conditions. Children in the forward conditions were less likely to trade the pencil selected for a sticker. Implications to the classical conditioning framework and to children's advertising are discussed.

INTRODUCTION

A basic principle of learning, classical conditioning, has received increased attention in the marketing literature (Nord and Peter 1980; Gorn 1982; McSweeney and Bierley 1984; Gresham and Shimp 1985; Allen and Madden, forthcoming). As Nord and Peter (1980) pointed out, advertisers appear to make frequent use of the principles of classical conditioning when repeatedly pairing exciting or sexy voices, music, etc. with a product.

Pavlov's (1927) work stands as the basic paradigm for classical conditioning. Pavlov arbitrarily selected a stimulus, a metronome, to be presented to a hungry dog. The metronome, labeled the conditioned stimulus (CS), was then followed by food, the unconditioned stimulus (US). The US, the food, automatically caused the dog to salivate, the unconditioned response (UR). With pairings of the metronome (CS) with the food (US), the dog responded to the metronome (CS) by salivating, now called the conditioned response (CR), even in the absence of food. The events are illustrated as follows:

CS(metronome)/US (food) ---------------------- UR (salivation)

CS(metronome) ----------------------------------- CR (salivation).

The above ordering is referred to as forward or traditional conditioning. The CS and US can be presented together (simultaneous conditioning), or the US can precede the CS (backward conditioning).

McSweeney and Bierley (1984) presented an excellent review of recent developments in classical conditioning. They reported changes in assumptions about the form of the CR and the temporal arrangement of the CS and US needed to establish a CR. In terms of form, the CR does not have to resemble the UR. Second, the sequencing of the CS and US is less restrictive. The traditional, temporal priority view suggested that a CS must precede a US in time for conditioning to occur. That is, the metronome must precede the food or a product must precede the music in an ad (forward conditioning). Rescorla (1967) argued for a predictiveness view suggesting that the CS (the metronome) must predict the US (the food), but the CS must not necessarily precede the US in time. His reformulation requires the use of a control group to ascertain whether true conditioning has occurred as distinct from CS or US familiarity or interaction. [The reader is referred to McSweeney and Bierley (1984) or Moore and Gormezano (1977) for a fuller explanation.]

McSweeney and Bierley (1984) cautioned consumer researchers against assuming that people respond like rats. They argued that consumers behave in more complicated settings than do animals in the laboratory. Many studies are needed before "translations" from the laboratory work with animals, upon which this literature is built, are made to consumer behavior.

The potential usefulness to consumer behavior of the classical conditioning paradigm appears encouraging by the work of Staats and Staats (1957, 1958, 1959) who suggested that the learning paradigm is applicable to humans. The Staats' work resulted in evaluative meaning being successfully conditioned. For example, Staats and Staats (1957) associated visually presented nonsense syllables with auditorily presented evaluative adjectives over a number of trials. They reported that the nonsense syllables that had been paired with positive meaning (e.g., happy) were considered more pleasant than those nonsense syllables paired with negative evaluation (e.g., ugly). Their work offers promise that the underlying mechanisms for automatic transfer are applicable to human settings that involve stimuli more approximate to advertising applications.

Empirical work on the classical conditioning paradigm in a marketing context is in the pioneering stage. Gorn s (1982) offered support for the usefulness of the classical conditioning paradigm. He reported two experiments testing whether the framework could account for the impact of music, a background feature, on product choice. Subjects saw a slide of either a blue or beige pen and listened to one minute of either pleasant or unpleasant music. They were asked to select a pen "for their help." Gorn reported a clear-cut impact of the music (pleasant or unpleasant) affecting the selection of the pen. Those subjects who heard the pleasant music were more likely to select the color of pen they saw on the slide, while those listening to the unpleasant music selected more often the color they had not seen. In the second experiment, Gorn (1982) manipulated the decision-making context and argued that classical conditioning accounted for choice behavior when the decision-making conteXt was uninvolved

Allen and Madden (forthcoming) questioned whether Gorn's findings would be replicated with a stronger experimental design. While acknowledging Gorn's important introduction of the neglected topic to the marketing literature, Allen and Madden (forthcoming) suggested that Gorn's processing of subjects in groups in a classroom setting could have resulted in either group interaction and/or awareness of other experimental conditions.

Allen and Madden (forthcoming) designed an experiment to be administered on an individual basis in which pleasant or unpleasant humor was paired with a color slide of a pen (green or black). They did not find support for the classical conditioning paradigm. In both the pleasant and unpleasant treatment groups, a majority ended up selecting the color viewed on the screen.

Allen and Madden (forthcoming) extended the theoretical inquiry by designing a clever buy-back procedure to assess whether a conditioned response, affect, would manifest itself in a decision that engaged more active cognitive processes. A statistically significant difference appeared between the pleasant and unpleasant humor groups with a greater percentage of pens sold back by those who listened to the unpleasant humor. They explained the results by suggesting that humor created feeling states in which those in pleasant humor condition were more likely to generate positive thoughts. Thus, they were more resistant to the buy-back attempt. Although no direct measures of the thought processes were taken, Allen and Madden (forthcoming) interpreted their finding as congruent with the "mood" literature (cf., Bower 1981; Clark and Isen 1982) suggesting that an affective state can influence people's judgments and behaviors indirectly by prompting and biasing cognitive activity.

Therefore, the two empirical efforts in a marketing context, both involving simultaneous conditioning, provide contradictory indications for the applicability of the classical conditioning framework to marketing phenomenon. Gorn's (1982) work with music offered strong support. However, Allen and Madden (forthcoming) reported none. In fact, the results from their buy-back measure, which may have produced a mood that biased cognitive activity, advised that no automatic responses occurred as expected under the classical conditioning paradigm.

The current research to be described consisted of another effort to witness conditioning effects. Both forward (or traditioning) conditioning and simultaneous conditioning were tested. An advantage of the present study was the use of young children as subjects. A major problem for consumer researchers to consider is subject awareness (Allen and Madden, forthcoming). Because young children are less likely to intuit a researcher's purpose, concern for this potential demand artifact is minimized.

RESEARCH HYPOTHESES

An experiment was designed to assess whether a favorable source paired with a product would directly affect children's product preferences. Two temporal arrangements between the CS and US were tested. First, it was hypothesized that children who saw a picture of a pencil together with a picture of a favorable character (simultaneous conditioning) would not evidence choice preference based on the simultaneous pairing. The second hypothesis predicted, however, that children who first saw a picture of the pencil followed by a picture of the favorable character (forward conditioning) would be more likely to select the pencil "advertised".

The rationale for the two hypotheses is as follows. The research literature grounding the first hypothesis is contradictory in terms of anticipated results. In general, laboratory work with animals suggests that simultaneous conditioning does not provide optimal conditioning (Smith, Coleman, and Gormezano 1969). Moore and Cormezano (1977) asserted that instances of simultaneous conditioning that have appeared have failed to include proper controls or have employed stimuli of too long a duration.

In terms of human conditioning, Gorn's (1982) work indicated success. As discussed, Allen and Madden (forthcoming) did not find effects. Thus, the first hypothesis is stated in a fashion consistent with the work based on animals and with Allen and Madden's (forthcoming) failure to find an effect with people.

In terms of the second hypothesis, effects from the forward conditions were predicted in accord with the literature based on animal literature. None of the human studies have included forward conditions, although Staats and Staats (1957, 1958, 1959) found effects from forward delay conditioning of adult subjects. As discussed in the introduction to this paper, Rescorla (1967) argued for a "predictiveness" view rather than just a temporally-based one. It is generally agreed that both views dictate the use of a control group to ascertain true conditioning. It should be noted that the issue of what constitutes an appropriate control has received considerable attention. An unpaired control group procedure assumes randomized sequencing of CS (alone) and US (alone) presentations at intervals larger than that of the conditioning group(s). Rescorla (1967) argued for a truly random control procedure in which CS and US were independently programmed so that the US would have equal probabilities of occurring in the presence or absence of the CS. Gormezano and Kehoe (1975) pointed out both empirical and theoretical difficulties with the truly random control procedure. In recognition of the problems (Holland and Rescorla 1975), the more traditional, unpaired control has once again become favored. Procedures will be described in the next section of this paper that incorporated an unpaired control for the current study.

It should be noted that designing an acceptable control procedure is particularly challenging when dealing with human subjects who may find unpaired or truly random exposures absurd and/or who may intuit the researcher's purpose. Subject awareness, or lack thereof, is a major consideration for the researcher. As discussed in the introduction of this paper, one advantage of working with preschoolers is that they are probably less likely to determine the researcher's purpose as compared to college students who may become "sophisticated" in their participation in projects.

METHOD

A 3 x 2 factor design was created to test for evidence of classical conditioning effects on children. Three levels of the first factor, temporal condition, were as follows: 1) simultaneous conditioning, CS and US presented together, 2) forward conditioning, CS presented sequentially followed by US, and 3) unpaired control, CS and US presented alone, randomly and in a long interval in relation to each other. The reader will notice that the forward conditioning reflects the traditional view of temporal spacing between the CS and US.

The second factor consisted of two levels of color of the product: 1) an orange pencil and 2) a yellow pencil. This factor was included to mock Gorn's (1982) and Allen and Madden's (forthcoming) behavioral measure of a pen selection.

Therefore, the six groups were as follows: 1) simultaneous conditioning - orange pencil, 2) simultaneous conditioning - yellow pencil, 3) forward conditioning - orange pencil, 4) forward conditioning yellow pencil, 5) unpaired control - orange pencil, and 6) unpaired control - yellow pencil.

Subjects

Arrangements were made with the directors of a daycare center and a preschool in two adjacent suburban communities of a large, midwestern city. Eighty-four subjects were recruited and randomly assigned to the conditions. The subjects were 4 and 5 years old (44 were four years, 40 were 5 years; mean age = 58.1 months). Boys and girls were approximately equal (43 boys, 41 females), and they were mostly white (77 white, 7 black). The subjects could best be described as coming from suburban, middle to upper-middle income homes surrounding an outerbelt highway of the city.

Stimuli

Brightly colored pencils were selected as the CS. Fifteen preschoolers, from similar socioeconomic backgrounds as the test children, compared four colors (black, blue, yellow, and orange) in side-by-side tests of actual pencils (the four colors were narrowed down from an original pool of eight). Orange and yellow were decided from the pairwise comparisons on an equal preference basis (53.32 preferred orange and 46.6% preferred yellow).

The source for pairing was also pretested. The same fifteen children rated nine characters on a four-point interval felt board illustrating smiling to frowning faces. The source rating task was systematically rotated with the pencil comparisons.

A picture of Smurf was clearly favored (n = 15, mean rating = 4.0 on a 1 to 4 scale). Indeed, the ceiling score indicated strong popularity. Other characters evaluated included Cabbage Patch's Otis, Cabbage Jack, Care Bear, Gargamel, Pink Panther. Prune Face, A Big Bad Wolf, and Pie Man who received the lowest score of 1.8.

Pictures of Smurf and the pencils (color photocopies of the actual pencils) were transferred to individual 9" by 12" laminated posters. Multiple copies were made for ease of administration of the three exposures included for each child in each condition

Procedure

Individuals were randomly assigned to one of the six treatment groups. Each child saw the poster of Smurf associated with a poster of either an orange or yellow pencil. As described earlier, the three temporal arrangements of the associations were either simultaneous, forward, or unpaired (control). Three exposures were included to strengthen the manipulation and were consistent with previous efforts in psychology (McSweeney and Bierly 1984). All interviews were conducted on an individual basis.

Children were told that they would be asked to answer some questions and play a game with the experimenter. In fact, the three exposures to the CS and US were buried in another study about children's understanding of advertising. At intervals appropriate to the conditioning treatment, children were asked to look at some pictures the experimenter "brought along." The subjects in the simultaneous conditions viewed the posters held side-by-side. The three exposures were approximately equally spaced during the 20 minute session. Children in the forward conditions saw the poster of the pencil immediately followed by the poster of Smurf, once again. at intervals approximately equally spaced. Children in the unpaired control condition saw one poster at six different breaks (approximately equal) in the 20 minute session. The six posters used in the unpaired control conditions were shuffled before each child entered the test room. The length of exposures, five seconds, was held constant across all conditions.

Each child rated his/her liking for Smurf on a four-point interval faces scale at the end of his/her session. Smiling to frowning faces were illustrated on a felt board which had also been used for the pretest. In addition, each child answered how good he/she thought Smurf was, and how good a feeling Smurf gave him/her. A four-point interval scale was vertically displayed on a board with the following anchors GOOD good bad BAD. Each child received at least two practice trials. The experimenter asked the child to "try the boards out" by indicating answers about such other familiar characters as Santa Claus, Strawberry Shortcake, Luke Skywalker, Pie Man, and so forth. The experimenter asked about Smurf after she judged the child to be proficient in the use of the boards.

The key measure for the study was taken at the very end of the session. The experimenter said that she wanted to thank the child for being in the study. She showed the child a yellow and orange pencil held side by side and asked the child to select one. (Pretests of the procedures indicated the importance of presenting the pencils precisely side by side. During one pretest session, the experimenter asked the child why he picked a particular color. Re responded, "because it is longer." The experimenter was careful to hold the pencils directly next to one another.)

A buy-back procedure was modified from Allen and Madden (forthcoming). The experimenter said to the child after his/her selection, "I can only give you one thing, but I want to be sure you get something you like - the pencil is yours now. But, would you like to trade me for a sticker?" The experimenter then opened a small box filled with a variety of stickers that were very popular during the time of the study.

This "buy back" procedure was included with the rationale provided by Allen and Madden (forthcoming). Whereas the choice between an orange and yellow pencil (that were equally desirable as based on a pretest) would probably not stimulate much cognitive activity, a decision to trade for a sticker would probably evoke some thought.

The experimenter placed each child's selection in an envelope that was sealed in his/her presence noting, "so it won't be lost." Thus, the child's selection was kept out of view from the other children. The child then returned to his/her classroom and joined in the ongoing activities. Teachers were asked not to discuss the study with the children but to listen for comments. Few comments were relayed; however, teachers reported such as the following, wit's fun to talk to the lady; you get to play a game, etc." No teacher reported children's awareness of the study's purpose; indeed, few comments were expressed bs the children.

RESULTS

Pencil color selection

The data were analyzed in a fashion consistent with prior work in marketing (Gorn 1982; Allen and Madden (forthcoming). As can be seen in Tables 1 and 2, the treatments did not yield statistical differences in color selection.

The results depicted in Table 1 suggest that there may have been an overall color preference for orange. Of the 14 control subjects who saw yellow, 10 selected the orange pencil. However, only 6 out of the 14 orange control group picked orange.

TABLE 1

ANALYSIS OF SUBJECTS' PENCIL COLOR CHOICES SIMULTANEOUS VERSUS FORWARD VERSUS CONTROL

More importantly, Table 2 illustrates the data in Gorn's (1982) fashion. The key concern is not what color was picked per se, but whether the source influenced the choice based on the color paired. Table 2 displays a cross-tabulation of the pencil selected ("advertised" versus "nonadvertised") by condition. No statistically significant difference appeared (X = .686, df = 2, n.s.).

Thus, the first hypothesis received support. Children who viewed Smurf simultaneously with the pencil evidenced no choice preference based on the pairing (simultaneous condition). The second hypothesis did not receive support, however. It will be remembered that children who first saw the pencil then viewed Smurf were predicted to show choice preference for the "advertised" pencil (forward condition). The data offered no support. Therefore, the form of conditioning, simultaneous or forward, provided for no statistical differences from the control group who saw Smurf and the pencil in an unpaired fashion.

TABLE 2

ANALYSIS OF SUBJECTS' ADVERTISED/NONADVERTISED CHOICES AND CONDITION

In short, no support was indicated for automatic behavioral responses to a favorable character, Smurf, being associated with a product. This finding is consistent with Allen and Madden's (forthcoming) failure to establish a simultaneous conditioning effect with adult females. It is inconsistent with Gorn's (1982) results in favor of the simultaneous conditioning hypothesis.

The trading results

Allen and Madden's (forthcoming) buy back procedure was mocked in that children were provided the opportunity to trade the pencil selected for a sticker. This measure was included to gauge whether Smurf would create a positive feeling state that would bias the cognitive evaluation of the pencils. In the procedure section of this paper, it was argued that a trade for a popular item would be more likely to stimulate thought as compared to a choice between two pencils that varied only in color.

The percentages who traded the sticker for a pencil were compared among the simultaneous, forward, and control groups. The following trades were made: simultaneous group, 82.1% (23 out of 28): forward group. 60.2: (17 out of 28); and control subjects, 89.3: (25 out of 28). The difference between the forward group and the control group (z = -2.18, p < .05) was statistically significant, and the one between the traditional and the simultaneous group (z = -1.53, p < .07) was nearly so at the traditional .05 level. Nearly the same percentage of the simultaneous (82.1%) and control subjects (89.3%) traded, z = -.71, n.s. Therefore, the treatments yielded one statistical difference in the children's willingness to trade. Apparently, the children who saw the pencil, then Smurf. were more reluctant to trade the selected pencil for a sticker. It will be argued that this procedure biased positive evaluations. Although no direct measures of thought processes were taken, children in the forward group appeared more resistant to trade.

At first glance, the trading results seem to lend support to the conditioning mechanism. As one would anticipate with conditioning. children in the forward group appeared more resistant to trade. For this interpretation to hold, however, one would expect that within the forward group, those children who selected the color of pencil paired with Smurf (consistent selectors), would have been more resistant to trade as compared to those children who selected the color they had not seen paired. Yet, within the forward group, 8 (61.5%) traded of the 13 who appeared consistent with their exposure, while 9 (60%) traded of the 15 who picked the color that they had not been shown. Thus, the difference in percentages of traders was not statistically significant (z = .06, n.s.), as one would expect with 8 true conditioning result. In addition, differences within the other groups were statistically insignificant [simultaneous (z = -.74, n.s.): 9 (75%) traded of 12 consistent selectors versus 14 (87.5%) traded of the 16 inconsistent selectors; control (z = -.95, n.s.): 8 (80%) traded of the 10 consistent selectors versus 17 (94.7%) traded of the 18 inconsistent selectors. Therefore, the color selected in terms of its exposure did not appear to relate to the trading variable.

The interpretation for cognitive activity is consistent with Allen and Madden's (forthcoming) finding with adults. They reported that pleasant humor, simultaneously presented with a slide of a pen, resulted in subjects' reluctance to accept money for the pen selected. Allen and Madden (forthcoming) argued that their finding was bolstered from indications from the "mood" literature. Humor resulted in an affective state that had an impact on subjects' judgments and behaviors indirectly by biasing cognitive evaluations of the pens.

The same argument is relevant to the current finding. Forever, only children who saw the items in a traditional temporal arrangement (CS before US) were cognitively stimulated. The procedure used in the current study may have precluded a "cognitive" result from those in the simultaneous condition. It will be remembered that all exposures were held constant at five seconds. Therefore, the children who saw the pencil and Smurf held together had less time, as compared to children in other conditions, to view the stimuli. If they had been provided a longer exposure, then a trading result may have been significant. On the other hand, as compared to the control children, the ones in the forward group appeared more reluctant to trade. The length of exposure during the control and forward conditions was equivalent. Nevertheless, the trading results are best viewed as offering no support for conditioning in absence of a difference within the forward group in terms of resistance to trade and consistency of selection. In short, children who selected a color consistent with their exposure should have been more reluctant to trade, but were not. Thus, no support is offered for a conditioning result in absence of such a difference to trade within the forward group.

DISCUSSION

The research findings offered no support for automatic responses to classical conditioning manipulations on children. In fact, results from an offer to trade suggested that cognitive activity may have been biased among children who saw a traditional, forward pairing. Limitations on the confidence of this interpretation include several methodological concerns in addition to the length of exposure already discussed.

First, it should be stressed that the current study contributes only one small piece of evidence on conditioning effects, or more precisely, the lack thereof. Many empirical studies are needed. Cell sizes of fourteen in the current study are not large. Moreover, children completed this exercise during a twenty minute session. Although the procedures probably minimized awareness, fatigue may have occurred (other tasks across subjects were held constant. however). Therefore, as is wise when interpreting any findings from one experimental setting, caution is warranted.

In addition, a US pre-exposure effect may have occurred. Familiarity has been found to decrease the effectiveness of the stimulus as a US (Mis and Moore 1973; Rescorla 1973). Smurf was not only familiar to the children but highly liked (pretest subjects' mean = 4.0, all experimental subjects mean = 3.8; 4-point scale). One could argue that an unfamiliar character should be paired before concluding a failure of conditioning. On the other hand, findings from the trade measure indicated that a forward pairing did have an effect- on cognitive evaluation, which is the antithesis of an anticipated automatic response. Therefore, an interpretation for a pre-exposure effect is questionable.

The current research provides some implications for those interested in issues of children's advertising. Little empirical work has been conducted on source effects. Yet, critics of television advertising allege that the use of characters takes unfair advantage of the younger viewer. Rossiter (1980) presented an excellent overview and research agenda on source effects in children's advertising. Be pointed out that NAB restrictions on endorsements to children can be interpreted as a reflection of concern for the allegedly negative effects.

The current research findings cannot be interpreted as support for the critics' concern. Children in the traditional, forward conditioning group were more reluctant to trade; however, one can argue that Smurf stimulated positive evaluations through active thought processing. Based on the data, one cannot argue for children's automatic responses to Smurf. Thus, a passive view of children's viewing, which is reflected in critics' concern, received no support.

Until recently, attention to television has been conceptualized as reactive in nature with the child being passively controlled by the medium. More recent work in children's television viewing does not lend support to this passive view. Instead. children appear to be active processors (c.f. Anderson and Lorch 1983; Wright and Houston 1981). It should be stressed, however, that the current research addresses these broader issues in only a most peripheral way. The present study consisted of a limited manipulation of sequencing of posters with likenesses of pencils and Smurf. Needless to say, television advertising consists of vastly more complex stimuli.

Nevertheless, the current findings lead to future research questions. Audiovisual manipulations of source effects are needed to decipher critics' concerns. The classical conditioning framework provides one path of inquiry to study whether children's responses are automatic or more active in nature

To conclude, no empirical support was indicated for behavioral effects when pictures of pencils were paired with Smurf in three temporal arrangements (simultaneous, forward, and unpaired control). Simultaneous effects were not predicted (first hypothesis); however, children in the forward conditions were predicted to exhibit automatic responses (second hypothesis). Results from a trade-back measure were more equivocal. This study contributes just one small piece of evidence in an area of research worthy of attention. Future research efforts should include manipulations of media, background features, temporal arrangements, and research settings. Hopefully, evidence will accumulate on the viability of the classical conditioning framework in marketing context including its application, or lack thereof, to issues of children's advertising.

REFERENCES

Allen, Chris T. and Thomas J. Madden (forthcoming), "A Closer Look at Classical Conditioning," Journal of Consumer Research, in press.

Anderson, Daniel R. and Elizabeth Pugzles Lorch (1983), "Looking at Television: Action or Reaction?", in Children's Understanding of Television, J. Bryant and D. R. Anderson (eds.), New York: Academic Press.

Bower. Gordon L. (1981), 'Mood and Memory," American Psychologist, 36 (February), 129-48.

Clark, Margaret S. and Alice M. Isen (1982), "Toward Understanding the Relationship Between Feeling States and Social Behavior," in Cognitive Social Psychology, A. B. Hastorf and A. M. Isen (eds.), Nev York, NY: Elsevier/North-Holland, 73-108.

Gormezano, I. and J. Kehoe (1975), 'Classical Conditioning: Some Methodological-Conceptual Issues," in Handbook of Learning and Cognitive Processes, Hillsdale, NJ: Lawrence Erlbaum Associates, Inc., 143-179

Gorn, Gerald J. (1982), 'The Effects of Music in Advertising on-Choice Behavior: A Classical Conditioning Approach," Journal of Marketing, 46 (Winter), 94-101.

Gresham, Larry G. and Terence A. Shimp (1985), 'Attitude Toward the Advertisement and Brand Attitudes: A Classical Conditioning Perspective," Journal of Advertising, 14 (1), 10-17.

Holland, Peter C. and Robert A. Rescorla (1975), "Second-Order Conditioning with Food Unconditioned Stimulus,' Journal of Comparative and Physiological Psychology, 88 (1), 459-467.

McSweeney, Frances K. and Calvin Bierley (1984), 'Recent Developments in Classical Conditioning," Journal of Consumer Research, 11 (September), 619-631.

Mis, Frederick W. and John W. Moore (1973), 'Effect of Preacquisition USC Exposure on Classical Conditioning of the Rabbit's Nictitating Membrane Response," Learning and Motivation, 4 (1), 108-114.

Moore, John W. and I. Gormezano (1977), "Classical Conditioning,' in Fundamentals and Applications of Learning, M. Marks and M. Bunch (eds.), New York MacMillan, 87-120.

Nord, Walter R. and J. Paul Peter (1980), "A Behavior Modification Perspective on Marketing," Journal of Marketing 44 (Spring), 36-47.

Pavlov, Ivan P. (1927), Conditioned Reflexes, Oxford, England: Oxford University Press.

Rescorla, Robert A. (1967), 'Pavlovian Conditioning and Its Proper Control Procedures," Psychological Review, 74 (1), 71-80.

Rescorla, Robert A. (1973), "Effect of US Habituation Following Conditioning,' Journal of Comparative and Physiological Psychology, 82 (1), 137-143.

Rossiter, John R. (1980), 'Source Effects and Self-Concept Appeals in Children's Television Advertising,' The Effects of Television Advertising on Children, R. P. Adler et al. (eds.), Lexington, VA: Lexington Books.

Smith, Marius C., Steven R. Coleman, and I. Gormezano (1969), "Classical Conditioning of the Rabbit's Nictitating Membrane Response at Backward, Simultaneous, and Forward CS-US Intervals,' Journal of Comparative and Physiological Psychology, 69 (2), 226-231.

Staats, Carolyn K. and Arthur W. Staats (1957), "Meaning Established by Classical Conditioning," Journal of Experimental Psychology, 54 (July), 74-80.

Staats, Carolyn K. (1958), "Attitudes Established by Classical Conditioning," Journal of Abnormal and Social Psychology, 57 (July), 37-40.

Staats, Carolyn K. (1959), 'Effect of Number of Trials on the Language Conditioning of Meaning,' Journal of General Psychology, 61, 211-223.

Wright, John C. and Aletha C. Huston (1981), "Children's Understanding of the Forms of Television, New Directions for Child Development: Viewing Children Through Television, L. Kelly and E. Gardner (eds.), San Francisco, CA: Jossey-Bass Inc.. Publishers.

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